Laps haber bi led a iakenony pepo nerer Pee een art eg adie ata tie Fae ts weeny Mtg tea teteeed rites aes rebeer tote pr te rtcepecet cies tae genee tere os re i. e AMERICAN A513 (| MALACOLOGICAL re BULLETIN VOLUME 14 NUMBER 1 Journal of the American Malacological Union CONTENTS Cryptostrakon corcovadensis, a new species of semislug from Costa Rica (Helicoidea: Xanthonychidae) with comments on the systematic position of the genus. MARIA GABRID LA CUBZ ZG oui viccsssenicdeciicves sebetnstacedan tee veueiansesiecastas tan cbininbisisavatansbives Gustasaneonienieiss I Chilina megastoma Hylton Scott, 1958 (Pulmonata: Basommatophora): a study on topotypic specimens. CRISTIAN F. ITUARTE 00... ecccceecneseeeeeeeneeeeesaeseeeseeseesesteeeeeseeeeeeanenes 9 The role of prey mobility in the population ecology of the nudibranch Cuthona nana (Gastropoda: Opisthobranchia). NADINE C. FOLINO ..00....cccccccececeeeeseceeeeseeseeesesseessesseeaeesseeaeees 17 Temporal and spatial patterns of abundance in the gastropod assemblage of a macrophyte bed. KENNETH M. BROWN .00.oo oc eeceeeceenesneceeetaeeneeeeeeaeeeeeeeseeesseeaeeeaeseeeseeegees af Sympatric speciation of freshwater mussels (Bivalvia: Unionoidea): a model. MAINES Me MRA soho ces in tty cnc es sxe pice asada aan oem asa asap caaddeen dda ap aed cdaereaa aes 35 Freshwater mussels (Bivalvia: Unionidae) in the Verdigris, Neosho, and Spring River Basins of Kansas and Missouri, with emphasis on species of concern. BRIAN K. OBERMEYER, DAVID R. EDDS, CARL W. PROPHET, erred PCD VEIN DERE ia occas clecaed cass nc cd cate eanataiangeasmcautvedl esac econ snnthasersvvintaie teen Bapeaddeabas 4] Growth and survival of juvenile mussels, Villosa iris (Lea, 1829) (Bivalvia: Unionidae), reared on algal diets and sediment. CATHERINE M. GATENBY, BRUCE C. PARKER, and RICHARD J. NEVES 00..........ccccccccccccccccesscesssecsssccessecesseeseneteeesesens 57 Zebra mussel induced mortality of unionids in firm substrata of western Lake Erie and a habitat for survival. DON W. SCHLOESSER, R. D. SMITHEE, G. D. LONGTON, and W. P. KOVALAK ooo. ccccccccccceessesesscseescesscstessussescssseusasesassevausausansnaserss 67 Effects of quarantine times on glycogen levels of native freshwater mussels (Bivalvia: Unionidae) previously infested with zebra mussels. MATTHEW A. PATTERSON, BRUCE C. PARKER, and RICHARD J. NEVES 2.0.0.0... cccccccccceccescesceseesseseeees 75 Research Note: Observations on the reproductive biology of the octopod Eledone gaucha Haimovici, 1988, in southern Brazil. JOSE ANGEL A. PEREZ, MANUEL HAIMOVICI, and ROBERTA AGUIAR DOS SANTOS 0ooooocccecccccccescecescsseseeccseees 81 ANMOUNCEMENL .......ccececccescecceseesccsscesscecaucsscsussacesecatesscsueeuees Gy eset daca hance ncesesnasgieeesanaeataieacoasaeeuo tu: 85 In Memoriam...........cececeesescseecesscecceececeaseeeesauscerses DELS Eeeh sebanc stele aah amas staat uadaaceaticleateateee eave suas des Sew ancoattareieteatecs 86 AMERICAN MALACOLOGICAL BULLETIN BOARD OF EDITORS RONALD B. TOLL, Editor-in-Chief PAULA M. MIKKELSEN, Managing Editor Department of Biology Department of Invertebrates Wesleyan College American Museum of Natural History Macon, Georgia 31297-4299 New York, New York 10024-5192 MELBOURNE R. CARRIKER College of Marine Studies ASSOCIATE EDITORS University of Delaware Lewes, Delaware 19958 GEORGE M. DAVIS Department of Malacology The Academy of Natural Sciences Philadelphia, Pennsylvania 19103 JOHN A, ALLEN Millport, United Kingdom JOHN M. ARNOLD Honolulu, Hawaii, U.S.A. JOSEPH C. BRITTON Fort Worth, Texas, U.S.A. JOHN B. BURCH Ann Arbor, Michigan, U.S.A. EDWIN W. CAKE, JR. Ocean Springs, Mississippi, U.S.A. PETER CALOW Sheffield, United Kingdom JOSEPH G. CARTER Chapel Hill, North Carolina, U.S.A. ARTHUR L. CLARKE Portland, Texas, U.S.A. CLEMENT L. COUNTS, III Wallops Island, Virginia, U.S.A. THOMAS DIETZ Baton Rouge, Louisiana, U.S.A. WILLIAM K. EMERSON New York, New York, U.S.A. DOROTHEA FRANZEN Bloomington, Illinois, U.S.A. ROGER HANLON Woods Hole, Massachusetts, U.S.A. JOSEPH HELLER Jerusalem, Israel ROBERT HERSHLER, Ex Officio Department of Invertebrate Zoology Smithsonian Institution Washington, D. C. 29560 BOARD OF REVIEWERS ROBERT E. HILLMAN Duxbury, Massachusetts, U.S.A. K. ELAINE HOAGLAND Washington, D.C., U.S.A. VICTOR S. KENNEDY Cambridge, Maryland, U.S.A. ALAN J. KOHN Seattle, Washington, U.S.A. LOUISE RUSSERT KRAEMER Fayetteville, Arkansas, U.S.A. JOHN N. KRAEUTER Baltimore, Maryland, U.S.A. ALAN M. KUZIRIAN Woods Hole, Massachusetts, U.S.A. RICHARD A. LUTZ Piscataway, New Jersey, U.S.A. GERALD L. MACKIE Guelph, Ontario, Canada EMILE A. MALEK New Orleans, Louisiana, U.S.A. MICHAEL MAZURKIEWICZ Portland, Maine, U.S.A. JAMES H. McLEAN Los Angeles, California, U.S.A. ROBERT F. MCMAHON Arlington, Texas, U.S.A. ANDREW C. MILLER Vicksburg, Mississippi, U.S.A. W. D. RUSSELL-HUNTER Department of Biology Syracuse University Syracuse, New York 13210 THOMAS R. WALLER Department of Paleobiology Smithsonian Institution Washington, D. C. 29560 BRIAN MORTON Hong Kong JAMES J. MURRAY, JR. Charlottesville, Virginia, U.S.A. RICHARD NEVES Blacksburg, Virginia, U.S.A. JAMES W. NYBAKKEN Moss Landing, California, U.S.A. A. RICHARD PALMER Edmonton, Canada WINSTON F. PONDER Sydney, Australia ROBERT S. PREZANT Indiana, Pennsylvania, U.S.A. CLYDE F. E. ROPER Washington, D.C., U.S.A. NORMAN W. RUNHAM Bangor, United Kingdom AMELIE SCHELTEMA Woods Hole, Massachusetts, U.S.A. DAVID H. STANSBERY Columbus, Ohio, U.S.A. FRED G. THOMPSON Gainesville, Florida, U.S.A. Cover: Io fluvialis (Say, 1825) is the logo of the American Malacological Union. THE AMERICAN MALACOLOGICAL BULLETIN is the official journal publication of the American Malacological Union. AMER. MALAC. BULL. 14(1) ISSN 0740-2783 AMERICAN MALACOLOGICAL BULLETIN VOLUME 14 NUMBER 1 Journal of the American Malacological Union CONTENTS Cryptostrakon corcovadensis, a new species of semislug from Costa Rica (Helicoidea: Xanthonychidae) with comments on the systematic position of the genus. IMARTAIG ARRIBA CUR ZZ Os ices csc oi. aise oer cra sasha osastbtataan vs Lave biden estat sue veidee lees rene evedadtants Chilina megastoma Hylton Scott, 1958 (Pulmonata: Basommatophora): a study on topotypic: specimens. CRISTIAN Fi. FEUAR TE ciccicrraeseicaiiatelvesscastareiansauans sativennleenisvaeneateaseancnneraaes The role of prey mobility in the population ecology of the nudibranch Cuthona nana (Gastropoda: Opisthobranchia). NADINE C. FOMUNO 2i2issiisietsccsscacsesstecunanactinsveavedsesssurenseetevensaeciverese Temporal and spatial patterns of abundance in the gastropod assemblage of a macrophyte bed. KENNICIH Mi. BROWN occiccsiscccccscssetierargecaatsscsscovssosseqsensetuctulUeienensestiteadavewenedeece Sympatric speciation of freshwater mussels (Bivalvia: Unionoidea): a model. | De Nop 0 0) Wa] Cs 5 CL re 0 ae en ce ae Pr ea eer eee re err Pees > Freshwater mussels (Bivalvia: Unionidae) in the Verdigris, Neosho, and Spring River Basins of Kansas and Missouri, with emphasis on species of concern. BRIAN K. OBERMEYER, DAVID R. EDDS, CARL W. PROPHET, SAUCERS NV UND oy LENT MOIR eer tes aca st gape occ eneloop ne Siuhs ties taudeagen detvabae teececteess te eeeaycameand ncaa iawine Growth and survival of juvenile mussels, Villosa iris (Lea, 1829) (Bivalvia: Unionidae), reared on algal diets and sediment. CATHERINE M. GATENBY, BRUCE C. PARKER, and RICHARD J. NEVES. o0.......cccccccccccccccessssccessssecssssecesesssceeeestseeeeeeaas Zebra mussel induced mortality of unionids in firm substrata of western Lake Erie and a habitat for survival. DON W. SCHLOESSER, R. D. SMITHEE, G: DELONGTON, and Wo Pa KOVALAKG siccc. ctecccsciecisedinyonihsconecescsnsssrisiadets dadceressdisatebosecadooavenstavavan Effects of quarantine times on glycogen levels of native freshwater mussels (Bivalvia: Unionidae) previously infested with zebra mussels. MATTHEW A. PATTERSON, BRUCE C. PARKER, and RICHARD J. NEVES ...000.0..00cccccccccccccesccescesectseeseens Research Note: Observations on the reproductive biology of the octopod Eledone gaucha Haimovici, 1988, in southern Brazil. JOSE ANGEL A. PEREZ, MANUEL HAIMOVICI, and ROBERTA AGUIAR DOS SANTOS ......000.0...occcccccccceceseeeeeeeeees PATIMIOUNGE INEM terrence neers eh ree aaa cree aa Eee Sa CoE Ee Ee es en JIFoX 109 (STO GVO a ETAT eee sesseseeceee meee eee eee ere cap ur een LIBRARIES Cryptostrakon corcovadensis, a new species of semislug from Costa Rica (Helicoidea: Xanthonychidae) with comments on the systematic position of the genus Maria Gabriela Cuezzo College of Agriculture & Life Sciences, Comstock Hall, Cornell University, U. S. A., and Instituto de Invertebrados, Fundaci6n Miguel Lillo, Miguel Lillo 251, 4000 Tucumén, Argentina Abstract: A new species of the semislug genus Cryptostrakon Binney, 1879, is described. A description of the external morphology as well as the pal- lial, reproductive, digestive, and nervous systems is carried out. Based on the new anatomical evidence the systematic position of the genus within the Xanthonychidae is reaffirmed. Comparisons among the semislug genera of the Xanthonychidae are also provided. Morphologically, Cryptostrakon appears to be more related to Xanthonyx and Metostracon than to any other xanthonychid genera. The phylogenetic relationships of Cryptostrakon are, however, still controversial and their clarification will have to wait until a testable phylogeny of the Xanthonychidae is proposed. Key words: anatomy, systematics, taxonomy, Central America The genus Cryptostrakon was described by Binney (1879) based upon a species collected by W. Gabb in Costa Rica. C. gabbi Binney is the type species by original desig- nation (Binney, 1879; Baker, 1963) and until now the only component of the genus. Unfortunately, the type material upon which the species description was based is dried, as stated by Pilsbry (1900). Although Baker (1963) designated a lectotype and several paralectotypes, anatomical observa- tions are not possible based upon them because of the poor preservation of their bodies. Consequently, the descriptions that Binney (1879) provided on the jaw, radula, shell, and external morphology of the body of C. gabbi remain as the only data for the genus. For this reason, until now the inter- nal anatomy of the different systems in this genus were completely unknown. The purpose of this study is to describe a new species of the genus Cryptostrakon, to compare this genus with other semislug genera, and to clarify its systematic position. MATERIAL AND METHODS The following material was studied (ANSP = Academy of Natural Sciences of Philadelphia; INBIO = Instituto de Biodiversidad de Costa Rica): Cryptostracon gabbi Binney, 1879 ANSP 57992: Costa Rica (Gabb), flanks of Pico Blanco, 5000-7000 ft elevation. TYPE. ANSP 246310, A9639: Costa Rica. Det. H. B. Baker. LECTOTYPE. ANSP 394149: Costa Rica. PARALECTOTYPES. Cryptostracon corcovadensis sp. nov. INBIO 468087: Costa Rica, Madrigal, Prov. Punta, P. N. Corcovado, Estacion Sirena, Sendero a Rio Los Patos, 10 m. LS 270600_508300. 14 August 1994, Marianella Segura! HOLOTYPE. INBIO 468080: Costa Rica, Madrigal, Prov. Punta, P. N. Corcovado, Estacion Sirena, Sendero a Rio Los Patos, 10 m. LS270600_507800. 18 August 1994. Marianella Segura! INBIO 468059: Costa Rica, Prov. Punta, P. N. Corcovado, Estacion Sirena, Sendero Los Espaveles, 20 m. LS270600_507800. 12 August 1994. Marianella Segura! Bunnya bernardinae Baker, 1942 ANSP A16728: ruins of monastery 20 km SW of Mexico City, El Desierto de Los Leones to La Venta, Distrito Federal, Mexico. July 1926. H. B. Baker! Metostracon mima Pilsbry, 1900 ANSP 77245, A9636: Morelia, Michoacan, Mexico. American Malacological Bulletin, Vol. 14(1) (1997):1-8 2 AMER. MALAC. BULL. 14(1) (1997) S. N. Rhoads! HOLOTYPE. ANSP A9635F: Uruapam del Progreso, Michoacan, Mexico. S. N. Rhoads! PARATYPE. ANSP A9410D: near Alvarez at km 53, San Luis de Potosi, Mexico. July 1934. H. A. Pilsbry! ANSP A9411A: km 42, Potosi and Rio Verde Railroad, San Luis de Potosi, Mexico. August 1934. H. A. Pilsbry! Xanthonyx sp. ANSP A16735: 14 km from Cordoba toward Orizaba on Mexican railroad, hills southeast of town, on opposite side of Canyon Sumidero, Vera Cruz State, Mexico. June 1926. H. B. Baker! The alcohol preserved material was dissected under a Wild MSA microscope and illustrations were made with the aid of a camera lucida. A Cambridge 200 scanning elec- tron microscope at ANSP was used for the observation of the radula and jaws; the preparation procedure followed was as described by Ploeger & Breure (1977). Terminology for the anatomical descriptions follows Tompa (1984). The alcohol preserved material of Xanthonyx, Metostracon, and Bunnya spp. were used for comparative proposes among the semislug xanthonychid genera. Formal descriptions on the anatomy of these taxa were not carried out; only the differences found with the original descrip- tions are pointed out. RESULTS Subclass Pulmonata Order Stylommatophora Superfamily Helicoidea Family Xanthonychidae Strebel & Pfeiffer, 1879 Genus Cryptostrakon Binney, 1879 Cryptostrakon corcovadensis sp. nov. Diagnostic characters Shell depressed with spiral nucleus, entirely con- cealed by mantle. Visceral mass reduced with pneu- mostome in middle right margin. Tail mid-dorsal groove well-developed, lined with thin pigmented band. Jaw smooth, presenting only shallow vertical grooves in medial zone. Mucus glands of terminal genitalia composed of mul- tiple thin blind sacs (tubules) enveloped by same tissue ending in one common duct inserted in medial zone of the dart sac. Atrium with thin transverse ridges in its interior. Type locality: (see Materials and Methods). Etymology: Named for Corcovado National Park in Costa Rica, type locality of this species. External morphology Semislug with visceral mass reduced and packed in medial dorsal portion of animal. Visceral mass protected by shell, completely concealed by mantle. Pneumostome wide, located in medial right margin of visceral mass (Figs. 1, 2). Openings of rectum and pneumostome clearly visible. Tail muscular with well-developed mid-dorsal groove marked by thin dorsal pigmented band that narrows, eventually dis- appearing toward tip of tail (Fig. 1). Foot sole not divided. Holotype 22 mm total preserved length. Internal morphology Shell Shell depressed, concealed by mantle, fragile to membranous. Two whorls without septa. Aperture wide, ovate. Concentric lines of growth noticeable on last whorl. Peristome lined with thin membrane reflected toward inte- rior of aperture. Measurements of holotype (Figs. 3, 4): Major diameter, 8 mm; minor diameter, 6.5 mm; aperture diameter, 5.5 mm; height, 4 mm. Body cavity Body cavity not extending into posterior part of foot, ending immediately behind visceral mass. Anterior part of foot containing part of digestive, reproductive, and nervous systems. Pallial system Pallial system lying under mantle and immediately beneath shell, thus protected by both shell and mantle (Fig. 2), composed of lung, kidney, ureters, and pericardium, occupying anterior left half of reduced visceral mass. Kidney, heart, and rectum perpendicular to longitudinal axis of the body. Kidney partially surrounding pericardium, with axis directed toward pneumostome. Kidney short, half the length of lung, not compact, with four inferior lobes with internal lamellae converging to kidney pore (Fig. 5). Primary ureter starting at right side of pericardium and run- ning transversely across to rectum (Figs. 2, 5). Secondary ureter running parallel to rectum, closely until ureteric interramus where it splits into two branches. Ureteric inter- ramus shallowly excavated. Adrectal branch of secondary ureter continuing straight to mantle collar while abrectal branch turns obliquely exiting in a separate pore. Lung short with alveoli on upper wall parallel to rectum (upper right margin of visceral mass). Reproductive system Distal portion of reproductive system (terminal gen- italia), in its natural position, running laterally (right mar- gin) to the digestive system. Reproductive pore located on right side of cephalic region, under right ocular tentacle. Medial and proximal portions of system extending dorsally CUEZZO: NEW COSTA RICAN SEMISLUG 3 Figs. 1-5. Cryptostrakon corcovadensis sp. nov., gross morphology and pallial complex. Fig. 1. Dorsal view of the animal. Fig. 2. Ventral view of the visceral mass. Figs. 3-4. Dorsal and ventral views, respectively, of shell of holotype. Fig. 5. Detail of pallial complex. (AG, albumen gland; AL, alveoli; DG, digestive gland; HD, hermaphroditic duct; I, intestine; K, kidney; KL, kidney lobes; M, mantle; MC, mantle collar; MD, mid-dorsal groove; O, ovotestis; PE, pericardi- um; PN, pneumostome; PU, primary ureter; R, rectum; ST, stomach; UI, ureteric interramus). Scale bars = 2.3 mm (5), 5 mm (2), 8 mm (1, 3-4). 4 AMER. MALAC. BULL. 14(1) (1997) Figs. 6-10. Cryptostrakon corcovadensis sp. nov., reproductive system. Fig. 6. Dorsal view of reproductive system, excluding ovotestis and hermaphroditic duct; penis illustrated in part. Fig. 7. Ventral view of reproductive system as in Fig. 6. Fig. 8. Detail of terminal genitalia. Fig. 9. Penial complex. Fig. 10. Detail of verge. (A, atrium; AG, albumen gland; BC, bursa copulatrix; BCD, bursa copulatrix duct; D, diverticulum; DP, dart papillae; DS, dart sac; F, fla- gelllum; FO, free oviduct; L, lappets; MG, mucus gland; MGD, mucus gland duct; MS, muscular strands; P, penis; PO, penial orifice; PR, penial retractor; V, vagina; VD, vas deferens; VE, verge). Scale bars = 2.5 mm (6-7), 1 mm (8, 10), 1.5 mm (9). CUEZZO: NEW COSTA RICAN SEMISLUG 5 Figs. 11-14. Cryptostrakon corcovadensis sp. nov., radula. Fig. 11. General view. Fig. 12. Detail of central and lateral teeth. Fig. 13. Dorsolateral view of marginal teeth. Fig. 14. Ventral view of marginal tricupsid teeth. Scale bars = 20 pm (12-14), 200 pm (11). over esophageal crop; albumen gland located under reduced visceral mass. Almost all of system’s wall bright white, contrasting with dark color of esophageal crop content to which it is in part appressed. Proximal and Medial Genitalia (Figs. 2, 6, 7): Ovotestis composed of multiple acini embedded in lobes of proximal digestive gland in visceral mass (Fig. 2). Color of ovotestis pale yellow. Hermaphroditic duct thin and convo- luted, descending from ovotestis to Fertilization Pouch- Spermathecal Complex (FPSC) in base of albumen gland, which is small and white. Spermoviduct extending from albumen gland spreading over esophageal crop. Bursa cop- ulatrix duct longer than spermoviduct and running parallel to it. Diverticulum short and thin. Bursa copulatrix sac nar- row, and about same diameter as duct. Terminal Genitalia (Figs. 6, 7, 8): Vagina short, with internal thin longitudinal folds. One dart sac, short and cylindrical, seated on inferior zone of vagina. Internal dart papilla extending from middle to basal part of dart sac (Fig. 8). Mucus gland composed of multiple thin tubules, finger- shaped, opening into one thicker and convoluted duct that ends in middle zone of dart sac. Tubules of mucus gland enveloped by common membranous and transparent tissue giving compact appearance to organ. Tubules of mucus gland folded into two equal portions (Fig. 7). Thin muscular strands connecting dart sac with free oviduct and mucus gland duct. Penial complex composed of thin slender penis continued by short epiphallum (Fig. 9). Verge present, located in upper portion of penis occupying half penial length. External structure of verge consisting of two opposed columns of oblique plaques (Fig. 10). Penial pore (= penial orifice) subterminal, lateral. Penial muscular 6 AMER. MALAC. BULL. 14(1) (1997) retractor long and thin, inserting in medial zone of epiphal- lum. Flagellum as long as penis being coiled over itself. Vas deferens thin, after passing trough angle formed by dart sac and mucus gland duct, inserting on penial complex. Atrium short with internal thin transversal ridges (Fig. 8). Digestive system Radula: Central teeth narrow, tricuspid, with tall mesocone, 9-10 bicuspid laterals, and 25 wide, tricuspid marginals (Figs. 11-14). Jaw: Arched, apparently smooth but fine parallel striae evident (by SEM observations) in medial zone, with- out ribs; both margins lined by fine denticles. Buccal mass round. Esophagus short and slender, opening dorsally from anterior buccal mass. Internal orna- mentation consisting of longitudinal ridges. Esophageal crop well-developed, appressing terminal portion of repro- ductive system toward right body wall. Esophageal crop running backward to end of body cavity. Wall of esophageal crop thin, almost transparent to whitish, with irregularly distributed internal pustules. Two salivary glands laterally appressed to esophageal crop. Paired sali- vary ducts entering buccal mass on dorso-lateral surface on either side of esophagus. Last portion of salivary glands not joining in a common body as observed in other xanthony- chids. Stomach and esophageal crop connected by thin tubular portion of digestive tract. Stomach embedded in digestive gland. System ending at rectum in mantle collar. Central nervous system Central nervous system forming ring around esoph- agus. Anterior nerve ring composed of clearly spherical cerebral ganglia. Cerebral commissure short (0.5 mm in length in Holotype). Cerebropedal connectives short (0.75 mm in length in Holotype). Pedal ganglia elongated per- pendicular to foot longitudinal axis. Pleuropedal connec- tives inserting on posterior external surface of pedal gan- glia. Visceral chain compact. Visceral ganglion fused with both parietal ganglia. Pleural ganglia in close contact with parietal ganglia. DISCUSSION Semislugs are snails in which shell reduction has proceeded so far that the esophageal crop is at least partly contained in the foot cavity and the animal cannot retract inside the shell (Tillier, 1984). Other particular characters found in Crytostrakon corcovadensis that, according to Tillier (1983, 1984), could be related with the semislug condition are: (1) stomach retained in the upper visceral cavity; (2) shortening of the lung and development of sec- ondary repiratory structures (alveoli); (3) reproductive sys- tem almost entirely contained in the foot cavity (with the exception of ovotestis and hermaphroditic duct); (4) diges- tive gland occupying most of the reduced visceral cavity and invading the lung cavity; (5) rotation of the axis of the kidney so that this axis tends to be perpendicular to the lon- gitudinal axis of the foot; (6) kidney U-shaped with basal foldings or lobes; and (7) mantle completely concealing the shell. Although some authors (Solem, 1978; Tillier, 1983, 1984, 1989) suggested that these kinds of characters are consequences of the semislug condition, their true signifi- cance should be explored only when based upon a phyloge- netic hypothesis constructed by strict methodology (Henning, 1966; Farris, 1983). When Binney established the genus Cryptostrakon, he compared it with Mariella, Gaeotis, and Parmella, stressing that differences were especially strong in the jaw and radula. Besides those comparisons he did not clarify the systematic position of Cryptostrakon. Pilsbry (1900: 29) described the genus Metostracon and affirmed that “I am somewhat disposed to think that it [Cryptostrakon] will be found to agree in the main with Xanthonyx or possibly Metostracon.” However, he stressed that as the shell of Peltella and Gaeotis (Bulimulidae) were very similar to that of Cryptostrakon, the systematic position of Cryptostrakon would remain doubtful until at least its geni- talia were known. According to Breure (1974) (who described the anatomy of the genus Gaeotis), Peltella, Gaeotis, Pellicula, and Amphibulima constitute a natural group within the Bulimulidae. The anatomical differences with Cryptostrakon are so strong that it would be impossi- ble to include it within the Bulimulidae. Concerning fami- lies of Helicoidea that also present semislug genera such as Helicarionidae and Vitrinidae, Cryptostrakon could not be assigned within these families because: (1) it does not have an esophageal crop (Tillier, 1989), (2) it does have an aula- copod foot and caudal gland, (3) the Vitrinidae are also car- nivorous showing an elongated buccal mass, and (4) it pre- sents a substantially different genital system. Based on the new internal anatomical data provided in this paper, especially the presence of a dart sac and mucus glands homologous (by position) to those in Xanthonychidae, the systematic position of the genus Cryptostrakon within the Xanthonychidae is confirmed. Unfortunately because there is no material of C. gabbi presently available (other than the type material, that, as stated before, is in poor shape for anatomical studies), it is not possible to point to several anatomical differences between the two species, but only to compare the few char- acters described by Binney in his original description. The species Cryptostrakon corcovadensis, presently described, was included in the genus because it agrees with Binney’s (1879: 258) generic diagnosis, especially in gen- eral body shape (“animal slug-like, cylindrical, attenuated CUEZZO: NEW COSTA RICAN SEMISLUG 7 behind”), disposition of the mantle (“mantle thin, small, oval, entirely covering the shell’), radula (“central teeth tri- cuspid, laterals bicuspid, marginals quadrate”’), and shell (“shell rudimentary, membranous, without distinction of septa, a spiral arrangement indicated above by depressed lines, below by raised ridges”). Cryptostrakon corcovadensis differs from C. gabbi, the other species of the genus, in its body size being small- er, and in the jaw described by Binney (1879: 260) as fol- lows: “Jaw strongly arcuate, ends blunt but little attenuated; anterior surface with two stout decided ribs, denticulating either margin and several other subobsolete ribs.” In C. cor- covadensis, the jaw is almost smooth without ribs. C. cor- covadensis also differs in shell size. These characters justi- fy the creation of a new species that without doubt belongs to Cryptostrakon. No other xanthonychid genera would have received this species. Comparison with other xantho- nychid semislug and slug genera revealed that (Table 1): C. corcovadensis differs from Xanthonyx in: (a) absence of terminal tail horn, (b) shell smalier and more depressed, (c) Table 1. Selected comparative characters among species of xanthonychid semislug genera. bernardinae corcovadensis a with 2-1/2 whorls, vitriniform, 18 mm in diameter excavating the tail not excavating the not excavating the not excavating the tail tail tail ee aan ribs to the left of the to the left of the partially surrounded kidney kidney by the kidney with 2 whorls, depressed, 8 mm in diameter one whorl, oblong- with 2 whorls, oval, 11 mm long spiral between two lobes of the kidney deeply excavated shallowly excavated shallowly excavated shallowly excavated with alveoli near the pneumostome with alveoli near the pneumostome with alveoli near the pneumostome with alveoli along the rectum and near the pneumostome short, with 3 terminal as long as the penis as long as penis, with half the penis length with 2 projections 3 flattened projections terminal projections 2, inserted in the vagina where it joins several long finger glands the dart sac inserted in dart sac one, round, muscular, one, inserted in vagina 3, double, round, inserted | one, cylindrical, seated inserted in vagina in vagina in the vagina straight, in pedal cavity bilobate, in visceral cavity | bilobate, in visceral cavity | curved, in visceral cavity fused with both visceral and left pleural 3, inserted in one duct with the vagina 2, inserted in apical portion of dart sac fused with visceral ganglion in contact, not fused with visceral ganglion and pleural ganglion position of visceral ganglion | on the left side on the left side on the left side respect the medial plane of pedal ganglia fused with visceral ganglion and right parietal left parietal ganglion 8 AMER. MALAC. BULL. 14(1) (1997) jaw not ribbed, (d) buccal mass globular, (e) heart partially surrounded by kidney, (f) kidney with lobes, (g) smaller verge, (h) shape of mucus glands, (i) albumen gland not bilobed, (j) visceral ganglion fused with both parietal gan- glia, and (k) general size of the body. C. corcovadensis is similar to Metostracon in that both genera present lobes in the kidney, an albumen gland not bilobed, a globular buccal mass, and fusion of ganglia in the central nervous system. The differences between C. corcovadensis and Bunnya appear to be stronger than those with the other two men- tioned genera. Many assumptions have been made about the phy- logenetic relationships of Cryptostrakon. Pilsbry (1900) stated that Cryptostrakon is morphologically closer to Xanthonyx than to any other xanthonychid semislug genera. He also suggested that the genera Xanthonyx, Metostracon, and Cryptostrakon should be grouped together in the same subfamily due to their common condition as semislug taxa. This idea was subsequently rejected by Nordsieck (1987) who proposed a new subfamily containing only the genus Metostracon which has incompletely reached the slug stage, separate from the Xanthonychinae which contains Xanthonyx. Nordsieck also stressed that the differences between Metostracon and Xanthonyx are so important that parallel evolution of these two groups must be assumed. Traditionally the lack of an objective, testable analysis on the phylogenetic relationships of the Xanthonychidae genera has caused confusion in their clas- sification. For this reason, to eliminate all arbitrary analy- sis, only the formulation of cladistic hypotheses based on clearly described characters will produce a predictive clas- sification. As stated by Wheeler (1995: 31), “Such phyloge- netic classifications combine the best of descriptive taxono- my and phylogenetic analysis providing the historical per- spective essential to a biology that is truly evolutionary.” ACKNOWLEDGMENTS This study was partially completed while the author was support- ed by a Jessup Fellowship awarded by ANSP. I thank George Davis for his support and encouragement. I am grateful to Caryl Hesterman and David Robinson who provided much help with the collection at ANSP and the use of the scanning electron microscope. Thanks are extended to Zaidett Barrientos (INBIO) for lending the specimens for study. I am indebted to Quentin Wheeler for providing laboratory facilities at Cornell University and for his unconditional permanent support. LITERATURE CITED Baker, H. B. 1963. Type land snails in the Academy of Natural Sciences of Philadelphia. Part II. Land Pulmonata, exclusive of North America north of Mexico. Proceedings of the Academy of Natural Sciences of Philadelphia 115(8):191-259. Binney, W. G. 1879. On the jaw and lingual dentition of certain Costa Rica land shells collected by Dr. William M. Gabb. Annals of the New York Academy of Sciences 1(9):257-262. Breure, A. S. H. 1974. Notes on the genus Gaeotis Shuttleworth, 1854 (Mollusca, Gastropoda, Bulimulidae). Netherlands Journal of Zoology 24(3):236-252. Farris, J. S. 1983. The logical basis of phylogenetic analysis. In: Advances in Cladistics, 2, N. Platnick and V. Funk, eds. pp. 7-36. Columbia University Press, New York. Hennig, W. 1966. Phylogenetic Systematics. University of Illinois Press, Urbana. 263 pp. Nordsieck, H. 1987. Systematic revision of the Helicoidea (Gastropoda: Stylommatophora). Archiv fiir Molluskenkunde 118:9-S0. Pilsbry, H. A. 1900. Metostracon, a new slug-like genus of dart-bearing Helicidae. Proceedings of the Malacological Society of London 4(1):24-30, pl. 3. Ploeger, S. and A. S. H. Breure. 1977. A rapid procedure for preparation of radulae for routine research with the scanning electron micro- scope. Basteria 41:47-52. Solem, A. 1978. Classification of the land Mollusca. In: Pulmonates, 2A, V. Fretter and J. Peake, eds. pp. 49-97. Academic Press, London and New York. Tillier, S. 1983. Structures respiratoires et excrétrices secondaires des Limaces (Gastropoda: Pulmonata: Stylommatophora). Bulletin de la Société Zoologique de France 108(1):9-19. Tillier, S. 1984. Patterns of digestive tract morphology in the limacisation of the helicarionid, succineid and athoracophorid snails and slugs (Mollusca: Pulmonata). Malacologia 25(1):173-192. Tillier, S. 1989. Comparative morphology, phylogeny and classification of land snails and slugs (Gastropoda: Pulmonata: Stylommatophora). Malacologia 30(1-2):1-303. Tompa, A. 1984. Land snails (Stylommatophora). In: The Mollusca, Vol. 7. Reproduction, K. M. Wilbur, ed. pp. 47-140. Academic Press, New York. Wheeler, Q. D. 1995. The “old systematics”: classification and phylogeny. In: Biology, Phylogeny, and Classification of Coleoptera, J. Pakaluk and S. A. Slipinski, eds. Muzeum i Instytut Zoologii PAN, Warszawa. Date of manuscript acceptance: 17 October 1996 Chilina megastoma Hylton Scott, 1958 (Pulmonata: Basommatophora): a study on topotypic specimens Cristian F. Ituarte Department of Invertebrate Zoology, Museo de La Plata, 1900 La Plata, Buenos Aires, Argentina, cituarte@isis.unlp.edu.ar Abstract: This is a contribution to the knowledge of Chilina megastoma Hylton Scott, 1958 (Pulmonata: Basommatophora: Chilinidae), based on the study of topotypic specimens. This species is endemic to the Iguazu Falls, a frontier region between Argentina and Brazil. C. megastoma is known only from a brief original description, which was based exclusively on the shell morphology of a single specimen. The relevant results include the following: the radula is formed by 40-50 transverse rows of teeth in “chevron” arrangement; the central tooth is bicuspid with a reduced or rudimentary third cusp; the first lateral is tricuspid with a wide and short base, and the remaining laterals and marginals (40-44) each have four cusps and a long attachment base, which is bent in the marginals. The secondary or accessory seminal receptacle, in the female genital system, is strongly reduced and does not seems to be functional. The penis is as long as the penis sheath, and the penis sheath is two or three times longer than the prepuce. Key words: Chilina, morphology, topotypes, endemic species, Chilinidae The family Chilinidae comprises only one genus, Chilina Gray, 1828, which is distributed exclusively in South America (Castellanos and Gaillard, 1981; Castellanos and Miquel, 1991). About 17 Chilina species have been recognized as distributed in Argentina (Castellanos and Gaillard, 1981; Castellanos and Landoni, 1995). Previous knowledge on the soft-part anatomy of members of the genus mainly comes from the studies of Haeckel (1911) and Duncan (1960a, b; 1975). Recently, Valdovinos and Stuardo (1995), based upon study of the radula, soft-part anatomy, and shell morphology of the 36 species of the genus described from Chile, proposed a new genus, Archaeochilina, to include Chilina angusta (Philippi, 1860), and recommended the rearrangement of the remaining Chilean species in three new subgenera: Chilina s. s., Mesochilina, and Neochilina. The authors pro- posed the nervous system organization as providing the only reliable character to afford a basis for supraspecific arrangement. Argentine species of Chilina (16 species with three subspecies) have been studied by Hylton Scott (1958), Castellanos and Gaillard (1981), Castellanos and Miquel (1980), and Miquel (1984; 1986a, b). Chilina megastoma Hylton Scott, 1958, was described from the Iguazt Falls (25°35’ S, 54°35’ W), Misiones Province, Argentina. The species is known only from the type locality (Castellanos and Landoni, 1995). The original description was based on the shell morphology of a single specimen, thus, details on soft-part anatomy, espe- cially the nervous system, genital system, and radular fea- tures, are completely unknown. In the present paper, the results of the study on topotypic specimens of Chilina megastoma, focusing on its shell morphometry, radular ultrastructure, anatomy of the male and female genital systems, and nervous system, are given. In addition, C. fluminea (Maton, 1809), a species widely distributed in Argentina, was used for comparative analysis of several anatomical features. MATERIALS AND METHODS Two samples of Chilina megastoma were obtained, one from the vertical cliff behind the Salto Dos Hermanas waterfall on the Argentine side of the Iguazu Falls, Misiones, Argentina, and the other from a vertical cliff per- manently swept by a small waterfall at the Brazilian side (the Iguazt River is in the northeastern part of Argentina, and it constitutes the natural geographical border with Brazil at that site). Two other samples were collected from Salto San Martin at the Iguazu Falls and from a small waterfall at the Arrechea Rivulet, the latter 2 km away from the type locality. American Malacological Bulletin, Vol. 14(1) (1997):9-15 9 10 AMER. MALAC. BULL. 14(1) (1997) The specimens were fixed in 10% formalin after having been partly relaxed by immersion in warm water (55°C) for a few minutes. After relaxation and fixation in formalin, the specimens used for anatomical studies were treated and preserved in a modified Raillet-Henry’s solu- tion (distilled water 93%, acetic acid 2%, formalin 5%). The radulae were dissected from the buccal mass, then treated with 10% sodium hydroxide (12 hrs), rinsed in dis- tilled water and properly mounted and coated with gold for scanning electronic microscopic observation. The type specimen of Chilina megastoma, deposit- ed at the Museo de la Plata (MLP, without registration number), has also been studied. Specimens of C. fluminea used for comparative purposes were collected from the sandy shore at the Rio de La Plata River and from Miguelin Rivulet, both at Punta Lara, Ensenada, Buenos Aires. Shell measurements (total shell length, last whorl length, major and minor diameter of the shell aperture) were taken under a stereoscopic microscope provided with a micrometer eyepiece. Voucher specimens were deposited at the malaco- logical collection at Museo de La Plata (MLP nos. 5098, 5099, 5128, 5129, 5130, 5246, and 5261). RESULTS SHELL: The diagnostic features of Chilina megas- toma are shape and development of the spire (extremely low) and the great development of the last whorl of the shell, resulting in a large aperture of the shell (Fig. 1). The ratio of last whorl to shell length ranged from 0.92-0.97 (mean 0.95, standard deviation 0.01). The aperture is oval; minor diameter to major diameter ratio ranged from 0.53- 0.87 (mean 0.67, standard deviation 0.08). The shell is yel- lowish, olivaceous, or in some cases dark brown. The sur- face is smooth, only sulcated with weak radial periostracal striae and helicoidal stripes, particularly evident near the suture, which is deeply marked (Fig.2). The columellar bor- der is straight or slightly curved, somewhat concave, with a marked fold at its upper end (usually referred to as the col- umellar tooth). The parietal area has a thin white callus. A weak fold of the parietal area forms an indistinct lamella, also referred to in the literature as a “parietal tooth” (Hylton Scott, 1958; Castellanos and Gaillard, 1981), lying slightly over the columellar fold (Fig. 3). The outer lip is sharp and evenly curved. DIGESTIVE SYSTEM: The general morphology of the digestive system follows the pattern already known in the genus (Haeckel, 1911; Harry, 1964). The posterior end of the radular sac exceeds the posterior end of the buc- cal bulb, forming a short tube slightly bent dorsally. The salivary glands are not fused at their posterior end. The radula is composed of approximately 40 sharply oblique rows of teeth (Fig. 4), that is, the tooth rows are arranged “in chevron.” The angle between the right and left halves of each row is about 90°. The central tooth is asymmetrical, bicuspid, with an additional third cusp, greatly reduced, and partly overlapping the major cusp (Fig. 5). The attachment base of the central tooth is elongated, deltoid in shape, with the proximal end concave. Figs. 1-3. Shell of Chilina megastoma Hylton Scott, 1958. Fig. 1. Apertural view (MLP 5098). Fig. 2. Dorsal view, showing the periostracal sculpture (MLP 5098). Fig. 3. Detail of the columellar area, columellar fold, and parietal lamella (MLP 5099). Scale bars = 2 mm. ITUARTE: CHILINA MEGASTOMA 11 Figs. 4-11. Radula of Chilina megastoma Hylton Scott, 1958. Fig. 4. General view showing the typical arrangement of teeth. Fig. 5. Central tooth and first lateral teeth. Figs. 6, 8, 10. Lateral teeth (from central toward marginals). Figs. 7, 9, 11. Details for Figs. 6, 8, 10, respectively. Scale bars = 10 pm (Fig. 5), 25 pm (Figs. 7, 9, 11), 50 pm (Figs. 6, 8, 10), 250 pm (Fig. 4). The major cusp is relatively short, cylindrical, and pointed at the tip. The minor cusps are very short and thorn-like. The first lateral tooth is tricuspid, with a short, wide, and strong base. The mesocone has a short dagger-like shape, somewhat directed toward the central tooth. The entocone and ectocone are short, triangular, and nearly symmetrical. The entocone is somewhat larger than the ectocone (Figs. 5- 6). The second lateral tooth is also tricuspid, but with a slender, longer, and slightly asymmetric attachment base. The mesocone is the largest cusp, and the ectocone marked- ly smaller than the entocone (Fig. 6). From the third lateral tooth to the marginal teeth, the morphology of the teeth is relatively similar. The free plate of the teeth is palmate with four cusps; the attachment base is slender, and it joins the free plate at a point slightly shifted toward the minor cusps. In the last marginal teeth, the attachment base becomes slender, obliquely inserted, and slightly bent (Figs. 7-11). REPRODUCTIVE SYSTEM: Regarding the gen- eral aspects, the features of the female and male genital sys- tems are coincident with those previously described for the genus (Haeckel, 1911; Duncan, 1960a, b, 1975). Thus, only the diagnostic characteristics for Chilina megastoma will be described. Female genital system: The seminal receptacle or spermatheca is oval, located at the left side of the visceral mass, lying just below the ventricle. A long duct runs trans- versely across the visceral mass, passing over the utero- vaginal complex, to connect the seminal receptacle with the distal end of the vagina just before the female genital open- ing at the right side of the body (Figs. 12-13). Another duct arises posteriorly, just between the vagina and the uterus, —————| Fig. 12-13. Chilina megastoma Hylton Scott, 1958. Fig. 12. Lateral view of a partially dissected specimen showing the terminal portion of male and female genital systems, and the deferent duct (arrow heads) dissected from the body wall (MLP 5210). Fig. 13. Details of the terminal female genital system showing the spermathecal duct crossing over the vagina to open at its distal end (MLP 5210). (cg, cerebral ganglion; dd, deferent duct; f, foot, fp, female pore; mp, male pore; p, penis; sd, spermathecal duct; sg, salivary gland; v, vagina). 12 AMER. MALAC. BULL. 14(1) (1997) called a secondary or accessory seminal receptacle. This duct runs closely adhered to the wall of the uterus, and it is barely enlarged at the tip to form a nearly indistinguishable, rounded ampulla. Male genital system: The deferent duct arises as a separate male genital duct only after traversing the prostate, Ib rb 17 Figs. 14-17. Chilina megastoma Hylton Scott, 1958. Genital and nervous systems. Fig. 14. Camera lucida drawing of the distal portion of the female and male genital systems. Fig. 15. Camera lucida drawing with details of the penial complex. Fig. 16. Semi-diagramatic scheme of the anterior and posterior nerve rings. Fig. 17. Semi-diagramatic scheme of the innervation of the buccal mass. Scale bars = 1 mm. (dd, deferent duct; Ib, left buccal ganglion; Icg, left cerebral ganglion; Ipe, left pedal ganglion; Ipg, left parietal ganglion; Iplg, left pleural ganglion; Ir, lateral penis retractor; pp, prepuce; pr, penis retractor; ps, penis sheath; rb, right buccal ganglion, tcg, right cerebal ganglion; rpe, right pedal ganglion; rpg, right parietal ganglion; rplg, right pleural ganglion; sig, subintestinal ganglion, v, vagi- na; vg, visceral ganglion). running below the uterus and the vagina. Near the distal end of the vagina the deferent duct emerges from the haemocoelic space, sometimes forming a very small loop, entering into the muscular body wall (Figs. 12-13). The dis- tal portion of the deferent duct is very sinuous running along the right side of the body wall (Figs. 12-14), close to the surface and immediately below the external reproduc- tive groove which, arising just at side of the female genital pore, runs toward the base of the right tentacle close to the male genital system opening. At this point, the deferent duct diverges from the body wall coming into the cephalo- pedal haemocoel and turning posteriorly toward the posteri- or end of the penis sheath, the site at which it enters the penis (Fig. 15). The penis sheath is a slender and nearly cylindrical tube (Fig. 14). The penis is slender, as long as the penis sheath; its surface is rugose due to the presence of deep transverse furrows and padded ridges or short lamel- lae. At the anterior half of the penis, these ridges are more or less longitudinally arranged. The prepuce is a somewhat triangular or cordiform structure with thick muscular walls (Figs. 14-15). The prepuce is shorter than the penis sheath; its length is half or a third of that of the penis sheath. A powerful penis sheath retractor muscle arises from the pos- terior end of the penis sheath (Fig. 15), and it is attached at the columellar muscle. Another series of four or five mus- cular bundles are attached to the lateral wall of the prepuce. Two long muscular bundles are joined to the frontal wall of the prepuce running on each side, and connecting with the penis sheath retractor. The contraction of these two muscles determines the erection of the penial complex (Figs. 14-15). The deferent duct runs over the right-most muscular bun- dle, toward the proximal end of the penis (Figs. 14-15). NERVOUS SYSTEM: The cerebral and pedal gan- glia are joined by commissures and connectives to form the anterior nerve ring located at the anterior half of the buccal mass, just at or a little behind the origin of the esophagus (Figs. 16, 18-19). However, the location of the anterior nerve ring showed considerable variation according to the degree of retraction of the buccal mass, but it was always located at the anterior half of the buccal mass. The right pleural ganglion is joined by somewhat short connectives to both cerebral and pedal ganglia. A relatively long connec- tive joins the right pleural ganglion to the right parietal gan- glion (supraesophageal ganglion of Fretter, 1975), which gives off a large nerve that supplies the osphradium (Fig. 19), and a second very long and thin connective that runs to the visceral ganglion (also called abdominal [Fretter, 1975] or innominate [Harry, 1964]), at the end of the posterior nerve ring. On the left side of the posterior nerve ring, two short connectives join the left cerebral and pedal ganglia with the pleural ganglion, from which a relatively short ITUARTE: CHIILINA MEGASTOMA 13 Figs. 18-20. Chilina megastoma Hylton Scott, 1958 (MLP 5210). Fig. 18. Dorsal view of a dissected specimen to show the anterior and posterior nerve rings (see Fig. 16 for reference). Fig. 19. Lateral view showing the location of the cerebral commissure. Posteriorly, the cerebral ganglion, the right pleural ganglion (arrowhead), the right parietal ganglion and the pari- etovisceral connective, are shown. Fig. 20. View of the mantle roof show- ing the location of the osphradium close to the renal pore at the right end of kidney. Scale bars = 1 mm. (cc, cerebral commissure; cg, cerebral gan- glion; k, kidney; os, osphradium; pg, parietal ganglion; pvc, parietoviscer- al connective; rp, renal pore). connective arises and runs toward the left parietal ganglion. A long connective joins the left parietal ganglion to the subintestinal ganglion, which lies over the columellar mus- Cle, at its posterior half. A somewhat large nerve arises at a point approximately 2/3 the length of the left parietal- subintestinal connective. There is a slight swelling just at the site where that nerve originates; however, it might not Figs. 21-24. Chilina fluminea (Maton, 1809) (MLP 5246). Fig. 21. Dorsal view of a partially dissected specimen showing position of the cerebral commissure, and the long loops of the deferent duct (arrowheads). Fig. 22. Right lateral view. Fig. 23. Left lateral view. Fig. 24. Dorsal view of the nervous system. Scale bars = 1 mm. (cc, cerebral commissure; cg, cerebral ganglion; ddl, deferent duct loop; e, esophagus; fp, female pore; Ib, left buccal ganglion; Ipg, left parietal ganglion; Iplg, left pleural ganglion; p, penis; pg, parietal ganglion; rpg, right parietal ganglion; sd, spermathecal duct; sg, salivary gland; sig, subintestinal ganglion; st, spermatheca; u, uterus; v, vagina; vg, visceral ganglion). 14 AMER. MALAC. BULL. 14(1) (1997) definitely be considered a ganglion. Finally, a very short connective joins the subintestinal ganglion to the visceral one, closing the posterior nerve ring. From the latter, a large nerve arises, and shortly thereafter, it gives off two nerves which supply the visceral mass. From the subintesti- nal ganglion, a large nerve runs toward the right, passing through the columellar muscle to innervate the distal part of the vagina and the pneumostomal appendage. As is charac- teristic for the genus, the pleuro-visceral connectives show incomplete torsion (Figs. 16, 18). At each side of the buccal mass (dorsolaterally), far removed, and behind the origin of the esophagus, are the buccal ganglia. The ganglia are joined by a somewhat short commisure which passes below the esophagus. A nerve, arising from the middle of this commisure, sinks into the buccal mass shortly after its origin (Fig. 17). The buccal ganglia are connected to each cerebral ganglion by long connectives, forming an U-shaped open nerve ring. The osphradium, located on the roof of the mantle cavity, close to the renal pore, and at the anterior end of the pneu- mostome, is a flat oval-shaped organ formed by an slightly curved furrow, bordered by two elongated and inflated lips (Fig. 20). DISCUSSION From the present knowledge of the distribution of Chilina megastoma, the species seems to be endemic to a reduced number of environments closely related to high energy freshwater courses such as vertical cliffs and rocky walls permanently swept by winds and water trickles from nearby waterfalls. As reported for other gastropod species (Trussell et al., 1993), the extreme reduction of the spire, the globose shape of the last whorl and the wide aperture of the shell shown by C. megastoma represent adaptive responses to particular environmental conditions, such as dislodgement agents in high energy watercourses. Hylton Scott (1958), and Castellanos and Gaillard (1981) considered that Chilina megastoma approaches C. fluminea in shell morphology, pointing to the lack of periostracal sculpture as a differential character in the latter. With respect to the anatomy of soft parts, several differences should be pointed out. The general features of the genital system of Chilina megastoma are coincident with those described by Miquel (1984) for C. fluminea and by Harry (1964) for C. fluctuosa Gray, 1828. The following differences have been observed: in C. megastoma the duct of the seminal receptacle or spermatheca crosses, from left to right, above the lower portion of the female genital com- plex, just below the floor of the pallial cavity, while in C. fluminea, the duct passes over the anterior border of the distal part of the spermoviduct and prostate, and then turns downward, just at the point where the uterus connects with the vagina (Fig. 21). Thereafter, to traverse the vagina ven- trally, the duct reappears dorsally and immediately opens near the tip of the vagina. The same features (Fig. 16) have been illustrated by Harry (1964) for C. fluctuosa from Chile. Moreover, in C. megastoma, the accessory or sec- ondary seminal receptacle is barely enlarged at the tip, while in C. fluminea the tip of this organ is greatly enlarged into a large pear-shaped bulb. This difference in shape and size does not seem to be related to organ physiology. In the male portion of the genital system, several differences among species should be pointed out: while in Chilina megastoma the segment of the deferent duct which runs into the muscular body wall just below the genital groove is markedly sinuous (Figs. 12, 14), in C. fluminea it is nearly always straight or faintly sinuous (Fig. 22). In addition, in C. fluminea, when the deferent duct reenters into the haemocoelic space toward the penis, it develops two or three large loops prior to running along the penial retractor muscle and entering the penis (Figs. 21-22). Unlike this, in C. megastoma the deferent duct is strictly as long as the genital organs it accompanies (Figs. 12, 14-15). According to Valdovinos and Stuardo (1995), the nervous system is the only reliable feature to differentiate Chilean species and to build suitable supraspecific arrange- ments. According to Haeckel (1911), the structural pattern of the posterior nerve ring seems to be a specific character- istic, especially with respect to the number of ganglia and the length of the connectives between the left pleural and subintestinal ganglia. The general pattern of the nervous system of Chilina megastoma, here studied, does not differ from that described by Harry (1964) for C. fluctuosa. However, Harry (1964) was unable to find the left parietal ganglion already described by Haeckel (1911). In the present study, the left parietal ganglion was easily identified in both C. megastoma and C. fluminea (Figs. 18, 23-24). In C. megas- toma, the lengths of the left cerebropleural connective and left pleuroparietal connectives were similar. The left pari- etal-subintestinal connective was approximately three times as long the pleuroparietal connective (Fig. 18). In C. flu- minea the left pleural and parietal ganglia are closer than they are in C. megastoma, being, in some cases, nearly con- tiguous (Fig. 24). In C. megastoma the anterior nerve ring tends to be located more posteriorly than in C. fluminea, in which it has always been observed at the anterior end of the buccal bulb (Figs. 21-23). As in Chilina fluminea, the posterior end of the radular sac of C. megastoma is projected from the buccal mass, forming a short cylindrical tube bent slightly dorsal- ly. This fact is in contrast with the description given by Harry (1964) for C. fluctuosa from Chile. Several details in morphology of the radular teeth ITUARTE: CHIILINA MEGASTOMA 15 Figs. 25-26. Radula of Chilina fluminea (Maton, 1809). Fig. 25. Details of central and first lateral teeth. Fig. 26. Lateral teeth. Scale bars = 25 pm (Fig. 25), 50 pm (Fig. 26). seem to differ consistently among species. In C. megastoma, the central tooth has a definite triangular, elon- gated base, while in C. fluminea the attachment base is shorter and stronger. The lateral and marginal teeth of C. megastoma, have more slender attachment bases and their cusps are longer, with a more definite dagger-like shape than those of in C. fluminea (Figs. 25-26). In addition, the last marginal tooth of each row in C. megastoma has four cusps (Figs. 10-11), while in C. fluminea it usually has six or seven cusps (Castellanos and Gaillard, 1981). The num- ber of tooth rows also differs: 40 in C. megastoma and between 60 and 65 in C. fluminea (Castellanos and Gaillard, 1981). LITERATURE CITED Castellanos, Z. J. A. and M. C. Gaillard. 1981. Mollusca, Gasteropoda, Chilinidae. Fauna de Agua Dulce de la Republica Argentina 15(4):20-51. Castellanos, Z. J. A. and N. A. Landoni. 1995. Mollusca Pelecypoda y Gastropoda. In: Ecosistemas de Aguas Continentales, Metodologias Para su Estudio, T. 2, E. C. Lopretto and G. Tell, eds. pp. 759-801. Ediciones Sur, La Plata, Argentina. Castellanos, Z. J. A. and S. E. Miquel. 1980. Notas complementarias al género Chilina Gray (Mollusca, Pulmonata). Neotropica 26(76): 171-178. Castellanos, Z. J. A. and S. E. Miquel. 1991. Distribucion de los Pulmonata Basommatophora. Fauna de Agua Dulce de la Republica Argentina 15(9):1-11. Duncan, C. J. 1960a. The evolution of the pulmonate genital system. Proceedings of the Zoological Society of London 134:601-609. Duncan, C. J. 1960b. The genital systems of the freshwater Basommatophora. Proceedings of the Zoological Society of London 135:339-355. Duncan, C. J. 1975. Reproduction. Jn: Pulmonates, Vol. 1. Functional Anatomy and Physiology, V. Fretter and J. Peake, eds. pp. 309- 365. Academic Press, London. Fretter, V. 1975. Introduction. Jn: Pulmonates, Vol. 1. Functional Anatomy and Physiology, V. Fretter and J. Peake, eds. pp. xi-xx1x. Academic Press, London. Harry, W. H. 1964. The anatomy of Chilina fluctuosa Gray reexamined, with prolegomena on the phylogeny of the higher limnic Basommatophora (Gastropoda: Pulmonata). Malacologia 1(3):355-385. Haeckel, W. 1911. Beitrage zur Anatomie der Gattung Chilina. Zoologische Jahrbiicher, supplement, 13(4):89- 136. Hylton Scott, M. I. 1958. Nueva especie de Chilina del norte argentino (Moll. Pulm. Basommatophora). Neotropica 4(13):26-27. Miquel, S. E. 1984. Contribucién al Conocimiento Biolégico de Gasterépodos Pulmonados del Area Rioplatense, con Especial Referencia a Chilina fluminea (Maton). Doctoral Thesis, Facultad de Ciencias Naturales y Museo, Universidad Nacional de La Plata. 133 pp. Miquel, S. E. 1986a. El ciclo de vida y la evolucién gonadal de Chilina fluminea fluminea (Maton, 1809) (Gastropoda Basommatophora Chilinidae). Neotropica 32(87):23-34. Miquel, S. E. 1986b. Tipos celulares del tracto genital de Chilina fluminea fluminea (Maton, 1809) (Gastropoda Basommatophora Chilinidae). Neotropica 32(88): 133-138. Trussell, G. C., A. S. Johnson, S. G. Rudolph, and E. S. Gilfillan. 1993. Resistance to dislodgement: habitat and size-specific differences in morphology and tenacity in an intertidal snail. Marine Ecology Progress Series 100:135-144. Valdovinos, C. and J. Stuardo. 1995. Morfologia funcional de Chilina angusta (Philippi, 1860), y evolucion de Chilinidae. Resumos, II Congresso Latino-Americano de Malacologia, Porto Alegre, Brasil:43. Date of manuscript acceptance: 18 June 1997 The role of prey mobility in the population ecology of the nudibranch Cuthona nana (Gastropoda: Opisthobranchia) Nadine C. Folino* Department of Zoology, University of New Hampshire, Durham, New Hampshire 03824 U.S. A. Abstract: The aeolid nudibranch Cuthona nana (Alder and Hancock, 1842) was studied in relation to its specific prey, the hydroid, Hydractinia polycli- na (Agassiz, 1862). In this predator-prey association, the predator’s spatial patterns and behavior, along with prey mobility, could play an important role in maintaining this nudibranch population. The seasonal abundances of C. nana documented in this study agree with the sub-annual life cycle previously pre- sented. Both adult and juvenile nudibranchs were present on colonies during most months sampled in 1986-1988. The low prevalence of nudibranchs on hydroid colonies, the behavior of adult C. nana (leaving colonies periodically), regeneration in the hydroid, and prey mobility appear to be crucial in main- taining this unique species-specific, predator-prey association. Most nudibranch prey are non-mobile while H. polyclina grows on gastropod shells occupied by hermit crabs of the genus Pagurus. Adult C. nana repeatedly leave the hydroid colonies both in the field and laboratory to lay egg masses, while juve- niles spend extended periods of time on the colony and leave only when they approach sexual maturity. The adult behavior of leaving colonies to lay egg masses does not severely jeopardize the newly hatched nudibranchs’ probability of finding food. Hermit crab mobility is high and within a 24-hr period crabs frequently pass a given area containing juvenile nudibranchs. Juveniles encounter the hydroid’s streaming gastrozooids sweeping over the substratum. C. nana undergoes non-pelagic lecithotrophic development from yolk-rich eggs with individuals hatching as crawling juveniles. The predator-prey dynam- ics between C. nana and mobile colonies of H. polyclina are similar to those seen in host-parasite associations. Key words: Hydroidea, Nudibranchia, population ecology, predator-prey, Aeolidoidea, Hydractinia Few studies describe nudibranch predator popula- the shells occupied by P. arcuatus (Squires, 1964). No stud- tion parameters in relationship to their prey (Potts, 1970; ies have considered how hermit crab movement affects the Harris, 1973; Todd, 1979, 1981, 1983). Most nudibranch accessibility of predators to epifaunal prey organisms on population studies show seasonal population fluctuations their shells. Knowing the probability of a hermit crab pass- and suggest that physical (temperature, wave action) or bio- ing a given area on the benthos would provide information logical factors (prey availability, competition, predation) on prey availability for a predator and how that could control such fluctuations (reviews: Harris, 1973; Todd, impact the population ecology. 1981, 1983). Aeolid nudibranchs (Gastropoda, The aeolid nudibranch Cuthona nana (Alder and Opisthobranchia) often are abundant in seasonal hydroid Hancock, 1842) (Aeolidoidea) is a hermaphroditic opistho- communities; the communities are sessile and seasonal due branch that feeds specifically and exclusively on to heavy grazing by predators, or because of changes in Hydractinia polyclina. Sexually mature C. nana leave crab temperature (Miller, 1961; Thompson, 1964; Fager, 1971; shells bearing hydroid colonies to mate and lay egg masses Clark, 1975; MacLeod and Valiela, 1975; Nybakken, 1978; (Rivest, 1978; Folino, 1993). Relocating a colony for fur- Harris and Irons, 1982). ther feeding is a unique challenge for C. nana compared to Unlike most hydroids, the colonial gymnoblastic other hydroid feeders. Non-pelagic lecithotrophic larvae hydroid Hydractinia polyclina (Agassiz, 1862) (formally hatch and are picked up by the gastrozooids (feeding known as H. echinata Fleming, 1828; Buss and Yund, polyps) of the passing hydroid colony (Rivest, 1978). The 1989) is a persistent food source throughout the year and is movement of the hermit crab is therefore important in a mobile prey. In the southern Gulf of Maine, H. polyclina bringing the prey, H. polyclina, to the slow-moving preda- grows primarily on shells occupied by the hermit crab tor, C. nana. Pagurus acadianus (Benedict, 1901) and less frequently on This study documents the localized dynamics of Cuthona nana and its prey off of the coast of Maine and *Present Address: Biology Department, Wheaton College, Wheaton, expands on shorter previous studies by Rivest (1978) and Illinois 60187-5593 U. S. A., Nadine.C.Folino @wheaton.edu Folino (1985). The purpose of this study was three-fold: (1) American Malacological Bulletin, Vol. 14(1) (1997):17-25 jh 18 AMER. MALAC. BULL. 14(1) (1997) to describe the population structure of C. nana by docu- menting nudibranch densities, distributions, and size frequencies on shells with colonies from May 1986 to May 1988 at Gosport Harbor, Maine; (2) to compare juvenile and adult movements on and off Hydractinia polyclina colonies; i. e. how this behavior affects the location of mobile prey on hermit crab shells; and (3) to estimate prey accessibility for C. nana by determining the degree of movement by hermit crabs with hydroids on their shells within a given area. Does crab mobility have an impact on prey availability which could regulate C. nana popula- tions? METHODOLOGY Collection of animals Specimens of Cuthona nana and colonies of Hydractinia polyclina were collected using SCUBA from Gosport Harbor (Haley Cove, Isles of Shoals, Maine), ca. 9.5 km off the New Hampshire coast (42°59’ N; 70°36’ W). Hermit crabs occurred mainly in the sandy portions of the harbor at depths of 5-10 m. Population data for Cuthona nana was obtained from 23 monthly collections (May 1986 to May 1988) of 21 to 90 hermit crabs, each with colonies covering ca. 100% of the avaiable shell surface. The coverage is approx- imate because the shell scrapes the bottom when the crab crawls, preventing colony growth. Hydroid-covered shells were placed in individual containers in the field to insure that nudibranchs remained on their original colonies. The containers were either plastic jars (125 ml) or mesh con- tainers (Toby Tea-boys, mesh size ca. 164 um; Daniel Peikin Company, Silver Spring, Maryland). Although efforts were made to collect shells of various sizes with Hydractinia polyclina, larger shells were more visible, cre- ating a sampling bias towards larger shells. Colonies were later examined for nudibranchs using a dissecting micro- scope. Nudibranch population structure For each hydroid-covered shell (therefore for each hydroid colony) the density of nudibranchs and their sizes were recorded. The number of Cuthona nana was used to determine the abundance of nudibranchs per colony. Indices of dispersion were calculated to determine the degree of aggregation of nudibranchs on colonies. The dis- persion pattern was further analyzed by making compar- isons to a negative binomial distribution (Ludwig and Reynolds, 1988; Krebs, 1989). Co-occurrence of nudi- branchs was determined by scoring the size of each nudi- branch collected on a Hydractinia-covered shell versus the number of conspecifics present on the same colony. Individuals were scored as being alone, paired, or with three or more nudibranchs. The nudibranchs were divided into non-reproductive (< 9 mm) or reproductive (> 9 mm) groups based on anatomical and behavioral aspects of reproduction (Rivest, 1978; Folino, 1993). To determine if colony size affected the number of nudibranchs present on a colony, correlation coefficients were calculated for nudibranch number and colony size for each month sampled. The surface area of a given shell was determined using its linear dimensions following a tech- nique modified (see Folino, 1993) from that used by Shenk and Karlson (1986). Colony size was then estimated from shell surface area. In months when there were small num- bers of colonies with nudibranchs present, the data were pooled. Histograms of nudibranch length frequencies were used to examine the age-class structure by month and were compared for yearly differences using Kolmogorov- Smirnov tests (Sokal and Rohlf, 1981). Monthly size-class histograms (See Folino, 1989) were summarized by group- ing individuals into non-reproductive individuals (< 9 mm) and reproductive individuals (> 9 mm) as previously men- tioned. Measuring nudibranch movement A laboratory experiment was designed to determine differences between juvenile and adult Cuthona nana in movement on and off hydroid-colonized hermit crab shells. Hermit crabs were maintained in trays (76 x 64 x 9 cm) filled with sand from Gosport Harbor. Beetags (from Chr. Graze KG, West Germany) were used to label shells so that individual colonies could be assessed for the presence or absence of nudibranchs. Because nudibranchs are difficult to tag, tagging the shells was a way of monitoring nudi- branch movement (i. e. whether a nudibranch was present or absent since the previous observation). Each tray con- tained ten tagged hermit crabs and four nudibranchs. Tray densities were chosen from the highest field densities recorded from cofferdam samples taken in May 1987. [A cofferdam is a metal cylinder (0.153 m2) placed on the ben- thos to prevent hermit crabs in the enclosed area from escaping before being counted. ] Two trials with two replicate trays were conducted using adult nudibranchs of 12-20 mm. Each adult trial last- ed 12 days, and the shells were examined twice daily for the presence or absence of nudibranchs, once in early morning and once in late afternoon to account for night and day activity. Nudibranch movement was measured by cal- culating the mean number of moves (or change in nudi- branch number) on or off a colony per day. A similar exper- iment was conducted using juvenile (2-4 mm) nudibranchs. One trial of juveniles consisting of four replicated trays was conducted for 21 days. FOLINO: NUDIBRANCH-PREY ASSOCIATION 19 Measuring hermit crab mobility Hermit crab mobility was estimated by deploying pitfall traps (Uetz and Unzicker, 1976) on the bottom of Gosport Harbor to sample crabs passing a given area in a 24-hr period. The traps were plastic containers ca. 11 cm in diameter and 15 cm deep. Two grids (each measuring 5 x 5 m) were used to randomly position 20 pitfall traps; holes for the containers were dug in the sand using an airlift. Each container was marked with a numbered flag to insure relocation. Trials were performed monthly from January through May 1988; April was excluded due to rough seas. A given trial consisted of leaving the traps uncovered for 24 h from mid-morning to mid-morning. Traps were emptied and crabs from a given trap were placed in a mesh bag. The number of crabs caught per trap and the presence or absence of Hydractinia polyclina colonies on the shells of each crab were recorded. Traps were covered between tri- als; a few traps became filled with sand and could not be relocated during two of the four trials. RESULTS Nudibranch densities Monthly collections of crabs with hydroid-covered shells provided seasonal estimates for the population of Cuthona nana at Gosport Harbor. The mean number of nudibranchs per colony collected from May 1986 to May 1988 demonstrated seasonality of densities, with maxima in April, May, and August, and minima in October and November (Fig. 1). The greatest mean number per shell occurred in May 1987 (2.380; SE = 0.386) and most of those individuals (87%) were < 4 mm in length. The percentage of hydroid-covered shells with one Mean Number / Shell MY JU _JL_AG SP_OT NV OC JN FB MR ARMY JN JL AG ST OT NV OC JN FB MR ARMY 1986 1987 1988 Fig. 1. The mean number of Cuthona nana per hydroid-covered hermit crab shell for each month sampled from May 1986 to May 1988. The bar for each mean represents standard error. —s— % Shells ---@--- Temperature Temperature °C % Shells with Cuthona nana t) My JU JL_AG ST OT NV _OC JN FB MR ARMY JN JL AG SP OT NV_OC JN FB MR ARMY 1986 1987 1988 Fig. 2. The percentage of hermit crab shells covered with Hydractinia polyclina having one or more Cuthona nana present, plotted with tempera- ture. or more Cuthona nana fluctuated over the 23 mo sampled. The greatest percentages of occupied shells for 1986 were in July and August with 55% and 56%, respectively, while April and May showed the greatest percentage of colonies with nudibranchs in 1987 and 1988 (Fig. 2). The percentage of shells with nudibranchs declined in late summer and early fall in 1986 and 1987, and began to increase in November and December in both years. The percentage of shells with nudibranchs increased in the colder months sug- gesting an increase in population numbers (Fig. 2). Nudibranch distributions Nudibranchs demonstrated an aggregated rather than random dispersion pattern on colonies for all months sampled; tests could not be performed for October, November, and December 1986, and November 1987 because of small sample sizes (Table 1). All indices of dis- persion (variance/mean number of nudibranchs per colony ratios) were greater than 1.0, providing evidence for aggre- gated distributions (Krebs, 1989) (Table 1). A large number of colonies had no nudibranchs, and the ranges of distribu- tion varied by month. In May 1987 and April 1988 a greater range of frequencies was observed; some colonies had as many as 11-19 nudibranchs per colony. Eight out of nine- teen months fit the negative binomial distribution indicating strong patterns of aggregation during those months (Table 1). Numbers of nudibranchs and colony size were not significantly correlated for any of the years sampled (1986: r = -0.080, N = 84; 1987: r = 0.068, N = 241; 1988: r = 0.015, N = 128, p > 0.50 in all cases). Correlation coeffi- cients were also determined for shell size (i. e. colony size for completely covered shells) and the number of non- reproductive nudibranchs to test the assumption that larger colonies acquired more juvenile nudibranchs (from the ben- thos). Again, no significant correlations were obtained 20 AMER. MALAC. BULL. 14(1) (1997) Table 1. The indices of dispersion (variance:mean ratios) and the negative binomial distribution statistics for the number of hydroid-covered shells with 0-19 Cuthona nana from May 1986 to May 1988. (k, expo- nent of the negative binomial distribution; N, number of sample units or hermit crabs; P, probability for the calculated X? values of the negative binomial distribution; S* / X, variance to mean ratio; X + SE, mean plus or minus standard error; X?, goodness-of-fit of the negative binomial distribution; * , Significant fit to the negative binomial distribution). MONTH S*/ X N X+SE k X2 P (df =n - 3) 1986 MAY 1.50 28 0.607 + 0.181 0.73 1.45 P < 0.100 (1) JUL 1.25 38 0.816 + 0.164 16.0 9.17 P <0.010 (2)* AUG 1.23 45 0.978 + 0.164 2.0 0.91 P < 0.250 (2) SEP 1.48 84 0.262 + 0.068 0.37 1.54 P <0.100 (1) OCT 1.56 68 0.132 + 0.055 --- NOV 1.00 33 0.030 + 0.030 --- DEC 1.46 61 0.164 + 0.063 --- 1987 FEB 1.55 73 0.247 + 0.072 0.30 1.28 P < 0.250 (1) MAR 3.97 52 0.519 + 0.199 0.21 13.17 P < 0.025 (6)* APR 1.76 63 1.060 + 0.172 2.00 19.49 P <0.001 (5)* MAY 5.64 90 2.380 + 0.386 0.54 19.79 P < 0.100 (16) JUN 1.83 63 0.683 v 0.141 0.79 1.97 P < 0.250 (2) JUL 3.03 87 0.759 + 0.163 0.44 8.44 P <0.100 (6) AUG 2.96 65 0.477 + 0.147 0.28 14.70 P <0.010 (6)* SEP 1.34 98 0.337 + 0.068 1.10 2.94 P <0.100 (2) OCT 1.47 83 0.217 + 0.062 0.62 0.25 P < 0.500 (1) NOV 1.04 44 0.295 + 0.083 --- --- DEC 1.40 21 0.476 + 0.178 1.30 0.88 P < 0.250 (1) 1988 JAN 1.26 56 0.429 + 0.098 1.10 0.695 P < 0.250 (1) FEB 2.38 59 0.661 + 0.163 0.71 18.56 P < 0.001 (5)* MAR 1.77 56 0.714 + 0.150 0.93 6.70 P < 0.005 (1)* APR 3.62 57 1.140 + 0.269 1.34 137.8 P <0.001 (9)* MAY 9.33 63 0.921 + 0.181 1.76 167.7 P<0.001 (7)* (1986: r = -0.098, N = 60; 1987: r = 0.086, N = 160; 1988: r =-0.010, N = 84). Chi-square analyses indicated no significant pattern for the distribution of reproductive adults being alone, paired, or with three or more individuals on a given colony (X? = 4.41, P < 0.111, df = 2, N = 95). Of the sexually mature animals on shells, approximately equal numbers were found alone, paired, or with three or more nudi- branchs. The > 3 category showed the lowest percentage (Fig. 3). On the other hand, non-reproductive nudibranchs showed a significant pattern of aggregation (ee = 92.88, P < 0.0001, df = 2, N = 768), with 49% of the animals occur- ring in groups of three or more on a colony. Significant dif- ferences existed in the three categories between the repro- ductive and non-reproductive individuals, suggesting behavioral differences between juvenile and adult nudi- branchs (G-test, p < 0.001). Size Frequencies Mean size of Cuthona nana on colonies varied for each month over the 2.5 yr sampling period (Fig. 4). Mean size increased from July to October for both 1986 and 1987. In all three years, more non-reproductive than repro- ductive individuals were present each month (Fig. 5). Reproductive adults were present on colonies in all months for 1986 except July and November, and were absent in August, November, and December 1987 (Fig. 5). This does not mean that adults were absent from Gosport Harbor, but they were not present on the colonies collected. Adults were present on colonies in all five of the months sampled in 1988. The percentage of adults in the summer months varied for 1986 and 1987 and decreased in late summer and early fall, followed by an increase in October for both years. Nudibranch movement on and off of colonies The results of nudibranch movement on and off of Hydractinia polyclina colonies indicated that adults were more active; they averaged one to two moves on or off crabs with colonies per day (Fig. 6). These numbers are most likely underestimates, because more excursions from colonies could have occurred within the time of census. Adults (12-20 mm) were more active than juveniles in both trials; often during the adult trials, animals were observed FOLINO: NUDIBRANCH-PREY ASSOCIATION 2A PERCENTAGE NUDIBRANCH DISTRIBUTION Fig. 3. The percentage of non-reproductive (1-9 mm) and reproductive (> 9 mm) Cuthona nana scored as solitary, paired, or with three or more indi- viduals on a hydroid-covered hermit crab shell from May 1986 to May 1988. Chi-square tests indicated a significant difference among the three categories for non-reproductive animals (P < 0.0001, df = 2) and non-sig- nificant differences for reproductive individuals. Length (mm) t+) My JU JL AG SP OT NV OC JN FB MR AR MY JN JL AG ST OT NV OC JN FB MR AR MY 1986 1987 1988 Fig. 4. Mean size of Cuthona nana on hydroid-covered hermit crab shells for each month of collection. The bar for each mean represents standard error. mating and laying egg masses on rocks and the sides of the sea water tables. No egg masses were laid on colonies of H. polyclina during the experiment. Juvenile nudibranchs (2-4 mm) did not leave shells covered with H. polyclina under the experimental conditions employed (Fig. 6). No juvenile nudibranchs left the colonies where they were initially placed in any of the four trays. Hermit crab mobility Considerable hermit crab activity occurred in a 24-h period in view of the fact that this experiment was conduct- ed during the colder months of the year. In 24-h, the mean number of crabs caught per pitfall trap ranged from 15-26 (Jan.: 15.4 + 10.2 SD, N = 308; Feb.: 21.3 + 11.0 SD, N = 383; Mar.: 15.0 + 11.0 SD, N = 251; May: 26.7 + 13.9 SD, N = 533). The mean number of crabs caught that were colo- nized with H. polyclina ranged from 3-8 (Fig. 7). On aver- age, six crabs with H. polyclina passed a given area during a 24-h period. This supports the probability that crabs with hydroids were likely to pass by a given nudibranch within a 24-h period. The number of shells with (and without) H. polyclina increased from March to May. The four months sampled were during the time of year with low crab and hydroid densities, thus providing conservative estimates (Grant, 1963; Rivest, 1978). DISCUSSION Nudibranch population patterns The population patterns of Cuthona nana at Gosport Harbor indicate a sub-annual life cycle during which the species undergoes several generations in a year (Miller, 1962; Thompson, 1964; Harris, 1973, 1975; Todd, 1981, 1983). The results of this study paralleled those of Rivest (1978) and Folino (1985) but also provided information on summer abundances during months when data had not been previously obtained. The presence of juveniles throughout all months sampled, in conjunction with adults present in the summer and fall and continuous egg-laying, indicates the existence of overlapping generations. Although most species of nudibranchs with several generations per year feed on seasonal prey (Miller, 1962; Clark, 1975; Harris, 1973; Todd, 1981, 1983), two species, Phestilla sp. and Cuthona nana, do not (Harris, 1975; Rivest, 1978, and Folino, 1989, respectively). Prey avail- ability for C. nana depends upon crab location; although most crabs migrate to deeper water during colder months, there are still crabs present in shallower water with colonies available for food during the winter (Rivest, 1978; Folino, 1989). Thus, the population of C. nana at Gosport Harbor is able to persist throughout the year due to the presence of crabs with colonies. Partial predation This nudibranch-hydroid association is an example of partial predation on colonial organisms, a phenomenon that has received increased attention in recent years for both plants and animals (Jackson, 1985; Coates and Jackson, 1985; Harper, 1985; Harvell and Suchanek, 1987; Todd and Havenhand, 1989). Hydractinia polyclina regen- erates when damaged by predators (Christensen, 1967; Sutherland and Karlson, 1977; Karlson, 1978; Buss et al., 1984; Folino, 1985; McFadden, 1986). Based on previous grazing rates (Folino, 1985, 1993), a large Cuthona nana could graze approximately one-quarter (23%) of an aver- age-sized colony in less than a 24-h period and leave a sub- stantial portion for continued growth. These estimates of oe AMER. MALAC. BULL. 14(1) (1997) Mi 9 mm 1986 s) 80 60 40 20 x 0 Als a. 4987 =z lu > 80 oOo Lu [e"4 uw 60 Re = oO 40 a lw a 20 ¥ | t) 1001 1988 FB MR AR MY JN JL AG ST OT NW MONTH Fig. 5. Percent frequencies of non-reproductive (1-9 mm) and reproduc- tive (> 9 mm) Cuthona nana collected from May 1986 to May 1988. (*, sampling not possible due to rough seas). grazing consider only one large nudibranch on a colony; obviously two or more large animals would do more dam- age. Even so, grazing does not completely decimate the prey as a food source. Furthermore, the majority of colonies collected did not support nudibranchs; the percent- age of colonies (or shells) with nudibranchs did not exceed 60% (Fig. 2). This again suggests that C. nana is not limit- ed by prey availability. Laboratory data on colony growth and regeneration are ambiguous. In small colonies, predator consumption rates (large Cuthona nana eat 200-500 polyps in a 24-h period at 12°C; Folino, 1993) can outstrip hydroid growth. However, growth rates also increase with temperature and colony size (McFadden et al., 1984; Folino, 1985) suggest- ing that larger colonies produce polyps at a rate closer to that of polyp removal by predators. The presence of chitinous spines on most colonies of Hydractinia polyclina in this study prevents complete removal of polyps by Cuthona nana (see Folino, 1993), a situation analogous to that seen in bryozoan zooids where spines reduce nudibranch grazing rates (Yoshioka, 1982; Harvell, 1984). Polyps that have been partially eaten can clearly regenerate (Folino, 1993). Therefore, C. nana do not decimate prey as is true in other nudibranch-hydroid associations (Clark, 1975; Todd, 1981, 1983). Nudibranch distributions and movement The aggregation of sexually immature Cuthona nana on Hydractinia polyclina (and lack of aggregation of reproductive individuals) differs from studies of other species in which aggregation occurs in both sexually mature and immature individuals (Miller, 1962; Clark, 1975; Todd, 1981, 1983). Potts’ (1970) work with Onchidoris bilamellata (Linné, 1767) [= O. fusca (Miiller, 1776)] suggested that the nudibranchs probably remain on the rock where they initially settled because ample food and mates are available. Fieid observations by Todd (1978a, 1979) for O. muricata (Miiller, 1776) and O. bil- amellata showed increased aggregation during the breeding season, suggesting that animals stay within an area where food and other sexually mature individuals are present. The distribution pattern seen in C. nana populations could be produced by the behavior of the nudibranchs. As juveniles, they hatch from eggs laid on rocks, mussel shells, and Chondrus (Rivest, 1978; Folino, 1993). Juveniles appear to be picked up by H. polyclina gastrozooids that sweep along the ocean bottom, and they remain on the colony until sex- ually mature. Adults leave hydroid colonies and follow mucus trails to find potential mates, as do other mollusks (Lowe and Turner, 1976; Todd, 1978b, 1979; Gerhart, 1986; see review: Hadfield and Switzer-Dunlap, 1984). During the peak reproductive periods for C. nana from April through September, Rivest (1978) and Folino (1989) — JUVENILES @ ADULTS 4 Mean movements/day Juvenile trial (1) Adult trials (2) Fig. 6. Nudibranch movement on and off of Hydractinia colonies. One trial with four replicates (left) indicating that juvenile Cuthona nana (2-4 mm) did not leave colonies during a 21-d period. Alternatively, adults (12- 20 mm) moved on and off an average of 1-2 times per day (two 12-d trials, right). The bar for each mean represents standard deviation. FOLINO: NUDIBRANCH-PREY ASSOCIATION 2 30 E) With H. polyclina 23 Hi Without H. polyciina Mean # of Crabs FEBRUARY MARCH MAY JANUARY Fig. 7. Mean number of hermit crabs caught in pitfall traps in a 24-h peri- od during four months in 1988. Twenty traps were sampled in January and May while 18 and 17 traps were sampled in February and March, respec- tively. The bar for each mean represents standard deviation. observed adult C. nana more frequently crawling on the bottom of Gosport Harbor. Mating does occur on the hydroid colonies, but occurs more often off of them (Harris et al., 1975; Rivest, 1978; Folino, 1985, 1989). The results of the laboratory nudibranch movement experiment support these field observations of nudibranch behavior (Fig. 6). C. nana behavior differs from those of the dorid nudibranchs, Onchidoris spp., where encounters with mates are enhanced due to aggregation near stationary prey (Potts, 1970; Todd, 1978a, 1979). Movement of adult C. nana off of its mobile prey increases the chances of encountering mates. The life history of Cuthona nana differs from those of other hydroid-feeding aeolids. In most species, larvae are planktonic, and settle on sessile prey for growth through sexual reproduction (Todd, 1981, 1983). In contrast, C. nana at Gosport Harbor lacks a planktonic veliger and exploits a mobile prey. Once picked up by a passing crab, juveniles feed on basal mat tissue until they reach ca. 5-6 mm, when they are large enough to consume polyps (Folino, 1993). Furthermore, crab mobility could help dis- tribute juvenile nudibranchs over several colonies prevent- ing over-predation of some (especially small) colonies (Rivest, 1978). This decreases the degree of grazing on an individual colony (Folino, 1993) and could also promote genetic variation in the population by ‘mixing up’ cohorts. Cuthona nana behavior is similar to that of plant bugs (Miridae) (Price, 1980). Adult plant bugs are large ectoparasites and are mobile, whereas the immature stages spend all of their time on a single host. Juvenile C. nana showed a similar behavior and did not switch colonies in the laboratory. Thus differences exist between adult and juvenile residence time, which affect the degree of grazing on the prey. This behavior in C. nana seems to parallel the prudent parasite model because partial (rather than total) consumption of the prey is important to the predator’s sur- vival (Holmes, 1983). Hermit crab movement Cuthona nana at Gosport Harbor do not lay egg masses on colonies of Hydractinia polyclina, but rather on the ocean floor. Because hermit crab movement will bring prey to juveniles on the bottom, adult C. nana at Gosport Harbor do not jeopardize the juveniles’ probability of find- ing food by depositing egg masses off of the colonies (Rivest, 1978; Folino, 1993). Non-planktonic development in the C. nana population at Gosport Harbor could actually be an adaptation to a mobile prey and to trophic stability (Clark and Goetzfried, 1978). With yolk present at meta- morphosis, juveniles can survive for up to ten weeks with- out feeding at 4°C (Rivest, 1978), which is ample time for a crab to bring food (Fig. 7). The results of the pitfall experi- ment indicate that crabs with colonies of H. polyclina are fairly active over a 24-h period and provide sufficient opportunities for prey encounters. This experiment in con- junction with the monthly collections of shells with colonies indicates a non-seasonal food supply for C. nana. Hydractinia is continuously available in Gosport Harbor, especially at the time of metamorphosis, allowing for non- planktonic development. There are several similarities between the life histo- ry of Cuthona nana and that of a parasite (Price, 1980; Strand and Obrycki, 1996); such similarities shed insight on the maintenance of C. nana’s population. C. nana is much like a parasite in being a specialist on Hydractinina polyclina. The phenomenon of juveniles being picked up by their prey is similar to a host-parasite relationship, such as is seen with intermediate stages of parasitic trematodes, flukes, or hookworm larvae (Cheng, 1970). Juvenile C. nana (“parasites”) on the ocean floor are picked up by prey (“host”) passing by; alternatively adults can conceivably crawl onto a colony while a crab is temporarily stationary and is filter feeding (Gerlach et al., 1976; Rivest, 1978; pers obs.). Furthermore, similar to a parasite (Price, 1980), C. nana is not a fast-moving predator “chasing” mobile colonies of H. polyclina. Adult C. nana can crawl onto a colony while a crab is stationary or onto a colony without a crab in the shell. Thus, the movement and egg-laying behavior of C. nana adults at Gosport Harbor along with the hydroids’ ability to regenerate and the mobility of the prey on the crab shells all contribute to the persistence of prey availability throughout the year. Prey availability is very important in determining the population patterns of a specialized predator like C. nana. These factors allow for the sub-annual life cycle and perhaps non-planktonic devel- opment of C. nana with prey not being the limiting factor for the population ecology of this prey-specific nudibranch. 24 AMER. MALAC. BULL. 14(1) (1997) ACKNOWLEDGMENTS This paper is part of a dissertation submitted in partial fulfillment of the requirements for a Ph.D. in Zoology at the University of New Hampshire. I am grateful for comments and criticisms by Nancy Carpenter, Dorothy Chappell, Michael Gross, Walt Lambert, Amy Staska, Phil Yund, and several anonymous reviewers. I thank the following people for their diving assistance: L. Harris, M. Herman, L. Kintzing, W. Lambert, P. Lavoie, P. Martin, P. Pellitier, S. Truchon, and K. Verney. This research was supported by two University of New Hampshire research grants and three summer fellowship awards. LITERATURE CITED Buss, L. W. and P. O. Yund. 1989. A sibling species group of Hydractinia in the northern United States. Journal of the Marine Biological Association of he United Kingdom 69:857-874. Buss, L., C. S. McFadden, and D. R. Keene. 1984. Biology of hydractiniid hydroids. 2. Histocompatibility effector system/competitive mechanism mediated by nematocyst discharge. Biological Bulletin 167:139-158. Cheng, T. C. 1970. Symbiosis. Western Publishing Company, New York. 250 pp. Christensen, H. E. 1967. Ecology of Hydractinia echinata (Fleming) (Hydroidea: Athecata) I. Feeding biology. Ophelia 4:245-275. Clark, K. B. 1975. Nudibranch life cycles in the northwest Atlantic and their relationship to the ecology of fouling communities. Helgolander Meeresuntersuchungen 27:28-69. Clark, K. B. and A. Goetzfried. 1978. Zoogeographic influences on devel- opment patterns of North Atlantic Ascoglossa and Nudibranchia, with a discussion of factors affecting size and number. Journal of Molluscan Studies 44:283-294. Coates, A. G. and J. B. C. Jackson. 1985. Morphological themes in the evolution of clonal and aclonal marine invertebrates. In: Population Biology and Evolution of Clonal Organisms, J. B. C. Jackson, L. W. Buss, and R. E. Cook, eds. pp. 67-108. Yale University Press, New Haven and London. Fager, E. W. 1971. Patterns in the development of a marine community. Limnology and Oceanography 16:241-253. Folino, N. C. 1985. Effects of Predation by the Nudibranch Cuthona nana on Growth Patterns and Interactions Among Colonies of the Hydroid Hydractinia echinata. Master’s thesis, University of New Hampshire, Durham. 53 pp. Folino, N. C. 1989. The Dynamics of a Nudibranch-Hydroid Predator- prey Association: Cuthona nana and Hydractinia echinata. Doctoral dissertation, University of New Hampshire, Durham. 186 pp. Folino, N. C. 1993. Feeding and growth of the aeolid nudibranch Cuthona nana (Alder and Hancock, 1842). Journal of Molluscan Studies 59:15-22. Gerhart, D. J. 1986. Gregariousness in the gorgonian-eating gastropod Cyphoma gibbosum: tests of several possible causes. Marine Ecological Progress Series 31:255-263. Gerlach, S. A., D. K. Ekstrom, and P. B. Eckart. 1976. Filter feeding in the hermit crab, Pagurus bernhardus. Oecologia 24:257-264. Grant, W. C., Jr. 1963. Notes on the ecology and behavior of the hermit crab, Pagurus acadianus. Ecology 44:767-771. Hadfield, M. G. and M. Switzer-Dunlap. 1984. Opisthobranchs. In: The Mollusca, Vol. 7, Reproduction, A. S. Tompa, N. H. Verdonk, and J. A. M. Van Den Biggelaar, eds. pp. 209-350. Academic Press, London. Harper, J. L. 1985. Modules, branches and the capture of resources. In: Population Biology and Evolution of Clonal Organisms, J. B. C. Jackson, L. W. Buss, and R. E. Cook, eds. pp. 1-34. Yale University Press, New Haven and London. Harris, L. G. 1973. Nudibranch associations. In: Current Topics in Comparative Pathobiology, T. C. Cheng, ed. pp. 213-315. Academic Press, New York. Harris, L. G. 1975. Studies on the life history of two coral-eating nudi- branchs of the genus Phestilla. Biological Bulletin 149:539-550. Harmis, L. G. and K. P. Irons. 1982. Substrate angle and predation as deter- minants. Jn: Artificial Substrates, J. Cairns, ed. pp. 131-174. Ann Arbor Science Publishers, Ann Arbor, Michigan. Harris, L. G., L. W. Wright, and B. R. Rivest. 1975. Observations on the occurrence and biology of the aeolid nudibranch Cuthona nana in New England waters. The Veliger 17:264-268. Harvell, C. D. 1984. Predator induced defense in a marine bryozoan. Science 224:1357-1359. Harvell, C. D. and T. H. Suchanek. 1987. Partial predation on tropical gor- gonians by Cyphoma gibbosum (Gastropoda). Marine Ecology Progress Series 38:37-44. Holmes, J. C. 1983. Evolutionary relationships between parasitic helminths and their hosts. In: Coevolution, D. J. Futuyma and M. Slatkin, eds. pp. 161-185. Sinauer Association, Sunderland, Massachusetts. Jackson, J. B. C. 1985. Distribution and ecology of benthic invertebrates. In: Population Biology and Evolution of Clonal Organisms, J. B. C. Jackson, L. W. Buss, and R. E. Cook, eds. pp. 297-356. Yale University Press, New Haven and London. Karlson, R. 1978. Predation and space utilization patterns in a marine epi- faunal community. Journal of Experimental Marine Biology and Ecology 31:225-239. Krebs, C. W. 1989. Ecological Methodology. Harper and Row Publishers, New York. 654 pp. Lowe, E. F. and R. L. Turner. 1976. Aggregation and trail-following in juvenile Bursatella leachii pleii. The Veliger 19:153-155. Ludwig, J. A. and J. F. Reynolds. 1988. Statistical Ecology. John Wiley and Sons, New York. 337 pp. MacLeod, P. and I. Valiela. 1975. The effects of density and mutual inter- ference by a predator: a laboratory study of predation by the nudi- branch Coryphella rufibranchialis on the hydroid Tubularia lar- ynx. Hydrobiologia 47:339-346. McFadden, C. S. 1986. Laboratory observations for a size refuge in com- petitive interactions between the hydroids Hydractinia echinata (Fleming) and Podycoryne carnea. Biological Bulletin 175:161- 174. McFadden, C. S., M. J. McFarland, and L. W. Buss. 1984. Biology of hydractiniid hydroids. 1. Colony ontogeny in Hydractinia echina- ta (Fleming). Biological Bulletin 166:54-67. Miller, M. C. 1961. Distribution and food of the nudibranchiate Mollusca of the south of the Isle of Man. Journal of Animal Ecology 30:95- 116. Miller, M. C. 1962. Annual cycles of some manx nudibranchs, with a dis- cussion of the problem of migration. Journal of Animal Ecology 31:545-569. Nybakken, J. 1978. Abundance, diversity and temporal variability in a California intertidal nudibranch assemblage. Marine Biology 45:129-146. Potts, G. W. 1970. The ecology of Onchidoris fusca (Nudibranchia). Journal of the Marine Biological Association of the United Kingdom 50:269-292. Price, P. 1980. Evolutionary Biology of Parasites. Princeton University Press, Princeton, New Jersey. 237 pp. FOLINO: NUDIBRANCH-PREY ASSOCIATION bas) Rivest, B. R. 1978. Development of the aeolid nudibranch Cuthona nana (Alder and Hancock, 1842), and its relationship with a hydroid and hermit crab. Biological Bulletin 154:57-175. Shenk, M. A. and R. Karlson. 1986. Colonization of a shell resource by calyptraeid gastropods: tests of habitat selection and preemption models. Journal of Experimental Marine Biology and Ecology 16:269-311. Sokal, R. R. and F. G. Rohlf. 1981. Biometry. Freeman, San Francisco. 859 pp. Strand, M. R. and J. J. Obrycki. 1996. Host specificity of insect parasitoids and predators. BioScience 46(6):422-429. Sutherland, J. P. and R. H. Karlson. 1977. Development and stability of the fouling community at Beaufort, North Carolina. Ecological Monographs 47:425-446. Thompson, T. E. 1964. Grazing and the life cycles of British nudibranchs. In: Grazing in Terrestrial and Marine Environments, D. J. Crisp, ed. pp. 275-297. Blackwell, Oxford, England. Todd, C. D. 1978a. Changes in spatial pattern of intertidal population of the nudibranch mollusc Onchidoris muricata in relation to life- cycle, mortality and environmental heterogeneity. Journal of Animal Ecology 47:189-203. Todd, C. D. 1978b. Gonad development of Onchidoris muricata (Miiller) in relation to size, age and spawning (Gastropoda: Opisthobranchia). Journal of Molluscan Studies 44:190-199. Todd, C. D. 1979. The population ecology of Onchidoris bilamellata (L.) (Gastropoda: Nudibranchia). Journal of Experimental Marine Biology and Ecology 41:213-255. Todd, C. D. 1981. The ecology of nudibranch molluscs. Oceanography and Marine Biology Annual Review 19:141-234. Todd, C. D. 1983. Reproductive and trophic ecology of nudibranch mol- luscs. In: The Molluscs, Vol. 6, Ecology, W. D. Russell-Hunter, ed. pp. 225-259 Academic Press, London. Todd, C. D. and J. N. Havenhand. 1989. Nudibranch-bryozoan associa- tions: the quantification of ingestion and some observations on partial predation among Doridoidea. Journal of Molluscan Studies 55:245-259. Uetz, G. W. and J. D. Unzicker. 1976. Pitfall trapping in ecological studies of wandering spiders. Journal of Arachnology 3:101-111. Yoshioka, P. M. 1982. Predator-induced polymorphism in the bryozoan Membranipora membranacea (L.). Journal of Experimental Marine Biology and Ecology 61:233-242. Date of manuscript acceptance: 12 May 1997 Temporal and spatial patterns of abundance in the gastropod assemblage of a macrophyte bed Kenneth M. Brown Department of Biological Sciences, Louisiana State University, Baton Rouge, Louisiana U. S. A. 70803, zobrow@Isuvm.sncc.lsu.edu Abstract: Although gastropods are more common and diverse in macrophyte beds than in areas without littoral vegetation, little is known of fine-scale patterns of abundance within macrophyte beds, or changes in abundance of species through time. I present data here on such patterns in the gastropod assem- blage found in Carrol Lake, Wisconsin. The assemblage was quite diverse, with 13 species, four of which had mean densities greater than 100/m2. The assemblage was dominated by small, thick-shelled species, especially Amnicola limosa Say, 1817. Gastropod density decreased across two field seasons, although trends within each field season were for increases with time in the abundance of most species, especially in shallower habitats. At the micro-habitat scale, abundances of the most common species were positively correlated, perhaps because they prefer to colonize the same macrophytes, based on the results of laboratory macrophyte-choice experiments. Key words: gastropods, micro-distributions, macrophyte choice Gastropods are clearly more abundant and diverse in macrophyte beds than in littoral-zone habitats without vegetative cover (Brown et al., 1988; Brown, 1991; Brown and Lodge, 1993). However, few studies have considered fine-scale patterns in the abundance of gastropod assem- blages within macrophyte beds, or how such patterns vary through time or space. In an earlier paper (Brown and Lodge, 1993), we studied whether gastropod distributions at the macrohabitat (e. g. sand versus cobble versus macro- phytes) scale were determined by habitat choice. Specifically, we worked experimentally with several gas- tropod species common in northern Wisconsin lakes, and found that most species, on a surface-area specific basis, preferred cobble over macrophyte substrata. However, given the fact that total abundances and diversity of gas- tropods are still greater in macrophyte beds because of the greater substratum surface area per unit bottom area, I con- sider it still important to look at patterns of microdistribu- tion and habitat choice within macrophyte beds. Here I present such data describing temporal and depth-related trends in the abundance of a gastropod assem- blage in a macrophyte bed in a northern Wisconsin lake. I present data on changes in the abundances of gastropods over two seasons, and describe how abundances change through time within each season at different depths. I am also interested in whether gastropods show similar use of space at the micro-habitat (e. g. single sample) level. Finally, I test for differences in macrophyte choice, with several gastropod species, on several macrophytes common in northern Wisconsin lakes. The purpose of these experi- ments was to determine if the high levels of overlap of species at the micro-habitat level (see results) could be explained by similar patterns of macrophyte use. METHODS SAMPLING OF GASTROPODS Sampling was conducted in the near-shore littoral zone of Carrol Lake, a circum-neutral, mesotrophic lake in Vilas County, Wisconsin (for a general description of these lacustrine habitats, see Lodge et al., 1994). Temperatures ranged from 2-3°C in winter to a maximum of 20°C in late summer. Dissolved oxygen readings ranged from 10 mg/1] in summer to 5 mg/l in mid-winter, underneath the ice. Substrata changed from sand and cobble in the first few meters from shore to sparse macrophytes, and then a dense macrophyte bed at depths greater than 0.5 m (Fig. 1). Filamentous green algae were common on the cobble and sand, and macrophyte diversity increased with distance from the shoreline. The most common macrophyte species were Vallisneria americana Michaux, Elodea canadensis Michaux, Potamogeton robinsii Oakes, Sagittaria spp., Najas flexilis Rostkovias and Schmidt, Ceratophyllum American Malacological Bulletin, Vol. 14(1) (1997):27-33 27 28 AMER. MALAC. BULL. 14(1) (1997) COBBLE SPARSE MACROPHYTES DENSE MACROPHYTES Fig. 1. Transect of the sampling site, indicating depth, substratum type, and macrophyte density. demersum Sieber ex Chamisso, Megalodonta beckii Greene, Myriophyllum exalbescens Fernald, P. amplifolius Tuckerman, and P. richardsonii Rydberg. A 500 m? section of littoral zone (approximately 22.5 m along the shoreline and 22.5 m out into the lake) was marked off along the northeastern shoreline of Carrol Lake, and all sampling was conducted within this area. The site was sampled over two field seasons in 1990 and 1991. In each season, five sampling groups were completed. The first sampling group was within two weeks of ice-out (usu- ally in mid-May), and each successive sampling group was taken at approximate three-week intervals through the sum- mer, with final sampling in late September. Each sampling group consisted of eight benthic cores, stratified into three depth bands based on relative area within the 500 m2 area: inshore (less than 0.5 m depth, two samples), mid-depth (0.5-1 m depth, three samples), and deep (greater than | m depth, three samples). The 500 m2 area was divided into 3 x 3 m plots, and these 9 m2 plots were chosen randomly within each stratum at each date for sampling. A 0.3 m-long benthic corer, constructed of 15 cm- diameter PVC pipe (sampling area = 182 cm2), was insert- ed by two SCUBA divers 3 cm into the substratum so that a horizontal slit in the side of the corer was at least 1 cm below the sediment surface, and an aluminum restraining plate was inserted. The top 1 cm of sediment, overlaying macrophytes, and any snails in the sediments or macro- phytes were then collected in the corer. A large plastic bag was held over the corer as it was brought to the surface to ensure that macrophytes or gastropods were not lost from the sample. More details on the corer and basic sampling methods can be found in Klosiewski (1991) and Lodge et al. (1994) . Samples were processed through a series of sieves (ranging from 2 mm to 0.5 mm mesh) to remove fine sedi- ments but retain snails, which all have a minimum diameter greater than 0.5 mm (Lodge et al., 1987, 1994). Samples were then sorted in shallow enamel trays with the aid of a magnifying lens and light (Brown, 1991). Macrophytes were removed and vigorously shaken in a water-filled tub to dislodge gastropods (preliminary experiments revealed this technique increased snail recovery and minimized time spent sorting). Gastropod abundances were converted to a per m2 basis, and both total gastropod abundance and the abun- dances of the each of the five most common gastropods were analyzed in a three-way analysis of variance (two years times five sampling groups times three depth strata). I considered year, sampling group, and depth to be fixed effects. Abundances were log-transformed before analysis of variance because of mean-variance correlations, but raw data are portrayed in all figures. Thus, I could look for sea- sonal, yearly, or depth-related trends in the abundance of the whole gastropod assemblage, along with differences among the distributions of each of the five most common species. To assess more fine-grained differences in the distri- butions of the gastropods, I also looked for correlations in the abundance of the snails on a per sample basis. This is essentially equivalent to calculating niche-overlap values for each of the species pairs, because most niche-overlap metrics (for example, the traditional Levins-MacArthur overlap statistic) are similar to correlation coefficients (Brown, 1982). The Pearson product-moment correlations were calculated from data pooled over both years, all sam- pling groups, and depths. MACROPHYTE COLONIZATION EXPERIMENT This experiment was conducted indoors at the Trout Lake Biological Station of the University of Wisconsin in July 1987. The experimental units were circular plastic tubs, 50 cm in diameter and 15 cm deep, with a sand sub- stratum of 2 cm and a water depth of 12 cm. Water temper- ature was 20°C and lighting was on a 14L:10D schedule. Each pan had six stems each of four macrophytes: Ceratophyllum demersum, Elodea_ canadensis, Potamogeton robinsii, and P. richardsonii. These macrophyte species were chosen both because they are common species in northern Wisconsin lakes (Lodge et al., 1994), and because they vary in mor- phology, ranging from thin-leaved (Ceratophyllum and Elodea) to relatively broad-leaved (both Potamogeton spp.). The wet masses of the individual stems were chosen so that surface area was standardized among the four species at 64 cm2. These masses were 0.71 g (Elodea), 1.2 g (Ceratophyllum), 0.85 g (P. robinsii), and 0.93 g (P. richardsonii). Standardization of surface area was neces- sary because colonization rates are dependent on surface area (Kershner and Lodge, 1990). Stems were hap-hazardly inserted in the sand substratum, and weighed down with lead strips so that stems were vertical. Fifty snails of each of four gastropod species were BROWN: LITTORAL GASTROPOD ASSEMBLAGE 29 introduced in mono-specific populations to the tubs. Numbers of snails colonizing each macrophyte were recorded after 8 hr because preliminary experiments indi- cated colonization peaked by that interval. Four species were chosen either because they were abundant in Carrol Lake (Amnicola limosa, Say 1817; Physa gyrina Say, 1821, Helisoma anceps Menke, 1830), or to represent the range of pulmonate families common in these Wisconsin lakes (Lymnaea emarginata Say, 1821). There were five repli- cates for each gastropod species. Statistical analysis of the data was as a two-way analysis of variance (four snail species times four macrophyte species), with a split-plot arrangement of treatments (macrophyte species were pre- sent as sub-plot variables in tubs with each gastropod species). RESULTS GASTROPOD SAMPLING The gastropod assemblage of Carrol Lake was quite diverse and abundant (Fig. 2), with 13 species total. Four species had mean abundances (over all sampling groups in both years) greater than 100 individuals per m2: Amnicola limosa (a caenogastropod), Valvata tricarinata (Say, 1817) (a “lower heterobranch”), and Gyraulus parvus (Say, 1817) and Physa gyrina (both pulmonates). Four additional pul- monates had intermediate levels of abundance (greater than 5/m2): Helisoma anceps, H. companulatum (Say, 1821), G. hirsutis (Say, 1817), and Promenetus exacuous (Say, 1821), WA NUMBER L__] size 20 10 NO. PER M? MEAN SHELL LENGTH | IAI y A | oe AL VT GP PG HA GH HC PE CD HT FG VG LF GASTROPOD SPECIES Fig. 2. Abundance of gastropod species, averaged over all sampling groups, years, and depth strata, + standard errors, along with mean shell lengths + standard errors. See Table 1 for the acronyms of the five most common species. (CD, Campeloma decisum, FG, Fossaria galbana; GH, Gyraulus hirsutis, HC, Helisoma companulatum, HT, H. trivolvis;, LF, Laevapex fuscus, PE, Promenetus exacuous, VG, Viviparus georgianus). Table 1. Results of a three-way analysis of variance on total abundance of gastropods (TOTDEN), and the abundance of the five most common species. Values are F statistics. There were 44 error degrees of freedom, one degree of freedom for year, four degrees of freedom for sampling group, and two degrees of freedom for depth. (AL, Amnicola limosa; GP, Gyraulus parvus,; HA, Helisoma anceps; PG, Physa gyrina, VT, Valvata tricarinata; *, P < 0.05; **, P< 0.01). EFFECT TOTDEN AL VT GP PG HA Year 4.5* 9.2** 0.2 0.4 01 1.4 Sampling Group 6.4** 6.0** 6.7** 0.1 0.4 2.2 Yr x Group 1.7 0.2 0.8 0.7 2.1 0.6 Depth 8.1** 9.0** 257 0.2 4.0* 4.9% Yr x Depth 0.6 0.3 0.4 1.9 0.7 0.2 Group x Depth 3.0* 3.1* 2.8* 0.9 0.4 1.0 3-way Interaction 0.5, 0.7 0.4 0.8 0.8 0.5 along with the caenogastropod Campeloma decisum (Say, 1816). Four additional species were rare (e. g. collected in fewer than five of the sampling groups): three pulmonates [Helisoma trivolvis (Say, 1817), Fossaria galbana (Say, 1825), and Laevapex fuscus (Adams, 1841)], and one caenogastropod [Viviparus georgianus (Lea, 1834)]. Most of the common gastropods were fairly small (Fig. 2), with the six most common species averaging (again over all sampling groups) less than 10 mm in shell length (spiral- shelled species) or diameter (plano-spiral species). There were complicated changes in the total abun- dance of gastropods across years, sampling groups and depths (Table 1; Fig. 3). The significant year-effect in the analysis of variance was because of a decrease in gastropod abundance over all sampling groups and depths in 1991. The significant sampling-group effect was because of an increase in abundance of most species later in the field sea- son in both years. There was also a significant trend toward increased gastropod density and diversity in deeper strata. For example, the mean number of species per sample, over both years, was 4.3 for the shallow substratum, 6.3 at inter- mediate depths, and 6.4 at the deepest depths. Finally, there was a significant interaction between sampling group and depth for total gastropod density. The basic trend appears to be a greater rate of increase in numbers in the shallower depths with time in each season. The pattern of significance of the F-values in the analysis of variance for Amnicola limosa, the most abun- dant species, was the same as for total gastropod density. This probably indicates that changes in the dynamics of this species are driving the overall effects (Table 1; Fig. 4). Again, densities decreased in 1991 (mean density was lower in 13 of the 15 sampling group and depth combina- 30 AMER. MALAC. BULL. 14(1) (1997) ] 1991 12 10 oO NCCU UD UD NUMBER PER M? (THOUSANDS) SAMPLING GROUP Fig. 3. Total abundance of all gastropod species + standard errors in five sampling groups in two years, for (A) shallow, (B) intermediate, (C) deep depth strata. tions [Fig. 4], with a mean reduction in density of 60%). There was also a trend within both years for increased abundances later in the season, with density (averaged over both years and all depth strata) increasing from 473/m2 in the first sampling group to 1,765/m2 in the last sampling group. There was also a significant change in abundance with depth for Amnicola limosa (over both years and all sampling groups); densities increased from 388/m2 in the shallow stratum to a maximum of 1,360/m2 at intermediate depths, and dropped to 728/m2 in the deep substratum. Finally, there was a greater increase in abundance through time in the shallow and intermediate areas, while densities were more stable in the deep stratum, explaining the signif- icant sampling group by time interaction (Table 1). For example, densities, averaged over both years, increased from 170 /m2 in the first sampling group to 622/m2 in the last in the shallow stratum, and from 523 to 3,748/m2 in the intermediate stratum, but were quite constant in the deep stratum (starting density of 928/m2 and ending density of 924/m2). For Valvata tricarinata, only the sampling-group main effect and the sampling group x depth interaction were significant. Averaged over both years and all depth VA 1990 J 1991 27 A 1 0 1 2 3 4 5 NUMBER PER WN? (THOUSANDS) SAMPLING GROUP Fig. 4. Abundance of Amnicola limosa + standard errors in five sampling groups in two years, for (A) shallow, (B) intermediate, (C) deep depth Strata. BROWN: LITTORAL GASTROPOD ASSEMBLAGE 31 substrata, densities increased from 96/m2 in the first sam- pling group to 661/m? in the fifth sampling group. The sig- nificant sampling group x depth interaction probably occurred because densities increased (averaged over both years) from 0 to 354/m2 in the shallow stratum from the first to last sampling group, and from 85 to 1,347/m2 at intermediate depths. However, they were again fairly con- stant in the deepest stratum, changing from only 207 to 283/m2. For the remaining three species that had fairly high densities overall (Gyraulus parvus, Physa gyrina, and Helisoma anceps), there were no differences in density between years, or across sampling groups (Table 1). However, there was a significant depth effect for the latter two species. For P. gyrina, densities (averaged over all dates and both years) increased from 37 to 207/m2. For H. anceps, densities actually peaked (127/m2) at the intermedi- ate depth compared to the shallow (22/m2) and deep habi- tats (48/m2). At the level of the individual sample, there were significant positive correlations in the abundance of species (Table 2). The abundance of the most common species, Amnicola limosa, was positively correlated with each of the four other fairly common species. The abundance of Valvata tricarinata was positively correlated with Physa gyrina and Helisoma anceps. None of the other correlation coefficients were significant. Thus, on a per sample basis, “hot spots” for snail abundance existed, and many species co-occurred at these sites. MACROPHYTE COLONIZATION EXPERIMENT The results of this experiment clearly indicate that all four species of gastropod have the same rank-order of macrophyte preference (Fig. 5). The plant main effect was highly significant (F379 = 43.8, P < 0.0001), and the plant- snail interaction was not significant (F979 = 2.0, P = 0.07), indicating the preference rank was similar in each of the snail species. However, the snail main effect was also highly significant (F3 2 = 13.2, P = 0.0004), indicating dif- ferences in colonization rates among snails. Ceratophyllum demersum had the lowest colonization rates, averaging 1.6 snails per plant (over all snail species). Next came Elodea Table 2. Pearson product-moment correlations among the abundances of the five most common gastropods in the samples. See Table 1 for species acronyms and symbols for significance. AL VT GP PG HA AL 1 0.65** = 0.25* 0.37** = 0.51** VT 1 -0.02 0.29* 0.37** GP 1 0.17 0.09 PG 1 0.19 HA 1 canadensis with 3.5 snails per plant, Potamogeton robinsii with 8.0 snails per plant, and finally Potamogeton richard- sonii with 12.6 snails per plant. Most gastropods had fairly similar colonization rates, except Lymnaea emarginata which had the lowest colonization rates on each macro- phyte. This is not surprising, because this species is com- mon on periphyton-covered cobble in shallow water in Wisconsin lakes, not on macrophytes (Weber and Lodge, 1990; Brown and Lodge, 1993). In summary, most gas- tropods had fairly similar colonization rates on each of the macrophytes tested, and colonized broad-leaved Potamogeton species more readily than thinner-leaved species like Elodea or Ceratophyllum. DISCUSSION SPATIAL OVERLAP The gastropod assemblage of Carrol Lake was quite diverse, with 13 species total, 5-6 of which had fairly high abundances. This is a fairly representative sample of the species present on macrophytes in northern Wisconsin lakes, and is certainly not an exceptionally diverse assem- SNAIL SPECIES ESS AL Z7ZZHA BEMBLE (| PG 20 NUMBER PER PLANT ZF mB P. robinsil P. richardsonii Ceratophyilum Elodea MACROPHYTE SPECIES Fig. 5. Mean numbers of four gastropod species colonizing four macro- phyte species, + standard errors. (LE, Lymnaea emarginata; see Table 1 for other acronyms). 32 AMER. MALAC. BULL. 14(1) (1997) blage for the area (Brown and Lodge, 1993). The gastropod assemblage was dominated by small, thick-shelled species like Amnicola and Valvata. Larger species with thick shells, like Helisoma spp., Viviparus georgianus, and Campeloma decisum were present, but rare. The only relatively thin- shelled species (Stein et al., 1984) that attained higher abundances were Gyraulus parvus and Physa gyrina. The higher relative abundances of small, thick-shelled species could be the result of selective predation by fish, which tend to remove larger, thinner-shelled gastropods (Stein et al., 1984; Osenberg and Mittelbach, 1989; Klosiewski, 1991). The reduced gastropod abundances in the second year are unexplained, but could indicate significant yearly variation in abundances caused by some abiotic factor (harsh winter weather, etc.). I do not believe the reduced abundances were caused by sampling disturbance, because the area sampled per season was only 0.2 % of the 500 m2. However, seasonal increases in snail densities in late sum- mer are fairly easy to explain. Most of these gastropods have a spring and early-summer recruitment period, result- ing in population increases by the end of the summer (Brown, 1991). The increased densities with depth were probably the result of increased macrophyte biomass at the intermediate and deep strata, because macrophyte density is positively correlated with gastropod density and diversity (Brown and Lodge, 1993). The constancy of numbers through time in the deeper areas, and the increases in shal- low areas as the season progressed could be caused by annual migrations. Cheatum (1934) was the first to docu- ment movements of snails to deeper water in winter, and back to shallower water in spring, and several others have noticed the same trend (Clampitt, 1974; Boag, 1981). However, an alternative hypothesis would be that snail pop- ulations in shallower areas reproduce earlier and thus grow more rapidly because of higher water temperatures (Brown, 1991). Amnicola limosa was the most abundant species, and it is therefore not surprising that the same treatment effects were significant for this species and gastropod abun- dance as a whole. Again, densities built up into the field season, were lower in the second year, and increased more with time in shallow areas. For the species in the second tier of abundance, fewer treatment effects could be detect- ed. This could simply be because of their lower abundances and thus patchy distributions (Brown, 1991). The trend for increased abundance with depth, however, occurred for most of the abundant species, again indicating the impor- tance of macrophyte cover to snail abundance. OVERLAP ON MACROPHYTES At the level of the individual sample, the most com- mon species had positive correlations among themselves in abundance. This is probably because of the patchy distribu- tion of snails in general: certain sites had higher snail densi- ties of all species, probably because of variation in macro- phyte diversity and cover, etc., among samples. For exam- ple, certain species of macrophytes might have greater peri- phyton abundances or periphyton species of higher nutrient quality, differ in their accessibility to snail grazers, or offer more of a refuge from predation. All of these hypotheses deserve further study. These data do indicate, however, that there is little spatial partitioning at the scale of the micro- habitat among these snail species. In fact, the laboratory experiment suggests that there is almost a remarkable degree of similarity in colo- nization rates of a group of macrophytes among the four species of snails studied. The snails all prefer broader- leaved species like Potamogeton over thin-leaved species like Elodea or Ceratophyllum. Several mechanisms could produce such results. Either broader-leaved species are col- onized by a richer periphyton assemblage, or broader- leaved species could have growth forms more suitable for colonization, or broader-leaved species could provide more of a refuge from visually orienting predators. The first and last hypotheses remain unexplored, but the second hypothe- sis contradicts a laboratory study by Kershner and Lodge (1990). Kershner and Lodge found higher colonization rates by two snail species on more finely divided artificial macrophytes (e. g. that would mimic thinner-leaved species like Elodea or Ceratophyllum), but did conclude that the particular relationship of macrophyte growth form to colo- nization rate depended on the animal group under study. Whatever the underlaying mechanism, the data suggest there is little potential for differential use of macrophytes to be a mechanism of niche partitioning by snail species. In fact, the similar patterns of macrophyte choice among species could explain why there were positive correlations among the abundances of the most common species in the sampling data. ACKNOWLEDGMENTS I would like to thank Yvonne Vadeboncoeur and various workers for help in collecting benthic gastropod samples, and Edward Haight for help with the macrophyte choice experiments. Drs. Roy Stein and David Lodge gave helpful advice on sampling and experimental design. The work was funded by NSF grant 89-07690. LITERATURE CITED Boag, D. A. 1981. Differential depth distribution among freshwater snails subjected to cold temperatures. Canadian Journal of Zoology 59:733-737. BROWN: LITTORAL GASTROPOD ASSEMBLAGE 33 Brown, C.L., T. P. Poe, J. R. P. French, and D. W. Schoesser. 1988. Relationships of phytomacrofauna to surface area in naturally occurring macrophyte stands. Journal of the North American Benthological Society 7:129-139. Brown, K. M. 1982. Resource overlap and competition in pond snails: an experimental analysis. Ecology 63:412-422. Brown, K. M. 1991. Mollusca: Gastropoda. In: Ecology and Classification of North American Freshwater Invertebrates, J. Thorp and A. Covich, eds. pp. 285-314. Academic Press, New York. Brown, K. M. and D. M. Lodge. 1993. Gastropod abundance in vegetated habitats: the importance of specifying null models. Limnology and Oceanography 38:217-225. Cheatum, E. P. 1934. Limnological investigations on respiration, annual migratory cycle, and other related phenomena in freshwater pul- monate snails. Transactions of the American Microscopical Society 53:348-407. Clampitt, P. T. 1974. Seasonal migration cycle and related movements of the freshwater pulmonate snail, Physa integra. American Midland Naturalist 92:275-300. Kershner, M. W. and D. M. Lodge. 1990. Effect of substrate architecture on aquatic gastropod-substrate associations. Journal of the North American Benthological Society 9:319-326. Klosiewski, S. P. 1991. The Role of Pumpkinseed Sunfish in Structuring Snail Assemblages in Northern Wisconsin Lakes. Doctoral Dissertation, Ohio State University, Columbus, Ohio. 147 pp. Lodge, D. M., K. M. Brown, S. P. Klosiewski, R. A. Stein, A. P. Covich, B. K. Leathers, and C. Bronmark. 1987. Distribution of freshwa- ter snails: spatial scale and the relative importance of physico- chemical and biotic factors. American Malacological Bulletin 5:73-84. Lodge, D. M., M. W. Kershner, J. P. Aloi, and A. P. Covich. 1994. Effects of an omnivorous crayfish (Orconectes rusticus) on a freshwater littoral food web. Ecology 75:1265-1281. Osenberg, C. W. and G. C. Mittelbach. 1989. Effects of body size on the predator-prey interaction between pumpkinseed sunfish and gas- tropods. Ecological Monographs 59:405-432. Stein, R. A., C. G. Goodman, and E. A. Marschall. 1984. Using time and energetic measures of cost in estimating prey value for fish preda- tors. Ecology 65:702-715. Weber, L. M. and D. M. Lodge. 1990. Periphytic food and predatory crayfish: relative roles in determining snail distribution. Oecologia 82:33-39. Date of manuscript acceptance: 04 September 1997 Sympatric speciation of freshwater mussels (Bivalvia: Unionoidea): a model Daniel L. Graf* Biology Department, Northeastern University, 414 Mugar Building, Boston Massachusetts 02115 U.S. A. and Museum of Comparative Zoology, Harvard University, 26 Oxford Street, Cambridge Massachusetts 02138 U.S. A. Abstract: Speciation of freshwater mussels (Bivalvia: Unionoidea) can occur sympatrically (ecologically) via new glochidial host acquisition. Due to the characteristics of the unionoidean life cycle, this hypothesis overcomes the objections of the classical allopatric speciation paradigm, namely homogamy and linkage of mate and habitat preferences. Examples of freshwater mussel populations in various stages of the speciation process are provided from the literature. Key words: Unionoidea, host race, parasitism, sympatric speciation, speciation model Inclusion of a parasitic stage in the life cycle of a taxon influences the nature and rate of its speciation. Parasitic taxa can diverge in allopatry when peripheral pop- ulations of parasite and host co-evolve subsequent to isola- tion from their respective parent stocks (Mayr, 1970). Another mechanism of population subdivision available to Parasitic species is the formation of host races. Although confusion exists regarding the meaning of the term “host race” (see Mayr, 1970), Bush (1974:3) defined it to describe “... an infraspecific category generally applied to populations of a parasitic species which exhibit distinct genetically-based preferences [for certain hosts].” The con- cept of host race formation has not, in the strict sense, been applied to freshwater mussels (Bivalvia: Unionoidea). The impact of their glochidial hosts on the popula- tion structure of mussel species has been noted in the litera- ture as far back as Ortmann (1920). Kat (1983; also Kat and Davis, 1984), for example, determined that unionoideans that utilized anadromous hosts maintained a high degree of genetic similarity between widely separated demes, while subpopulations of mussels that infested territorial fish diverged more rapidly. Because the population structure of freshwater mussels is so dependent on the ecology of their hosts, Kat (1984) also suggested that an intraspecific change in host fish might precede the formation of new species. Kat cited Bush’s (1974) model of fruit fly specia- *Present Address: Museum of Zoology, University of Michigan, Ann Arbor, Michigan 48109 U. S. A., dgraf@umich.edu tion via host race formation as relevant to the Unionoidea, and I propose here to extend these ideas, in particular, their applicability to the concept of speciation without geograph- ic isolation. THE LIFE CYCLE OF FRESHWATER MUSSELS Although the mechanics of the unionoidean life cycle have been sufficiently detailed elsewhere (e. g. Coker et al., 1921; Kat, 1984), the biological implications of the various stages of this process have been largely ignored. To this end, an overview of mussel reproduction follows to emphasize certain points bearing upon the speci- ation model. The life cycle is here divided into four stages: spawning, brooding, encystment and dispersal, and adult- hood. Spawning Male mussels expel their sperm directly to the water, and these are eventually entrained in the respiratory current of the female. Work with sea urchins (Pennington, 1985) has shown that even under low flow conditions, sperm diffuse rapidly, and the probability of fertilization decreases to nearly zero for females only a meter down- stream. Unfortunately, such work has not been done for freshwater bivalves, but it is reasonable to expect a similar spatial-probability picture (Downing et al., 1993). The range of fertilization, though, could be slightly increased by American Malacological Bulletin, Vol. 14(1) (1997):35-40 35 36 AMER. MALAC. BULL. 14(1) (1997) the mussels’ greater ability to filter the water column and by the way they package their sperm (D. O Foighil, pers. comm.). Freshwater mussels, like oysters, possess sperma- tozeugmata (Edgar, 1965; Lynn, 1994) which deliver many sperm together rather than let them freely diffuse and be lost (see fe) Foighil, 1989). Brooding Mussel embryos develop into larvae, known as glochidia, within the female’s ctenidial marsupia. With maturity and appropriate environmental conditions, the glochidia are shed into the water. The extrinsic stimuli lead- ing to glochidial release vary from seasonal fluctuations in temperature and food availability (as in Margaritifera spp.; Ziuganov et al., 1994) to the presence of a potential host (as in Lampsilis spp.; Kraemer, 1970). There is intra- and interspecific variability in the timing of spawning, length of the brood period, and the time of glochidial release (e. g. Zale and Neves, 1982; Neves and Widlak, 1988). Encystment and Dispersal Glochidia undergo metamorphosis while encysted in the gill or fin epithelium of an appropriate fish (or, in one case, amphibian) host. Glochidia, incapable of selecting a suitable fish, reach their hosts passively and clamp to any tissue they contact (Lefevre and Curtis, 1910). Certain species, especially those of the genus Lampsilis, exhibit behaviors and morphologies capable of attracting fishes (Kat, 1984), increasing the probability of glochidia contact- ing a potential host. However, the actual specificity of attraction by mantle flaps and conglutinates has yet to be determined. After encysting to a potential host, those glochidia that do not evoke an immune response from the fish are able to complete their metamorphosis (Bauer and Vogel, 1987). During encystment (lasting days to months), the glochidia are dispersed by the fish to new sites. Adult mus- sels are capable of limited movement, but the distances traveled are generally short (Amyot and Downing, 1997) and the progress is erratic (figured in Baker, 1928, and Mathiak, 1979); significant dispersal is facilitated only by their hosts. Because the habitat of an adult mussel is a func- tion of where it excysted (Isley, 1911, 1914), it is expected to be found in the habitat preferred by its host. Mussels themselves have low habitat specificity (Strayer, 1981; Strayer and Ralley, 1993) and possess a plastic phenotype (e. g. Ortmann’s Law of Stream Distribution; Ortmann, 1920) enabling their adaptability to various stream condi- tions (Watters, 1994b). THE MODEL Some model elements are common to all Unionoidea, such as limited fertilization range and confor- mation to the host habitat. Other factors which might not apply to all mussel species need to be evoked for the model presented here. The model mussel population has high host speci- ficity. Initially, this specificity is limited to one or possibly a few of the fish species present over the range of the mus- sel population. Although certain mussel species are charac- terized by low host specificity, other unionoideans do pos- sess this type of high specificity (Hoggarth, 1992; Watters, 1994a). Host specificity is under genetic control, and through spontaneous mutation the mussel population has developed the ability to parasitize a new host. The basis of host specificity is immunological (Bauer and Vogel, 1987). If the immune defenses of the infested fish recognize the surficial molecules of the glochidium as foreign, the para- site will be sloughed off. Such glochidial molecules must be under genetic control. That the host specificity of a par- ticular mussel is under genetic control is also evidenced by the fact that mussels sharing a gene pool (i. e. species) tend to share host species. Congeners often employ different host species (Hoggarth, 1992). During the course of their divergence from the ancestral population, additional hosts must have been added to the repertoire of these species. The mecha- nism by which this new host is added to the population need not necessarily be mutation; any of the normal processes that increase the genetic diversity of a breeding population (i. e. recombination, hybridization, etc.) are also acceptable. The original fish host and the new host have dif- ferent habitat requirements, and these fish have a strong preference for these habitats. For example, one host prefers riffles, while the other prefers pools; or one might prefer lakes and the other rivers. This can most often be achieved when the new and old hosts belong to different genera. The fish are not ecologically excluded from moving through habitats other than their preferred one. The two fish hosts differ in their seasonal presence over the range of the mussel population. For instance, one host might be more prevalent late in the mussels’ breeding season while the other fish might occur in greater numbers early in the breeding season. The timing of glochidial release is heritable; it is ultimately under genetic control and can be acted upon by selection. Glochidial release (and spawning) can be trig- gered by an array of environmental cues such as day length, water temperature, and perhaps host presence. However, the basis for the mussels’ perception and recognition of these cues is doubtless the result of physiological characters encoded in the mollusks’ genome. Different mussel species exhibit different characteristic breeding periods GRAF: FRESHWATER MUSSEL SPECIATION by (Watters, 1994a), and one reasonable explanation is that the timing of glochidial release is under genetic control. Due to high host specificity and the population’s dependence on the fish for dispersal, the mussels are initial- ly distributed over the habitat of the original host. If the original host spends 90% of its time in its preferred habitat, then about 90% of the mussel population should also occur there. If the new host occurs, even infrequently, in the habitat of the mussels during their breeding period, these fish could serve as the host for mussels that possess the mutant phenotype (i. e. those able to parasitize the new host). Such mussels will begin slowly to accumulate (over many generations) in the new host’s habitat, and the result will be at least partial habitat separation of the two pheno- types. Because of the limited range of fertilization (see above), the tendency would be for adjacent mussels (i. e. within the same habitat) to interbreed. Thus, host specificity not only biases the habitat in which a mussel lives but also the phenotype of its mates. Mussels with a particular host specificity will mate more frequently with mussels sharing the same specificity (homogamy) because of their proximi- ty. Linkage of mate and habitat “preferences” is the primary assumption of theories proposing reproductive isolation without geographical separation (Maynard Smith, 1966), but it is also a major bone of contention of the classic allopatric-geographic speciation paradigm (Mayr, 1947). Temporal differences between the presence of the two fishes would contribute selection pressure towards the isolation of the two mussel phenotypes. Those mussels whose glochidial release coincides with peak host availabil- ity will have an obvious advantage over those that release their glochidia at other times; this would lead to synchro- nization of glochidial release with host presence. Kat (1984) has suggested that synchronization of mussel repro- duction and fish activity is among the least specialized adaptations that unionoideans have evolved to increase the probability of glochidia encountering an appropriate host. Fig. 1 shows an overview of the speciation model. At Stage A, the mussel population is distributed over the habitat range of the original host. New individuals are added to the population and tend to remain in the fish’s pre- ferred habitat. Between Stages A and B, a mutation appears that allows some members of the population to infect the new host. During Stage B, the mussels possessing the mutant phenotype accumulate in the habitat of the new host. Stage B can last many generations as the number of glochidia produced each generation that survive to repro- ductive age is very small. Successful mating can occur in the new habitat after the mussels have accumulated to the point that their density is conducive to fertilization (Downing et al., 1993); this marks the beginning of Stage C. Interbreeding occurs with- in but not between the habitats and there is a tendency for glochidia produced to remain in the habitat in which they were conceived. Thus, gene flow between the habitats is limited. Over time, selection synchronizes the breeding of these two phenotypes with the habits of their respective host fish. This further decreases the amount of gene flow between the two incipient host races. Selection can acceler- ate this process by contributing to the spawning asynchrony of the two host races. Eventually, Stage D is reached when the two breeding types have become completely isolated due to their opposite host and habitat affinities. The biological basis for the mutation in host speci- ficity is immunological. It might also be reasonable, how- ever, that the initial change in host specificity is due to a mutation that changes the time of glochidial release, with the habitat preferences of the host contributing selection towards separation. Conceptually, this would require a minor modification of the assumptions, but the same model remains applicable. EXAMPLES There are no unequivocal examples in the literature. Finding examples of populations in the early stages of this type of speciation process has proven especially difficult; a simultaneous examination of a mussel population’s genetics and host preferences has yet to be undertaken. Provided for illustration are two unionoidean populations that could be candidates for just such a future study. The first example involves Anodonta woodiana (Lea, 1834) in Japan. The “population” is composed of genetically and morphologically distinguishable sympatric morphs (A and B) (Tabe ef al., 1994). Further distinguish- ing these mussels is their breeding period. Morph A is tachytictic and releases its glochidia in late spring and early summer, while Morph B mussels are bradytictic with glochidial release occurring in the early spring (Fukuhara et al., 1994). The reported fish hosts for A. woodiana are a goby and at least one cyprinid, and it has been shown that different hosts of this species are associated with different periods of mussel gravidity (Watters, 1994a). Elliptio waccamawensis (Lea, 1863), a second example, is endemic to Lake Waccamaw and the Waccamaw River of the Atlantic Slope drainage of North Carolina (Johnson, 1970). Electrophoretic studies of mem- bers of the genus Elliptio indicate that this mussel is most closely related to Elliptio cistelliformis (Lea, 1863), which is also found in the lake as well as surrounding drainages (Davis et al., 1981). Davis and co-workers (1981) reported 38 Mutation Occurs Time Passes and —e Selection Acts AMER. MALAC. BULL. 14(1) (1997) Distribution of the New Host Glochidial Release, Metamorphosis, and Excystment The New Host Transports Glochidia with the New Phenotype Fig. 1. Overview of the model of mussel speciation via host race formation. See text for discussion. GRAF: FRESHWATER MUSSEL SPECIATION 39 a distinct ecological difference between the two species, and I would predict that the fish host of E. waccamawensis is one of the endemic fishes of the Waccamaw drainage. DISCUSSION AND CONCLUSIONS Besides the two examples cited above, many other unionoidean taxa meet the criteria of this model for sym- patric speciation via new host acquisition, including the limitation to one or a few host fish that share a common habitat preference. This appears contradictory to the wide- ly held notion (e. g. Kat, 1984:199) that “... host specificity among unionaceans seems to be rather low.” However, a critical examination of Hoggarth (1992) reveals that, of those associations confirmed by actual glochidial metamor- phosis, greater than 73% of the 60 unionids reviewed are known to parasitize fish belonging to two or fewer genera. That is, for the parasite-host relationships thus far deter- mined, the Unionoidea exhibit rather high fidelity to partic- ular host genera. Within many of the host genera implicat- ed, congeners share similar habitats and habits (K. Hartel, pers. comm.), and as it concerns the model, host specificity refers not so much to the number of species utilized but to the number of habitats frequented by those fishes. The shortcoming of most models of ecological spe- ciation is the difficulty of explaining how adapting to a new niche would lead to reproductive isolation. Under the assumption that the habitat preference of an organism could be changed by a mutation at a single locus, random inter- breeding would tend to swamp the effects of the gradual accumulation of ecological separation between the two phenotypes (Mayr, 1947, 1970). The model presented here, however, is not based on the progressive acquisition of iso- lation. The fish hosts possess genetically hard-wired habitat preferences, and the mussels can capitalize on the niche fidelity fine-tuned during the evolution of the fish. Short effective fertilization distance completes the picture. I do not argue that reproductive isolation via new host acquisition is a common mode of speciation in the Unionoidea. However, I would suggest that theories of spe- ciation by geographical separation alone fail satisfactorily to explain the zoogeography and diversity of all mussel species in the Mississippi basin; the allopatric paradigm has yet to be corroborated by vicariance with the diversity of other families of aquatic organisms which would presum- ably reflect the same isolating events. Further, the habitat separation of genotypes achieved through the action of a shift in host fish could contribute to allopatric speciation following the erection of extrinsic barriers that subdivide populations (e. g. stream capture, etc.). Sympatric specia- tion via new host acquisition should be considered a viable alternative to allopatric speciation in the Unionoidea, and it is a mechanism in need of further theoretical and experi- mental testing. ACKNOWLEDGMENTS The contributions of my peers and mentors to this endeavor are immeasurable. Besides the comments of P. Arnofsky and J. Hogan (Northeastern University, Boston, Massachusetts) and K. J. Boss, R. I. Johnson, and T. Kausch (Museum of Comparative Zoology [MCZ], Harvard University, Cambridge, Massachusetts), this work greatly benefit- ed from the efforts of G. S. Jones and E. Ruber of Northeastern University. Professor Jones met with me weekly for speciation discussions and acted as my designated skeptic. Numerous invaluable suggestions and allowance to make this detour (and many others) from my thesis research were granted by Professor Ruber. Additional insights during revision were provided by K. Hartel (MCZ), D. O Foighil (Museum of Zoology, University of Michigan, Ann Arbor) and an anonymous reviewer. Any errors, however, are solely my own. LITERATURE CITED Amyot, J.-P. and J. A. Downing. 1997. Seasonal variation in vertical and horizontal movement of the freshwater bivalve Elliptio complana- ta (Mollusca: Unionidae). Freshwater Biology 37:345-354. Baker, F. C. 1928. The fresh-water Mollusca of Wisconsin. Part 2. Pelecypoda. Wisconsin Geological and Natural History Survey General Series 1301. University of Wisconsin Bulletin 70(2): 495 PP- Bauer, G. and C. Vogel. 1987. The parasitic stage of the freshwater pearl mussel (Margaritifera margaritifera L.). 1. Host response to glochidiosis. Archiv fiir Hydrobiologie 76 (supplement):393-402. Bush, G. L. 1974. The mechanism of sympatric host race formation in the true fruit flies (Tephritidae). In: Genetic Analysis of Speciation in Insects, M. J. D. White, ed. pp. 3-23. Australia and New Zealand Book Co., Sydney. Coker, R. E., A. F. Shira, H. W. Clark, and A. D. Howard. 1921. Natural history and propagation of freshwater mussels. Bulletin of the Bureau of Fisheries 37:77-181. (Reprinted as Bureau of Fisheries Document No. 893.) Davis, G. M., W. H. Heard, S. L. H. Fuller, and C. Hesterman. 1981. Molecular genetics and speciation in Elliptio and its relationships to other taxa of North American Unionidae (Bivalvia). Biological Journal of the Linnean Society 15:131-150. Downing, J. A., Y. Rochon, and M. Pérusse. 1993. Spatial aggregation, body size, and reproductive success in the freshwater mussel Elliptio complanata. Journal of the North American Benthological Society 12:148-156. Edgar, A. L. 1965. Observations on the sperm of the pelecypod Anodontoides ferussacianus (Lea). Transactions of the American Microscopical Society 84:228-230. Fukuhara, S., M. Tabe, T. Kondo, and A. Kawamura. 1994. Difference in breeding season between two genetic types of Anodonta woodi- ana (Bivalvia: Unionidae) in a small creek. Venus 53:37-42. Hoggarth, M. A. 1992. An examination of the glochidia-host relationships reported in the literature for North American species of Unionacea (Mollusca: Bivalvia). Malacology Data Net 3:1-30. Isley, F. B. 1911. Preliminary note on the ecology of the early juvenile life 40 AMER. MALAC of the Unionidz. Biological Bulletin 20:77-80. Isley, F. B. 1914. Experimental study of the growth and migration of fresh-water mussels. Appendix III to the Report of the U. S. Commissioner of Fisheries for 1913 (Reprinted as U.S. Bureau of Fisheries Document No. 792). 24 pp., 3 pls. Johnson, R. I. 1970. The systematics and zoogeography of the Unionidae (Mollusca: Bivalvia) of the southern Atlantic Slope region. Bulletin of the Museum of Comparative Zoology 1 40:263-450. Kat, P. W. 1983. Morphologic divergence, genetics, and speciation among Lampsilis (Bivalvia: Unionidae). Journal of Molluscan Studies 49: 133-145. Kat, P. W. 1984. Parasitism and the Unionacea (Bivalvia). Biological Reviews 59: 189-207. Kat, P. W. and G. M. Davis. 1984. Molecular genetics of peripheral pop- ulations of Nova Scotian Unionidae (Mollusca: Bivalvia). Biological Journal of the Linnean Society 22: 157-185. Kraemer, L. R. 1970. The mantle flap in three species of Lampsilis (Pelecypoda: Unionidae). Malacologia 10:225-282. Lefevre, G. and W. C. Curtis. 1910. Reproduction and parasitism in the Unionidae. Journal of Experimental Zoology 9:79-115, 5 pls. Lynn, J. W. 1994. The ultrastructure of the sperm and motile sperma- tozeugmata released from the freshwater mussel Anodonta gran- dis (Mollusca: Bivalvia: Unionidae). Canadian Journal of Zoology 72:1452-1461. Mathiak, H. A. 1979. A River Survey of the Unionid Mussels of Wisconsin, 1973-1977. Sand Shell Press, Horicon, Wisconsin. 75 pp. Maynard Smith, J. 1966. Sympatric speciation. American Naturalist 100:637-650. Mayr, E. 1947. Ecological factors in speciation. Evolution 1:263-288. Mayr, E. 1970. Populations, Species, and Evolution. Belknap Press of Harvard University Press, Cambridge, Massachusetts. 453 pp. Neves, R. J. and J. C. Widlak. 1988. Occurrence of glochidia in stream drift and on fishes of the upper North Fork Holston River, Virginia. American Midland Naturalist 119:111-120. . BULL. 14(1) (1997) O Foighil, D. 1989. Role of spermatozeugmata in the spawning ecology of the brooding oyster Ostrea edulis. Gamete Research 24:219-228. Ortmann, A. E. 1920. Correlation of shape and station in fresh-water mus- sels (Naiades). Proceedings of the American Philosophical Society 59:269-312. Pennington, J. T. 1985. The ecology of fertilization of echinoid eggs: the consequences of sperm dilution, adult aggregation, and synchro- nous spawning. Biological Bulletin 169:417-430. Strayer, D. L. 1981. Notes on the microhabitats of unionid mussels in some Michigan streams. American Midland Naturalist 106:411- 415. Strayer, D. L. and J. Ralley. 1993. Microhabitat use by some assemblages of stream-dwelling unionaceans (Bivalvia), including two rare species of Alasmidonta. Journal of the North American Benthological Society 12:247-258. Tabe, M., S. Fukuhara, and Y. Nagata. 1994. Genetic difference between two types of swan mussel, in Japan. Venus 53:29-35. Watters, G. T. 1994a. An annotated bibliography of the reproduction and propagation of the Unionoidea (primarily of North America). Ohio Biological Survey Miscellaneous Contribution No. 1. 158 PP. Watters, G. T. 1994b. Form and function of unionoidean shell sculpture and shape (Bivalvia). American Malacological Bulletin 11:1-20. Zale, A. V. and R. J. Neves. 1982. Fish hosts of four species of lampsiline mussels (Mollusca: Unionidae) in Big Moccasin Creek, Virginia. Canadian Journal of Zoology 60:2535-2542. Ziuganov, V., A. Zotin, L. Nezlin, and V. Tretiakov. 1994. The Freshwater Pearly Mussels and Their Relationships with Salmonid Fish. Vniro Publishing House, Moscow. 104 pp. Date of manuscript acceptance: 18 June 1997 Freshwater mussels (Bivalvia: Unionidae) in the Verdigris, Neosho, and Spring River basins of Kansas and Missouri, with emphasis on species of concern Brian K. Obermeyer!*, David R. Edds!, Carl W. Prophet!, and Edwin J. Miller? IDivision of Biological Sciences, Emporia State University, Emporia, Kansas 66801, U.S. A. 2Kansas Department of Wildlife and Parks, P. O. Box 945, Independence, Kansas 67301, U.S. A. Abstract: ‘We examined freshwater mussel assemblages at 99 sites from 1993 to 1995 in the Arkansas River system of southeastern Kansas and south- western Missouri. Emphasis was placed on assessing the distribution, relative abundance, and habitat use of five unionid candidates for future federal listing (species of concern): Lampsilis rafinesqueana Frierson, 1927, Ptychobranchus occidentalis (Conrad, 1836), Cyprogenia aberti (Conrad, 1850), Quadrula cylindrica (Say, 1817), and Alasmidonta marginata Say, 1818. We collected a total of 15,068 mussels of 35 species, including 1,301 L. rafinesqueana, 83 P. occidentalis, 29 C. aberti, seven Q. cylindrica, and one A. marginata. The three most abundant species collected from our study were Amblema plicata (Say, 1817), Q. metanevra (Rafinesque, 1820), and Q. pustulosa (Lea, 1831). However, species abundance rankings varied from stream to stream; for exam- ple, L. rafinesqueana was the most abundant species collected in the Spring River. Habitat use by candidate species varied considerably between streams, however, they were consistently found in shallow riffles and runs (mean depths 25.0-33.7 cm), with stable and moderately compacted substratum, predomi- nantly gravel, with a minimum of silt. Key Words: Unionidae, species of concern, freshwater mussels, Arkansas River system Prompted by concern for diminishing freshwater mussel populations, the U. S. Fish and Wildlife Service (USFWS) listed six unionid species native to the Arkansas River system of southeastern Kansas and southwestern Missouri as candidates (“Category 2”) for possible addition to the list of U. S. Endangered and Threatened Wildlife (15 November 1996, Federal Register 59(219):58982-59028). [Note: USFWS has discontinued the listing of “Category 2” candidate species, and has since labeled them as species of concern (05 December 1996, Federal Register 61:64481- 64485)]. These species of concern are the Neosho mucket (Lampsilis rafinesqueana Frierson, 1927), Ouachita kid- neyshell [Ptychobranchus occidentalis (Conrad, 1836)], western fanshell [Cyprogenia aberti (Conrad, 1850)], rab- bitsfoot [Quadrula cylindrica (Say, 1817)], elktoe (Alasmidonta marginata Say, 1818), and purple lilliput [(Toxolasma lividus (Rafinesque, 1831)]. Lampsilis rafinesqueana is endemic to the Arkansas River system (Neosho, Spring, Elk, Illinois, and Verdigris River basins) in Kansas, Missouri, Oklahoma, and Arkansas (Gordon and Brown, 1980; Johnson, 1980; Oesch, 1984; Harris and Gordon, 1987; Mather, 1990; *Present address: Rt. 2, Box 141, Eureka, Kansas 67045, U.S. A. Stewart, 1992; Obermeyer et al.; 1995; Clarke and Obermeyer, 1996). Although populations of L. rafinesqueana persist within these states, its range has declined (Cope, 1979; Metcalf, 1980; Mather, 1990; Stewart, 1992; Clarke and Obermeyer, 1996; Obermeyer et al., 1995, 1997). Ptychobranchus occidentalis is confined to the Arkansas, Black, Red, St. Francis, and White River systems in Arkansas, Kansas, Missouri, and Oklahoma (Valentine and Stansbery, 1971; Johnson, 1980). Although Buchanan (1980) and Oesch (1984) reported P. occidentalis in the Meramec River basin of Missouri (i.e. Upper Meramec River near the mouth of Blue Springs Creek), both (A. C. Buchanan, pers. comm.; R. D. Oesch, pers. comm.) have discounted this locality account due to sus- pected specimen mislabeling. Cyprogenia aberti is native to the Arkansas, Black, St. Francis, Ouachita, and White River systems in Arkansas, Kansas, Missouri, and Oklahoma (Johnson, 1980; Oesch, 1984; Harris and Gordon, 1987; Stewart, 1994). Its previously reported pres- ence in the Meramec River basin of Missouri (e.g. Buchanan, 1980; Oesch, 1984) is now considered in error due to the same suspected mislabeling of specimens men- tioned for P. occidentalis. Cyprogenia aberti is currently found in 14 streams in Arkansas, five in Missouri, and three American Malacological Bulletin, Vol. 14(1) (1997):41-55 41 42 AMER. MALAC. BULL. 14(1) (1997) in Kansas (Stewart, 1994); it is considered extirpated in Oklahoma (Mather, 1990). Quadrula cylindrica is native to the Ozarkian and Cumberland faunal regions (Johnson, 1980) of 13 states (Williams et al., 1993), perhaps reaching its greatest abundance in the Black River system of Arkansas (D. H. Stansbery, pers. comm.). A subspecies, Q. cylindrica strigillata (Wright, 1898), which is considered by some as an ecomorph (e.g. Simpson, 1914; Gordon and Layzer, 1989; Clarke and Obermeyer, 1996) (but see Ortmann, 1920), occurs in the Clinch, Powell, and Holston Rivers of the Upper Tennessee River drainage (Ortmann, 1920; Bogan and Parmalee, 1983; Yeager and Neves, 1986). Alasmidonta marginata is widely distributed throughout eastern North America, being found in 22 states and one Canadian province (Clarke, 1981; Williams et al., 1993). Toxolasma lividus is found in the Ohioan, Cumberlandian, and Ozarkian faunal regions (Johnson, 1980) of 12 states (Williams et al., 1993). The primary objectives of this study were to assess the distribution, abundance, and habitat use of five of the six species of concern mentioned above (i.e. all except Toxolasma lividus) in the Arkansas drainage system in east- ern Kansas and southwestern Missouri (Neosho, Verdigris, and Spring River basins). However, we also noted the com- position of the total mussel assemblage. STUDY AREA The Neosho and Verdigris River basins are situated within the tallgrass prairie ecoregion in southeastern Kansas. Cross and Collins (1995) termed the lotic waters of these two basins as Ozark-border streams, and character- ized them as having the greatest habitat diversity for fishes in Kansas. The greatest richness of Kansas’ unionid fauna also occurs within these basins — 37 species (Obermeyer et al., 1997). Both basins are primarily agricultural, with native rangeland in many headwater reaches, whereas extensive cultivation occurs on and near the flood plains of tailwater reaches. Chert-gravel, derived of Permian and Pennsylvanian limestones (Wilson, 1984; Aber, 1992), is the dominant substratum of shallow riffle habitats. Principal streams of the Neosho and Verdigris River basins along with their respective drainage area (km2) in Kansas follow: the Neosho (15,000) and Cottonwood (4,940) Rivers of the Neosho River basin, and the Verdigris (8,690), Fall (2,290), and Elk (1,820) Rivers in the Verdigris River basin (Fig. 1). Despite their size, these streams are subject to periodic flow interruptions during severe droughts (Deacon, 1961; Geiger et al., 1995; Miller and Obermeyer, 1997). Recent flow disruptions have result- ed from the construction and operation of several federal flood-control impoundments: Council Grove Lake and John Redmond Reservoir (Neosho River), Marion Lake (Cottonwood River), Fall River Lake (Fall River), Toronto Lake (Verdigris River), and Elk City Lake (Elk River) (Fig. 1). Streams in the Spring River basin (Fig. 1), exclud- ing the North Fork Spring River, which is a prairie stream (Davis and Schumacher, 1992), originate from the north- western flank of the Ozark Uplift. The basin’s flow is gen- erally westward until reaching Kansas, where it turns southward into Oklahoma (Davis and Schumacher, 1992), eventually joining the Neosho River. The Spring River basin drains approximately 5,415 km2 of southwestern Missouri, and an additional 1,370 km2 in southeastern Kansas (Davis and Schumacher, 1992). Streams examined in the Spring River basin included the Spring and North Fork Spring Rivers, and Shoal and Center Creeks. These streams differ from Ozark-border streams by having lower turbidity, richer aquatic faunas (Cross and Collins, 1995), and flows sustained by headwater springs during droughts. Land use in several of these streams also differs from that in the Neosho and Verdigris basins in that a sizable propor- tion of the drainage area is forested (e. g. 45% for Shoal Creek; Davis and Schumacher, 1992). In addition, exten- sive lead and zinc mining has occurred in the Spring River basin, which has especially affected the lower Spring River and Shoal Creek in Kansas and Center Creek in Missouri (Kansas Department of Health and Environment, 1980; Davis and Schumacher, 1992). Furthermore, these streams lack the large flood-control impoundments that have altered streams in the Neosho and Verdigris basins (Obermeyer et al., 1997). The mussel assemblage of the Spring River basin differs from that of the Neosho and Verdigris River basins in having the following species: Alasmidonta mar- ginata, A. viridis (Rafinesque, 1820), Fusconaia ozarkensis (Call, 1887), Toxolasma lividus (Rafinesque, 1831), and Venustaconcha ellipsiformis (Conrad, 1836) (Gordon and Brown, 1980; Cope, 1985; Obermeyer et al., 1995). Also, four species present in the latter basins are absent from the Spring River basin: Ellipsaria lineolata (Rafinesque, 1820), Truncilla donaciformis (Lea, 1827), T. truncata Rafinesque, 1820, and Megalonaias nervosa (Rafinesque, 1820) (Cope, 1985; Obermeyer et al., 1995). METHODS SAMPLING Sampling sites were confined to streams in the Arkansas River Basin with known accounts of one or more of the targeted species. An attempt was made to space sam- ple sites evenly within each stream; however, unsuitable habitat in some stream stretches (e. g. unstable banks, bedrock substratum) and/or difficulty in securing legal access sometimes made this impossible. We also tried to OBERMEYER &T AL.: UNIONIDAE OF NEOSHO, VERDIGRIS, AND SPRING RIVER BASINS — 43 Council Grove John Redmond ee Res. mn in -O '¢ 1. ' N. Fork Spring R. Spring R. Center Cr. Fig. 1. Sampling sites in southeastern Kansas and southwestern Missouri. sample sites examined by previous surveyors; unfortunate- ly, many of these sites lacked exact locality data. To locate living mussels in shallow water (15 cm to < 1 m) with adequate visibility, we used a snorkel and face mask, whereas at depths exceeding 1 m, SCUBA was used. Mussels were located both by tactile cues (groping) and by visual cues during snorkeling and SCUBA searches. We also visually searched for mussels in shallow habitats as well as recently exposed substratum. Sampling was concen- trated in riffles and runs; however, runs and pools were also examined to assess usage of these habitats. All searches were timed to quantify sampling effort, which ranged from 40 min to 9 h, depending on quantity and quality of habitat. Weather conditions and water levels also influenced sam- pling effort. We also quantitatively examined 14 sites in Kansas (Neosho = 9, Spring = 2, Fall = 3) using a 1-m2 quadrat; a total of 505 quadrats was sampled at these sites. Quadrats were placed along measured coordinates chosen randomly, with the substratum excavated by hand to an approximate depth of 10-15 cm. To seek evidence of young recruits, we sampled substratum from habitats cited as being most often utilized by juveniles (Isely, 1911; Clarke, 1986; Neves and Widlak, 1987); substratum was dredged with a shovel and trans- ferred to a 1-m2 sieve (6-mm mesh) supported by a floating 15-cm PVC-pipe frame. Dredging ceased when the weight of the substratum caused the frame to sink. The substratum was then sieved in an attempt to locate small mussels. The number of sieve samples examined at each site varied from 0 to 21. Except for a few specimens collected for reference, living unionids were identified in the field, measured with either a dial caliper or an aluminum plate shell-sizer with openings of 2, 3, 4, 6, 8, 10, and 12 cm (Obermeyer, 1996a), and returned to their original location. Reference shells from sites sampled in 1994 are deposited in the Ohio State University Museum of Zoology in Columbus, Ohio, and vouchers from 1995 sites will be housed at the Kansas Biological Survey, University of Kansas, Lawrence. Nomenclature of unionids follows Turgeon et al. (1988); however, subgenera Utterbackia Baker, 1927, and Pyganodon Crosse and Fischer, 1893, are elevated to gener- ic status following Hoeh (1990), and Fusconaia ozarkensis and F. flava (Rafinesque, 1820) collected from the Spring River and Shoal Creek are listed as Fusconaia spp. due to identification uncertainties. Nomenclature of fishes follows Robins et al. (1991). HABITAT CHARACTERIZATION At specific locales where living individuals of can- didate species were found, we made visual estimates of three substratum variables: substratum compaction, percent composition of substratum types, and silt deposition on the 44 AMER. MALAC. BULL. 14(1) (1997) substratum. Substratum compaction was coded as 0, 1, or 2, with O being loose, 1 moderately compacted, and 2 very compacted. Substrata were divided into five approximate size classes: mud (< 0.8 mm), sand (0.8 - 4 mm), gravel (4 - 50 mm), cobble (50 - 290 mm), and boulder (> 290 mm) (modified from Platts et al., 1983). We coded the degree of silt deposition from 0 to 3, where O characterized a clean substratum, 1 had a detectable silt layer, 2 was moderately covered with silt, and 3 was heavily silt-laden. Current speed and water depth were measured for each candidate specimen with a pygmy Gurley current meter no. 625 at 60% depth and at the substratum-water interface (100% depth). RESULTS From a combined effort of 505 1-m?2 quadrats and approximately 200 h of qualitative sampling from 99 sites in the Arkansas River system (Neosho River basin = 30 sites; Verdigris River basin = 32 sites; Spring River basin = 37 sites), we collected 15,068 living mussels representing 35 species (Table 1). Corbicula fluminea (Miller, 1774), a recent bivalve invader (Corbiculidae), was also found in all streams. Over 9% of our collections consisted of species of concern, with 1,301 Lampsilis rafinesqueana, 83 Ptychobranchus occidentalis, 29 Cyprogenia aberti, seven Quadrula cylindrica, three Toxolasma lividus, and one Alasmidonta marginata collected. The most abundant species encountered during the survey was Amblema plica- ta (Say, 1817), comprising 18.9% of the total sample, fol- lowed by Quadrula metanevra (Rafinesque, 1820) and Q. pustulosa (Lea, 1831), representing 18.2% and 11.8%, respectively (Table 1). However, species rank varied among basins and streams; Q. metanevra, A. plicata, and Q. pustu- losa were the three most common species in the Neosho River Basin, A. plicata, Q. pustulosa, and Q. metanevra the most numerous in the Verdigris River basin, and L. rafinesqueana, Fusconaia spp., and Elliptio dilatata (Rafinesque, 1820) the most common in the Spring River basin (Table 1). Although Lampsilis rafinesqueana was the fourth most abundant species encountered in this study (8.6% of total sample), most of these individuals (1192 = 91.6%) were collected from the Spring River, representing 40.2% of the Spring River total (Table 1). This species was found alive at 13 of 20 Spring River sites, from just downstream of state Highway 97 bridge near Stott City, Lawrence County, Missouri, to the confluence of Turkey Creek, Kansas (Fig. 2). It was the most abundant species encoun- tered at 11 of these sites. In Shoal Creek, 26 L. rafinesqueana were collected at five of 11 sites (Table 1; Fig. 2), but only in the Missouri portion of this stream. Two of three North Fork Spring River sites yielded 12 L. rafinesqueana specimens (Table 1; Fig. 2). This species was not collected alive in Center Creek, but one recently dead specimen was recovered. In the Neosho River, 32 L. rafinesqueana were collected at six of 21 sites, representing 0.6% of this river’s collection (Table 1); these were all found downstream from John Redmond Reservoir (Fig. 2). In the Verdigris River, a total of five L. rafinesqueana was found at four of 14 sites (0.2% of the total Verdigris River sample; Table 1); all four of these sites were located down- stream from Toronto Lake and upstream from the conflu- ence of the Elk River (Fig. 2). Thirty-four L. rafinesqueana were collected at five of 12 Fall River sites between Fall River Lake and the confluence of the Verdigris River (Table 1; Fig. 2), representing 1.7% of the total sample from this stream. Although weathered shells of this species were observed at sites in the Cottonwood, Caney, and Elk Rivers, living or recently dead specimens were not found (Table 1). Young Lampsilis rafinesqueana, either living or freshly dead, were found at few sites. Based on external estimations of annuli, most Verdigris and Neosho basin specimens were over 20 years old; only three specimens collected in these two basins were estimated to be of young age (6-10 years). Spring River basin specimens were com- prised mostly of two or three cohorts between eight and 20 years of age; the youngest L. rafinesqueana specimens col- lected alive or as recently dead specimens were four years old, the smallest being a recently dead specimen from Shoal Creek that measured 49 x 32 x 16 mm (length, height, and width, respectively). In the Neosho and Verdigris River basins, mean length for caliper-measured L. rafinesqueana was 131.2 mm (SD = 12.96), with speci- mens ranging 94 - 163 mm. Spring River L. rafinesqueana, which were measured with the aluminum shell-sizer, aver- aged 110.8 mm (SD = 11.10), whereas caliper-measured Shoal Creek specimens were considerably smaller (mean = 72.5 mm, SD = 8.73). Lampsilis rafinesqueana was collected most often in shallow riffles and runs having predominantly gravel substratum (Table 2); however, there was a substantial dif- ference in habitat use by L. rafinesqueana in the Spring River and Shoal Creek compared to that in the Neosho, Fall, and Verdigris Rivers (Table 2). For instance, mean current speed at locales utilized by L. rafinesqueana was much higher in the Spring River basin than in prairie streams (Table 2). The mean coded value for silt deposi- tion at L. rafinesqueana sites in the Spring River was 0.2 (SD = 0.40) compared to 1.4 (SD = 0.50) in the Neosho, Verdigris, and Fall Rivers (Table 2). These data are likely variant due to the uniqueness of the Spring River compared to other Kansas streams (Cross and Collins, 1995), and because of greater L. rafinesqueana densities in the Spring 45 UNIONIDAE OF NEOSHO, VERDIGRIS, AND SPRING RIVER BASINS OBERMEYER ET AL IZ Sel@ L8L7 00E 9Sb ‘xajduroo (, 981 ‘[1@D) Sisuaysv20 *4 pue (078I ‘anbsouyey) vanjf DIDUOISNJ aIy syuasaidas ‘dds DIDUOISN 5 Fe eae ee ee ee ee ee Ee ESS SS v967 = HHT (4 6S ELLs 890ST STRIOL i ee EEE z I p — IT = Pp —_ IT — P I I -_ P pa P P P P P pas pm pas P P Pp — P P IT pm P P 4 ry P P pa pat I pa pm = P P P pa IT = P P I I y 1D Koued =—s-s IMO a P FT pa = rT pa pa 9 p 67 8 I v7 se 67 +8 Oe! 8L S8P 88€ - 9 ve P 887 8S9 = pM pa P L P pa 6l I 07 €@ 67 P 7" C pa 6 Ov 6 Lv eel - © 8 Pp v pm pa pM pa £ €7 ve pa pa pm 8 67 8L 9 0¢ 91 pa P pa pm ve SI = rat aa TT L pM S Ul Ls 19p 889 v cl vl U a a Med SUIBIPIOA uIseg SUBIPI9 A, 0c Pp Il 1D [eous L € € IQ ‘yu Buuds Jud YON ulseg Suuidg 19 0c - as = Pp = IT - _ os P Pp 9L 91 Pp vl = P 09 €s I tbe 0€ P I - pm SI a a c = = (4 I a Sb a pa I 9 I Ste a = pa 6 = IJ - - 4 (4 pa pM = pm ry] € Pp 87 a pa € oI Pp I S pM g 2 a céLT = af vs a Pp CLE - = Cl pa 087 = pa I - - €l a = v6 16 pa p =- - 4 I ID “A PAD Suuds | anaqey 4104 °S “‘PMUONOD OYSOdN ulseg OYSOoN = IJ 9 4 U a (QE8] ‘peTUOD) Siusofisdiyja DyQUOYyIDISNUuaA (6781 ‘Aes) sipioaquat vIy9Dq4au/) (1Egt ‘Aeg) snusvjo.yai snsamoiupy OZ8I ‘anbsaujey oipoun4 * 1 (8Z8I ‘Va ‘]) Siusofisouop ojpIIUNL (OZ8I ‘anbsauyey) vsoonssaA DIUOSOI14] (€Z8I ‘souseg) snaind *7 (LEgl ‘anbsouryey) snpiay] vusvjoxo | (LI8I ‘Aes) snivjnpun snuydous (0z8I ‘anbsauiyey) vjnsponb ‘O (LER ‘ex ‘[—) vsojnisnd ‘O (oZgI ‘anbsouyey) vivjnpou ‘O (0781 ‘anbsouyey) vuaauvjau ‘GO (LI8I ‘Aeg) va14pu1jXo vjnsponO (6781 ‘Aeg) sipuns3 uopouvskg (9€8I ‘peruoD) syjojuap1990 snyouDsqoyoiid (6181 ‘yoreure’]) snjoundand ‘gq (0781 ‘anbsouyey) sisuaiyo snjiumjiog (pEgI ‘peruOD) wnauidI209 Duagosnad OZ8I ‘anbsouyey oxapfas v1sonbyjqC (0zgI ‘anbsouyey) vsoasau spivucjp3ay (Teg ‘Aeg) vrwsssosqns 7 (6181 ‘yoreule’]) vided DIUNSIT (OZ8I ‘anbsauyey) s1j1304f vapoidaT] (0781 ‘onbsourey) 9101809 °7 (€ZRI ‘saweg) DIDUDjdwWO02 DUOSIUSDT (OZR ‘anbsauryey) sasai °7 (€Z8I ‘soweg) vapionbyis “7 LZ6| ‘uOsiau DuDanbsauifos “7 (0z8I ‘anbsouyey) wnipsvo sijisduv7] , dds pipuoosnJ (OZgI ‘anbsauyey) vanjf vivuossny (Oz8I ‘anbsauyey) vinwyip ondijq (OZ8I ‘anbsouyey) vivjoauy visvsdiy]q (OS8I ‘peruod) uaaqn viuasosdkD (LISI ‘Keg) o1v911d vuajquy (Oz8I ‘anbsouyey) sipizia “y 8IgI ‘Aes vivui840m Diuoprusyy ‘SOUS ‘ON 2 SUTLINS ‘[peop pasayeam ‘pm ‘plOdal siNjesoy] ‘IT {(WU900I) peap ‘p] ‘WRANS YORE JO UONNQUUOD sy) pur ‘LINOSSIJA| Wis}JSOMUINOS PUL sesULY Wa}sBoyINOS UT SUISeQ JOATY SLIZIPIOA Pure ‘BuLIdg ‘OYsOaN at) WOIJ C66 1-€661 Ul PeI99][09 sjassnui pruorUy) “fT eIqBL 46 AMER. MALAC. BULL. 14(1) (1997) Council Grove ()EIk R. OQ EkCityA LINOSSIJY N. Fork Spring R. Spring R. Center Cr. a Fig. 2. Range map of Lampsilis rafinesqueana in the Neosho, Spring, and Verdigris River basins in southeastern Kansas and southwestern Missouri. Solid circles indicate sites where living specimens were found, triangles represent recently dead specimens, open circles are sites yielding only weathered and/or relic valves, and small solid dots represent sites in which we did not find evidence of the species. River. For example, 67 L. rafinesqueana were collected in one 1-m2 quadrat (located at a depth of 28 cm with current speeds of 90 and 68 cm/s at 60 and 100% depth, respective- ly, in clean, moderately loose substratum consisting of 10% sand, 80% gravel, and 10% cobble); whereas the species was found only sporadically in other Kansas streams. Ptychobranchus occidentalis ranked nineteenth in relative abundance from collections in the three basins (Table 1). This species was not collected alive in the Neosho River, despite abundant weathered valves at several sites. In the Verdigris River, 11 P. occidentalis were found at four sites (Table 1; Fig. 3), representing 0.4% of this stream’s total sample. Nineteen specimens were collected at six Fall River sites (Table 1; Fig. 3), comprising 0.9% of unionids collected in this stream. In the Spring River, we collected 45 P. occidentalis at ten sites, representing 1.5% of the collection (Table 1; Fig. 3). We also found two speci- mens at one site in the North Fork of the Spring River, and six individuals at a Shoal Creek site in Missouri (Table 1; Fig. 3). In the Cottonwood, Elk, Caney, and South Fork Rivers, only weathered shells of this species were noted (Table 1). Most P. occidentalis specimens were over seven years old. The youngest noted was a recently dead three- year-old specimen (41 x 20 x 9 mm) collected from a Fall River site. Mean shell length for P. occidentalis in the Verdigris basin was 90.2 mm (SD = 20.74), whereas Spring River basin specimens were slightly larger (mean = 97.4 mm, SD = 16.17). Like Lampsilis rafinesqueana, P. occi- dentalis exhibited differences in habitat use among streams (Table 2); however, it was found predominantly in riffles. Cyprogenia aberti was collected alive in only three streams, representing 0.2% of the total sample. In the Verdigris River, we collected 11 C. aberti at five sites, in the Fall River five specimens were found at four sites, and in the Spring River we collected 13 specimens at six sites (Table 1; Fig. 4). We also found one relic C. aberti valve from the Elk River (Table 1; Fig. 4), which represents a new stream record. Although C. aberti was documented in the Neosho River by both Call (1885a) and Scammon (1906), we were unable to find evidence of this species, either recent or weathered valves, in this stream. Only four of the C. aberti we collected were less than five years old, all measuring less than 45 mm in length; the smallest mea- sured 34 x 26 x 16 mm. A Verdigris River site also yielded a young freshly dead specimen in exposed gravel, which measured 44 x 35 x 15 mm; we estimated this specimen to be three years old. Shell length of living specimens from the Spring, Fall, and Verdigris Rivers ranged 34 - 81 mm 47 UNIONIDAE OF NEOSHO, VERDIGRIS, AND SPRING RIVER BASINS OBERMEYER ET AL (v0) 770 (0°0) 0"! (9°0) €'0 ($0) $I (SO) 71 (0'0) 00 (S'0) €°0 (sO) €1 (90) 71 (Z'0) 1'0 (v0) 70 ($0) I ($0) 9'I (S'0) €1 uONneIIS (Z'0) 60 (0'0) 0'I (90) L’0 (p'0) 8°0 (00) 0'1 (€0) €1 (€°0) 60 (00) 0'1 (Z'0) 60 (v°0) 60 (0'0) O'1 (pO) 1'1 (0'0) 0'1 (9°0) 71 uonoeduiog - = (L°Z1) 0°08 (L’°Z1) 0'0@ = = (S'€) SZE (¢°8) 0°09 (Tr) OL (L'0) $0 = (€€€) €°89 (62) LI (O'S) O'OE = a (L°ST) VSL (19) EL (9°83) 9°71 (09) I'v (6° 1€) O'7Z ($62) USb (€°S7Z) P81 (ZS1) 74 ($0) 70 e (O'S) SL (9°€) 0°78 (p27) 811 (0°0) 00 = (9°9) v's (1p) 0°69 (Z'°SZ) 9'b7 (3°1) O'1 = (°L) 68 (68) TEL (SP) ESI (L°2) 97 (L'0@) 69 (VZD6EI (S61) O'79 (ZZ) €'S1 (€"€) 81 = (6°91) €'8 (S71) SPL (VO) ULI (TD €o = 7 (9°91) EPL (691) p91 (€°€) O'1 (rl) ty (9'PZ) 6SE (0'0@) €' Ir (ED 641 (9°9) €'€ = (ZL) O'LT (781) O'7S (L'S) O11 (COD OTT (Zr) SI (€°~Z) OLE ($°7Z) b'8b (CTD LI (3°1) 0 Japjnog 219999 [ELN 29) pues pny (9%) JoyoeIeYyO WINeNsqns (6'1€) 7°96 (1°6) 8°€Z (9°91) Trb ¢ "y suds (I'T€) O'8E (€°91) LZ (r'€l) ZI Zz “Y OYsoaN DI1Apuljxd DjnspONO (8°S€) $9 (O'LI) TLZ (LOI) ELE € "y suds (9°61) 602 (V81) VL1 (97) $97 6 "Y SUSIPIOA (171) 8°91 (6'L) v8 (9'L1) 9°62 ¢ ‘Urs N4aqgo viuasosdk (p'9) 1'L6 (I'L) 6'PE (O'r) SEL v “ID Teoys (6'LZ) bb (8°61) 8°97 (LI) O'lr ZI "y Suuds (pr) 9°81 (€01) Z€I (18) 0°61 6 "Y SUSIPIOA (9°01) I'v1 (1D 771 (8°71) SLI LI “Ueda Sijpjuap1g90 snysUuDIGOY ig (p82) T@ (STI) P07 (p'S1) b'6S 0 “ID jeoys (12) ZL (€°61) SEP (L'11) O'€€ 8Sz "y 3uuds (p°SZ) O'LZ (3'€1) O91 (Z'@) 9°6€ ze "y OYysoan (EL) 7S (Or) ZE (6°81) 797 ¢ “YU SUSIPI9A (€°8) Z'€1 (L'01) PZ (6°02) I've ve “Urea puvanbsaurfos syisduoy widop %09 «=: dap % 001 (wo) ndeq N weans 278 (S/d) poads yuauind “LINOSSIJ PU SeSULY UI SUaNs UISeg SesUeyYIY WOY sjassnu aJepipuvs Aq [(({S) Uo] Osn yeIIQeY PeAtIsSqC °Z 31GUL 48 AMER. MALAC. BULL. 14(1) (1997) Council Grove D Oklahoma LINOSSTJA N. Fork Spring R. Spring R. a Fig. 3. Range map of Prychobranchus occidentalis in the Neosho, Spring, and Verdigris River basins in southeastern Kansas and southwestern Missouri. Symbols as in Fig. 2. (mean = 61.0 mm, SD = 13.40). C. aberti specimens col- lected in this study were generally confined to shallow rif- fles and runs in predominantly clean, moderately compact- ed gravel-sand substrata (Table 2). Living representatives of Quadrula cylindrica were found in the Neosho and Spring Rivers (Table 1; Fig. 5), predominantly in shallow habitats with clean, moderately compacted gravel-sand substrata (Table 2). In the Neosho, we collected two living specimens from two sites (Table 1; Fig. 5), as well as two recently dead articulated specimens with desiccated softparts at one of these sites. Although freshly dead specimens of this species have been found in recent years in this stream (C. H. Cope, pers. comm.), these individuals are the first Q. cylindrica specimens reported alive from the Neosho River since 1912 (Isely, 1924). Relic Q. cylindrica valves were also found at nine addition- al Neosho River sites (Fig. 5). In the Spring River, we col- lected a total of five specimens at one Kansas and three Missouri sites (Table 1; Fig. 5). Relics were collected from one Shoal Creek site in Missouri, which represents the first evidence of this species in Shoal Creek. New stream records for Q. cylindrica were also made for the Fall and Cottonwood Rivers, with relic valves collected at two sites in each of these streams. Although Isely (1924) reported living Q. cylindrica in the Verdigris River in 1912, we found only relic valves of the species at eight of 14 sites (Fig. 5). We estimated that three Quadrula cylindrica speci- mens from the Neosho River (two recently dead and one alive) were in their sixth year of growth (78, 86, and 87 mm long); an additional living specimen was estimated to be in excess of ten years old (113 mm long). A rather large speci- men (recently dead valve) of this species, which measured 127 mm in length, was also recovered at one of these sites. In the Spring River, Q. cylindrica specimens ranged 74 - 109 mm in length ( mean = 93.0 mm; SD = 12.71). Only one Alasmidonta marginata, measuring 73 x 38 x 30 mm, was collected during the study, at a Kansas Spring River site (Table 1). It was found in riffle habitat with current speeds of 72 and 33 cm/s (60 and 100% depth, respectively) at 54 cm depth in predominantly cobble sub- stratum (6% sand, 15% gravel, and 79% cobble). Weathered shells of this species were recovered at two additional Spring River sites in Kansas; one weathered valve was found at a Shoal Creek site in Missouri, the first account of this species in that stream. Although Toxolasma lividus was not initially targeted in this study, we found three living individuals at one site in Shoal Creek in Missouri, as well as dead shells at three addi- tional Shoal Creek sites (Table 1). The three living speci- OBERMEYER &T AL.: UNIONIDAE OF NEOSHO, VERDIGRIS, AND SPRING RIVER BASINS = 49 Council Grove Oklahoma John Redmond N. Fork Spring R. LINOSSIJ Spring R. Center Cr. Fig. 4. Range map of Cyprogenia aberti in the Neosho, Spring, and Verdigris River basins in southeastern Kansas and southwestern Missouri. Symbols as in Fig. 2. mens were found in a shallow backwater pool, which con- sisted of silty sediments with no detectable current. DISCUSSION Habitat descriptions for freshwater mussels have often been generalized (Gordon and Layzer, 1989). For example, Quadrula cylindrica occurring in medium to large streams is cited as preferring sand-gravel substrata in 6-10 feet of water (Parmalee, 1967; Cummings and Mayer, 1992) with a detectable current (Parmalee, 1967). However, in smaller streams the species is considered a rif- fle species, being most often found near shore in cobble substratum with a slack current (Stansbery, 1974) or, as Gordon and Layzer (1989) reported, in close proximity to the swiftest flows. Anecdotal descriptions of habitat use by Ptychobranchus occidentalis are gravel substratum in rif- fles with depths between 2.5 and 75 cm in slow to moderate current (Buchanan, 1980; Oesch, 1984); Gordon and Layzer (1989) stated that two congeners, P. fasciolaris (Rafinesque, 1820) and P. subtentum (Say, 1825), seem to prefer shallow riffles in moderate to swift currents. Habitat use by Cyprogenia aberti is described as shallow water (7- 45 cm), with mud, sand, and gravel substrata (Murray and Leonard, 1962; Buchanan, 1980; Oesch, 1984). Alasmidonta marginata is reported to prefer riffles in cob- ble-gravel and gravel-sand substrata in medium to large rivers (Clarke and Berg, 1959; Clarke, 1981; Cummings and Mayer, 1992), with a preference for moderate to swift currents (Clarke and Berg, 1959; Gordon and Layzer, 1989). Oesch (1984) described the habitat use of Lampsilis rafinesqueana in Missouri as shallow water with a moder- ate current in fine to medium gravel. More detailed habitat descriptions for unionids are difficult because of broad microhabitat tolerances (Strayer, 1981; Kat, 1982; Gordon and Layzer, 1989; Holland- Bartels, 1990; Strayer and Ralley, 1993; Strayer et al., 1994), and because of site-specific preferences (Strayer, 1981) due to macro-scale variation (e. g. hydrologic vari- ability) among sites (Strayer and Ralley, 1993). Habitat use by mussels collected in the present study was also variable when compared among different streams (Table 2), which made it difficult to extrapolate habitat suitability indices from one stream to another. Habitat use on a broader scale, however, was more predictable; that is, mussels were most often found in shallow riffles and runs at depths less than one meter, with stable and moderately compacted substra- tum, predominantly gravel, with a minimum of silt. Examination of deeper, more silt-laden habitats (i. e. pools) revealed a decrease in species richness and abundance of 50 AMER. MALAC. BULL. 14(1) (1997) Council Grove Oklahoma Ok John Redmond LINOSSTY N. Fork Spring R. Spring R. a Fig. 5. Range map of Quadrula cylindrica in the Neosho, Spring, and Verdigris River basins in southeastern Kansas and southwestern Missouri. Symbols as in Fig. 2. unionids, including the absence of candidate species. Sites that were unstable (i. e. loose, shifting substrata) were espe- cially low in unionid numbers. Of the five species targeted in this study, habitat use by Lampsilis rafinesqueana was especially intriguing. In the Neosho, Verdigris, Spring, and North Fork Spring Rivers, L. rafinesqueana was found most often in riffle habitat, usually in a swift current. However, in Shoal Creek, the species was often found in habitats near shore or out of the strongest current. Mather (1990) and C. C. Vaughn (pers. comm.) found a similar trend in another Ozarkian stream, the Illinois River in Oklahoma, and described the habitat most often associated with L. rafinesqueana in this stream as backwater areas. The rarity of L. rafinesqueana in mid-channel flows in Shoal Creek might be due to greater disruptions of substratum, especially during spates, than in other streams. Despite L. rafinesqueana’s apparent inability to colonize extremely unstable habitats, this species seemed more adapted to unstable habitats than most other unionids. We observed that individuals of L. rafinesqueana in the Spring River and Shoal Creek often had their foot well extended into the substratum, especially in loose gravel. Kat (1982) similarly noted that Elliptio complanata (Lightfoot, 1786) used its foot to maintain a viable position in unstable habitats (i. e. soupy mud). Foot extension of L. rafinesqueana was seldom observed, how- ever, in prairie streams of the Neosho and Verdigris basins and in the North Fork Spring River. Foot anchoring by L. rafinesqueana in the Spring River basin was probably due to swifter average current speeds in the Spring River basin versus the Neosho and Verdigris River basins (Table 2), and because substrata in the Spring River and Shoal Creek were less compacted than those in the prairie streams mentioned. Lampsilis rafinesqueana has apparently become extirpated from six Kansas streams: the Elk, Caney, Cottonwood, and South Fork of the Cottonwood Rivers, and Shoal and Middle Creeks (Cope, 1979, 1985; Metcalf, 1980; Obermeyer, 1996b; Obermeyer et al., 1997), and cur- rently remains in the Verdigris, Fall, Neosho, and Spring Rivers (Cope, 1979, 1983, 1985; Miller, 1993; Obermeyer et al., 1995; 1997) (Fig. 2). In addition to its range decline in Kansas, L. rafinesqueana seems to have become less abundant (Obermeyer et al., 1997). It is presently listed in Kansas as endangered. In Missouri, L. rafinesqueana is confined to the Spring and Elk River basins (Gordon and Brown, 1980; Johnson, 1980; Oesch, 1984; Clarke and Obermeyer, 1996), and is state-listed as rare (Anonymous, 1994). Despite the decline of L. rafinesqueana in Kansas, the relative abundance of this species could have increased in the Spring River since Branson’s (1967) survey. For OBERMEYER &£T AL.: UNIONIDAE OF NEOSHO, VERDIGRIS, AND SPRING RIVER BASINS 51 example, Branson collected only 15 “muckets” at one of our Spring River sites in Kansas, whereas we found 112 living L. rafinesqueana [Note: His identifications of Actinonaias ligamentina (Barnes, 1823) were likely L. rafinesqueana]. Presently, the species is probably more abundant in the Spring River from Carthage, Missouri, to near the confluence of Center Creek, Kansas, than any- where else throughout its range. Although L. rafinesqueana appears to remain within most of its historic range in Missouri, Branson (1967) recovered the species at sites far- ther upstream in both the Spring River and Shoal Creek, perhaps indicating a slight decrease in range. In the Kansas portion of the Spring River, L. rafinesqueana, as well as most other riverine mussel species, is apparently extirpated downstream from Turkey Creek (Fig. 2). Although L. rafinesqueana was previously reported as absent in the Spring River downstream from the confluence of Center Creek (Cope, 1985; Stewart, 1992; Obermeyer et al., 1995), we collected this species at two riffle sites immedi- ately downstream from Center Creek, which may indicate improving stream conditions. Lampsilis rafinesqueana is a bradytictic breeder (Barnhart and Roberts, 1997), and females, like other lamp- silines, attract potential hosts with a mantle lure (Johnson, 1980; Oesch, 1984; Barnhart and Roberts, 1997), with July and August being the period of most frequent mantle dis- play (B. K. Obermeyer, pers. obs.). Two potential hosts have been identified for L. rafinesqueana: smallmouth bass (Micropterus dolomieu Lacépéde, 1802) and largemouth bass [M. salmoides (Lacépéde, 1802)] (Barnhart and Roberts, 1997); M. C. Barnhart (pers. comm.) suspects that spotted bass [M. punctulatus (Rafinesque, 1819)] would also serve as a host, although this species has not been tested. Ptychobranchus occidentalis has experienced the largest reduction in range in Kansas of the five candidate species targeted in this study (Obermeyer et al., 1997). Ten Kansas streams have historic records for this species: the Cottonwood, South Fork of the Cottonwood, Elk, Fall, Caney, Neosho, Spring, and Verdigris Rivers, and Otter (Greenwood County) and Cedar (Chase County) Creeks (Popenoe, 1885; Call, 1885c, 1885d, 1886; Scammon, 1906; Isely, 1924; Cope, 1979, 1985; Metcalf, 1980; Obermeyer et al., 1995, 1997). However, extant representa- tives have been recovered recently from only four Kansas streams: the Neosho, Spring, Fall, and Verdigris Rivers (Branson, 1966a, 1967; Frazier, 1977; Liechti and Huggins, 1977; Schuster, 1979; Schuster and DuBois, 1979; Cope, 1979, 1983, 1985; Miller, 1993; Obermeyer et al., 1995, 1997). Furthermore, P. occidentalis might have recently become extirpated from the Neosho River (Obermeyer et al., 1995). In the Spring River basin, P. occidentalis was uncommon. This observation agrees with the contention of Buchanan (1980) and Oesch (1984) that P. occidentalis, although widely distributed in the southern half of Missouri, is uncommon at any one locale. Presently, P. occidentalis is listed as threatened in Kansas and as a watch species (i.e. a species of concern) in Missouri (Anonymous, 1994). Ptychobranchus occidentalis is a bradytictic breeder (Johnson, 1980; Barnhart and Roberts, 1997) that releases mimetic larval packets from pleated marsupial gills in early spring (Barnhart and Roberts, 1997). The orangethroat [Etheostoma spectabile (Agassiz, 1854)], greenside (E. blennioides Rafinesque, 1819), yoke (E. juliae Meek, 1891), and rainbow darter (E. caeruleum Storer, 1845) have been identified as potential hosts (Barnhart and Roberts, 1997). Of these four species, only two species are found in the study area; the greenside darter is found in the Spring River basin, whereas the orangethroat darter is found throughout the study area (Pflieger, 1975; Cross and Collins, 1995). Cyprogenia aberti occurred historically in Kansas in the Fall, Elk, Verdigris, Neosho, Spring Rivers (Popenoe, 1885; Call, 1885a, b, 1886, 1887a; Scammon, 1906; Murray and Leonard, 1962; Obermeyer et al., 1995, 1997); however, it has been found recently in only the Verdigris, Fall, and Spring Rivers (Branson, 1966b; Liechti and Huggins, 1977; Cope, 1979, 1985; Miller, 1993; Ober- meyer et al., 1995, 1997), and is considered rare with a patchy distribution (Obermeyer et al., 1997). Oesch (1984) and Harris and Gordon (1987) reported that in Ozarkian streams of Missouri and Arkansas, C. aberti was locally abundant, especially in the Spring (White River system) and Caddo (Ouachita River system) Rivers in Arkansas (Harris and Gordon, 1987). However, we found the species uncommon in the Spring River basin. C. aberti is presently listed as endangered in Kansas and as rare in Missouri (Anonymous, 1994). Gravid females of Cyprogenia aberti have been found in all stages of reproductive development throughout the year (Call, 1887b); however, Chamberlain (1934) observed that the species released conglutinates in late win- ter, and M.C. Barnhart (pers. comm.) observed the periodic release of conglutinates during winter and spring months. Preliminary findings by M. C. Barnhart (pers. comm.) indi- cate that, from a total of 24 species infected, the banded sculpin [Cottus carolinae (Gill, 1861)], log perch [Percina caprodes (Rafinesque, 1818)], and fantail darter (Etheostoma flabellare Rafinesque, 1819) are suitable hosts, based on Spring River C. aberti specimens. Another cited host for C. aberti is the goldfish [Carassius auratus (Linnaeus, 1758)] (Watters, 1994), based on Chamberlain (1934); however, although glochidial cysts on goldfish were noted for up to five hours following that fish’s inges- tion of C. aberti conglutinates, subsequent examinations of glochidial cysts (i. e. after 5 h) were not made to determine 32 AMER. MALAC. BULL. 14(1) (1997) host suitability. Oesch (1984) reported that Quadrula cylindrica is restricted in Missouri to the Black, St. Francis, and Spring Rivers; it also occurred historically in Center Creek (Utterback, 1915) as well as in Shoal Creek (Clarke and Obermeyer, 1996). In Kansas, Obermeyer et al. (1997) stated that Q. cylindrica’s continued persistence in the state is questionable, with extant representatives limited to a few locales in the Spring and Neosho Rivers (Cope, 1985; Obermeyer et al., 1997). Its current distribution contrasts greatly with its past presence in the Neosho, Cottonwood, Spring, Verdigris, and Fall Rivers as well as in Shoal Creek (Popenoe, 1885; Call, 1885b, d; Scammon, 1906; Isely, 1924; Obermeyer et al., 1997). Clarke and Obermeyer (1996) remarked that the species has exhibited a similar trend of decline throughout most of its range in eastern North America. Q. cylindrica is presently listed as endan- gered in both Kansas and Missouri. Knowledge of the reproductive biology of Quadrula cylindrica is based largely on Tennessee populations of Q. c. Strigillata, except for brief breeding records by Utterback (1915) and Ortmann (1919). Yeager and Neves (1986) found Q. c. strigillata to be tachytictic, with the bigeye chub [Notropis amblops (Rafinesque, 1820)], spotfin shiner [Cyprinella spiloptera (Cope, 1868)], and whitetail shiner [C. galactura (Cope, 1868)] potential hosts based on artifi- cial infestations. Although differences between Q. c. cylin- drica and Q. c. strigillata could be due to phenotypic plas- ticity (Gordon and Layzer, 1989; Clarke and Obermeyer, 1996), it is possible that host specificity varies between eastern populations, especially of Q. c. strigillata, and those further west, such as in Kansas and Missouri. Further evi- dence of host differences is suspected because in the Neosho River, where small populations of this species remain, suitable hosts identified by Yeager and Neves (1986) are believed to be absent (Cross, 1967; F. B. Cross, pers. comm.). Oesch (1984) described Alasmidonta marginata as being widely distributed in the southern half of Missouri, but noted it is uncommon at any one locale. A. marginata was first documented in Kansas by Branson (1966a), who found three living specimens in the Spring River in 1964. Although additional specimens have since been collected in the Spring River (Obermeyer et al., 1995), the only other stream record for A. marginata in Kansas is from the Marais des Cygnes River in east-central Kansas, based on a recently dead specimen in Franklin County collected in 1983 (Distler and Bleam, 1987) and a relic valve found in Osage County (Obermeyer, 1996b). The recovery of only one living A. marginata and the rarity of fresh shell materi- al in the present study raises concern for the species because earlier surveyors (Branson, 1967; Cope, 1985) found the species in the Spring River in Kansas more fre- quently and at more sites. It is listed in Kansas as endan- gered, but is not currently listed in Missouri. Five potential hosts have been identified for Alasmidonta marginata (fide Howard and Anson, 1923), which is a bradytictic breeder (Ortmann, 1919; Oesch, 1984; Watters, 1994). These are the northern hog sucker [Hypentelium nigricans (Lesueur, 1817)], rock bass [Ambloplites rupestris (Rafinesque, 1817)], shorthead red- horse [Moxostoma macrolepidotum (Lesueur, 1817)], war- mouth [Lepomis gulosus (Cuvier, 1829)], and white sucker [Catostomus commersoni (Lacépéde, 1803)], all of which occur in the Spring River basin (Pflieger, 1975; Cross and Collins, 1995). CONCLUSIONS Clarke and Obermeyer (1996) rcommended federal listing status for four of the five species targeted in this study; that is, all except Alasmidonta marginata, which is on the western edge of its range in the study area and is widely distributed throughout much of eastern North America (Clarke, 1981). Clarke and Obermeyer (1996) stressed the need for habitat conservation as the key ele- ment for future protection of these species, and regarded preservation of the Spring River mussel fauna a top priority because of its diverse mussel assemblage. Furthermore, they considered the Spring River a possible refuge from the impending threat of Dreissena polymorpha (Pallas, 1771) (French, 1990; Ludyanskiy et al., 1993) because of the rari- ty of headwater impoundments, which can function as upstream sources for Dreissena veligers (McMahon, 1991). They also regarded the Spring River an important refuge because of its tolerance of droughts due to spring-fed flows. We agree that the Spring River is an important resource to concerve, and believe that basin-wide recovery/protection plans [e. g. U. S. Fish and Wildlife Service (1994)] should be considered to help protect remaining mussel assem- blages in not the Spring River, but in the Neosho, Verdigris, and Fall River basins as well. Because populations of some of these unionids examined in this study could be ecologi- cally distinct despite exhibiting similar shell morphology with other populations (e. g. possible differences in host specificity for Quadrula cylindrica), unique biological enti- ties are perhaps being overlooked. Therefore, we believe it is important to identify and protect these disjunct and/or distinct populations. We also recommend additional testing of potential fish hosts for these and other species, perhaps on a drainage-by-drainage basis. ACKNOWLEDGMENTS Principal financial support came from the Kansas Department of OBERMEYER EFT AL.: UNIONIDAE OF NEOSHO, VERDIGRIS, AND SPRING RIVER BASINS — 53 Wildlife and Parks (KDWP) through USFWS Section-6 monies; however, additional support was provided by ECOSEARCH, Inc., KDWP Chickadee Checkoff funds, Kansas commercial mussel harvest permit monies, an Emporia State University Faculty Research and Creativity Grant, and the Kansas Biological Survey. The following helped with labor-intensive field sampling: Lewis Anderson, Bill Browning, Bill Busby, Myron Frans, Bob Funke, Linda Fuselier, Don George, Eugene Goff, Jerry Horak, Jim Minnerath, Dan Mulhern, Bernadine Obermeyer, Jim Peterson, Kevin Ricke, Tom Swan, Vernon Tabor, Rick Tush, Dan VanLeeuwen, Kenny Whitehead, Isaac Whitehead, and participants of the 1993 KDWP mussel sampling workshop. Searches of Kansas unionid records were performed by Bob Angelo, Craig Thompson, Bill Busby, and Jerry Horak. Helpful advice and suggestions for this project was generous- ly provided by, to name a few, Chris Barnhart, Bill Busby, Arthur H. Clarke, Charles Cope, Don Distler, Tom Mosher, Richard Neves, Larry Scott, David H. Stansbery, David Strayer, and Caryn Vaughn. Voucher specimens were accepted by David H. Stansbery and Kathy Borror of Ohio State University. Ken Brunson, Lanny Jones, Juanita Bartley, and Pam Fillmore assisted in grant administration. Equipment loans were pro- vided by Roger Ferguson, Jim Peterson, John Bills, Carson Cox, Jerry Horak, and Rick Tush. LITERATURE CITED Aber, J. S. 1992. Chert gravel, drainage development, and sinkholes in the Walnut Basin, southcentral Kansas. Transactions of the Kansas Academy of Science 95:109-121. Anonymous. 1994. Rare and Endangered Species Checklist of Missouri. Missouri Department of Conservation, Natural History Database, Jefferson City, Missouri. 31 pp. Barmhart, M. C. and A. Roberts. 1997. Reproduction and fish hosts of unionids from the Ozark Uplifts. Jn: Conservation and Management of Freshwater Mussels II: Initiatives for the Future, K. S. Cummings, A. C. Buchanan, C. A. Mayer, and T. J. Naimo, eds. Proceedings of UMRCC [Upper Mississippi River Conservation Committee] Symposium, 16-18 October 1995, St. Louis, Missouri. pp. 15-20. Upper Mississippi River Conservation Committee, Rock Island, Illinois. Bogan, A. E. and P. W. Parmalee. 1983. Tennessee’s Rare Wildlife, Vol. 2, The Mollusks. Tennessee Wildlife Resources Agency, Nashville. 123 pp. Branson, B. A. 1966a. Alasmidonta marginata and Ptychobranchus fasci- olaris in Kansas. The Nautilus 80:21-24. Branson, B. A. 1966b. Unionid records from Kansas, Arkansas, and Louisiana. Sterkiana 23:7-8. Branson, B. A. 1967. A partial survey of the Spring River in Kansas, Oklahoma, and Missouri, Part I: collecting sites, basic limnologi- cal data, and mollusks. Transactions of the Kansas Academy of Science 69:242-293. Buchanan, A. C. 1980. Mussels (Naiades) of the Meramec River Basin, Missouri. Aquatic Series 17, Missouri Department of Conservation, Jefferson City, Missouri. 68 pp. Call, R. E. 1885a. Description of a new species of Unio from Kansas. Bulletin of the Washburn College Laboratory of Natural History 1(1):48-49. Call, R. E. 1885b. Contributions to a knowledge of fresh-water Mollusca of Kansas, I: freshwater bivalves. Bulletin of the Washburn College Laboratory of Natural History 1(2):48-51. Call, R. E. 1885c. Contributions to a knowledge of fresh-water Mollusca of Kansas, III: fresh-water bivalves. Bulletin of the Washburn College Laboratory of Natural History 1(3):93-97. Call, R. E. 1885d. Contributions to a knowledge of fresh-water Mollusca of Kansas, IV. Bulletin of the Washburn College Laboratory of Natural History \(4):115-124. Call, R. E. 1886. Fifth contribution to a knowledge of fresh-water Mollusca of Kansas. Bulletin of the Washburn College Laboratory of Natural History \(6):177-184. Call, R. E. 1887a. Sixth contribution to a knowledge of fresh-water Mollusca of Kansas. Bulletin of the Washburn College Laboratory of Natural History 2(8):11-25. Call, R. E. 1887b. Note on the ctenidium of Unio aberti Conrad. American Naturalist 21:857-860. Chamberlain, T. K. 1934. The glochidial congluninates of the Arkansas fanshell, Cyprogenia aberti (Conrad). Biological Bulletin 66:55- 61. Clarke, A. H. 1981. The tribe Alasmidontini (Unionidae: Anodontinae), Part 1: Pegias, Alasmidonta, and Arcidens. Smithsonian Contributions to Zoology 326. 101 pp. Clarke, A. H. 1986. The mesoconch: a record of juvenile life in Unionidae. Malacology Data Net 1:21-36. Clarke, A. H. and C. O. Berg. 1959. The freshwater mussels of central New York with an illustrated key to the species of northeastern North America. Cornell University Agricultural Experimental Station Memoir 367: 1-79. Clarke, A. H. and B. K. Obermeyer. 1996. A Survey of Rare and Possibly Endangered Freshwater Mussels (Mollusca: Unionidae) of the Spring River Basin (with Observations on the Elk River Basin) in Missouri. Unpublished, U. S. Fish and Wildlife Service, Contract No. 60181-2-1621. 34 pp. + appendices. Cope, C. H. 1979. Survey of the Unionidae Considered for Conservation Status in Kansas. Unpublished report, Kansas Fish and Game Commission, Pratt. 39 pp. Cope, C. H. 1983. Kansas Freshwater Mussel Investigation, NMFS Project 2-378-R. Unpublished report, Kansas Fish and Game Commission, Pratt. 126 pp. Cope, C. H. 1985. The Spring River Drainage Basin: Kansas Resource in Need of a Management Plan. Unpublished report, Kansas Fish and Game Commission, Pratt. 96 pp. Cross, F. B. 1967. Handbook of Fishes of Kansas. Museum of Natural History Publication 45, University of Kansas, Lawrence. 357 pp. Cross, F. B. and J. T. Collins. 1995. Fishes in Kansas, 2nd ed., rev. Museum of Natural History, University of Kansas, Lawrence. Public Education Series 14. 315 pp. Cummings, K. S. and C. A. Mayer. 1992. Field Guide to Freshwater Mussels of the Midwest. Manual 5, Illinois Natural History Survey, Champaign. 194 pp. Davis, J. V. and J. G. Schumacher. 1992. Water-Quality Characterization of the Spring River Basin, Southwestern Missouri and Southeastern Kansas. Water-Resources Investigations Report 90- 4176, U. S. Geological Survey, Rolla, Missouri. 112 pp. Deacon, J. E. 1961. Fish populations, following a drought, in the Neosho and Marais des Cygnes rivers of Kansas. University of Kansas Publication, Museum of Natural History 13:359-427. Distler, D. A. and D. E. Bleam. 1987. Records of two invertebrate species in Kansas. Transactions of the Kansas Academy of Science 90:158. Frazier, J. A. 1977. Unionid Mussels of the Neosho River Drainage. Master’s Thesis, Emporia State University, Emporia, Kansas. 32 pp. French, J. R. P. 1990. The exotic zebra mussel - a new threat to endan- gered freshwater mussels. Endangered Species Technical Bulletin 15:3-4, Geiger, C. O., D. L. Lacock, D. R. Schneider, M. D. Carlson, and B. J. Dague. 1995. Water Resources Data - Kansas, Water Year 1994. 54 AMER. MALAC. BULL. 14(1) (1997) U. S. Geological Survey Water-Data Report KS-94-1, Lawrence, Kansas. 479 pp. Gordon, M. E. and A. V. Brown. 1980. Significant additions to the mol- luscan fauna of the Illinois River, Arkansas. Proceedings of the Arkansas Academy of Science 34:113-114. Gordon, M. E. and J. B. Layzer. 1989. Mussels (Bivalvia: Unionoidea) of the Cumberland River: review of life histories and ecological rela- tionships. U. S. Fish and Wildlife Service Biological Report 89(15), 99 pp. Harris, J. L. and M. E. Gordon. 1987. Distribution and status of rare and endangered mussels (Mollusca: Margaritiferidae, Unionidae) in Arkansas. Proceedings of the Arkansas Academy of Science 41:49-56. Hoeh, W. R. 1990. Phylogenetic relationships among eastern North American Anodonta (Bivalvia: Unionidae). Malacological Review 23:63-82. Holland-Bartels, L. E. 1990. Physical factors and their influence on the mussel fauna of a main channel border habitat of the upper Mississippi River. Journal of the North American Benthological Society 9:327-335. Howard, A. D. and B. J. Anson. 1923. Phases in the parasitism of the Unionidae. Journal of Parasitology 9:68-82. Isely, F. B. 1911. Preliminary note on the ecology of the early juvenile life of the Unionidae. Biological Bulletin 20:77-80. Isely, F. B. 1924. The fresh-water mussel fauna of eastern Oklahoma. Proceedings of the Oklahoma Academy of Science 4.43-118. Johnson, R. I. 1980. Zoogeography of North American Unionacea (Mollusca: Bivalvia) north of the maximum Pleistocene glacia- tion. Bulletin of the Museum of Comparative Zoology 149:77- 189. Kansas Department of Health and Environment. 1980. Water Quality Investigations of Lead-Zinc Mine Drainage Effects on the Spring River and Associated Tributaries in Kansas, 1978-1979. Water Quality Management Section, Topeka, Kansas. 42 pp. Kat, P. W. 1982. Effects of population density and substratum type on growth and migration of Elliptio complanata (Bivalvia: Unioni- dae). Malacological Review 15:119-127. Liechti, P. M. and D. G. Huggins. 1977. Unionacean mussels of Kansas. Technical Publications of the State Biological Survey of Kansas 4:17-30. Ludyanskiy, M. L., D. McDonald, and D. MacNeill. 1993. Impact of the zebra mussel, a bivalve invader. BioScience 43:533-542. Mather, C. M. 1990. Status Survey of the Western Fanshell and the Neosho Mucket in Oklahoma. Unpublished report, Oklahoma Department of Wildlife Conservation, Oklahoma City. 22 pp. + appendices. McMahon, R. F. 1991. Mollusca: Bivalvia. Jn: Ecology and Classification of North American Freshwater Invertebrates, J. H. Thorp and A. P. Covich, eds. pp. 315-399. Academic Press, San Diego, California. Metcalf, A. L. 1980. Unionacean mussels, past and present, from six streams in Kansas and Oklahoma. Transactions of the Kansas Academy of Science 83:1-19. Miller, E. J. 1993. Evaluation of Verdigris River, Kansas, freshwater mus- sel refuge. In: Conservation and Management of Freshwater Mussels, K. S. Cummings, A. C. Buchanan, and L. M. Koch, eds. Proceedings of a UMRCC Symposium, 12-14 October 1992, St. Louis, Missouri. pp. 56-60. Upper Mississippi River Conservation Committee, Rock Island, Illinois. Miller, E. J. and B. K. Obermeyer. 1997. Population increase of Quadrula metanevra in southeast Kansas. In: Conservation and Manage- ment of Freshwater Mussels II: Initiatives for the Future, K. S. Cummings, A. C. Buchanan, C. A. Mayer, and T. J. Naimo, eds. Proceedings a 2nd UMRCC Symposium, 16-18 October 1995, St. Louis, Missouri. pp. 30-36. Upper Mississippi River Conservation Committee, Rock Island, Illinois. Murray, H. D. and A. B. Leonard. 1962. Handbook of the Unionid Mussels in Kansas. University of Kansas, Lawrence, Museum of Natural History, Miscellaneous Publication 28. 184 pp. Neves, R. J. and J. C. Widlak. 1987. Habitat ecology of juvenile freshwa- ter mussels (Bivalvia: Unionidae) in a headwater stream in Virginia. American Malacological Bulletin 5(1):1-7. Obermeyer, B. K. 1996a. Unionidae (Bivalvia) of the Arkansas River System of SE Kansas and SW Missouri: Species of Concern, Historical Change, Commercial Harvest, and Sampling Methods. Master’s Thesis, Emporia State University, Emporia, Kansas. 131 PP: Obermeyer, B. K. 1996b. Freshwater Mussels (Bivalvia: Unionidae) Collected by KDWP Stream Survey Teams, 1995-1996. Un- published report, Kansas Department of Wildlife and Parks, Pratt. 15 pp. + appendices Obermeyer, B. K., D. R. Edds, and C. W. Prophet. 1995. Distribution and Abundance of Federal “Candidate” Mussels (Unionidae) in Southeast Kansas. Unpublished report, Kansas Department of Wildlife and Parks, Pratt. 76 pp. + appendices. Obermeyer, B. K., D. R. Edds, E. J. Miller, and C. W. Prophet. 1997. Range reductions of southeast Kansas unionids. Jn: Conservation and Management of Freshwater Mussels II: Initiatives for the Future, K. S. Cummings, A. C. Buchanan, C. A. Mayer, and T. J. Naimo, eds. Proceedings of a UMRCC Symposium, 16-18 October 1995, St. Louis, Missouri. pp. 108-116. Upper Mississippi River Conservation Committee, Rock Island, Illinois. Oesch, R. D. 1984. Missouri Naiades: a Guide to the Mussels of Missouri. Missouri Department of Conservation, Jefferson City. vii + 270 Pp. Ortmann, A. E. 1919. Monograph of the naiades of Pennsylvania, Part III: systematic account of the genera and species. Memoirs of the Carnegie Museum 8:1-385, 21 pls. Ortmann, A. E. 1920. Correlation of shape and station in freshwater mus- sels (Naiades). American Philosophical Society Proceedings 59:269-312. Parmalee, P. W. 1967. The Fresh-Water Mussels of Illinois. Illinois State Museum Popular Science Series 8. 108 pp. Pflieger, W. L. 1975. The Fishes of Missouri. Missouri Department of Conservation, Jefferson City. 343 pp. Platts, W. S., W. F. Megahan, and G. W. Minshall. 1983. Methods for Evaluating Stream, Riparian, and Biotic Conditions. Forest Service Technical Report 580-584, U. S. Government Printing Office. 70 pp. Popenoe, E. A. 1885. List of Unionidae collected in Kansas rivers, with localities. Transactions of the Kansas Academy of Science 9:78- 79. Robins, C. R., R. M. Bailey, C. E. Bond, J. R. Brooker, E. A. Lachner, R. N. Lea, and W. B. Scott. 1991. Common and Scientific Names of Fishes from the United States and Canada, 5th ed. American Fisheries Society Special Publication 20, Bethesda, Maryland. 183 pp. Scammon, R. E. 1906. The Unionidae of Kansas, Part I: an illustrated cat- alogue of the Kansas Unionidae. University of Kansas Science Bulletin 3:279-373, pls. 52-86. Schuster, G. A. 1979. Notes on the freshwater mussel fauna of the Verdigris River system in Kansas. Transactions of the Kansas Academy of Science 82:1 1-24. Schuster, G. A. and M. B. DuBois. 1979. Additional new records of fresh- water mussels (Bivalvia: Unionidae) from Kansas. Technical Publications of the State Biological Survey of Kansas 8:1-11. OBERMEYER &T AL.: UNIONIDAE OF NEOSHO, VERDIGRIS, AND SPRING RIVER BASINS — 55 Simpson, C. T. 1914. A Descriptive Catalogue of the Naiades, or Pearly Fresh-Water Mussels. Bryant Walker, Detroit, Michigan. 1540 Pp. Stansbery, D. H. 1974. An environmental survey of several groups of aquatic macroinvertebrates of the proposed Paint Creek impound- ment area. In: Environmental Analysis of the Paint Creek Lake Project, Ohio, D. H. Stansbery and C. E. Herdendorf, eds. pp. 195-252. Unpublished report DAC W69-73-C-0004, Department of Army, Huntington District, Corps of Engineers, Huntington, West Virginia. Stewart, J. H. 1992. Status Review of the Neosho Mucket, Lampsilis rafinesqueana. U. S. Fish and Wildlife Service, Jackson, Mississippi. 3 pp. Stewart, J. H. 1994. Status Review of the Western Fanshell, Cyprogenia aberti, U. S. Fish and Wildlife Service, Jackson, Mississippi. 3 Pp. Strayer, D. L. 1981. Notes on the microhabitats of unionid mussels in some Michigan streams. American Midland Naturalist 106:411- 415. Strayer, D. L. and J. Ralley. 1993. Microhabitat use by an assemblage of stream-dwelling Unionaceans (Bivalvia), including two rare species of Alasmidonta. Journal of the North American Benthological Society 12:247-258. Strayer, D. L., D. C. Hunter, L. C. Smith, and C. K. Borg. 1994. Distribution, abundance, and roles of freshwater clams (Bivalvia, Unionidae) in the freshwater tidal Hudson River. Freshwater Biology 31:239-248. Turgeon, D. D., A. E. Bogan, E. V. Coan, W. K. Emerson, W. G. Lyons, W. L. Pratt, C. F. E. Roper, A. Scheltema, F. G. Thompson, and J. D. Williams. 1988. Common and Scientific Names of Aquatic Invertebrates from the United States and Canada: Mollusks. American Fisheries Society Special Publication 16, Bethesda, Maryland. 277 pp. U. S. Fish and Wildlife Service. 1994. Technical/Agency Draft Mobile River Basin Ecosystem Recovery Plan. U.S. Fish and Wildlife Service, Jackson, Mississippi. 128 pp. Utterback, W. I. 1915. The naiades of Missouri-II. American Midland Naturalist 4:97-152. Valentine, B. D. and D. H. Stansbery. 1971. An introduction to the naiads of the Lake Texoma region, Oklahoma, with notes on the Red River fauna (Mollusca: Unionidae). Sterkiana 42:1-40. Watters, G. T. 1994. An Annotated Bibliography of the Reproduction and Propagation of the Unionoidea (Primarily of North America). Ohio Biological Survey Miscellaneous Contribution 1. 158 pp. Williams, J. D., M. L. Warren, K. S. Cummings, J. L. Harris, and R. J. Neves. 1993. Conservation status of freshwater mussels of the United States and Canada. Fisheries 18:6-22. Wilson, F. 1984. Landscapes: a geologic diary. In: Kansas Geology: an Introduction to Landscapes, Rocks, Minerals, and Fossils. R. Buchanan, ed. pp. 9-39. University Press of Kansas, Lawrence. Yeager, B. L. and R. J. Neves. 1986. Reproductive cycle and fish hosts of the rabbit’s foot mussel, Quadrula cylindrica strigillata (Mollusca: Unionidae) in the upper Tennessee River drainage. American Midland Naturalist 116:329-340. Date of manuscript acceptance: 12 May 1997 Growth and survival of juvenile rainbow mussels, Villosa iris (Lea, 1829) (Bivalvia: Unionidae), reared on algal diets and sediment Catherine M. Gatenby!, Bruce C. Parker,! and Richard J. Neves2 1Department of Biology, Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24061, U.S. A. 2National Biological Service, Virginia Cooperative Fish and Wildlife Research Unit, Department of Fisheries and Wildlife Sciences, Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24061, U.S. A. Abstract: To develop a suitable diet for rearing recently metamorphosed freshwater mussels, nine species of algae and one live bacterium were tested in various trialgal or bialgal and bacterium combinations. A substratum was included in the culture chambers with each treatment to facilitate pedal-feeding by juvenile mussels. Kaolin, an artificial substratum, and fine sediment were tested in combination with algae and without algae. An unfed (no algae or sub- stratum) control treatment also was tested. Juveniles fed a freshwater trialgal diet consisting of Neochloris oleoabundans Chantanachat and Bold, 1962, Bracteacoccus grandis Bischoff and Bold, 1963, Phaeodactylum tricornutum Bohlin, 1897, with fine sediment showed the best growth over time (140 d). These individuals achieved a mean shell length of 1747 um, and had 30.0% survival after 140 d. Other trialgal mixtures containing N. oleoabundans, Nitzschia acicularis (Kiitzing) Wm. Smith, 1853, and Cyclotella meneghiniana Kiitzing, 1844, enhanced growth more than a green trialgal mixture of Chlorella vulgaris Beyernik, 1890, Ankistrodesmus falcatus (Corda) Ralfs, 1848, and Chlamydomonas reinhardtii Dangeard, 1888. A diet of fine sediment alone sustained juveniles through 140 d; however, an additional food source such as algae was necessary to increase survival. Bacteria did not contribute appreciably to juvenile growth and survival. Juvenile mussels reared on commercial yeast diets survived only 8 d; juveniles reared on kaolin and algae survived 60 d. Key Words: juvenile freshwater mussels, algal diets, fine sediment, kaolin With nearly 70% of native freshwater mussel 1971; Webb and Chu, 1983; Tan Tiu et al., 1989), and more species (Unionidae) in decline and another dozen commer- importantly, that polyunsaturated fatty acids (PUFA’s) are cial species threatened by the invasion of the zebra mussel, essential to the early larval and juvenile life stages of most research aimed at the conservation of freshwater mussels fish and shellfish (Ando, 1968; Castell, 1970; Ackman, has become a priority in the United States (Williams et al., 1983; Watanabe et al., 1983; Webb and Chu, 1983; 1993). Culture and propagation of freshwater mussels for Napolitano et al., 1988, 1990). It is generally thought that stock enhancement, preservation of endangered species, PUFA’s are important functional components of cells and and creation of refugia for native populations from the membranes (Ackman, 1983), and that lipid storage prod- invading zebra mussel, exemplify current conservation ucts provide a cheap energy source. In addition, mixed efforts. Research reported here is from an ongoing study on diets of different species of algae provided better growth in the culture and propagation of freshwater mussels. marine bivalves than did quantities of any single food test- Previous studies in the United States and Europe on cultur- ed (Davis and Guillard, 1958; Walne, 1970; Epifanio, 1979; ing freshwater mussels have reported difficulties in main- Romberger and Epifanio, 1981; Enright et al., 1986a). taining juveniles beyond 4 wk (Lefevre and Curtis, 1910; Resident bacteria in aquatic systems also have been impli- Howard, 1917, 1922). One reason could be that the food cated as an important ingredient in the juvenile bivalve diet source used to rear juvenile mussels was inadequate. (Urban and Langdon, 1984; Crosby et al., 1990; Baldwin Unfortunately, little information is available on the suitabil- and Newell, 1991). ity of various foods for unionids other than the experiments Feeding in the sediments with foot ciliation is of Coker et al. (1921), Imlay and Paige (1972), and Hudson common among bivalves, especially in the juvenile life and Isom (1984) who showed that algae, detritus, and com- stage, until their gills are fully developed for filter-feeding mercial fish food (as dissolved nutrients in the water) were (Allen, 1961, 1985; Morton, 1976; Lasee, 1991; Yeager et potential food sources for freshwater mussels. al., 1994). Thus, juvenile bivalves can derive nutrition Numerous studies in aquaculture have shown that from detrital material adhered to or adsorbed on the surface algae are the principal food of marine shellfish (Ukeles, of sediment particles. In an earlier study, pedal-feeding American Malacological Bulletin, Vol. 14(1) (1997):57-66 57 58 AMER. MALAC. BULL. 14(1) (1997) behavior by juveniles of two freshwater mussel species was reported by Gatenby et al. (1996). Specific research objec- tives of our study, therefore, were to identify a suitable diet for rearing juvenile freshwater mussels, to examine the nutritional role of riverine sediment and bacteria, and to compare growth and survival of juvenile mussels reared on different diets after metamorphosis. MATERIALS AND METHODS The rainbow mussel, Villosa iris (Lea, 1829), was selected for this study because it is common in Virginia, and like most endangered freshwater mussel species, is found in fast-moving riverine environments. Recently metamorphosed juveniles were fed various combinations of nine species of algae, including six green algae and three diatom species, one bacterium, yeast, kaolin, and riverine sediment (Table 1). Juvenile mussels can contain fatty reserves from their parasitic life stage that can allow them to live 2 wk post-metamorphosis without food supplementa- tion (Lasee, 1991). We decided, therefore, that we would monitor growth and survival in mussels reared on our diets for at least 45 d post-metamorphosis. We were able, howev- er, to maintain juvenile mussels in laboratory culture much longer than 45 d. Thus, three tests were performed on data collected at 60, 100, and 140 d post-metamorphosis. Gut contents of juveniles and river sediment were examined for algae and bacteria. Pedal-feeding behaviors also were observed, and the number of days post-metamorphosis that juveniles employed pedal-feeding was estimated. Newly metamorphosed mussels were obtained from host-fish infestations in the laboratory (Zale and Neves, 1982; Gatenby et al., 1996). Gravid Villosa iris were collected from Copper Creek, Clinch River drainage, Scott County, Virginia. The host fish [rock bass, Ambloplites rupestris (Rafinesque, 1817), and largemouth bass, Micropterus salmoides (Lacépéde, 1802)], 15-30 cm in length (Table 1) were collected from Tom’s Creek, New River drainage, Montgomery County, Virginia, treated for parasites, and acclimated in the laboratory prior to infesta- tion with glochidia (Zale and Neves, 1982; Neves et al., 1985). River water and sediment were collected from the New River, Montgomery County, Virginia. Water hardness (CaCO3) was 55.0 mg/l, and pH was 7.6. River water was filtered using a 4.25 cm diameter Whatman Glass Microfiber Filter to remove particles > 0.45 um. Sediment was maintained in the laboratory under aeration, and when needed, it was sieved through a 130 ym mesh screen before being added to the juvenile culture chamber. Thus, the sieved sediment contained the natural assemblage of microorganisms and other organic material associated with Table 1. Summary of diets tested for rearing juvenile freshwater mussels of Villosa iris. (CAC = Chlorella vulgaris, Ankistrodesmus falcatus, and Chlamydomonas reinhardtii; NNiC = Neochloris oleoabundans, Nitszchia acicularis, and Cyclotella meneghiniana; NOC = N. oleoabun- dans, Oocystis marsonii, and C. menenghiana; NPB = N. oleoabundans, Phaeodactylum tricornutum, and Bracteacoccus grandis; NPE = N. oleoabundans, P. tricornutum, and Enterobacter aerogenes; Sed = fine sediment). Treatment (replicates) Juveniles per replicate Unfed (1) 45 Kaolin-only (3) 50, 50, 50 Sed-only (2) 71, 70 Yeast/Sed (3) 100, 100, 100 CAC/Kaolin (2) 92, 100 CAC/Sed (3) 200, 200, 215 NOC/Sed (3) 100, 100, 100 NNiC/Sed (3) 111, 100, 103 NPB/Sed (3) 100, 101, 108 NPE/Sed (3) 100, 103, 100 riverine sediments. Newly metamorphosed juveniles were transferred to replicated glassculture dishes, 8 cm in diameter, 5 cm in height, and filled with 175 ml of filtered river water. A slow stream of air was introduced into each of the static chambers by fixing an Eppendorf pipette to the end of vinyl air tubing. Approximately 25-40 ml (4 g dry wt) of fine sediment was added to the static chambers, with the exception of those treatments receiving kaolin. Fine sedi- ment covered an area of 50 cm2, was 1.0-4.0 mm in depth and loosely packed in the culture chamber. Sediment parti- cles we term fine” were defined as “fine sand to clay” by Wentworth (1922). Approximately 2-3 g of kaolin were added to the kaolin treatments at the onset of the experi- ment. Water temperature was monitored daily, water was changed weekly, and chambers received new sediment or kaolin weekly. About two-thirds of the substratum in any given culture dish was discarded and replaced with fresh substratum. Algae were cultured in media ideal for growth of each species (Tables 2, 3, and 4), under continuous cool white fluorescent light of 60-100 »mol/m2/s photon flux at 20 = 1°C. All media were autoclaved at 121°C, >15 psi, for 20 min, then cooled prior to being innoculated with algae. Unialgal cultures were not axenic; however, the culture flasks were capped with cotton, cheese-cloth stoppers to prevent contamination and allow aeration through a sterile filter system. At near-maximum stationary phase, algal cells were counted in a hemacytometer. Cultures were then centrifuged at 7000-10,000 rpm for 25 min and the medium decanted. Algae were resuspended in tap water and kept under dark refrigeration for up to 2 wk. Approximately 1 g of a manipulated yeast diet (Artemia Reference Center, Ghent, Belgium) was mixed in GATENBY ET AL.: GROWTH OF VILLOSA IRIS 59 Table 2. Algae and bacteria species tested in feeding experiment, species sources, and growth medium used to culture species. (BMB = Bold’s Modified Bristol’s; OCM = Our Chlorella Medium). Species tested Species source Growth medium! Neochloris oleoabundans Chantanachat and Bold, 1962 Martek BioSciences Corp., Columbia, Maryland Neochloris” Bracteacoccus grandis Bischoff and Bold, 1963 University of Texas, Austin, Texas OCM Ankistrodesmus falcatus (Corda) Ralfs, 1848 Carolina Biological Supply, Burlington, North Carolina BMB4 Chlorella vulgaris Beyernik, 1890 Carolina Biological Supply BMB4 Oocystis marsonii Lemmermann, 1898 University of Texas ocm3 Chlamydomonas reinhardtii Dangeard, 1888 University of Texas BMB4+ Cyclotella meneghiniana Kiitzing, 1844 Nitzschia acicularis (Kiitzing) Wm. Smith, 1853 Phaeodactylum tricornutum Bohlin, 1897 Enterobacter aerogenes Kruse, 1896 1 Media formulations given in Table 3. 2McArdle et al., 1994. 3Behrens et al., 1989. 4Nichols, 1973. SDIFCO, Detroit, Michigan. 500 ml of tap water, the equivalent of 5-10 1 of live algae or 1.0 x 107 - 2.0.x 107 cells/ml. The bacterial culture was axenic and grown in nutri- ent broth media (Table 2), under continuous 60-100 umol/m2/s photon flux of white fluorescent light. The air temperature was approximately 20 + 1°C. Bacterial cell concentration was determined by plate counts. Bacteria were concentrated by centrifugation at 7000 rpm for 20 min; cells were separated from culture media, resuspended in tap water, and then kept alive under refrigeration for the duration of the experiment, with growth arrested at a con- centration of 1.0 x 108 cells/ml. Concentrated algae were administered daily to achieve treatment diets of 3.0 x 105 to 5.0 x 105 cells/ml in juvenile culture dishes. Nine treatments tested whether riverine sediment with associated native bacteria and other organic material provided nutritional value to juvenile mussels, whether bac- teria added to the culture diet enhanced growth and sur- vival, and whether various algae or a commercial yeast diet known to be high in lipids, could enhance growth and sur- vival of juvenile mussels. Three replicates of most treat- ments were tested; however, due to constraints on availabil- ity of juveniles, fewer replicates of some treatments were necessary (Table 1). The treatments included an unfed con- trol treatment (Unfed), Kaolin-only (USP K2-500, a fine clay product of Fisher Chemical), fine sediment (Sed-only), yeast with fine sediment (Yeast/Sed), Chlorella vulgaris, Ankistrodesmus falcatus, Chlamydomonas reinhardtii with kaolin (CAC/Kaolin), CAC with fine sediment (CAC/Sed), Neochloris oleoabundans, Oocystis marsonii, C. menegh- iniana with fine sediment (NOC/Sed), N. oleoabundans, Nitzschia acicularis, Cyclotella meneghiniana with fine sediment (NNiC/Sed), N. oleoabundans, Bracteacoccus grandis, Phaeodactylum tricornutum with fine sediment Loras College, Dubuque, Iowa Loras College University of Texas Biology Dept., Virginia Tech Bi-phasic/BMB4 Bi-phasic/BMB*+ Soil extract/BMB4 Nutrient Broth® Table 3. Neochloris (McArdle et al., 1994) and OCM (Our Chlorella Medium) (Behrens et al., 1989) growth medium formulations. Neochloris OCM Component Concentration Component Concentration NaCl 5.82 g/l KNO3 1.0 g/l MgSO,4°7H 0 2.47 g/l MsSO,4°7H70 5.0 ml, 100 g/l KNO3 1.0 g/l Ca(NO3)9°4H70 ~—. 2.5 ml, 25 g/l KCL 0.75 g/l KyHOP4 2.0 ml, 50 g/l FeCl, 1.0 ml, 0.81 g1 KyHOP, 1.5 ml, 50 g/l CaCly 2.0 ml, 43.9 g/l_ Fe - versenate* 2.0 ml/ Na EDTA 2.0 ml, 11.17 g/l *NayEDTA 4 g/l H3B03 1.0 ml, 12.36 g/l *FeSO4*7H 20 5 g/l Metals* 10.0 ml/ *Distilled HyO toll *FeCl3 20 ml, 33 g/l Metals* 1.0 m/l *ZnClo 20ml,1.1 gM **ZnSO4*7H,0 222 mg/ml *CuSO4°5SH70 2.0 ml, 5 g/l **CuSO4°5H,0 —- 79 mg/ml *H3B03 20 ml, 6 g/l **H3BO 2.86 g/l *MnCly°4H50 20 ml, 16 g/l **MnCly*4H70 1.81 gl *NayMoO04°2H7O 20mi,1.2g/ **NayMoO4*2H»0 390 mg/ml *CoCly*6H 0 2.0 ml, 4 g/l **Co(NO3)7°6H20 49.4 mg/ml *Na EDTA 4.0 g/ **Distilled HyO toll *Distilled HO toll Distilled HyO toll Vitamins** (add 50 ml to 1 1 media after autoclaving) **KH>PO4 2.0 ml, 54.4 g/l **ThiaminesHCl 0.1 ml, 1 mg/ml **Biotin 2.5 ml, 0.2 mg/ml **Vitamin B12 2.5 ml, 0.2 mg/ml Distilled HyO toll (NPB/Sed), and N. oleoabundans, P. tricornutum, Enterobacter aerogenes with fine sediment (NPE/Sed) (Table 1). Neither sediment nor kaolin was included in the particle concentration calculation; however, the bacterium E. aerogenes was included in the particle concentration for the diet NPE/Sed. All experiments were conducted on a 12:12 hr light:dark cycle, and each chamber contained 50- 100 juveniles (Table 1). 60 AMER. MALAC. BULL. 14(1) (1997) We examined growth and survival at 60 d and 100 d, and growth over 140 d post-metamorphosis. Juveniles in a culture dish were sieved out and removed to a Petri dish for measuring and assessing survival, weekly. Shell lengths of a random sample of 15 to 25 juveniles in each dish were measured using a calibrated ocular micrometer on a dissect- ing microscope. Randomness was achieved by moving the Petri dish on the dissecting stage and measuring the first animal that came into view. Juveniles were measured, counted, and returned to their culture dish. Thus, no juve- nile was measured twice, nor counted twice. Shell lengths at 60 and 100 d were compared using one-way analysis of variance (ANOVA) with nested replicates, followed by multiple contrast tests (Sokal and Rohlf, 1981). The survival data were compared by ANOVA (without nested replicates), followed by Tukey- Kramer multiple comparison. Arcsine transformation was applied to the percent survival data to satisfy the normality assumption. Multiple comparisons and contrast tests were evaluated at « < 0.01 (0.003 and 0.005 for the 60 and 100 d nested ANOVA’s) to control the overall experiment-wise error of a = 0.05 (Zar 1974; Sokal and Rohlf, 1981). Growth rates at 140 d post-metamorphosis were compared using analysis of covariance (ANCOVA), followed by a sequential analysis of the slopes to determine differences among treatments. Multiple comparisons were evaluated at Table 4. Bold’s Modified Bristol’s (BMB) (Nichols, 1973) and bi-phasic growth media formulations, and soil extract preparation. Component Concentration NaCl 10 ml, 1 g/400 ml MgSO,°7H70 10 ml, 3 g/400 ml NaNO3 10 ml, 3 g/400 ml K»yHPO4 10 ml, 3 g/400 ml KH P04 10 ml, 7 g/400 ml CaCly°2H70 10 ml, 1 g/400 ml Na ,EDTA*2H70 1 ml, 50 g/l KOH 1 ml, 31 g/l FeSO4°7H20 1 ml, 5 g/ H S04 1 ml, 1 ml/ H3BO03 1 ml, 11.42 g/l ZnSO4°7H70 1 ml, 8.82 g/l MnCl 5°4H50 1 ml, 1.44 g/l Mo03 1 ml, 0.71 g/l CuSO4*5H70 1 ml, 1.57 g/ Co(NO3)9°6H»0 1 ml, 0.49 g/l Soil extract 40 ml/l Autoclave soil (with water overlay) two times, at 121°C, > 15 psi, for 20 min. Decant soil water; preferably using Whatman 1 Filter. If still turbid, centrifuge to achieve a clear extract. Biphasic Place autoclaved river soil to a depth of about 3 cm, with Bold’s Basal media overlay. a = 0.01, so that the overall experimental error also was controlled at & = 0.05. Treatments without survivors were not included in any of the analyses. All statistical tests were calculated using JMP 2.0 (SAS Institute, Inc., 1991). Fluorescence microscopy, which induces a dis- tinctive red fluorescence of chlorophyll under blue light, was used to verify that juveniles were ingesting algae, and that the algae were well distributed throughout the substra- tum. Juveniles were placed on microscope slides and squashed under cover slips for examination. Locomotory and feeding behaviors of Villosa iris were observed using a dissecting microscope at the time of shell measurement and survival tabulation. Approximately 20 juveniles from each culture dish were observed for about 20 min each week. RESULTS The Unfed, Kaolin-only, Yeast/Sed, and NOC/Sed diets did not support growth beyond 35-40 d, and thus, were not included in the statistical analyses. Shell lengths of juveniles reared on all other diets were not significantly different (p = 0.4104) at 60 d. Shell lengths of juveniles at 100 d also were not significantly different (0.0830); howev- er, all juveniles reared on CAC/Kaolin were dead. Subsequent contrast t-tests at 100 d (a = 0.005) indicated that mean shell lengths of juveniles fed NPB/Sed (1354 um) was significantly greater (p < 0.0018) than those of juveniles fed on all other treatments. At 100 d, mean shell lengths of juveniles fed CAC/Sed (874 pm) and Sed-only (828 um) were similar (p = 0.6183), and mean lengths of juveniles fed CAC/Sed, NPE/Sed (978 um) and NNiC/Sed (1009 um) also were similar (Table 5). Juveniles fed NNiC/Sed and NPE/Sed, however, exhibited more growth at 100 d than juveniles in Sed-only (p = 0.0661 and 0.0556, respectively), although not significant at the contrast test level of a = 0.005. Individuals fed the better diet (NPB/Sed) achieved a maximum shell length of 1077 um at 60 d, and a maximum shell length of 1795 um at 100 d (Table 5). An analysis of covariance after 140 d confirmed that juveniles fed various diets exhibited significantly dif- ferent growth rates over time (p < 0.001) (Fig. 1). A sequential analysis of estimated slopes indicated that the growth rate of juveniles fed NPB/Sed was significantly greater than the growth rates recorded for all other treat- ments, and that all algal diets produced greater growth than a diet of fine sediment only (Table 6). Growth over time was similar for juveniles fed NPE/Sed and NNiC/Sed (p = 0.7626), and both diets marginally improved growth over the green algal diet, CAC/Sed (p = 0.0656 for NPE/Sed versus CAC/Sed and p = 0.0177 for NNiC/Sed versus GATENBY ET AL.: GROWTH OF VILLOSA IRIS 61 Table 5. Mean shell length (+ SD, with minimum-maximum in parentheses) and percent survival of Villosa iris juveniles fed various diets for 60 and 100 d post-metamorhosis. Initial lengths of juveniles were 263 + 37 um. p-values are given for the ANOVA test for treatment effect (a = 0.05). Means with the same superscripts (a, b) were similar according to contrast t-tests and Tukey-Kramer test; however, due to high variability within treat- ments in survival data, the ANOVA test at 60 d was sufficiently robust to detect differences among treatments. (Abbreviations as in Table 1). 60 Days Diet Length (um) Unfed total mortality Kaolin-only total mortality Sed-only 7104 + 108 (462-897) 50.8 Yeast/Sed total mortality CAC/Kaolin 7442 + 69 (641-795) 43 CAC/Sed 6584 +108 (500-886) 51.3 NOC/Sed total mortality NNiC/Sed 6384 + 69 (513-769) 13.6 NPB/Sed 7348 + 153 (564-1077) 33.7 NPE/Sed 6554 + 105 (513-923) 48.6 p-value 0.4104 0.4591 CAC/Sed) (Table 6). After 140 d, individuals fed NPB/Sed achieved a mean length of 1747 um and maximum length of 2359 um (Table 7). Survival at 60 d ranged from 4.4% for CAC/ Kaolin to 51.3% for CAC/Sed, and variability within a treatment was high (Table 5). Although 4.4% survival is drastically different from 51.3% survival and is probably different from survival for all other diets, except perhaps NNiC/Sed (13.6%), the ANOVA was of insufficient robust- ness to detect differences among treatments. Mean percent survival at 100 d for all treatments ranged from 20.7- 36.0%, and an ANOVA indicated no difference in percent survival among treatments (p = 0.7768) (Table 5). How- ever, one to two replicates within each treatment showed 0% survival; thus, with so few experimental units, no statis- tical test was sufficiently robust to adequately analyze our survival data. After 140 d, algae in combination with sedi- ment appeared to gain significance in enhancing survival. Mean survival at 140 d for individuals reared on algae with fine sediment was 24.5%, whereas mean survival for indi- viduals in fine sediment-only was 10% (Table 7). The Unfed treatment was unable to maintain juve- niles beyond 37 d, and the Kaolin-only treatment had no survivors at 38 d. The Yeast/Sed diet did not support juve- niles beyond 8 d, and all individuals fed NOC/Sed were dead by 46 d, with replicates 1 and 2 dead at 24 d. Almost all juveniles died in two replicates of NNiC/Sed between 49 and 65 d. Chemical analysis of the prepared diets indi- cated that NOC and NNiC contained 89.2 and 55.9 mg/I of potassium, respectively, and these levels could have been toxic to mussels (Imlay, 1973). Two tanks of NPB/Sed were lost at 80 d due to a waterline break. Juveniles were very active during the 20 min they Survival (%) 100 Days Length (um) Survival (%) total mortality total mortality 8280 + 75 (667-949) 31.12 total mortality total mortality 8745 + 138 (643-1143) 36.08 total mortality 1009 + 155 (769-1281) 20.78 13543 + 296 (1026-1795) 32.58 9780 + 231 (564-1795) 34.28 0.0830 0.7768 were observed each week. By protracting and retracting the foot, the juveniles moved through the substratum in a “tumble-like” fashion. While lying valve-side down, juve- niles continually used a foot-sweeping movement which carried a current of particles toward the pedal gape. The foot was highly ciliated, and some particles adhered to it, while larger particles were lost as the juvenile retracted its foot inside the valves. These behaviors have previously been described by Reid et al. (1992) and Yeager et al. (1994) for several marine and freshwater bivalves. Fluorescence microscopy indicated that juveniles reared on sediment/algae and kaolin/algae had ingested algae (Fig. 2A), and that the algae were well distributed throughout the substratum (Fig. 2B). Subsequent gut squashes showed par- NPB/Sed_ y = 193.07 + 10.443x RA2 = 0.943 NPE/Sed yy = 353.34 + 6.3067x R42 = 0.982 NNiC/Sed y = 373.29 + 5.5565x RA2 = 0.954 A CAC/Sed_ y = 309.46 + 5.2599x RA2 = 0.973 Sed y = 376.06 + 3.9148x RA2 = 0.947 ry a) e A a Shell Length (um) Time (d) Fig. 1. Comparison of growth equations of juvenile Villosa iris fed various diets for 140 d post-metamorphosis. (Abbreviations as in Table 1). 62 AMER. MALAC. BULL. 14(1) (1997) Table 6. Comparison of growth equations of Villosa iris juveniles fed various diets for 140 d post-metamorphosis. The diet with the largest slope (NPB) was compared to all other diets first. Diets found to be significantly different from the diet in ques- tion were then excluded sequentially from each test until all desired comparisons were made. p-values are given for the multiple comparisons of slope estimates from a sequential analysis of covariance. Slopes are signficantly different at a < 0.01, thus, the overall experimental error is controlled at a = 0.05. (* = the slope of this diet is significantly different from the diet to which all other diets are being compared in that particular test; na = not applicable; other abbreviations as in Table 1). Sequential Comparison of Slopes NPB/Sed with all other Growth Equation diets NPB/Sed y = 193.07 + 10.443x na NPE/Sed y = 353.34 + 6.3067x 0.1729 NNiC/Sed y = 373.29 * 5.5565x 0.4952 CAC/Sed y = 309.46 + 5.2599x 0.0000* Sed-only y = 376.06 + 3.9148x 0.0000* Table 7. Mean shell length (+ SD, with minimum-maximum in parenthe- ses) and percent survival of Villosa iris juveniles in each culture chamber, and pooled survival for all juveniles reared on algae and sediment or sedi- ment-only for 140 d post-metamorphosis. [nd = no data (not considered in the pooled survival calculation); other abbreviations as in Table 1). Diet Length (um) Survival (%) Sed-only 1 914 + 120 (769-1180) 15.5 Sed-only 2 855 + 74 (769-897) 43 Pooled survival for Sed-only 10.0 CAC/Sed 1 nd3 nd CAC/Sed 2 1146 + 248 (718-1590) 24.5 CAC/Sed 3 nd3 nd3 NNic/Sed 1 1149 + 283 (564-1667) 15.3 NNic/Sed 2 nd! nd! NNic/Sed 3 nd! nd! NPB/Sed 1 nd2 nd2 NPB/Sed 2 nd2 nd2 NPB/Sed 3 1747 + 301 (1282-2359) 29.6 NPE/Sed 1 954 + 249 (610-1439) 30.0 NPE/Sed 2 1369 + 436 (897-2051) 9.7 NPE/Sed 3 1410 + 308 (897-2051) 38.0 Pooled survival for algae and sediment 24.5 INNic could have contained toxic levels of potassium (from algal medi- um). 2Burst water line caused mortality. Chironomid larvae might have preyed upon juveniles. tially digested algal cells, or ghost cells lacking chloroplas- ts, as well as colloidal particles, indicating that algae asso- ciated with the fine sediment were utilized as food by the juvenile mussels. Villosa iris exhibited locomotory and pedal-sweep movements for approximately 140 d. After 272 d, V. iris were observed positioned anterior-end in the NPB/Sed NPE/Sed NPE/Sed NNic/Sed with with all with with, NNic/Sed, except NNic/Sed, CAC/Sed, NPE/Sed NPB/Sed CAC/Sed Sed-only na na na na 0.0040* na 0.7626 na 0.0080* 0.6925 na na na 0.1431 0.0656 0.0177 na 0.0007* na 0.0045* sediment; apertures were visible and siphoning. DISCUSSION The fatty acid composition of many algae has been characterized, although most of the work has focused on marine algae (Ben-Amotz et al., 1985; Cranwell et al., 1988; Ahlgren et al., 1992). Thus, although we did not per- form lipid analyses on our diets, it was anticipated that cer- tain freshwater algae high in lipids, especially PUFA’s, would enhance growth in juvenile freshwater mussels. Because the culture environment affects the lipid content of algae (Spoehr and Milner, 1949; Fogg, 1959; Shifrin and Chisholm, 1981; Enright et al., 1986b), we harvested our algal cultures at the late-logarithmic, early-stationary phase when algae tend to produce more unsaturated than saturated fatty acids (Chu and Dupuy, 1980; Ahlgren et al., 1992; Dunstan et al., 1993). Freshwater Chlorophyceae generally lack substan- tial amounts of long-chained PUFA’s and, unlike other marine algae, are able to synthesize only minimal amounts of polyunsaturated fatty acids (Pohl and Zurheide, 1979; Ahlgren et al., 1992). It was not surprising, therefore, that the entirely green algal diet of Chlorella, Ankistrodesmus, and Chlamydomonas produced poorer growth than the other algal diets tested. It should be noted, however, that these algal species are not without some nutritional value. Total lipids (as percent dry weight) reported for C. vulgaris, Chlamydomonas sp., and Ankistrodesmus sp. were 13, 20, and 18%, respectively (Shifrin and Chisholm, 1981; Ben- Amotz et al., 1985), and this trialgal mixture produced bet- ter growth than a diet of fine sediment only. We suspected that juvenile mussels fed diets con- taining the diatoms WNitzschia, Cyclotella, and Phaeodactylum would show the best growth because GATENBY ET AL.: GROWTH OF VILLOSA IRIS 63 Fig. 2. A. Red fluorescing chlorophyll from ingested algae as seen in gut area of live 30 d juvenile Villosa iris. B. Red fluorescing chlorophyll from algae well-distributed in kaolin substratum. diatoms are characterized by a high percentage of unsatu- rated fatty acids and generally store oils when nutrients are limited (Erwin, 1973; Werner, 1978; Pohl and Zurheide, 1979; Ahlgren et al., 1992). In addition, species of Nitzschia and Cyclotella contain more lipids and have a significantly greater proportion of unsaturated fatty acids than green algae (Shifrin and Chisholm, 1981). P. tricornu- tum is abundant in PUFA’s (Reitan et al., 1994), and its del- icate, siliceous walls are probably loosened easily by grind- ing or enzymatic action of the digestive system of juvenile mussels. Unlike most green algae, the oil-producing poten- tial of Neochloris oleoabundans has been well described (Lien, 1981), and it could also contain PUFA’s (Tornabene et al., 1983). Bracteacoccus grandis, another green alga, is abundant in lipids (Bishoff and Bold, 1963). Thus, juve- niles reared on Neochloris, Bracteacoccus, and Phaeodactylum showed the best growth over time, and juveniles fed the Neochloris, Nitzschia, and Cyclotella diet also showed very good growth when compared to those reared under the other diets tested. Unfortunately, almost all juveniles died in two repli- cates of NNiC/Sed between 49 and 65 d. This sudden mor- tality likely resulted from the laboratory culture environ- ment, and not the nutritional quality of the diet. Imlay (1973) reported that exposure of freshwater mussels to potassium levels of 4-7 mg/I over 7 d was lethal. Chemical analysis of the NNiC diet showed high levels of potassium (55.9 mg/l). The algal medium contained high levels of potassium (Tables 3 and 4). Due to the bi-phasic culture methods used to culture Cyclotella and Nitzschia, soil parti- cles were often extracted along with the algal cells when preparing diets that included these two diatoms. Even though centrifuged cells were rinsed several times, soil par- ticles could have retained more of the algal medium than expected. As judged by the volume of the culture chamber and the volume of algae administered, potassium levels could have reached 6.7 mg/l after 7 d and 9.6 mg/l after 10 d of administering the diet. 64 AMER. MALAC. BULL. 14(1) (1997) All individuals fed NOC/Sed were dead by 46 d, with replicates 1 and 2 dead by 24 d. A chemical analysis of the NOC diet also showed high potassium levels (89.2 mg/l). The water and sediment in each juvenile culture dish was usually changed every 7-10 d. The algae (NOC) were administered daily to the juvenile cultures; however, the water and sediment was not changed during the first 14 d. Thus, the level of potassium in the mussel culture dishes could have reached 21.4 mg/l after 14 d. The causes for high levels of potassium in the diet are possibly two-fold. First, Oocystis marsonii is characterized by thick cell walls with large cellulose microfibrils and embedded with hemi- celluloses (Preston, 1974). As this was the only alga used in our studies that had such a thick wall, it is possible that more of the algal medium (and excess potassium) was retained by the cells than was desirable. Secondly, soil par- ticles associated with the preparation of this diet also could have retained excessive amounts of the algal medium. The addition of a cultured live bacterium, Enterobacter aerogenes, in combination with the algae Neochloris and Phaeodactylum and fine sediment, did not enhance growth over the trialgal diets NPB/Sed and NNiC/Sed. This was not surprising because bacteria are generally considered “‘poor food” for zooplankton due to a lack of PUFA (Wood, 1974). Crosby et al. (1990) reported that bacterial mediation enabled the eastern oyster, Crassostrea virginica (Gmelin, 1791), to make more effi- cient use of refractory detrital carbon and nitrogen. Thus, bacteria can break down previously unavailable food sources, but they are not a primary source of nutrition. Juveniles reared in fine sediment only, however, showed a 330% increase in shell length after 140 d. The success of a sediment only diet could be attributed to microbial and organic content because fine sediments contain an array of digestible material (Swain, 1970). Essential minerals, amino acids, and vitamins are also adsorbed onto soil parti- cles creating a storehouse of nutrients (Kelley, 1942; Weiss, 1969; Swain, 1970; Baldwin and Newell, 1991). Juveniles fed algae with kaolin exhibited growth similar to juveniles fed algae with sediment at 60 d. Thus, resident bacteria in sediments did not enhance growth and are likely not essential to digestion or superior in nutrition to algae. However, survival of juveniles fed algae with kaolin was very low. We attribute this mortality to the fine particle size (1-2 um) of kaolin. The kaolin became paste- like over time, and we suspect this could have inhibited pedal-feeding or suffocated the juveniles. The fine sedi- ment in our static chambers remained loose in composition and survival of juveniles was high. Yeager et al. (1994) observed suspension feeding by juvenile Villosa iris in the interstitial spaces of the sediment. More recently, research on the development of gills in juvenile mussels indicates that juveniles filter-feed in combination with pedal-feeding, until gill development is complete (R. Tankersley, pers. comm.). Particle size and composition of the substratum, therefore, seems important for effective pedal-feeding. The Yeast/Sed diet did not support juveniles beyond 8 d. Preliminary trials using other commercial yeast-based diets also did not support juvenile mussels. These yeast- based diets were developed for mariculture of oysters, clams, and other invertebrates (Coutteau and Sorgeloos, 1992). Because juvenile mussels were short-lived in this treatment, when compared to the sediment only treatment, we believe that this diet could contain a toxic component that is otherwise not toxic to marine organisms. Perhaps the commercial yeast diet will be suitable if used in concen- trations lower than that used in our experiments. Although sediment provides minerals and some organic food material, an additional food source such as algae increased survival and growth. No specific algal diet produced statistically better survival over another. Thus, many algal diets administered with fine sediment could support juvenile mussels for at least several months. For long-term propagation, the more nutritious algae should produce the best growth and survival. Information on the early life history requirements of freshwater mussels is minimal, with their nutritional requirements probably the least understood. Early investi- gators believed that “finding suitable nutrition for the first month or so of free life” was critical to the success of rear- ing mussels (Lefevre and Curtis, 1910, 1912; Coker et al., 1921; Howard, 1922). This study has shown that the trial- gal mixture of Neochloris oleoabundans, Phaeodactylum tricornutum, and Bracteacoccus grandis was a suitable diet for rearing juvenile mussels, and provided adequate nutri- tion for growth and survival through 140 d post-metamor- phosis. The algae Nitszchia acicularis, and Cyclotella meneghiniana in combination with N. oleoabundans also was suitable for rearing juvenile mussels to 140 d post- metamorphosis. However, juveniles reared on N. oleabun- dans, P. tricornutum, and B. grandis showed the best growth over time. In addition, adult broodstock of various freshwater species have been maintained on N. oleoabun- dans to 10 mo at Virginia Tech’s Aquaculture Center (F. O’Beirn, pers. comm.). Research in aquaculture indicates that fish and marine bivalves have different requirements for protein, carbohydrates, and polyunsaturated fatty acids as they mature (Walne, 1973; Watanabe et al., 1983; Hawkins and Bayne, 1991). It is likely, therefore, that certain algae might be more suitable to enhance growth in juvenile mussels, and other algae are more suitable for growth of adult mus- sels. The future success of propagating freshwater mussels will depend on a better understanding of the feeding ecolo- gy and environmental requirements of the juvenile life stage. GATENBY ET AL.: GROWTH OF VILLOSA IRIS 65 ACKNOWLEDGMENTS We thank Michael Latham for designing our laboratory and assisting in juvenile mussel culture, Myron Beaty for assistance in algae culturing, and Mary Clark and Richard Baptiste (Harbor Branch Oceanographic Institution, Ft. Pierce, Florida) for providing training in bivalve culture. Our extreme appreciation goes to the Martek Biosciences Corporation for stock algal cultures and references crucial to this research. This research was sponsored by the Tennessee Wildlife Resources Agency, Virginia Department of Game and Inland Fisheries, and the National Biological Service. LITERATURE CITED Ackman, R. G. 1983. Fatty acid metabolism in bivalves. In: Proceedings of the Second International Conference on Aquaculture Nutrition: Biochemical and Physiological Approaches to Shellfish Nutrition, G. D. Pruder, C. J. Langdon, and D. E. Conklin, eds. World Mariculture Society 2: 358-376. Louisiana State University Press, Baton Rouge. Ahlgren, G., I. Gustafsson, and M. Boberg. 1992. Fatty acid content and chemical composition of freshwater microalgae. Journal of Phycology 28: 37-50. Allen, J. A. 1961. The development of Pandora inaequivalvis (Linné). Journal of Embryology and Experimental Morphology 9:252-268. Allen, J. A. 1985. The recent Bivalvia: their form and function. In: The Mollusca, 10, Evolution, E. R. Trueman and M. R. Clarke, eds. pp. 337-403. Academic Press, New York. Ando, K. 1968. Biochemical studies on the lipids of cultured fishes. Journal of Tokyo University Fisheries 54:61-98. Baldwin, B. S. and R. I. E. Newell. 1991. Omnivorous feeding by plank- totrophic larvae of the eastern oyster Crassostrea virginica. Marine Ecology Progress Series 78:285-301. Behrens, P. W., S. E. Bingham, S. D. Hoeksema, D. L. Kohoon, and J. C. Cox. 1989. Studies on the incorporation of CO into starch by Chlorella vulgaris. Journal of Applied Phycology 1:123-130. Ben-Amotz, A., T. G. Tornabene, and W. H. Thomas. 1985. Chemical profile of selected species of microalgae with emphasis on lipids. Journal of Phycology 21:72-81. Bishoff, H. W. and H. C. Bold. 1963. Phycological studies IV: some soil algae from Enchanted Rock and related algal species. University of Texas Publication 6318:1-95. Castell, J. D. 1970. The Essential Fatty Acid Requirements of Rainbow Trout (Salmo gairdneri). Doctoral Dissertation, Oregon State University, Corvallis, Oregon. 116 pp. Chu, F.-L. and J. L. Dupuy. 1980. The fatty acid composition of three unicellular algal species used as food sources for larvae of the American oyster (Crassostrea virginica). Lipids 15:356-364. Coker, R. E., A. F. Shira, H. W. Clark, and A. D. Howard. 1921. Natural history and propagation of freshwater mussels. Bulletin of the United States Bureau of Fisheries 37:79-133. Coutteau, P. and P. Sorgeloos. 1992. The use of algal substitutes and the requirement for live algae in the hatchery and nursery rearing of bivalve molluscs: an international survey. Journal of Shellfish Research 11(2):1-10. Cranwell, P. A., M. E. Creighton, and G. H. M. Jaworski. 1988. Lipids of four species of freshwater chrysophytes. Phytochemistry 27:1053-1059. Crosby, M. P., R. I. E. Newell, and C. J. Langdon. 1990. Bacterial media- tion in the utilization of carbon and nitrogen from detrital com- plexes by Crassostrea virginica. Limnology and Oceanography 35(3):725-639. Davis, H. C. and R. R. Guillard. 1958. Relative value of ten genera of microorganisms: as foods for oyster and clam larvae. United States Fish and Wildlife Service, Fisheries Bulletin 136 (58):290- 304. Dunstan, G. A., J. K. Volkman, S. M. Barrett, and C. D. Garland. 1993. Changes in the lipid composition and maximisation of the polyun- saturated fatty acid content of three microalgae grown in mass culture. Journal of Applied Phycology 5:71-83. Enright, C. T., G. F. Newkirk, J. S. Craigie, and J. D. Castell. 1986a. Evaluation of phytoplankton diets for juvenile Ostrea edulis L. Journal of Experimental Marine Biology and Ecology 96: 1-13. Enright, C. T., G. F. Newkirk, J. S. Craigie, and J. D. Castell. 1986b. Growth of juvenile Ostrea edulis L. fed Chaetoceros gracilis Schutt of varied chemical composition. Journal of Experimental Marine Biology and Ecology 96:15-26. Epifanio, C. E. 1979. Growth in bivalve molluscs: nutritional effects of two or more species of algae in diets fed to the American oyster Crassostrea virginica (Gmelin) and the hard clam Mercenaria mercenaria (L.). Aquaculture 18:1-12. Erwin, J. A. 1973. Comparative biochemistry of fatty acids in eukaryotic microorganisms. Jn: Lipids and Biomembranes of Eukaryotic Organisms, J. A. Erwin, ed. pp. 42-143. Academic Press, New York. Fogg, G. E. 1959. Nitrogen nutrition and metabolic patterns in algae. Symposia of the Society of Experimental Biology 13:106-125. Gatenby, C. M., R. J. Neves, and B. C. Parker. 1996. Influence of sedi- ment and algal food on cultured juvenile freshwater mussels. Journal of the North America Benthological Society 15(4):597- 609. Hawkins, A. J. and B. L. Bayne. 1991. Nutrition of marine mussels: fac- tors influencing the relative utilizations of protein and energy. Aquaculture 94:177-196. Howard, A. D. 1917. A second generation of artificially reared fresh-water mussels. Transactions of the American Fisheries Society 46:90- 92. Howard, A. D. 1922. Experiments in the culture of freshwater mussels. Bulletin of the United States Bureau of Fisheries 38:63-89. Hudson, R. G. and B. G. Isom. 1984. Rearing of juveniles of the freshwa- ter mussels (Unionidae) in a laboratory setting. The Nautilus 98:129-135. Imlay, M. J. 1973. Effects of potassium on survival and distribution of freshwater mussels. Malacologia 1(1): 97-113. Imlay, M. J. and M. L. Paige. 1972. Laboratory growth of freshwater sponges, unionid mussels, and sphaerid clams. Progressive Fish- Culturist 34(4):210-216. Kelley, W. P. 1942. Modern clay researches in relation to agriculture. Journal of Geology 50(3):307-319. Lasee, B. A. 1991. Histological and Ultrastructural Studies of Larval and Juvenile Lampsilis (Bivalvia) from the Upper Mississippi River. Doctoral Dissertation, Iowa State University, Ames, Iowa. 146 pp. Lefevre, G. and W. C. Curtis. 1910. Experiments in the artificial propaga- tion of freshwater mussels. Bulletin of the United States Bureau of Fisheries 28: 615-626. Lefevre, G. and W. C. Curtis. 1912. Studies on the reproduction and arti- ficial propagation of freshwater mussels. Bulletin of the United States Bureau of Fisheries 30: 105-201. Lien, S. 1981. Photobiological production of fuels by microalgae. In: Proceedings of the Brighton International Conference on Energy from Biomass, W. Palz, P. Chartier, and D. O. Hall, eds. pp. 697- 702. Applied Science Publishers, London. McArdle, R. N., J. Lasekan, R. Engel, and P. Behrens. 1994. Protein and 66 AMER. MALAC. BULL. 14(1) (1997) amino acid content of selected microalgae species. Lebensmittel Wissenschaften von Technologie 27:249-252. Morton, B. 1976. The biology and functional morphology of the southeast Asian mangrove bivalve, Polymesoda (Geloina) erosa (Solander, 1786) (Bivalvia: Corbiculidae). Canadian Journal of Zoology 54:482-500. Napolitano, G E., W. M. N. Ratnayake, and R. G. Ackman. 1988. Fatty acid composition of larval Ostrea edulis (Linnaeus): importance of triacylglycerols as a fatty acid reserve. Comparative Biochemistry and Physiology 90B (4):875-883. Napolitano, G. E., R. G. Ackman, and W. M.N. Ratnayake. 1990. Fatty acid composition of three cultured algal species (Isochrysis gal- bana, Chaetoceros gracilis and Chaetoceros calcitrans) used as food for bivalve larvae. Journal of the World Aquaculture Society 21(2):122-130. Neves, R. J., L. R. Weaver, and A. V. Zale. 1985. An evaluation of host fish suitability for glochidia of Villosa vanuxemi and V. nebulosa (Pelecypoda: Unionidae). American Midland Naturalist 113:13- 19. Nichols, H. W. 1973. Growth media - freshwater. In: Handbook of Phycological Methods. J. R. Stein, ed. pp. 13-14. Cambridge University Press, New York. Pohl, P. and F. Zurheide. 1979. Fatty acids and lipids of marine algae and the control of their biosynthesis by environmental factors. Jn: Marine Algae in Pharmaceutical Science, H. A. Hoppe et al., eds. pp. 473-523. Walter de Gruyter, New York. Preston, R. D. 1974. The Physical Biology of Plant Cell Walls. John Wiley and Sons, New York. 491 pp. Reid, R. G. B., R. F. McMahon, D. O Foighil, and R. Finnigan. 1992. Anterior inhalant currents and pedal feeding in bivalves. The Veliger 35:93-104. Reitan, K. I., J. R. Rainuzzo, and Y. Olsen. 1994. Effect of nutrient limita- tion on fatty acid and lipid content of marine microalgae. Journal of Phycology 30:972-979. Romberger, H. P. and C. E. Epifanio. 1981. Comparative effects of diets consisting of one or two algal species upon assimilation efficien- cies and growth of juvenile oyster, Crassostrea virginica (Gmelin). Aquaculture 25:77-87. SAS Institute. 1991. JMP 2.0. Cary, North Carolina. Shifrin, N. S. and S. W. Chisholm. 1981. Phytoplankton lipids: interspe- cific differences and effects of nitrate, silicate, and light-dark cycles. Journal of Phycology 17: 374-384. Sokal, R. R. and F. J. Rohlf. 1981. Biometry, 2nd ed. W. H. Freeman and Co., San Francisco, Calilfornia. 859 pp. Spoehr, H. A. and H. W. Milner. 1949. The chemical composition of Chlorella. Effect of environmental conditions. Plant Physiology 24: 120-149. Swain, F. M. 1970. Non-marine Organic Geochemistry. Cambridge University Press, New York. 445 pp. Tan Tiu, A., D. Vaughan, T. Chiles, and K. Bird. 1989, Food value of eurytopic microalgae to bivalve larvae of Cyrtopleura costata (Linnaeus, 1758), Crassostrea virginica (Gmelin, 1791) and Mercenaria mercenaria (Linnaeus, 1758). Journal of Shellfish Research 8(2): 99-405. Tornabene, T. G., G. Holzer, S. Lien, and N. Burris. 1983. Lipid compo- sition of the nitrogen starved green alga Neochloris oleoabun- dans. Enzyme Microbial Technology 5:435-440. Ukeles, R. 1971. Nutritional requirements in shellfish culture. Jn: Proceedings of the Conference on Artificial Propagation of Commercially Valuable Shellfish, K. S. Price and D. L. Maurer, eds. pp. 42-64. College of Marine Studies, University of Delaware, Newark. Urban E. R. and C. J. Langdon. 1984. Reduction in costs of diets for the American oyster (Crassostrea virginica) (Gmelin), by the use of non-algalsupplements. Aquaculture 38:277-291. Walne, P. R. 1970. Studies on the food value of nineteen genera of algae to juvenile bivalves of the genera Ostrea, Crassostrea, Mercenaria, and Mytilus. Fishery Investigations, London Series, 2, 26(5):1-62. Walne, P. R. 1973. Growth rates and nitrogen and carbohydrate contents ° of juvenile clams, Sazidonus giganteus fed three species of algae. Journal of the Fisheries Research Board of Canada 30:1825- 1830. Watanabe, T., C. Kitajima, and S. Fujita. 1983. Nutritional values of live organisms used in Japan for mass propagation of fish: a review. Aquaculture 34:115-143. Webb, K. L. and F. L. Chu. 1983. Phytoplankton as a food source for bivalve larvae. In: Proceedings of the Second International Conference on Aquaculture Nutrition: Biochemical and Physiological Approaches to Shellfish Nutrition, G. D. Pruder, C. J. Langdon, and D. E. Conklin, eds. World Mariculture Society 2: 272-291. Louisiana State University Press, Baton Rouge. Weiss, A. 1969. Organic derivatives of clay minerals, zeolites, and related minerals. In: Organic Geochemistry, Methods and Results, G. Elington and M. T. J. Murphy, eds. pp. 737-775. Springer- Verlag, New York. Wentworth, C. K. 1922. A scale of grade and class terms for clastic sedi- ments. Journal of Geology 30:377-392. Werner, D. 1978. Regulation of metabolism by silicate in diatoms. Jn: Biochemistry of Silicon and Related Problems, G. Vendz and I. Lindquist, eds. pp. 149-176. Plenum, New York. Williams, J. D., M. L. Warren, Jr., K. S. Cummings, J. L. Harris, and R. J. Neves. 1993. Conservation status of freshwater mussels of the United States and Canada. Fisheries 18:6-22. Wood, B. J. 1974. Fatty acids and saponifiable lipids. Jn: Algal Physiology and Biochemistry, W. D. P. Stewart, ed. pp. 236-265. Blackwell, London. Yeager, M. M., D.S. Cherry, and R. J. Neves. 1994. Feeding and burrow- ing behaviors of juvenile rainbow mussels, Villosa iris (Bivalvia: Unionidae). Journal of the North American Benthological Society 13(2):217-222. Zale, A. V. and R. J. Neves. 1982. Fish hosts of four species of lampsiline mussels (Mollusca: Unionidae) in Big Moccasin Creek, Virginia. Canadian Journal of Zoology 60:2535-2542. Zar, J. H. 1974. Biostatistical Analysis. Prentice-Hall, Englewood Cliffs, New Jersey. 620 pp. Date of manuscript acceptance: 13 February 1997 Zebra mussel induced mortality of unionids in firm substrata of western Lake Erie and a habitat for survival Don W. Schloesser!, R. D. Smithee2, G. D. Longton2, and W. P. Kovalak2 1U. ]S. Geological Service, Biological Resources Division, Great Lakes Science Center, 1451 Green Road, Ann Arbor, Michigan 48105 U.S.A. 2Environmental Assessment, Detroit Edison Company, 6100 W. Warren, Detroit, Michigan 48226 U.S. A. Abstract: The present study was conducted to determine impacts of zebra mussel [Dreissena polymorpha (Pallas, 1771); Dreissenidae] infestation on - unionids in firm substrata in western Lake Erie. Unionid mollusks were collected at a total of 15 stations on three offshore depth contours (2, 3, and 4 m) in 1983 (before zebra mussel infestation), in 1990 and 1993 (after zebra mussel infestation), and at one station on a nearshore 2-m depth contour and along one transect on a nearshore 1-m depth contour in 1993. Numbers of living unionids on substrata along offshore contours remained similar between 1983 and 1990 and then decreased from 97 individuals in 1990 to only five individuals in 1993. In addition, the number of species decreased from nine to four between 1990 and 1993. In contrast, on nearshore contours 85 living individuals representing nine species were found in 1993. About 48% of the living and 79% of the dead unionids at the two nearshore locations were covered with byssal threads of dreissenid mussels, but were not actively infested by mussels. The presence of living unionids on nearshore contours of western Lake Erie in 1993 indicates that survival of unionids in the presence of abundant zebra mussel populations can be possible in firm substrata and that these habitats can provide natural “refugia” for unionid populations. At present, we do not know what allows unionids to survive in the presence of zebra mussel colonization, but believe that water-level fluctuations and waves could contribute to the removal of mussels from unionids. This information could be of major concern in the mitigation of impacts of infestation on unionids in waters through- out North America. Key words: Dreissena, Unionidae, refugia, Great Lakes, unionid mortality Zebra mussels [Dreissena polymorpha (Pallas, umented in soft-mud substrata of open waters (> 6 m deep) 1771); Dreissenidae] have been shown to be ectoparasites (Schloesser and Nalepa, 1994). Soft substrata are believed causing reduced fitness, shell deformities, and mortality of to suffocate unionids that cannot maintain themselves at the unionids (Bivalvia: Unionidae) in waters of Europe and substrate-water interface due to the added weight of infest- North America (Ricciardi et al., 1995; Schloesser et al., ing zebra mussels (Schloesser and Nalepa, 1994). However, 1996). In the Laurentian Great Lakes, infestation of union- unionid mortality caused by infestation has occurred in ids by zebra mussels was one of the first and most visible some firm-substratum habitats (e. g. sand and gravel) where ecological impacts of the early invasion of mussels into suffocation caused by sediments is unlikely (Nalepa, 1994; North America (Hebert et al., 1989, 1991; Mackie, 1991; Tucker, 1994; Schloesser et al., 1996), but observations Schloesser and Kovalak, 1991; Hunter and Bailey, 1992; indicate that unionids do survive in the presence of zebra Haag et al., 1993; Nalepa and Schloesser, 1993). Unionid mussels on some firm substrata (DWS: unpub. data; D. mortality increased as a result of infestation and unionid Blodgett, Illinois Natural History Survey, Havana, Illinois, populations were nearly eliminated in some areas (Gillis pers. comm.). and Mackie, 1994; Schloesser and Nalepa, 1994; Nalepa et The present study was performed to determine al., 1996). In addition, zebra mussels have invaded and are changes in the unionid population in firm compacted sand believed to be causing unionid mortality in major rivers of substrata along the perimeter of western Lake Erie in 1983, North America, such as the Detroit, St. Lawrence, before zebra mussels were present, and 1991 and 1993, Mississippi, Ohio, and Tennessee Rivers (Tucker, 1994; after mussels were present. We hypothesized that if smoth- Ricciardi et al., 1996; Strayer and Smith, 1996; reviewed ering in soft substrata is the only mechanism causing by Schloesser et al., 1996). unionid mortality in western Lake Erie, then unionids In western Lake Erie, increased mortality of union- should survive infestation in firm substrata along the ids attributed to zebra mussel infestation has only been doc- perimeter of western Lake Erie. American Malacological Bulletin, Vol. 14(1) (1997):67-74 67 68 AMER. MALAC. BULL. 14(1) (1997) METHODS Unionid bivalve mollusks and infesting zebra mus- sels were collected at 15 stations located on three offshore depth contours (2, 3, and 4 m depths) along the perimeter of western Lake Erie between 23 August and 15 October 1983, 31 October and 11 November 1990, and 8 September and 27 October 1993 (Fig. 1). Unionids were also collected at one station on a nearshore 2-m depth contour, 9 September 1993 and in a transect area on a nearshore 1-m depth contour, 21 October 1993. The offshore 2-m and nearshore 2-m depth contours were separated by a deeper 3-m contour. Fluctuations of daily water levels were cor- rected based on the frame of reference of low water data at Father Point, Quebec (Great Lakes Basin Commission, 1975). In general, sampling depths were about 1 m deeper than low water datum, except for the period of time when the nearshore transect was sampled during a seiche. All liv- ing and dead unionids (represented by both valves) were collected and used in the present study. In addition, total Western Lake Erie 1km Fig. 1. Locations in western Lake Erie where unionid mollusks were sam- pled: at 15 stations (@) on three offshore depth contours (dark dashed lines; 2-, 3-, and 4-m) in 1983, 1990, and 1993, and at one station (C) on a nearshore 2-m depth contour (light-dashed line) and along one transect (C)) on the nearshore 1-m depth contour (light-dashed line) in 1993. counts of valves and shell pieces of unionids were deter- mined in 1990 and 1993. Unionids at stations on offshore contours and the nearshore 2-m depth contour were collected by SCUBA (Fig. 1). At each station, random searching was conducted for 30 min within a 50-m diameter circle. Unionids were removed from the water, individually separated, and taken to the laboratory. Unionids in the transect area (7 m by 150 m) along the nearshore 1-m depth contour were collected by walking and manually removing exposed specimens (Fig. 1). This was possible because water in western Lake Erie was pushed away from shore by westerly winds causing a seiche, thereby dewatering substrata on the nearshore 1-m depth contour for a period of about 6 hr. In the laboratory, up to ten randomly selected unionids and attached zebra mussels from one station on the offshore 4-m depth contour, the station on the nearshore 2-m depth contour, and the transect on the nearshore 1-m depth contour were analyzed. Living infested unionids were selected, except from the offshore, 4-m depth contour in 1993 when only dead unionids were present. Schloesser and Nalepa (1994) have shown that length-frequency distri- butions of zebra mussels from living and dead unionids in western Lake Erie are similar. Infesting zebra mussels were removed from individual host unionids and retained in a U. S. Standard Number 60 sieve (0.25-mm mesh). Length-fre- quency distributions of infesting zebra mussels were con- structed from shell-length measurements in 1-m size class- es. All mussels or a randomly selected sub-sample of 200- 300 mussels (< 6-mm long per unionid) and all mussels > 7-mm long were measured in each sampling period. All mussels were identified and counted. All mussels were zebra mussels (Dreissena polymorpha); no quagga mussels (Dreissena bugensis Andrusov, 1897) were present in west- ern Lake Erie (Mills et al., 1993; MaclIsaac, 1994). Length- frequency distributions of unmeasured mussels < 6-mm long were based on the proportions of measured sub-sam- pled mussels in each whole millimeter size group. This pro- cedure has been shown to adequately determine length-fre- quency distributions of zebra mussels in western Lake Erie (Schloesser and Kovalak, 1991; Schloesser and Nalepa, 1994). Dry weights (in g at 105°C for 48 hr) of individual unionids and infesting mussels (a measure of infestation intensity; Schloesser and Kovalak, 1991) were determined. Unionids were identified following Clarke (1981) and by comparison with bivalve taxonomic reference col- lections (Detroit Edison Company, Detroit, Michigan, and Great Lakes Science Center, Ann Arbor, Michigan). Taxonomic nomenclature follows Williams et al. (1993) with the exception that Lampsilis radiata radiata (Gmelin, 1791) was combined with L. siliqguoidea because these two species are believed to interbreed in the Great Lakes SCHLOESSER ET AL.: ZEBRA MUSSEL INDUCED MORTALITY 69 (Clarke, 1981). Mean numbers of living and dead unionids per station at 15 stations on the three offshore depth con- tours were tested by Student’s t-test after log(io) + 1 trans- formation (Snedecor and Cockran, 1967). RESULTS UNIONIDS ON OFFSHORE CONTOURS Numbers of living and dead unionids changed sub- stantially on the three offshore depth contours in western Lake Erie between 1983 and 1993 (Table 1). In 1983, total mean number of living unionids per station was significant- ly greater (paired t-tests, P < 0.05) than dead unionids. In 1990 and 1993, total mean numbers of dead unionids were significantly higher than living unionids. Numbers of living unionids remained about the same between 1983 and 1990: 85 and 97 individuals, respectively. Then they decreased significantly from 97 in 1990 to only five in 1993. Numbers of dead unionids increased significantly from nine in 1983 to 360 in 1990, then decreased to 157 in 1993. In 1983, mean numbers of living individuals were significantly greater than numbers of dead individuals on the 2-m and 4- m depth contours; in 1990, there were significantly more numbers of dead than living individuals on the 3-m and 4-m depth contours. In 1993, the numbers of dead individuals was greater than the number of living on all contours. Between 1983 and 1990, the number of living unionids increased on the 2-m and 3-m depth contours and decreased on the 4-m depth contour. The greatest increase in the num- ber of dead unionids occurred on the 4-m depth contour between 1983 and 1990. Numbers of living unionid species decreased on off- shore depth contours in western Lake Erie between 1983 and 1993 (Table 2). Twelve living species were found in 1983, nine species in 1990, and four species in 1993. Between 1983 and 1990, the number of species on the 2-m and 3-m depth contours increased from seven to nine, whereas the number on the 4-m depth contour decreased from 12 to six species. In 1993, only four species (Amblema plicata plicata, Lampsilis siliquoidea, Potamilus alatus, and Table 1. Numbers of living and dead unionids collected at 15 stations on three offshore depth contours in western Lake Erie, 1983, 1990, and 1993. (*; significant difference [P < 0.05] in mean number of unionids per sta- tion). Depth 1983 1990 1993 (m) Living Dead Living Dead Living Dead 2 19 * 1 31 26 2 20 3 12 3 27 * 59 3 41 4 54 * 5 39 * 1275 0 * 96 Total 85 * 9 97 * 360 5 * 157 Isignificant (P = 0.060). Quadrula quadrula) represented by five individuals were found on the 2-m and 3-m depth contours. UNIONIDS ON NEARSHORE CONTOURS Relatively large numbers of living unionids occurred in sampled areas on the nearshore 2-m and 1-m depth contours (Table 3). A total of 85 living unionids rep- resented by nine species occurred at nearshore locations. Only living unionids (55 individuals, represented by eight species) were collected at the nearshore 2-m depth station. Both living (30 individuals, represented by five species) and dead (29 individuals, represented by ten species) union- ids were found in the transect area on the nearshore 1-m depth contour. Only 15% (17 of 114 individuals) of the unionid shells collected on nearshore contours were infested with zebra mussels (Table 3). In addition, 70% (80 of 114) only showed evidence of past infestation (i. e. only byssal threads present on unionid shells). Evidence of infestation was absent on 15% (17) of individual shells (10% [11] of these were living and 5% [6] were dead unionids). INFESTING ZEBRA MUSSELS Mean dry weights of infesting zebra mussels were substantially higher on the offshore 4-m depth contour than on the nearshore 2-m depth contour; no infestation occurred on the 1-m nearshore depth contour (Table 4). On the off- shore contour in 1990, mean infestation weights on living and dead unionids were 41.1 and 36.3 g/unionid, respec- tively; in 1993, infestation weights decreased to 32.9 and 27.6 g/unionid, respectively. In general, weights of infesting zebra mussels were about equal to host unionid weights on the offshore contour. In nearshore areas, weights of infest- ing zebra mussels ranged between 0.5 and 4.4 g/unionid, and mean infestation weights were substantially less than mean unionid weights. Length-frequency distributions of zebra mussels removed from unionids collected on the offshore 4-m depth contour and the nearshore 2-m depth contour in Fall 1993 were substantially different (Fig. 2). Mussels on the off- shore contour were larger (mean = 12 mm long) than those on the nearshore contour (mean = 6 mm long). Peak distrib- ution of mussels on unionids from the offshore contour was between 12 and 16 mm, whereas the peak distribution of mussels on the nearshore contour was between 3 and 7 mm. DISCUSSION UNIONID MORTALITY ON OFFSHORE CONTOURS The present study, that by Nalepa (1994) in Lake St. Clair immediately upstream of western Lake Erie, observa- tions by Tucker (1994) in the Mississippi River, and data of 70 AMER. MALAC. BULL. 14(1) (1997) Table 2. Numbers of living unionids collected at 15 stations on three offshore depth contours in western Lake Erie, 1983, 1990, and 1993. (-, none collected). Species 1983 1990 1993 Depth (m) Depth (m) Depth (m) 2 3 4 2 3 4 2 3 4 Amblema plicata plicata (Say, 1817) 1 1 2 12 3 12 - 2 - Elliptio dilata (Rafinesque, 1820) - - 1 - - - = = a Fusconaia flava (Rafinesque, 1820) 2 1 16 1 8 - - - Lampsilis ovata (Say, 1817) - - 2 - 1 - - - - L. siliquoidea (Barnes, 1823) 6 6 14 7 16 9 1 - - Leptodea fragilis (Rafinesque, 1820) 1 2 1 3 1 1 - - - Ligumia recta (Lamarck, 1819) - - 1 - - - — = ga4 Obliquaria reflexa (Rafinesque, 1820) 1 - 1 2. - 1 - - = Potamilus alatus (Say, 1817) - - 2 - 1 - - Pyganodon grandis (Say, 1829) - - 6 - 1 - - - - Quadrula pustulosa pustulosa (Lea, 1831) 6 7 1 8 - - - Q. quadrula (Rafinesque, 1820) 2 1 1 2 _ 1 = Total Species 7 6 12 7 8 6 2 2 0 Schloesser (unpub. data) indicate that high unionid mortali- ty can occur on firm substrata where smothering by soft substrata is unlikely. In addition, recent data indicate that high unionid mortality could be occurring in areas where infestation does not exist, but zebra mussels are found on surrounding substrata (Ricciardi et al., 1996; Strayer and Smith, 1996). In the present study, zebra mussel infestation caused substantial unionid mortality (94%) in firm-sand substrata located on offshore contours of western Lake Erie between 1983 and 1993. Mortality occurred between 1990 and 1993 after five years (1989-1993) of zebra mussel colonization of nearshore waters; not between 1983 and 1990 (after one year of zebra mussel colonization) when densities of union- ids actually increased. In the fall of 1989, zebra mussels became very abundant in western Lake Erie reaching densi- ties in excess of 340,000/m2 in open waters (> 6 m depth) and 700,000/m2 in nearshore waters (2-3 m depth) and infestation intensities in excess of 10,000/unionid (Schloesser and Kovalak, 1991; Kovalak et al., 1993; Table 3. Numbers of living and dead unionid species with attached zebra mussels and with attached byssal threads of zebra mussels collected at one station on a nearshore 2-m depth contour, 8 September 1993, and in one transect area on a nearshore 1-m depth contour, 21 October 1993, in western Lake Erie. Station on Nearshore 2-m Depth Contour! Living Unionids Attached Zebra Mussels Amblema plicata plicata (Say, 1817) 13 Fusconaia flava (Rafinesque, 1820) 1 Lampsilis ovata (Say, 1817) - L. siliquoidea (Barnes, 1823) - Leptodea fragilis (Rafinesque, 1820) - Obliquaria reflexa (Rafinesque, 1820) - Pleurobema cordatum (Rafinesque, 1820) - Potamilus alatus (Say, 1817) 1 Pyganodon grandis (Say, 1829) - Quadrula pustulosa pustulosa (Lea, 1831) 2 Q. quadrula (Rafinesque, 1820) - Total Number 17 Total Species 4 Transect on Nearshore 1-m Depth Contour2 Living Unionids Dead Unionids Byssal Threads Byssal Threads Byssal Threads Present Absent Present Absent Present Absent 12 2 13 3 8 - 2 - 2 2 1 1 7 : - : 1 - 1 - 1 2 2 1 - - - 2 - 1 - - - 1 - 1 1 - - - - 7 - 3 1 - - - - 2 2 7 1 3 1 2 3 - 4 1 1 - 34 4 23 7 23 8 3 5 4 10 4 INo dead unionids. 2No attached zebra mussels. SCHLOESSER ET AL.: ZEBRA MUSSEL INDUCED MORTALITY 71 Leach, 1993; Schloesser and Nalepa, 1994). To date, mor- tality of unionids induced by abundant densities of zebra mussels has been documented in Europe in the mid-1930s, waters of the Great Lakes, the Hudson River and, possibly, the Mississippi River in the early-1990s (Sebestyen, 1938; Gillis and Mackie, 1994; Schloesser and Nalepa, 1994; Tucker, 1994; Nalepa et al., 1996; reviewed by Schloesser et al., 1996; Ricciardi et al., 1996; Strayer and Smith, 1996; DWS, unpub. data). Changes in the number of dead unionids in the study area are partly attributable to the movement of dead shells into and out of the study area. The large increase in the number of dead unionids (nine to 360) between 1983 and 1990 suggests that dead unionid shells were entering the sampled area from other areas of western Lake Erie prior to 1990. Schloesser and Nalepa (1994) have shown that at one station in open waters (> 6m deep) of western Lake Erie, mortality of infested unionids increased between Table 4. Mean (+ S.E. per unionid) and range of dry weights (g) of infest- ing zebra mussels and, in parentheses, mean (+ S.E.) and range of dry weights of unionids at stations on offshore and nearshore depth contours in western Lake Erie 1990 and 1993. Location/Date Living Unionids Dead Unionids Offshore 4-m depth contour 1 November 1990! n= 102 n=10 41.1+2.9 36.3+4.9 29.7 - 61.2 17.3 - 67.3 (50.5 + 9.2) (46.1 + 3.1) (21.8 - 119.9) (29.0 - 62.1) 8 September 1993 n=53 n=10 32.9 + 7.6 27.6+2.8 6.8 - 49.0 15.1 - 40.3 (39.8 + 18.0) (73.3 + 10.5) (32.8 - 118.1) (27.5 - 123.0) Nearshore 2-m depth contour 9 September 1993 n=10 n=0 4.4+2.0 0.5 - 19.0 (106.9 + 12.4) (61.7 - 188.0) Nearshore 1-m depth contour 21 October 1993 n= 30 n=29 0 0 0 0 (85.7 + 8.0) (79.8 + 7.2) (28.4 - 158.5) (24.4 - 164.6) IMost southwestern station (Fig. 1, open circle) on offshore, 4-m depth contour. 2Number of unionids. 30ffshore 2-m and 3-m depth contour; no living unionids along the off- shore 4-m depth contour. Nearshore 2-m Depth Contour Station Number zebra mussels per unlonid 5 10 15 20 25 Zebra mussel shell length (mm) Offshore 4-m Depth Contour Station Number zebra mussels per unionid 5 10 15 20 25 Zebra mussel shell length (mm) Fig. 2 . Length-frequency distributions of zebra mussels infesting unionid mollusks collected on the nearshore 2-m depth contour and the most west- ern station on the offshore 4-m depth contour in western Lake Erie, 8 September 1993. Fall 1989 and Spring 1990, and by Fall 1990 was 100%. This period of time corresponds with the large increase in the number of dead unionids in the study area, especially on the offshore, 4-m depth contour in Fall 1990. In 1990, most shells appeared fresh-dead; some contained decaying tissues, many exhibited pearly nacre, and few were decalci- fied. In 1993, however, few shells appeared fresh-dead; no decaying tissues were found, few exhibited a pearly nacre, and most were decalcified and difficult to identify. Loss of pearly nacre is believed to occur within 6-12 mo after death (Schloesser and Nalepa, 1994; D. Neves, National Biological Service, Blacksburg, Virginia, pers. comm.). In addition, less than 5% of the valves (n = 252) collected in 1990 were described as pieces, whereas about 25% of the valves (n = 256) collected in 1993 were described as pieces. The decrease in the number of dead unionids between 1990 and 1993 is attributed to transport of shells out of the study area into shallow bays of the lake. Substrata in shallow water bays near the sampling area were largely covered with shells of dead unionids in 1993 and 1994 (DWS, GDL, RS, pers. obs.). Several of these areas (50 m by 1000 m transects) were described as “union- i AMER. MALAC. BULL. 14(1) (1997) id graveyards” with a paved-like bottom of unionid shells. Large accumulations of dead unionids has rarely been seen in the Great Lakes prior to colonization by zebra mussels (Neves, 1987; WPK, Michigan, pers. obs.). Survival of a few unionids on offshore contours in western Lake Erie supports the data of Schloesser and Nalepa (1994) and Nalepa et al. (1996) that infestation on unionids by zebra mussels does not cause 100% mortality of all unionids, but does reduce the population to < 5% of pre-zebra mussel colonization. The survival of a few union- ids along the offshore 2-m and 3-m depth contours (five living of 157 total unionids) in 1993 is similar to that (four living of 191 total unionids) found in open waters of west- ern Lake Erie in 1991 (Schloesser and Nalepa, 1994). However, the long-term viability of unionid populations at low densities has been questioned, even in the absence of infestation (Lefevre and Curtis, 1910; Downing and Downing, 1992; Downing et al., 1993). To date, artificial maintenance of infested unionids has shown that survival of individual unionids in waters of western Lake Erie is possi- ble (Schloesser, 1996). Mean weights of zebra mussels infesting unionids on offshore contours (27.6-41.1 g/unionid) were similar to weights of infesting mussels that have caused nearly 100% mortality of unionids in other studies in western Lake Erie (Schloesser and Kovalak, 1991; Schloesser and Nalepa, 1994; GDL and RS, pers. obs.). In 1989, immediately after zebra mussels increased exponentially in western Lake Erie, weights of infesting mussels were between 30.0 and 54.9 g/unionid in nearshore waters and between 9.0 and 75.3 g/unionid in open waters (Schloesser and Kovalak, 1991; Schloesser and Nalepa, 1994). In 1990, weights in open waters ranged from 2.3 to 40.8 g/unionid (Schloesser and Nalepa, 1994). These data indicate that weights of infestation that equal or exceed host unionid weights cause severe mortality of unionids in western Lake Erie (similarly noted as mean:mass ratios by Ricciardi et al., 1996). UNIONID SURVIVAL IN NEARSHORE WATERS Relatively large numbers of living infested unionids and living unionids showing evidence of past infestation on nearshore-depth contours indicates that zebra mussel induced mortality of unionids could be minimal or does not exist in some areas along the perimeter of western Lake Erie. The high incidence of infestation (90%) of living unionids on nearshore contours indicates that infestation did occur and that mortality of unionids could have occurred. However, temporal variation in the number of mussels per unionid and densities of mussels colonizing adjacent bottom substrata were not assessed in the present study, and these factors are believed to influence zebra mussel induced mortality of unionids (Schloesser and Kovalak, 1991; Ricciardi et al., 1995; Schloesser et al., 1996). At present, the reason unionids survive in the nearshore area of western Lake Erie is not known. Several studies have suggested that some species of unionids are more likely to survive zebra mussel infestation than other species (Haag et al., 1993; Nalepa, 1994; Gillis and Mackie, 1994; Tucker, 1994). Possible explanations for these observations include; unionid sex (males less impact- ed), robustness of shells (unionid species with robust shells less impacted), and length of brooding time (unionid species with short brooding periods less impacted) (reviewed by Schloesser et al., 1996). In the present study, robust species with short brooding times (Amblema plicata plicata, Fusconaia flava, Quadrula pustulosa pustulosa, Q. quadrula) accounted for about one-half (57%) the individu- als on offshore contours in 1990. By 1993, nearly all union- id individuals were extirpated from offshore contours and of those individuals remaining, three of five were robust, short-term brooder species. However, robust species with short brooding times (four above and Pleurobema corda- tum) accounted for 87% of living unionids on nearshore contours in 1993. These data support the belief that robust shelled, short-term brooders appear to survive longer than thin-shelled, long-term brooders. But, to date, a total of 31 species of unionids have been infested by zebra mussels in North America, and none appear to be immune to impacts (reviewed by Schloesser et al., 1996). Possible explanations for survival of unionids in shallow waters of western Lake Erie is that infesting zebra mussels could be removed from unionids and surrounding substrata by predators such as fish and ducks, or mussels could release from substrata in response to unfavorable habitat conditions (Stanczykowska, 1977; French and Bur, 1993; Mitchell and Carlson, 1993; Stanczykowska and Lewandowski, 1993; Hazlett, 1994). Length-frequency distributions of zebra mussels infesting unionids and SCUBA observations from offshore and nearshore areas indicate that removal of mussels in the nearshore area by predators is not the likely cause for low infestation of unionids in the nearshore area. Length-fre- quency distributions indicate that mostly small, young mus- sels were present on unionids in the nearshore area, where- as mostly larger, older mussels were present in the offshore area. This indicates that mussels probably leave unionids between their first and second year of life because peak modes often correspond to year-classes of mussels (Morton, 1969; Stanczykowska, 1977; Schloesser and Kovalak, 1991) and typically, length-frequency distribu- tions of zebra mussels in Fall in western Lake Erie are bimodal (Griffiths et al., 1991; Schloesser and Kovalak, 1991; Schloesser and Nalepa, 1994). In addition, similar observation by SCUBA indicate that zebra mussels attached to substrata in the nearshore area were of low den- SCHLOESSER ET AL.: ZEBRA MUSSEL INDUCED MORTALITY 73 sity and small size compared to mussels in the offshore area (GDL and RS, pers. obs.). Because mussels have been shown to have great adhesive strength and predators tend to remove smaller mussels rather than larger mussels (Ackerman et al., 1993; Eckroat et al., 1993; French and Bur, 1993; Hamilton et al., 1994), we believe that mussels were not removed by predators in the nearshore area. Rather, zebra mussels voluntarily released from unionids, perhaps because of factors such as waves, water-level fluc- tuations, and ice scour in nearshore areas (< 2-m depth). In Lake Erie, mortality of unionids has occurred in another nearshore area (2-3 m depth) where no waves, exposure due to water-level fluctuations, or ice scour occur (Schloesser and Kovalak, 1991; Schloesser 1996; WPK, unpub. data). Movement en masse of zebra mussels has been observed and attributed to fall storms, wave action, and ice scour in Europe and western Lake Erie (Ehrenberg, 1957; Lewandowski, 1976; Griffiths et al., 1991; Nalepa and Schloesser, 1993). In the Illinois River, Tucker and Atwood (1995) suggested that changing water levels (among other factors) could have contributed to substantial decreases in numbers of zebra mussels in an area. Areas where unionids survive in the presence of zebra mussel colonization appear to be natural “refugia” where long-term survival of unionids could be possible. Although long-term studies have not been completed, sev- eral sites have been found where unionids continue to live in the presence of zebra mussel colonization and large year- to-year fluctuations in the density of mussels in surround- ing areas (Tucker and Atwood, 1995; Schloesser et al., 1996; DWS, unpub. data; D. Blodgett, Illinois Natural History Survey, Havana, Illinois, and D. Miller, U. S. Army Corps of Engineers, Vicksburg, Mississippi, pers. comm.). Observations such as these lead Clarke, (1992), Masteller et al., (1993), Tucker and Atwood (1995), and Schloesser (1996 and unpub. data) to identify possible refugia and/or suggest the need for the establishment of managed refugia to save unionid populations. These refugia would only be needed in a few areas where unique unionid species and populations are found because many waters in North America are unlikely to be heavily colonized by zebra mus- sels (Strayer, 1991). Indeed, dreissenid mussels in rivers appear to be a threat to unionids immediately below impoundments and lakes that provide mussel veligers con- tributing to infestation (Schloesser et al., unpub. data, DWS and D. Hunter, Oakland University, Rochester, Michigan, unpub. data). CONCLUSIONS Results of the present study, conducted in firm-sub- Stratum areas along the perimeter of western Lake Erie, were similar to those of Schloesser and Nalepa (1994) who studied soft-substratum areas in open waters of western Lake Erie and found infestation of unionids by zebra mus- sels can equal 100% unionid mortality at individual sites, but not 100% mortality of unionids throughout the study area. However, the presence of living unionids in two nearshore areas in the present study indicate that natural “refugia” could exist in beach/littoral areas of western Lake Erie. ACKNOWLEDGMENTS This is Contribution Number 985 of the Great Lakes Science Center, U. S. Geological Survey, Biological Resources Division, Ann Arbor, Michigan. LITERATURE CITED Ackerman, J. D., C. R. Ethier, D. G. Allen, and J. K. Spelt. 1993. The bio- mechanics of adhesion in zebra mussels (Dreissena polymorpha): tests with a rotating disk. In: Zebra Mussels: Biology, Impacts, and Control. T. F. Nalepa and D. W. Schloesser, eds. pp. 265- 281. Lewis/CRC Press, Boca Raton, Florida. Clarke, A. H. 1981. The Freshwater Molluscs of Canada. National Museum of Natural Sciences, Ottawa, Canada. 446 pp. Clarke, A. H. 1992. Ontario’s Sydenham River, an important refugium for native freshwater mussels against competition from the zebra mussel, Dreissena polymorpha. Malacology Data Net 3(1-4):43-55. Downing, J. A. and W. L. Downing. 1992. Spatial aggregation, precision and power in surveys of freshwater mussel populations. Canadian Journal of Fisheries and Aquatic Sciences 49:985-991. Downing, J. A., Y. Rochon, and M. Perusse. 1993. Spatial aggregation, body size, and reproductive success in the freshwater mussel Elliptio complanata. Journal of the North American Benthological Society 12(2):148-156. Eckroat, L. R., E. C. Masteller, J. C. Shaffer, and L. M. Steele. 1993. The byssus of the zebra mussel (Dreissena polymorpha): morphology, byssal thread formation, and detachment. Jn: Zebra Mussels: Biology, Impacts, and Control. T. F. Nalepa and D. W. Schloesser, eds. pp. 239-263. Lewis/CRC Press, Boca Raton, Florida. Ehrenberg, H. 1957. Die Steinfauna der Brandungsufer ostholsteinrischer Seen. Archives de Hydrobiologia 53:87-159. French, J. R. P. and M. T. Bur. 1993. Predation by the zebra mussel (Dreissena polymorpha) by freshwater drum in western Lake Erie. In: Zebra Mussels: Biology, Impacts, and Control. T. F. Nalepa and D.W. Schloesser, eds. pp. 453-463. Lewis/CRC Press, Boca Raton, Florida. Gillis, P. L. and G. L. Mackie. 1994. Impact of the zebra mussel, Dreissena polymorpha, on populations of Unionidae (Bivalvia) in Lake St. Clair. Canadian Journal of Zoology 72:1260-1271. Great Lakes Basin Commission. 1975. Great Lakes Basin Framework Study: Appendix 11. Levels and Flows. Great Lakes Basin Commission, Ann Arbor, Michigan. 206 pp. Griffiths, R. W., D. W. Schloesser, J. H. Leach, and W. P. Kovalak. 1991. Distribution and dispersal of the zebra mussel (Dreissena poly- morpha) in the Great Lakes. Canadian Journal of Fisheries and Aquatic Sciences 48:1381-1388. Haag, W. R., D. J. Berg, D. W. Garton, and J. L. Farris. 1993. Reduced survival and fitness in native bivalves in response to fouling by the introduced zebra mussel (Dreissena polymorpha) in western 74 AMER. MALAC. BULL. 14(1) (1997) Lake Erie. Canadian Journal of Fisheries and Aquatic Science 50:13-19. Hamilton, D. J., C. D. Ankney, and R. C. Bailey. 1994. Predation of zebra mussels by diving ducks—an exclosure study. Ecology 75(2):521- 531. Hazlett, B. A. 1994. Crayfish feeding responses to zebra mussels depend on microorganisms and learning. Journal of Chemical Ecology 20(10):2623-2630. Hebert, P. D. N., B. W. Muncaster, and G. L. Mackie. 1989. Ecological and genetic studies on Dreissena polymorpha (Pallas): a new mol- lusc in the Great Lakes. Canadian Journal of Fisheries and Aquatic Science 46:1587-1591. Hebert, P. D. N., C. G. Wilson, M. H. Murdoch, and R. Lazar. 1991. Demography and ecological impacts of the invading mollusc Dreissena polymorpha. Canadian Journal of Zoology 69(2):405- 409. Hunter, R. D. and J. F. Bailey. 1992. Dreissena polymorpha (zebra mus- sel): colonization of soft substrata and some effects on unionid bivalves. The Nautilus 106(2):60-67. Kovalak, W. P., G. D. Longton, and R. D. Smithee. 1993. Infestation of power plant water systems by the zebra mussel (Dreissena poly- morpha). In: Zebra Mussels: Biology, Impacts, and Control., T. F. Nalepa and D. W. Schloesser, eds. pp. 359-380. Lewis/CRC Press, Boca Raton, Florida. Leach, J. H. 1993. Impacts of the zebra mussel (Dreissena polymorpha) on water quality and fish spawning reefs in western Lake Erie. In: Zebra Mussels: Biology, Impacts, and Control, T. F. Nalepa and D. W. Schloesser, eds. pp. 381-397. Lewis/CRC Press, Boca Raton, Florida. Lefevre, G. and W. C. Curtis. 1910. Studies on the reproduction and artifi- cial propagation of freshwater mussels. Bulletin of the United States Bureau of Fisheries 30:105-202. Lewandowski, K. 1976. Unionidae as a substratum for Dreissena poly- morpha Pall. Polskie Archiwum Hydrobiologii 23(3):409-420. Mackie, G. L. 1991. Biology of the exotic zebra mussel, Dreissena poly- morpha, in relation to native bivalves and its potential impact in Lake St. Clair. Hydrobiologia 219:251-268. Masteller, E. C., K. R. Maleski, and D. W. Schloesser. 1993. Unionid bivalves (Mollusca: Bivalvia: Unionidae) of Presque Isle Bay, Erie, Pennsylvania. Journal of Pennsylvania Academy of Science 67(3):120-126. Maclsaac, H. J. 1994. Comparative growth and survival of Dreissena polymorpha and Dreissena bugensis, exotic molluscs introduced to the Great Lakes. Journal of Great Lakes Research 20(4):783- 790. Mills, E. L., R. M. Dermott, E. F. Roseman, D. Dustin, E. Mellina, D. B. Conn, and A. P. Spidle. 1993. Colonization, ecology, and popula- tion structure of the quagga mussel (Bivalvia, Dreissenidae) in the lower Great Lakes. Canadian Journal Fisheries and Aquatic Sciences 50(11):2305-2314. Mitchell, C. A. and J. Carlson. 1993. Lesser scaup forage on zebra mussels at Cook Nuclear Plant, Michigan. Journal of Field Ornithology 64(2):219-222. Morton, B. S. 1969. Studies on the biology of Dreissena polymorpha Pall. III. Population dynamics. Proceedings of the Malacological Society of London 38:417-482. Nalepa, T. F. 1994. Decline of native bivalves (Unionidae: Bivalvia) in Lake St. Clair after infestation by the zebra mussel, Dreissena polymorpha. Canadian Journal Fisheries and Aquatic Science 51:2227-2233. Nalepa, T. F., D. J. Hartson, G. W. Gostenik, D. L. Fanslow, and G. A. Lang. 1996. Changes in the freshwater mussel community of Lake St. Clair: from Unionidae to Dreissena polymorpha in eight years. Journal of Great Lakes Research 22(2):354-369. Nalepa, T. F. and D. W. Schloesser, eds. 1993. Zebra Mussels: Biology, Impacts, and Control. Lewis/CRC Press, Boca Raton, Florida. 810 pp. Neves, R. J. 1987. Recent die-offs of freshwater mussels in the United States: an overview. In: Proceedings of a Workshop on Die-Offs of Freshwater Mussels in the United States. R. J. Neves, ed. pp. 7- 18. U. S. Fish and Wildlife Service/Upper Mississippi River Conservation Committee, Arlington, Virginia. Ricciardi, A., F.G. Whoriskey, and J. B. Rasmussen. 1995. Predicting the intensity and impact of Dreissena infestation on native bivalves from Dreissena field density. Canadian Journal of Fisheries and Aquatic Science 52(7):1449-1461. Ricciardi, A., F. G. Whoriskey, and J. B. Rasmussen. 1996. Impact of the Dreissena invasion on native unionid bivalves in the upper St. Lawrence River. Canadian Journal of Fisheries and Aquatic Sciences 53:1434-1444. Schloesser, D. W. 1996. Mitigation of unionid mortality caused by zebra mussel infestation: cleaning of unionids. North American Journal of Fisheries Management 16:942-946. Schloesser, D. W. and W. P. Kovalak. 1991. Infestation of unionids by Dreissena polymorpha in a power plant canal in Lake Erie. Journal of Shellfish Research 10(2):355-359. Schloesser, D. W. and T. F. Nalepa. 1994. Dramatic decline of unionid bivalves in off shore waters of western Lake Erie after infestation by the zebra mussel, Dreissena polymorpha. Canadian Journal of Fisheries and Aquatic Science 51:2234-2242. Schloesser, D. W., T. F. Nalepa, and G. L. Mackie. 1996. Zebra mussel infestation of unionid bivalves (Unionidae) in North America. American Zoologist 36:300-310. Sebestyen, O. 1938. Colonization of two new fauna-elements of Ponto- origin (Dreissena polymorpha Pall. and Corophium curvispinum G. O. Sars from a devium Wundsch) in Lake Balaton. Internationale Vereinigung fiir Theoretische und Angewandte Limnologie Verhandlungen 8:169-182. Snedecor, G. W. and W. G. Cockran. 1967. Statistical Methods. lowa State University Press, Ames, Iowa. 593 pp. Stanczykowska, A. 1977. Ecology of Dreissena polymorpha (Pallas) (Bivalvia) in lakes. Polskie Archiwum Hydrobiologii 24:461-530. Stanczykowska, A. and K. Lewandowski. 1993. Thirty years of studies of Dreissena polymorpha ecology in Mazurian Lakes of Northern Poland. In: Zebra Mussels: Biology, Impacts, and Control. T. F. Nalepa and D. W.Schloesser, eds. pp. 3-37. Lewis/CRC Press, Boca Raton, Florida. Strayer, D. L. 1991. Projected distribution of the zebra mussel, Dreissena polymorpha, in North America. Canadian Journal Fisheries and Aquatic Sciences 48:1389-1395. Strayer, D. A. and L. C. Smith. 1996. Relationships between zebra mus- sels (Dreissena polymorpha) and unionid clams during the early stages of the zebra mussel invasion of the Hudson River. Freshwater Biology 36:771-779. Tucker, J. K. 1994. Colonization of unionid bivalves by the zebra mussel, Dreissena polymorpha, in pool 26 of the Mississippi River. Journal of Freshwater Ecology 9(2):129-134. Tucker, J. K. and E. L. Atwood. 1995. Continuous backwater lakes as pos- sible refugia for unionid mussels in areas of heavy zebra mussel (Dreissena polymorpha) colonization. Journal of Freshwater Ecology 10(1):43-47. Williams, J. D., M. L. Warren, K. S. Cummings, J. L. Harris and R. J. Neves. 1993. Conservation status of freshwater mussels of the United States and Canada. Fisheries 18(9):6-22. Date of manuscript acceptance: 06 June 1997 Effects of quarantine times on glycogen levels of native freshwater mussels (Bivalvia: Unionidae) previously infested with zebra mussels Matthew A. Patterson!, Bruce C. Parker!, and Richard J. Neves2 1Department of Biology, Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24061, U.S. A. 2Virginia Cooperative Fish and Wildlife Research Unit, Department of Fisheries and Wildlife Sciences, Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24061, U.S. A. Abstract: The effects of zebra mussel infestation and subsequent quarantine on three mussel species were evaluated through glycogen analyses of man- tle tissue. Specimens of Amblema plicata (Say, 1817) and Quadrula pustulosa (1. Lea, 1831), collected from a heavily infested (> 350 zebra mussels/m2) reach of the Ohio River in 1996, had significantly lower glycogen levels (2.73 and 1.84 mg/g, respectively) than specimens collected from a lightly infested (< 5 zebra mussels/m2) reach upstream (8.08 and 6.20 mg/g, respectively). Levels of glycogen after 7, 14, and 30 d of quarantine in tanks declined dramati- cally with length of quarantine. After 30 d without supplemental feeding, mean glycogen levels of A. plicata collected from the low density reach had dropped to 15% of that of wild caught specimens (1.22 versus 8.08 mg/g, respectively). After 30 d, mean glycogen levels of Q. pustulosa also dropped sig- nificantly to 30% of that of wild caught specimens (1.90 versus 6.20 mg/g, respectively). Mean glycogen levels of Fusconaia ebena (I. Lea, 1831), collected from the heavily infested reach of the Ohio River dropped to extremely low levels (from 2.75 to 0.53 mg/g) after 30 d of quarantine. Specimens of F. ebena were quarantined for an additional 100 d because zebra mussels were found on unionids after 30 and 60 d of quarantine. Feeding every three days between 30-130 d of quarantine was insufficient to allow for recovery after 100 d (0.30 mg/g) or 130 d (0.34 mg/g). A 30 d quarantine of unionids removed from zebra mussel-infested waters causes a significant reduction in glycogen levels which are further reduced if additional quarantine time is required. Feeding of unionids is necessary to maintain their condition during lengthy quarantine, or more effective methods are needed to remove zebra mussels and thus shorten the required quarantine period. Key words: glycogen, zebra mussels, Unionidae, Ohio River, quarantine Freshwater mussels of the family Unionidae reach 1995, the Ohio River Valley Ecosystem Team identified the their greatest diversity in North America with nearly 300 potential decline of native aquatic mollusks as its top man- species (Williams et al., 1993). However, increased habitat agement priority (USFWS, 1995). alteration, siltation, and pollution have caused dramatic Resource agencies and universities are currently declines in both species diversity and richness (Bogan, testing removal and translocation as a management tool to 1993). Populations of freshwater mussels are now at further conserve declining numbers of unionids from zebra mussel- risk of extirpation or extinction from the exotic zebra mus- infested waters. The current protocol in West Virginia for sel, Dreissena polymorpha (Pallas, 1771). Since the zebra unionid salvage from zebra mussel infested waters requires mussel’s introduction into Lake St. Clair around 1985, this that all unionids be thoroughly scrubbed to remove zebra exotic mollusk has decimated local populations of freshwa- mussels (J. Clayton, pers. comm.). Cleaned unionids are ter mussels throughout the Great Lakes (Hunter and Bailey, then hand-inspected before being placed in aerated quaran- 1992; Gillis and Mackie, 1994; Schloesser and Nalepa, tine tanks for a minimum of 30 d to allow juvenile zebra 1994). The pelagic veliger stage has enabled zebra mussels mussels missed during the scrubbing procedure to become to colonize many of the large river systems of the south- visible. During quarantine, water quality parameters (i. e. eastern United States, and extremely high fecundities have temperature, dissolved oxygen, and pH) are monitored to allowed populations to increase exponentially after settle- provide suitable conditions for unionid survival. At the end ment of veligers (Sprung, 1991). Adult zebra mussels were of 30 d, individual unionids are inspected under 10X mag- first collected from the lower Ohio River in 1991 nification, and if zebra mussels are found, all specimens (USACOE, 1993), and densities reached nearly 100,000/m2 must be rescrubbed, placed in clean tanks, and quarantined in the lower river by July 1994 (A. Miller, pers. comm.). In for an additional 30 d. Finally, translocation can occur only American Malacological Bulletin, Vol. 14(1) (1997):75-79 75 76 AMER. MALAC. BULL. 14(1) (1997) when the native mussels are certified free of zebra mussels. Zebra mussel infestations in combination with col- lection, transport, and handling during quarantine, can lead to increased stress in freshwater mussels. Glycogen, an important energy reserve for animals, especially bivalves (de Zwann and Zandee, 1972; Barber and Blake, 1981; Bayne and Newell, 1983; Haag et al., 1993), has been shown to change in response to environmental perturba- tions such as temperature extremes, anaerobiosis, pollu- tants, or starvation (de Zwann and Wijsmann, 1976; Hummel et al., 1989). In marine bivalves, glycogen levels also have been shown to change seasonally (Hummel et al., 1988) in response to such factors as gametogenesis (Gabbott, 1983) and winter food shortages (Gade, 1983). Haag et al. (1993) showed that the mean glycogen content of Amblema plicata (Say, 1817) and Lampsilis radiata (Gmelin, 1791) from Lake Erie was significantly lower in zebra mussel-encrusted versus unencrusted control speci- mens. Thus, a glycogen assay was used in this experiment to assess the impact of zebra mussel infestation, removal, and 30 d quarantine on the physiological condition of fresh- water mussels collected from the Ohio River. Specific research objectives were to (1) quantify the glycogen levels of freshwater mussels infested with zebra mussels in high versus low density areas, and (2) assess the change in glycogen levels of unionids during quarantine periods rang- ing from 30 to 130 d. METHODOLOGY The effect of zebra mussel infestation on unionid glycogen levels was compared between specimens collect- ed from high versus low zebra mussel-infested sites on the Ohio River. To minimize the natural, seasonal fluctuation in glycogen levels, specimens were collected from the study sites between 23 July and 21 August 1996. Ten specimens each of Amblema plicata and Quadrula pustulosa (I. Lea, 1831) were collected from Ohio River Mile (ORM) 175.5 on 23 July 1996. This low infestation site near Parkersburg, West Virginia, had a mean density of 0.3 zebra mussels/m2, and a maximum of one zebra mussel/unionid (P. Morrison, pers. comm.). On 16 August 1996, ten specimens of A. pli- cata were collected from ORM 967. This heavily infested site near Paducah, Kentucky, had a mean density of 3,600 zebra mussels/m2 (A. Miller, pers. comm.). Because Q. pustulosa was uncommon at ORM 967, ten specimens were collected from ORM 397 on 21 August 1996. Zebra mussel densities at this site near Maysville, Kentucky, increased 30-fold between 1995 and 1996. With a mean density of 360 zebra mussels/m2 and a maximum of 92 zebra mus- sels/unionid (P. Morrison, pers. comm.), ORM 397 also was considered to be a heavily infested site. All specimens collected in the field were sacrificed on the day of collec- tion, shucked, weighed, preserved in 95% ethanol, and transported to the laboratory for analysis. To assess the effect of quarantine on unionid con- dition, additional specimens of Amblema plicata and Quadrula pustulosa (250 and 80, respectively) were col- lected from ORM 175.5. All specimens were aged, mea- sured, tagged, and transported in well water to 300 1, aerat- ed quarantine tanks on Middle Island, Ohio River Islands National Wildlife Refuge, in St. Mary’s, West Virginia. Because the quarantine tanks did not provide flow-through conditions, tank water was drained and filled with well water every 2 d. Specimens of A. plicata were placed in individual quarantine tanks at densities of 250/m2 and 65/m2, respectively, to determine possible density effects on glycogen stores. The 80 specimens of Q. pustulosa were placed in a third tank. During the 30 d quarantine, unionids were not fed, simulating likely conditions during recovery, quarantine, and relocation of threatened unionids. Ten spec- imens of each species were sacrificed from each tank at 7, 14, and 30 d of quarantine, and preserved in 95% ethanol for subsequent glycogen analysis. Heavily infested individuals of Amblema plicata and Quadrula pustulosa could not be used to monitor glycogen levels during quarantine because specimens could not be collected in sufficient numbers from the lower Ohio River. Instead, 250 specimens of Fusconaia ebena (I. Lea, 1831) were collected from ORM 967 on 16 August 1996, to determine the effect of quarantine on heavily infested unionids. Ten specimens were sacrificed in the field, and the remainder transported to the quarantine site. Again, ten specimens were sacrificed after 7, 14, and 30 d of quaran- tine, and preserved in 95% ethanol for subsequent glycogen analysis. At the end of 30 d, zebra mussels (3 mm in length) were discovered attached to the umbonal region of five F; ebena in quarantine. All specimens were removed, rescrubbed, hand-inspected, and placed in clean quarantine tanks for an additional 30 d. After the initial 30 d, mussels were fed from a fertilized algae tank every 3 d. After 60 d, zebra mussels again were found attached to the umbonal region of five specimens of F: ebena, and all specimens were rescrubbed, inspected, and placed in clean quarantine tanks. At the end of 100 d, no zebra mussels were found during inspection but an additional 30 d period was required to assure that no zebra mussels would be trans- ported out of quarantine. After 130 d, unionids were certi- fied free of zebra mussels and removed from quarantine. To assess the effect of this long-term quarantine period, ten specimens were sacrificed after 100 d and 130 d and pre- served in 95% ethanol for subsequent glycogen analysis. The glycogen content of all preserved specimens was determined using the technique described by Keppler PATTERSON ET AL.: GLYCOGEN LEVELS OF UNIONIDS 77 and Decker (1974). A 50-100 mg sample of preserved man- tle tissue was dissected, blotted dry to remove the ethanol and weighed. Tissue samples were homogenized for 2 hr in 3M perchloric acid and neutralized with 2M KHCO3. Glycogen was converted to glucose with amyloglucosidase (Sigma Chemical Co., St. Louis, Missouri), combined with a dye solution containing o-dianisidine dihydrochloride, and absorbance measured in a spectrophotometer at 450 nm. Total glycogen was determined from a standard curve of glycogen extracted from the blue mussel, Mytilus edulis (Linné, 1758), and expressed in milligrams glycogen/gram preserved mantle tissue. It should be noted that 95% ethanol dehydrates tissue and preserved tissue weights like- ly underestimate wet tissue weights. However, dehydration also reduces error that can result from any change in tissue water levels during stress. Mean glycogen levels were not standardized by total body weight because simple regres- sion revealed no correlation between wet weight and glyco- gen content (r2 < 0.10). The mean glycogen levels of all treatments (high versus low zebra mussel density and 7-30 d of quarantine) were normally distributed according to the Kolmogorov-Smirnov goodness of fit test (@ = 0.05). However, zebra mussel infestation and starvation during quarantine uniformly decreased the glycogen levels of all specimens and consequently decreased overall variance. Following Lentner (1993), the sample variances were equalized using the square root of each individual glycogen value. Converted mean glycogen levels were then com- pared using ANOVA. If significant differences were detect- ed, Scheffe F-test was used to determine the statistical sig- nificance of individual treatments. RESULTS Initial mean glycogen levels of Amblema plicata collected from the heavily infested site (ORM 967) were significantly lower (p < 0.05) than those collected from the upper river at ORM 175.5 (2.73 + 2.81 mg/g versus 8.08 + 4.26 mg/g, respectively). The initial mean glycogen level of Quadrula pustulosa collected from ORM 397 also was sig- nificantly lower (p < 0.05) than that collected from ORM 175.5 (1.84 + 1.23 mg/g versus 6.20 + 2.89 mg/g, respec- tively). During quarantine, the mean glycogen level of A. plicata collected from ORM 175.5 dropped significantly (p < 0.05) after 7 d (Fig. 1). While significant differences were not observed between 7 and 14 d (p > 0.3), the mean glyco- gen level continued to drop significantly (p < 0.05) between 14 and 30 d until reaching 15% of that measured in wild- caught specimens (Fig. 1). The mean glycogen level of Q. pustulosa collected from ORM 175.5 also dropped signifi- cantly (p < 0.05) after 7 d of quarantine (Fig. 1). Between 7 and 14 d, the mean glycogen level increased; however, the increase was not statistically significant (p > 0.1). At 30 d, the mean glycogen level dropped significantly (p < 0.05) to only 31% of that measured in wild-caught specimens (Fig. 1). There was no significant difference (p > 0.3) in mean glycogen levels between mussels held at 250/m2 and 65/m2 after 7 d (3.56 + 1.78 mg/g and 4.09 + 2.18 mg/g, respec- tively) or 14 d (3.27 + 1.74 mg/g and 3.10 + 1.57 mg/g, respectively). Specimens of Fusconaia ebena collected from ORM 967 also showed a significant decline (p < 0.05) in the mean glycogen level after 7 d of quarantine (Fig. 2). However, significant changes were not detected for the remainder of the quarantine period. After 30 d, the mean glycogen level was only 20% of that measured in wild- caught specimens (Fig. 2). Feeding of unionids every 3 d between 30 and 130 d was not sufficient to allow unionid glycogen levels to recover. After 130 d, the mean glycogen level was only 12% of that measured in wild-caught speci- mens (Fig. 2). DISCUSSION While different densities (up to 250 unionids/m2) in quarantine had no significant effect on the glycogen stores of Amblema plicata, it is clear that previous levels of zebra mussel infestation and starvation during quarantine signifi- cantly reduce unionid energy stores. By attaching in great densities to the outer shell of living unionids, zebra mussels reduce glycogen stores, presumably by reducing vital food resources, disrupting proper feeding and respiration, and preventing valve opening and closing (Mackie, 1991). Glycogen (mg/g) Time (Days) Fig. 1. Glycogen levels (mg/g) of Amblema plicata and Quadrula pustu- losa at 1, 7, 14, and 30 d of quarantine (n = 10/sampling period). 78 AMER. MALAC. BULL. 14(1) (1997) ——F. ebena 25 a 2 i") =) = 5 1.5 8 a) o | 05 0 1 yi 14 c 100 130 Time (Days) Fig. 2. Glycogen levels (mg/g) of Fusconaia ebena at 1, 7, 14, 30, 100, and 130 d of quarantine (n = 10/sampling period). Thus, energy stores were already at low levels when union- ids entered the 30 d quarantine period. Assay results from Fusconaia ebena revealed that low glycogen levels of unionids removed from areas with high densities of zebra mussels reach dangerously low levels during a 30 d quaran- tine period. It is unclear whether a threshold level of glyco- gen is required to cause mortality, but low energy stores after quarantine can decrease the likelihood that unionids will survive the relocation process. Unionids collected from high quality habitat with low zebra mussel densities can have sufficient energy stores to survive a quarantine period and subsequent translocation, however, unionids in areas with high densities of zebra mussels are the primary candi- dates for relocation. Thus, when unionids from zebra mus- sel-infested waters are translocated, a major limiting factor could be the physiological condition and energy reserves of unionids at the time of relocation. In a review of the literature, Cope and Waller (1995) reported that survival of translocated unionids is typically low (< 50%) and is influenced by many factors. Factors affecting translocation success such as habitat suit- ability, numbers of individuals released, and the frequency of release have been given significant attention in recent years for both terrestrial and aquatic organisms (Griffith et al., 1989; Cope and Waller, 1995). However, no attention has been given to the physiological condition or energy reserves of relocated organisms. In order to reduce the like- lihood of latent mortality of mussels salvaged from zebra mussel-infested waters, it is necessary to either provide suf- ficient food and favorable water quality conditions during quarantine or to have a brief quarantine period to ensure that unionids have sufficient energy stores to recover from the stressful relocation to new environments. As judged by the energy reserves in specimens of Fusconaia ebena from 30-130 d of quarantine, starved unionids can reach a point where supplemental feeding contributes little to the recovery of energy reserves. Thus, a detailed study to determine the amount of food required to maintain unionid condition during quarantine is needed. In addition to maintaining unionid condition, food supple- ments also will increase the growth rate of juvenile zebra mussels that are missed during the scrubbing procedure. It is evident that small zebra mussels can avoid extensive scrubbing and inspection, possibly by residing in the crevices of damaged shells. Because the purpose of quaran- tine is to guarantee the absence of zebra mussels, increased growth rates would enhance detection and justify a reduc- tion in the quarantine period. More effective techniques of zebra mussel removal also should be developed to reduce or perhaps eliminate the need for a lengthy quarantine peri- od. Assay results from this study reveal that the glyco- gen levels of all three species decreased significantly after 7 d, and then stabilized between 7 and 14 d. Thus, a reduc- tion in the quarantine period from 30 to 15 d would greatly improve the overall condition of unionids prior to translo- cation. However, under current protocol standards, unionids must endure a minimum of 30 d of quarantine and a total of 60 d if zebra mussels are detected, which can cause glyco- gen levels of unionids to decline to life-threatening levels. Thus, one of the greatest concerns during the salvage of zebra mussel-infested unionids should be the physiological condition of unionids at the time of their final relocation. ACKNOWLEDGMENTS We would like to thank Patty Morrison, Mitch Ellis, and every- one else at the Ohio River Islands National Wildlife Refuge Office as well as Dr. Andrew Miller and the United States Army Corps of Engineers for their help in collecting mussels from the Ohio River. We would also like to thank the many volunteers that assisted with collection and processing of mussels in the field. Thanks also go to Li-Yen Chen for assistance in developing the glycogen assay and Catherine Gatenby for building the quarantine facility, assisting in the field, and assisting with manuscript revisions. Finally, very special thanks go to Jim Dotson for his long hours in the field and his incredible ability to fix anything in quarantine. This study was funded by Quick Response Funds of the National Biological Service. LITERATURE CITED Barber, B. J. and N. B. Blake. 1981. Energy storage and utilization in rela- tion to gametogenesis in Argopecten irridians concentricus (Say). Journal of Experimental Marine Biology and Ecology 52:121- 134. Bayne, B. L. and R. C. Newell. 1983. Physiological energetics of marine PATTERSON ET AL.: GLYCOGEN LEVELS OF UNIONIDS a? molluscs. In: The Mollusca, Vol. 4, Physiology, Part 1, A. S. M. Saleuddin and K. W. Wilbur, eds. pp. 407-515. Academic Press, New York. Bogan, A. E. 1993. Freshwater bivalve extinctions (Mollusca: Unionoida): a search for causes. American Zoologist 33(6):599-609. Cope, G. W. and D. L. Waller. 1995. Evaluation of freshwater mussel relocation as a conservation and management strategy. Regulated Rivers: Research and Management 11(2):147-155. de Zwann, A. and D. I. Zandee. 1972. Body distribution and seasonal changes in the glycogen content of the common sea mussel Mytilus edulis. Comparative Biochemistry and Physiology 43A:53-58. de Zwann, A. and T. C. M. Wijsman. 1976. Anaerobic metabolism in Bivalvia (Mollusca). Characteristics of anaerobic metabolism. Comparative Biochemistry and Physiology 54B:313-324. Gabbott, P. A. 1983. Developmental and seasonal metabolic activities in marine molluscs. In: The Mollusca, Vol. 2, Environmental Biochemistry and Physiology, P. W. Hochachka, ed. pp. 165-217. Academic Press, New York. Gade, G. 1983. Energy metabolism of arthropods and mollusks during environmental and functional anaerobiosis. Journal of Experimental Zoology 228:415-429. Gillis, P. L. and G. L. Mackie. 1994. Impact of the zebra mussel, Dreissena polymorpha, on populations of Unionidae (Bivalvia) in Lake St. Clair. Canadian Journal of Zoology 72(7):1260-1271. Griffith, B., J. M. Scott, J. W. Carpenter, and C. Reed. 1989. Translocation as a species conservation tool: status and strategy. Science 245:477-480. Haag, W. R., D. L. Berg, D. W. Garton, and J. L. Farris. 1993. Reduced survival and fitness in native bivalves in response to fouling by the introduced zebra mussel (Dreissena polymorpha) in western Lake Erie. Canadian Journal of Fisheries and Aquatic Sciences 50 (1):13-19. Hummel, H., L. de Wolf, and A. W. Fortuin. 1988. The annual cycle of glycogen in estuarine benthic animals. Hydrobiological Bulletin 22:199-202. Hummel, H., L. de Wolf, W. Zurburg, L. Apon, R. H. Bogaards, and M. Van Ruitenburg. 1989. The glycogen content in stressed marine bivalves: the initial absence of a decrease. Comparative Biochemistry and Physiology 94B(4):729-733. Hunter, R. D. and J. F. Bailey. 1992. Dreissena polymorpha (zebra mus- sel): colonization of soft substrata and some effects on unionid bivalves. The Nautilus 106(2):60-67. Keppler, D. and K. Decker. 1974. Glycogen determination with amy- loglucosidase. In: Methods of Enzymatic Analysis, H. U. Bergmeyer, ed. pp. 11-17. Academic Press, New York. Lentner, M. 1993. Experimental Design and Analysis. Valley Book Company, Blacksburg. 585 pp. Mackie, G. L. 1991. Biology of the exotic zebra mussel, Dreissena poly- morpha, in relation to native bivalves and its potential impact in Lake St. Clair. Hydrobiologia 219:251-268. Schloesser, D. W. and T. F. Nalepa. 1994. Dramatic decline of unionid bivalves in offshore waters of western Lake Erie after infestation by the zebra mussel, Dreissena polymorpha. Canadian Journal of Fisheries and Aquatic Sciences 51(10):2234-2242. Sprung, M. 1991. Costs of reproduction: a study on metabolic require- ments of the gonads and fecundity of the bivalve Dreissena polymorpha. Malacalogia 33(1-2):63-70. United States Army Corps of Engineers (USACOE). 1993. Final Supplement 1, Final Environmental Impact Statement for the Replacement of Locks and Dams 52 and 53 (Olmsted Locks and Dams) Lower Ohio River, IL-KY. U. S. Army Corps of Engineers District, Louisville, Kentucky. 368 pp. United States Fish and Wildlife Service (USFWS). 1995. Strategic Plan for Conservation of Fish and Wildlife Service Trust Resources in the Ohio River Valley Ecosystem. U. S. Government Printing Office, Washington, D. C. 13 pp. Williams, J. D., M. L. Warren, Jr., K. S. Cummings, J. L. Harris, and R. J. Neves. 1993. Conservation status of freshwater mussels of the Unites States and Canada. Fisheries 18(9):6-22. Date of manuscript acceptance: 06 August 1997 Research Note Observations on the reproductive biology of the octopod Eledone gaucha Haimovici, 1988, in southern Brazil Jose Angel A. Perez!, Manuel Haimovici2, and Roberta Aguiar dos Santos2 1Faculdade de Ciéncias do Mar, Universidade do Vale do Itajai, Cx. Postal 360, Itajaf, SC, 88.302-202, Brazil 2Departamento de Oceanografia, Fundagao Universidade do Rio Grande, Cx. Postal 474, Rio Grande, RS, 96500-900, Brazil Abstract: Maturation, fecundity and reproductive cycle of the octopod Eledone gaucha Haimovici, 1988, were studied based on preserved samples col- lected during groundfish surveys conducted off southern Brazil. The patterns of sexual maturation of males and females and low fecundity were similar to the congeneric and sympatric E. massyae Voss, 1964. It is suggested that E. gaucha does not exhibit a marked seasonal reproductive cycle and more than one breeding group overlap on the continental shelf. Key words: Octopoda, Eledone, sexual maturation, Brazil The small congeneric octopods, Eledone massyae Voss, 1964, and E. gaucha Haimovici, 1988, coexist year- round on the sandy and muddy bottoms of the middle and outer continental shelf off southern Brazil (Haimovici and Andriguetto, 1986; Haimovici and Perez, 1991a, b). A series of groundfish surveys conducted between 1982 and 1987 by the R/V “Atlantico Sul” of the Rio Grande University (Table 1), collected both species between 33°45’ and 30°43’ S and the isobaths of 40 and 160 m. Specimens of Eledone were fresh-frozen or preserved on board, for comparative genetic and morphological studies (Haimovici, 1988; Levy et al., 1988) and basic descriptions of their pop- ulation structure, reproductive biology, and ecology (Perez et al., 1990; Perez and Haimovici, 1991, 1995). A more detailed study on maturation and reproductive cycle was conducted on the larger and more commonly trawled E. massyae (fide Perez and Haimovici, 1991). Less informa- tion was available on the smaller and scarcer E. gaucha. Our observations on the reproductive biology of this species are summarized in this note. A total of 88 males and 95 females of Eledone gaucha (Table 1) were fixed in 10% formalin and preserved in 70% ethanol. All specimens were weighed, measured (ML, dorsal mantle length) in millimeters, and their gonads and gonoducts dissected out and weighed to the nearest 0.1 g. Macroscopic maturity stages were assigned to males and females according to the scale proposed for E. massyae (fide Perez and Haimovici, 1991). For females the stages were: I, immature; II, early maturation; III, maturing; IV, advanced maturity. The female maturity scale does not con- sider fully mature or spawned ovaries, because these were not found in the samples. For males the stages were: I, immature; II, maturing; III, mature (with functional sper- matophores in the spermatophore sac); IV, spent. In females, all eggs were counted and length measured to the nearest 0.1 mm and the presence of sperm sacs (evaginated spermatophores) and free sperm within the ovary were recorded as evidence of mating. In males, all sper- matophores stored in the spermatophore sac were counted. 3, 2.5 | ? Bad : 3 2 ile o 5151 ; S ; iy £ 0.5 > © 0 ———— 0 5 10 15 20 25 Body Weight (g) Fig. 1. Ovary and oviduct weight as a function of total body weight of female Eledone gaucha collected off southern Brazil. American Malacological Bulletin, Vol. 14(1) (1997):81-84 82 AMER. MALAC. BULL. 14(1) (1997) Stage I 1) 5 t+ + tt 015 3 45 6 75 9 ae a ee 45 6 75 9 : 015 3 4. c Egg length (mm) Fig. 2. Length-frequency distribution of ovarian eggs of female Eledone gaucha pooled by maturity stages: I, immature; II, early maturation; III, maturing; IV, advanced maturity. Females ranged from 14 to 55 mm ML and 0.9 to 25.5 g wet weight (Table 1). The ovary and oviducts showed enlargement in females heavier than 5.0 g and reached a maximum of 17% of total weight, varying greatly at a given body size (Fig. 1). In the examined ovaries, most of the eggs enlarged at about the same time throughout maturation (Fig. 2). A bimodal distribution was observed in females at stage III. In the most advanced stage IV, most eggs concentrated around a single mode (5.0 mm). The largest eggs were ca. 7.4 mm long as observed in a female (14.4 g wet weight) caught in June 1980. Even these eggs, however, were longitudinally striated indicating that they were not fully mature. Neither mature nor spent females were caught during this study. Mated females, with free sperm and sperm sacks in the ovary, were caught in all sampled months mostly in females in advanced maturity stages. The smallest maturing female with evidence of mating was 16.6 mm ML with mean egg length of 0.84 mm. The number of eggs present in the ovary ranged from 5 to 58 (n = 91, mean 35 + 12.7 SD), and the relative fecundity (eggs per gram of total weight) ranged from 0.68 to 51.11 (n = 91, mean 7.4 + 8.7 SD). No significant linear correlation was found between the number of eggs and female weight (r = 0.031; p > 0.50) or mantle length (r = 0.035; p > 0.50). Males ranged from 14 to 47 mm ML and 1.2 to 17.8 g wet weight (Table 1). Genital bag (testis, sper- matophore sac, and spermatophore glandular systems) enlargement had begun in males heavier than 3.0 g and larger than 22.0 mm ML, and reached a maximum of 9.5% of the total weight (Fig. 3). Spermatophores were observed in males as small as 20 mm ML. Half of the males carrying spermatophores in the spermatophore sac, however, were larger than 31 mm ML. A significant positive correlation 0.8 cc) =) 0.6 e S bo 0.4 ==} = 0.2 S e) 0 0 5 10 15 20 Body Weight (g) Fig. 3. Genital bag weight (testis, spermatophore glands, and sper- matophore sac) as a function of total body weight of male Eledone gaucha collected off southern Brazil. PEREZ ET AL.: REPRODUCTION OF ELEDONE GAUCHA IN SOUTHERN BRAZIL 83 Table 1. Summary of Eledone gaucha samples collected in 11 surveys conducted from 1980 and 1985 off southern Brazil. n, sample size; ML, mantle length. Males Females Year Month Depth n ML (mm) Weight (g) n ML (mm) Weight (g) (m) min - max min-max min-max min-max 1980 June 11-140 y) 26 5.2 -6.1 6 32 - 55 3.1 - 19.0 1980 July 40-120 2 30 - 34 8.9-14.8 0 1981 April 13-89 6 24 -34 6.3 - 10.0 0 1982 January 10-63 20 14-44 1.2 - 13.7 20 14 - 35 0.9 - 10.0 1983 April 13-122 17 18 - 34 2.6 - 10.0 17 15 - 41 14-144 1983 July — 1 44 13.7 0 1983 August 14-110 21 26 - 45 3.4-14.8 23 25 - 44 1.9 - 25.5 1983 November 10-100 4 17 - 47 1.2-17.8 7 31 - 47 1.9- 18.0 1984 January 12-200 2 24 - 35 6.0- 11.8 4 30 - 35 10.5 - 15.0 1984 November 10-197 8 20 - 43 2.2 - 16.8 7 41-49 14.2 - 25.5 1985 September — 1 37 12.9 11 25 - 34 5.2 - 12.8 Totals 88 14 - 47 1.2 - 17.8 95 14-55 0.9 - 25.5 was found between the number of spermatophores stored in 0.2 the spermatophore sac and male weight (r = 0.522; p < 0.001), mantle length (r = 0.303; p < 0.005) and genital bag . weight (r = 0.446; p < 0.001). The maximum number of oe 0.15 | r dian, 4 spermatophores observed in a single male was 82 (mean a : 4 4 23.9 + 3.6 SD, n = 86). Males with spermatophores = 4 4 occurred in all sampled months. z 0.1 ' “a 4 The maturation cycle was assessed by the variation 5 P ] 4 4 of the gonad index (gonad weight expressed as a proportion 0 0.05 | 3 4 j : a5 4h of the total wet weight) of individuals collected during all : i, BP a 4 surveys pooled by month (Figs. 4 and 5). Octopods in both | Be 3 Z 3 x2 initial and advanced stages of maturation co-occurred in all i ee eee ener oe aus peels eh) _ Z sampled months. The pattern was less clear among males 0 2 4 6 8 10 2 because young individuals were probably not well repre- Month sented in the samples (Perez and Haimovici, 1995). Whereas the data suggest an overlap of generations within a year, the duration of the maturation cycle could not be defined. The pattern observed, however, could be an arti- fact of pooling individuals collected in different years by month, specially if the reproductive cycle is shorter than one year, such as that of other small octopodids (Forsythe, 1984). Because the data set available is scarce and sparsely distributed within the five-year period of sampling, an ade- quate test of the latter hypothesis was not possible. The maturation processes of Eledone gaucha are consistent with those of E. massyae (fide Perez and Haimovici, 1991) and in the European species, E. cirrhosa (Lamarck, 1798) and E. moschata (Lamarck, 1798) (Mangold, 1983; Boyle and Knobloch, 1983; Moriyasu, 1988). Females reach maturity later and at a wider range of sizes than males. Females can mate while still not fully mature and store sperm in the apical filaments of the eggs until vitellogenesis is completed (Perez et al., 1990). Mature and spent females were not vulnerable to the trawl Fig. 4. Monthly variation of gonad indices of female Eledone gaucha col- lected off southern Brazil. The numbers represent maturity stages assigned for each individual: 1, immature; 2, early maturation; 3, maturing; 4, advanced maturity. net. As pointed out in previous studies (Perez and Haimovici, 1991, 1995), this could relate to migration off- shore, out of the study area, towards deep rocky bottoms suitable for spawning. Although few animals were present in each sample of E. gaucha, there was no evidence of sea- sonality in the reproductive cycle in contrast to that of E. massyae, in which mating and spawning were concentrated in the spring and autumn, respectively (Perez and Haimovici, 1991). In addition, as discussed in a previous study (Perez and Haimovici, 1995), more than one cohort of E. gaucha seem to overlap throughout the year, each of them with reproductive cycles of uncertain duration but possibly sub-annual. If confirmed, this feature plus dietary differences during the adult phase (Perez and Haimovici, 84 AMER. MALAC. BULL. 14(1) (1997) 0.1 i 3 0.08 2 s { 20.06 + 2 %3 = 7 2 33 3 0.04 re ei 1® 2 = U. T ? “53 eae 4 Se ge om 0.02 43 %, " 0 +: +—_+——+ + + aera +t +4 0 2 4 6 8 10 12 Month Fig. 5. Monthly variation of gonad indices of male Eledone gaucha col- lected off southern Brazil. The numbers represent maturity stages assigned for each individual: 1, immature; 2, early maturation; 3, mature; 4, spent. 1995) could allow ecological niche divergence between these two sympatric octopods. ACKNOWLEDGMENTS We thank C. M. Vooren, N. Brunetti, R. K. O’Dor, and J. Voight for critically reviewing previous versions of this manuscript. Two of the authors (JAAP and RAS) were supported by a CAPES scholarship (Ministry of Education, Government of Brazil). LITERATURE CITED Boyle, P. R. and D. Knobloch. 1983. The female reproductive cycle of the octopus Eledone cirrhosa (Lamarck). Journal of the Marine Biology Association of the United Kingdom 63:71-83. Forsythe, J. W. 1984. Octopus joubini (Mollusca: Cephalopoda): a detailed study of growth through the full life cycle in a closed sea- water system. Journal of Zoology, London 202: 393-417. Haimovici, M. 1988. Eledone gaucha, a new species of eledonid octopod (Cephalopoda: Octopodidae) from southern Brazil. The Nautilus 102:82-87. Haimovici, M. and J. M. Andriguetto Fo. 1986. Cefalépodes costeiros cap- turados na pesca de arrasto do litoral sul do Brasil. Arquivos de Biologia e Tecnologia, Curitiba 29(3): 473-495. Haimovici, M. and J. A. A. Perez. 1991a. Abundancia e distribuigado de cefalépodes em cruzeiros de prospec¢4o pesqueira demersal na plataforma externa e talude continental do sul do Brasil. Atlantica 13(1):189-200. Haimovici, M. and J. A. A. Perez. 1991b. Coastal cephalopod fauna of southern Brazil. Bulletin of Marine Science 49(1-2):221-230. Levy, J. A., M. Haimovici, and M. B. Conceigéo, 1988. Genetic evi- dences for two species of the genus Eledone (Cephalopoda: Octopodidae) from southern Brazil. Comparative Biochemistry and Physiology 90B:275-277. Mangold, K. 1983. Eledone moschata. In: Cephalopod Life Cycles, Vol. I, Species Accounts, P. R. Boyle, ed., pp. 387-400. Academic Press, London. Moriyasu, M. 1988. Analyse de la maturation sexuelle d’Eledone cirrosa (Cephalopoda: Octopoda) du Golfe du Lion. Aquatic Living Resources 1:59-65. Perez, J. A. A. and M. Haimovici. 1991. Sexual maturation and reproduc- tive cycle of Eledone massyae Voss, 1964 (Cephalopoda: Octopodidae) in southern Brazil. Bulletin of Marine Science 49(1-2):270-279. Perez, J. A. A. and M. Haimovici. 1995. The descriptive ecology of two South American eledonids (Cephalopoda: Octopodidae). Bulletin of Marine Science 53(3):757-771. Perez, J. A. A., M. Haimovici, and J. C. B. Cousin. 1990. Some observa- tions on sperm storing mechanisms and fertilization in eledonids from Brazil. Malacologia 32(1):147-154. Date of manuscript acceptance: 05 March 1996 64th ANNUAL MEETING THE AMERICAN MALACOLOGICAL UNION WASHINGTON, D. C. JULY 25, 1998 In 1998 the American Malacological Union will meet jointly with Unitas Malacologica and the Western Society of Malacology in Washington, D. C. The meeting site is the Smithsonian Institution on the National Mall. A large number of malacologists from outside of North America are expected to attend, and you are urged to participate in what surely will be a unique meeting. A detailed meeting announcement, registration, materials, and the call for papers and posters can be downloaded from the AMU web site (http://erato.acnatsci.org/amu). Three plenary symposia are scheduled in addition to contributed paper sessions: 1) Refining Molluscan Characters Organizers: Tim Collins (collins@servms.fiu.edu), Gerhard Haszprunar, Diana Lipscomb, and Winston Ponder. 2) Interactions between Molluscs and Humans Organizers: Philippe Bouchet, George Davis (davis@say.acnatsci.org), Eric Hochberg, and Gerrado Vasta 3) Bridging Temporal Scales in Malacology: Uniting the Living and the Dead Organizers: Satoshi Chiba, Douglas Erwin (erwin.doug@nmnh.si.edu) and David Reid There will also be an opportunity to present research in a poster format. Due to the potentially large num- ber of contributed papers expected, you are urged to submit abstracts in a timely manner as these will be processed on a first-come, first-served basis. Three social events are planned, highlighted by a concluding evening cruise along the Potomac River. Please note that the single registration fee includes all of the social events! During open evenings partici- pants will be encouraged to take advantage of the fine dining opportunities in the D. C. area. Washington is a veritable tourist mecca and participants also should plan on spending time visiting some of the museums, galleries, monuments, federal buildings,and parks. For further information please contact: Robert Hershler, President, AMU Department of Invertebrate Zoology, NHB Stop 118 National Museum of Natural History Smithsonian Institution Washington, D. C. 29560 e-mail: HERSHLER.ROBERT@NMNH.SILEDU 85 IN MEMORIAM Donald R. Moore Gale G. Sphon, Jr. Margaret C. Teskey 86 1 oy a y \y iy CONTRIBUTOR INFORMATION The American Malacological Bulletin serves as an outlet for reporting notable contributions in malacological research. Manuscripts concerning any aspect of original, unpublished research, important short reports, and detailed reviews dealing with molluscs will be considered for publi- cation. Each original manuscript and accompanying illustra- tions must be submitted with two additional copies for review purposes. Text must be typed on one side of 8-1/2 x 11 inch bond paper, double-spaced, and all pages num- bered consecutively with numbers appearing in the upper right hand corner of each page. Leave ample margins on all sides. Form of the manuscript should follow that outlined in the Council of Biology Editors Style Manual (sixth edition, 1994). This can be purchased from the CBE, 11 S. LaSalle Street, Suite 1400, Chicago, IL 60603, U.S.A. Text, when appropriate, should be arranged in sections as follows: 1. Cover page with title, author(s) and address(es), and suggested running title of no more than 50 characters and spaces. Authors should also supply five key words, placed at the base of this page, for indexing purposes. Abstract (less than 5% of manuscript length) 3. Text of manuscript starting with a brief introduction followed by methodology, results, and discussion. Separate sections of text with centered subtitles in cap- ital letters. Acknowledgments 5. Literature cited 6. Figure captions All binomens must include the author and date attrib- uted to that taxon the first time the name appears in the manuscript [e. g. Crassostrea virginica (Gmelin, 1791)]. This includes nonmolluscan taxa. The full generic name along with specific epithet should be written out the first time that taxon is referred to in each paragragh. The gener- ic name can be abbreviated in the remainder of the para- graph as follows: C. virginica. References should be cited within text as follows: Hillis (1989) or (Hillis, 1989). Dual authorship should be cited as follows: Yonge and Thompson (1976) or (Yonge and Thompson, 1976). Multiple authors of a single article should be cited as follows: Beattie et al. (1980) or (Beattie et al., 1980). In the literature cited section of the manuscript refer- ences must also be typed double spaced. All authors must be fully identified, listed alphabetically and journal titles must be unabbreviated. Citations should appear as follows: Beattie, J. H, K. K. Chew, and W. K. Hershberger. 1980. Differential survival of selected strains of Pacific oys- ters (Crassostrea gigas) during summer mortality. Proceedings of the National Shellfisheries Association 70(2):184-189. Hillis, D. M. 1989. Genetic consequences of partial self fertilization on population of Liguus fasciatus (Mollusca: Pulmonata: Bulimulidae). American Malacological Bulletin 7(1):7-12. Seed, R. 1980. Shell growth and form in the Bivalvia. In: Skeletal Growth of Aquatic Organisms, D. C. Rhoads and R. A. Lutz, eds. pp. 23-67. Plenum Press, New York. Yonge, C. M. and T. E. Thompson. 1976. Living Marine Molluscs. William Collins Son and Co., Ltd., London. 288 pp. Illustrations should be clearly detailed and readily reproducible. All line drawings should be in black, high quality ink. Photographs must be on glossy, high contrast paper. All diagrams must be numbered in the lower right hand corners and adequately labeled with sufficiently large labels to prevent obscurance with reduction by one half. Magnification bars must appear on the figure, or the cap- tion must read Horizontal field width = x mm or x pm. All measurements must be in metric units. All illustrations submitted for publication must be fully cropped, mounted on a firm white backing ready for reproduction, and have author’s name, paper title, and figure number on the back. All figures in plates must be nearly contiguous. Additional figures submitted for review purposes must be of high quality reproduction. Xerographic reproduction of pho- tomicrographs or detailed photographs will not be accept- ablefor review. Abbreviations used in figures should occur in the figure caption. Indicate in text margins the appropri- ate location in which figures should appear. Color illustra- tions can be included at extra cost to the author. Original illustrations will be returned to author if requested. Any manuscript not conforming to AMB format will be returned to the author for revision. Final submission of accepted, revised manuscripts should include two typed copies of the text, tables, etc. and an additional copy in electronic form on 3.5” diskette. The electronic version should be readable as non-formatted ASCII files. New Taxa. The Bulletin welcomes complete descriptions of new molluscan taxa. Establishment of new taxa must conform with the International Code of Zoological Nomenclature (1985). Descriptions of new species-level taxa must include the following information in the order as given: higher taxon designation as needed for clarity; fami- ly name with author and date; generic name with author and date; Genus species author sp. nov. followed by numeration of all figures and tables; complete synonymy (if any); listing of type material with holotype and any other type material clearly designated along with complete museum catalogue or accession information; listing of all additional non-type material also with full museum deposi- tion information; type locality; diagnosis and full descrip- tion of material done in telegraphic style including measurements and zoogeographic distribution as neces- sary; discussion; etymology. Descriptions of new supraspe- cific taxa should include type species (for new genus) or type genus (for new family), diagnosis and full description done in telegraphic style, and list of included taxa. Proofs. Page proofs will be sent to the author and must be checked for printer’s errors and returned to the printer within a three day period. Significant changes in text, other than printer errors, will produce publishing charges that will be billed to the author. Mailing. All overseas mailing must be done via airmail. The American Malacological Union will not be responsible for deferred publication of manuscripts delayed in surface mail. Charges. There are no mandatory page costs to authors lacking financial support. Authors with institutional, grant or other research support will be billed for page charges. The current rate is $35.00 per printed page. Acceptance and ultimate publication is in no way based on ability to pay page costs. Reprints. Order forms and reprint cost information will be sent with page proofs. The author receiving the order form is responsible for insuring that orders for any coau- thors are also placed at that time. Submission. Submit all manuscripts to Dr. Ronald B. Toll, Editor-in-Chief, American Malacological Bulletin, Depart- ment of Biology, Wesleyan College, 4760 Forsyth Road, Macon, Georgia 31297-4299 U.S.A. (e-mail: ron_toll@ post.wesleyan-college.edu). Subscription Costs. Institutional subscriptions are avail- able at a cost of $48.00 per volume. Membership in the American Malacological Union, which includes personal subscriptions to the Bulletin, is available for $35.00 ($15.00 for students, $45 for affiliated clubs). All prices quoted are in U.S. funds. Outside the U.S. postal zones, add $5.00 seamail and $10.00 airmail per volume within North America, or $10.00 seamail and $20.00 airmail per volume in other locations. For subscription and back-issue information contact Dr. Paula M. Mikkelsen, Managing Editor, American Malacological Bulletin, Department of Invertebrates, American Museum of Natural History, Central Park West at 79th Street, New York, New York 10024-5192, U.S.A. (e-mail: mikkel@amnh.org). NT 3 9088 01276 4908 AMERICAN MALACOLOGICAL BULLETIN VOLUME 14 1998 NUMBER 2 Journal of the American Malacological Society CONTENTS Boucardicus victorhernandezi, a new, endangered species of cyclophorid land snail from Madagascar. KENNETH C. EMBERTOON....000..cccccceccccscssescsesesessssssecseeseseeesessasseseassesseeaseasees 87 Differences in the ecology and distribution of lotic pulmonate and prosobranch gastropods. KENNETH M. BROWN, JAMES E. ALEXANDER, eum DAVIES, FTO ois occas opted coh Seas eat pasete cae Seaddvnc fh acsdamtennnnneeeesebnddadeseenseecagnnavevess 91 An analysis of the community structure of subtidal and intertidal benthic mollusks of the Inlet of Bafio (Ria de Ferrol) (northwestern Spain). CELIA OLABARRIA, VICTORIANO URGORRI, and JESUS S. TRONCOSO 0000... eicceecssseeeeeeseeseeeeseseesaesaceseeenees 103 Predatory gastropod traces: a comparison of verified shallow-water and _ presumed deep-sea boreholes. MELBOURNE R. CARRIKER...................... featbyoendatesmannanseaseeseaus 121 Redescription of Atys macandrewii E. A. Smith, 1872, an amphiatlantic cephalaspidean. E. MARTINEZ and J. ORTEA ...0.0...... eee eeeeesceecescesceeceseceseenscrsccenacessesceseesaeseaes 133 Induction of polyploidy and embryonic development of the abalone, Haliotis diversicolor, with temperature treatment. HONG-SHIIT YANG, HON-CHENG CHEN, sar TUIN= VAIN TING a ts cds aah cacnctstnc ines cde eat cada eedaoac eeslea gut bceec ts naunte suaanepanaccdeos estes 139 The gametogenic cycle of Brachidontes exustus (Linné, 1758) (Bivalvia: Mytilidae) at Wassaw Island, Georgia. MARY L. SWEENEY and IRANDAE L... WA EIRER 5csc5s535sncccponencdacieadupeannsiesouacbevioreanse icsss eiavasaqaotosauobsnensdanenesvapbadasaciarddsiecveinsts 149 Aspects of gametogenesis of Diplodon rotundus gratus (Wagner, 1827) (Bivalvia: Hyriidae) in Brazil. WAGNER EUSTAQUIO PAIVA AVELAR and SONIA HELENA SANTESSO T. DE MENDONCA ....0000...cccescssssessssesceseseesteaesseseeseseesceesaees 157 Survival and growth of juvenile freshwater mussels (Unionidae) in a recirculating acquaculture system. FRANCIS X. O’BEIRN, RICHARD J. NEVES, and MICHELLE Be STIG ooi2s5.s 300} B 2 200+ 38 5 Fe 1007 A PLZLLELA 0 Physella Elimia Helisoma Pleurocera Lithasia SPECIES 0 2 a | a Lu OW ” Oo = =a! S < ~ O Physelia Elimia Helisoma Pteurocera Lithasia SPECIES Fig. 1. Comparison of turnover time (righting response time) in seconds (top, mean + SE, N above histogram) and average crawling speed in mil- limeters per second (bottom, mean + SE, N above histogram) for two pul- monates (Physella and Helisoma) and three prosobranch snails. Means with different letters are significantly different (P < 0.05), based on Tukey’s a posteriori ranges. 199 La, b). DISCUSSION Differences in physiological adaptations A considerable amount of work has been done on the physiological ecology of freshwater snail species, although few comparative studies of lotic species have been undertaken. We therefore summarize the arguments and data presented in reviews of pulmonates (McMahon, 1983) and prosobranchs (Aldridge, 1983). Both authors main- tained that pulmonates, which were typically found in shal- low, physically and chemically more-variable environ- ments, had broader tolerance adaptations to temperature. However, the ranges reported in field studies were fairly similar (mean range + SE for pulmonates of 28.2 + 1.4°C, N = 12 populations, versus a mean range of 26°C for the two prosobranch studies). The mean upper temperature reported for the pulmonates was 30.9°C versus 27°C for the prosobranchs (data presented by McMahon, 1983). However, although the two thermal maxima were similar, pulmonates have been reported in earlier studies to have some of the highest thermal maxima in multicellular ani- mals (Van der Schalie and Berry, 1973; Russell-Hunter, 1978; McMahon, 1983, 1985). For example, physid species are capable of activity under ice cover, and have also been reported in warm springs. Pulmonates clearly show greater capacity adapta- tions to changing temperatures. Laboratory studies have suggested that the metabolic rates of pulmonates change less with increasing temperature than those of proso- branchs. The experimental studies, summarized in Table 3, indicate significantly lower Qj9 values for metabolic rates across the same temperature range (t = 2.1, 0.02 < P< 0.05, DF = 28). With Qj9 values nearer 2.0 (measured in their ambient temperature range), pulmonates can better regulate acute temperature change effects on their metabolic rates (McMahon, 1983). Six of ten pulmonates, versus zero of seven prosobranchs, had Qj9 values below 2.0. This ability WA ALONE |__| MIXED 0.20 0.16 0.12 0.08 MEAN GROWTH (GRAMS) 0.04 0.00 LITHASIA HELISOMA SPECIES Fig. 2. Results of a month-long competition experiment in artificial streams where each species was reared alone and compared to a mixed population in an additive design. Data are mean increases in individual biomass per stream + SE, N = 6. BROWN ET AL.: LOTIC GASTROPOD ECOLOGY 97 to regulate metabolic rate during acute temperature changes may be an adaptation to life in fluctuating, shallow-water habitats. Prosobranchs may not need metabolic regulation, because they are common in deeper, more permanent, and more thermally stable habitats (McMahon, 1983). McMahon (1983) also pointed out that most pulmonates can reproduce across a broader temperature range than can prosobranchs, again a valuable adaptation in variable aquat- ic habitats. Differences between pulmonates and prosobranchs in resistance and capacity adaptations to hypoxia are more complicated. In regard to tolerance, Aldridge (1983) noted that prosobranchs require higher oxygen levels in water than do pulmonates (Boycott, 1936; Palmieri et al., 1980), but this could simply be because pulmonates can use pul- monary (i. e. aerial) respiration. Prosobranchs are more resistant to hypoxia in laboratory experiments when aerial respiration is not possible (Von Brand et al., 1950; Hawkins and Ultsch, 1979). There is also variation in resistance to hypoxia within pulmonates. For example, planorbids with their respiratory pigments are more tolerant (and more aquatic) than lymnaeids and physids; the latter two groups rely more on pulmonary respiration and are more amphibi- ous (McMahon, 1983; Alexander and Covich, 1991b; Covich er al., 1994). Pulmonates clearly have greater capacity adapta- tions to variation in Oxygen concentration, although some background on experimental methods is necessary to illus- trate this. In Table 3, the degree of metabolic regulation in response to declining oxygen levels is compared between the two groups by fitting standardized metabolic rate versus PO > (partial pressure of oxygen in water) in a quadratic equation (Mangum and van Winkle, 1973), with standard- ized uptake = a + b;(PO2) + b2(PO2)2. The quadratic coeffi- cient (b2, multiplied by 10°) is the dependent variable, and the degree of metabolic regulation is determined using sta- tistical methods (Mangum and van Winkle, 1973; McMahon, 1985). Values of b> (103) near -0.1 indicate nearly perfect regulation (the metabolic rate under hypoxia is very near that at air saturation), whereas values near zero indicate conformity, where metabolic rates are a linear function of oxygen concentration (Mangum and van Table 3. Differences in regulation of metabolic rates as a function of tem- perature (Qj) and as a function of partial pressure of oxygen (b2) between pulmonate and prosobranch snails. Data from McMahon (1983). See text for statistical tests of the differences between the two groups. Group Mean Qio + SE (N) b> (x 10°) + SE (N) Pulmonates 2.12 + 0.09 (18) -0.061 + 0.009 (12) Prosobranchs 2.76 + 0.29 (13) -0.014 + 0.012 (7) Winkle, 1973). Pulmonates do possess values of bo signifi- cantly less than those shown by the prosobranchs (t = -3.15, P < 0.01, DF = 17). This result is somewhat surprising (because pulmonates are evidently more sensitive to sus- tained hypoxia); however this could again reflect their com- mon occurrence in shallow-water environments that are fre- quently exposed to diurnal cycles of hypoxia. In summary, pulmonates evidently have smaller tol- erance adaptations to long-term hypoxia (they could be able to avoid this problem with aerial respiration) but greater capacity adaptations. Pulmonates are also better adapted to resist desiccation because they can use pulmonary respira- tion, store their nitrogen excretory products as urea (not ammonia), and can aestivate by forming an epiphragm cov- ering their aperture (McMahon, 1983; Brown, 1991) to minimize water loss. The fact that pulmonates show smaller changes in metabolic rates to changing temperature or oxygen tension could be critically important in an ecological sense. Gastropods with lower maintenance costs under extreme physical or chemical conditions would have more energy to partition into growth or reproduction, and thus show greater fitness and/or competitive ability. Differences between pul- monates and prosobranchs in shell thickness and reliance on aerial refugia (and aerial respiration) could be contrast- ing adaptations due at least in part to different life histories and evolutionary pathways in the two groups (Vermeij and Covich, 1978; Alexander and Covich, 1991a, b; Covich et al., 1994). Literature studies of predation In Table 4, we have summarized earlier studies of selectivity of fish and crayfish molluscivores in lentic and lotic systems. In the two studies of molluscivorous sunfish (Stein et al., 1984; Klosiewski, 1991), there was a clear preference for thin-shelled pulmonates like Physella or Helisoma over prosobranchs like Elimia or Campeloma. Experimental results on selectivity of orconectid crayfish in Wisconsin lakes (Brown, in press) also indicated thin- shelled species like Physella and Gyraulus are consumed Table 4. Summary of earlier studies of selectivity among molluscivores. Predator Habitat Preference Reference Red-ear sunfish Ohiorivers Physella > Helisoma > Stein et al., Elimia 1984 Pumpkinseed Wisconsin Lymnaea > Amnicola > Klosiewski, 1991 sunfish lakes Campeloma Orconectes Wisconsin Gyraulus > Physella> Brown, spp. lakes Amnicola in press 98 AMER. MALAC. BULL. 14(2) (1998) more frequently than Amnicola spp. Evaluation of hypotheses regarding pulmonates and prosobranchs The results of sampling the Salt River support our predictions of where pulmonates and prosobranchs should occur. Specifically, the lack of thin-shelled species like Physella in large rivers could be due to a greater diversity, abundance, or greater body size of fish predators, whereas thick-shelled species (e. g. Pleurocera canaliculata) could be absent from from the more temporary headwater streams due to desiccation or hypoxia resulting from low or inter- mittent flow, or because of relatively poor dispersal abili- ties. Pulmonates might also be feeding generalists or have more success at repopulating disturbed sites because of their “‘r-selected” life histories. Although pulmonates as a group are considered micro-herbivores, Physella and Helisoma are feeding generalists (Brown, 1982, 1991). Some aspects of these distribution patterns, howev- er, could be unique to the Salt River system. Little shallow- water macrophyte habitat occurs along the Salt River because of the constrained nature of the littoral zone, a function of the river eroding a steep channel in the lime- stone substratum. In river systems with extensive marsh- like littoral zones, pulmonates could be more common along the edges of rivers. However, we do not consider this type of horizontal distribution pattern in rivers as repudiat- ing our general hypothesis. These littoral macrophytes might provide shelter from predators, and shallow littoral habitats are again expected to vary more in physicochemi- cal variables than the main river channel. Similarly, many southeastern U. S. rivers originate from springs, or have montane headwater reaches that are quite turbulent. In these systems we might actually expect a preponderance of prosobranchs in headwaters, because turbulence increases dissolved oxygen levels, or because spring sources vary lit- tle in physicochemical variables (see data in Johnson et al., 1994). Pulmonate species might become more common downstream, because of the addition of more ephemeral lit- toral zone habitats, or greater variation in physicochemical variables (P. Johnson, Southeastern Aquatic Research Center, unpubl. data.; Foin and Stiven, 1970; Dillon and Benfield, 1982). Again, such distributional patterns would not contradict our general hypothesis. Finally, the proso- branch Elimia was common in both the headwaters and middle reaches of the Salt River, contrary to our prediction that pulmonates should dominate headwaters. Perhaps those particular headwaters are less susceptible to drying. The experimental results and literature review also suggest support for some of our hypotheses. An exception is that the prosobranch, Lithasia, a common denizen of higher-order rivers, appeared to be a better competitor than the headwater pulmonate, Helisoma. Growth declined less under interspecific competition in this large-river proso- branch than in the pulmonate. The result is again somewhat surprising in light of earlier research suggesting that pul- monates, as a group, crop periphyton resources more close- ly to the substratum (Barnese et al., 1990). However, the earlier study was completed in a lentic habitat, and only one species of prosobranch, Elimia livescens (Menke, 1830), was compared to several pulmonates. Also, there is considerable variation within pulmonates in grazing ability. For example, lymnaeids are, because of differences in radu- lar anatomy, better adapted to removing filamentous algae than are physids, whose complex teeth are better at remov- ing adnate diatoms (see references by Brown, 1991). Thus, differences in grazing ability could depend on the particular pulmonate or prosobranch family in question. Because Lithasia is most common on littoral-zone cobble along the Ohio River, we would not have a priori predicted it would do better than the pulmonate, which occurs in soft sediments such as those lining the bottom of the experimental channels. We would have also expected that the relatively short time-span of the experiment would have favored the pulmonate, because of its short life cycle. However, it is also important to note that Lithasia and Helisoma do not overlap in nature, being at opposite ends of the river continuum, and thus lack an extensive history of competition. The evidence for dispersal ability as a limiting fac- tor in field distributions is equivocal. Physella clearly is the most active species and disperses at a higher rate, as expected because it is most common in headwater environ- ments. However, Helisoma, which is also common in the same environments, moves no faster than any of the proso- branch species, including Lithasia and Pleurocera, which are found only in medium to large rivers. Of course, both pulmonates could have greater passive dispersal abilities, as mentioned earlier, a factor which we did not investigate, and pulmonates can also self-fertilize, a valuable aid in col- onizing new habitats. The available data in the literature do seem as a whole to indicate that pulmonates are better adapted physi- ologically to the large range of physical and chemical varia- tion occurring in ephemeral habitats like shallow littoral zones, or the seasonal headwater habitats in the Salt River. Conventional wisdom is that pulmonates show broader resistance and capacity adaptations to temperature. The capacity adaptations of pulmonates to variation in dissolved oxygen are at least as great as in prosobranchs, and pul- monates can respire aerial oxygen. Pulmonates are also bet- ter at resisting desiccation. In addition, prosobranchs do not have the ability to aestivate and survive over periods when habitats dry. Finally, both our experimental results and the litera- ture clearly suggest the importance of predation as an eco- BROWN ET AL.: LOTIC GASTROPOD ECOLOGY 99 logical factor determining lotic gastropod distributions. As stream order increases, it is likely that the abundance of crayfish (and probably other invertebrate predators) at first increases, and then decreases as fish become more abun- dant and diverse (Crowl, 1990). The thick shells of proso- branchs afford them better protection from fish (Stein et al., 1984; Klosiewski, 1991) and crayfish predation (this study). They also have an advantage against other shell- invading invertebrate predators like leeches because they possess an operculum (Bronmark and Malmgqvist, 1986; Brown and Strouse, 1988). Suggestions for future research This paper clearly demonstrates the need for further work on the community ecology of lotic gastropods. It has been our purpose to establish several hypotheses (that are not necessarily mutually exclusive) explaining gastropod distributions along river continua, rather than to definitively identify which hypotheses are most important. For exam- ple, future studies could examine the competitive ability of all species in the replacement series in a particular river system to determine, for example, if the competitive hierar- chy is transitive. These studies should also gauge the rela- tive importance of intraspecific competition (e. g. density dependence) and interspecific competition for both snail groups. Although such a comparative study might prove difficult, future researchers could also test the differential passive dispersal hypothesis by sampling birds and other vertebrates arriving at lotic sites to determine the relative abundance of phoretic juvenile gastropods. Future physio- logical research should test for differences in capacity and tolerance adaptations among co-occurring sets of gas- tropods found along river continua to rigorously test the conventional wisdom that pulmonates are better adapted to physically more extreme habitats. Also, although compara- tive work has been done on pulmonates, very little is known of differences in physiological tolerances among prosobranchs. In addition, little is known of the compara- tive tolerances of pulmonates and prosobranchs to other potentially important physiological stresses occurring at different points along the river continuum, such as spates, turbidity, and wave action. We hope that this paper will stimulate future research in understanding the community ecology of river- ine snail assemblages, and thus help preserve the rich gas- tropod assemblages found in many southeastern U. S. river systems. This is especially important because as many as one-third of the species face threats of extinction (Neves ef al., 1998). Although they are relatively unstudied, proso- branch gastropods like Elimia are important as grazers in southeastern U. S. lotic systems (Richardson et al., 1988; Hill et al., 1992; Rosemond et al., 1993). A vital first goal for conservation efforts is to more fully understand what ecological and evolutionary factors constrain the ranges and population biology of these lotic gastropods. ACKNOWLEDGEMENTS We are especially grateful to Debbie Vaughn for sampling many of the lotic gastropod populations, and also appreciate the help of Kim Greenwood, Janet Wolanin Alexander, Chad Peterson, Tara Abboud, and Sonya Pace for field and laboratory assistance. LITERATURE CITED American Public Health Association (APHA). 1985. Standard Methods for the Examination of Water and Wastewater, 16th ed. American Public Health Association, Washington, D. C. 1134 pp. Aldridge, D. W. 1983. Physiological ecology of freshwater prosobranchs. In: The Mollusca, Vol. 6, Ecology, W. D. Russell-Hunter, ed. pp. 329-358. Academic Press, Orlando, Florida. Alexander, J. E. and A. P. Covich. 1991a. Predator avoidance by the fresh- water snail Physella virgata in response to the crayfish Procambarus simulans. Oecologia 87:435-442. Alexander, J. E. and A. P. Covich. 1991b. Predation risk and avoidance behavior in two freshwater snails. Biological Bulletin 180:387- 393. Barnese, L. E., R. L. Lowe, and R. D. Hunter. 1990. Comparative grazing efficiency of pulmonate and prosobranch snails. Journal of the North American Benthological Society 9:35-44. Boycott, A. E. 1936. The habitats of freshwater Mollusca in Britain. Journal of Animal Ecology 5:116-186. Branson, B. A. 1987. Keys to the aquatic Gastropoda known from Kentucky. Transactions of the Kentucky Academy of Science 48:11-19. Branson, B. A. and D. L. Batch. 1983. Gastropod and spaeriacean clam records for streams west of the Kentucky river drainage, Kentucky. Transactions of the Kentucky Academy of Science 44:8-12. Bronmark, C. and B. Malmqvist. 1986. Interactions between the leech Glossophonia complanata and its gastropod prey. Oecologia 69:268-276. Bronmark, C., S. P. Klosiewski, and R. Stein. 1992. Indirect effects of pre- dation in a freshwater, benthic food chain. Ecology 73:1662-1674. Brown, K. M. 1982. Resource overlap and competition in pond snails: an experimental analysis. Ecology 63:412-422. Brown, K. M. 1983. Do life history tactics exist at the intraspecific level: data from freshwater snails. American Naturalist 121:871-879. Brown, K. M. 1991. Mollusca: Gastropoda. In: Ecology and Classification of North American Freshwater Invertebrates, J. Thorp and A. P. Covich, eds. pp. 285-314. Academic Press, Orlando, Florida. Brown, K. M. In press. The role of shell strength in foraging of crayfish for gastropod prey. Freshwater Biology. Brown, K. M. and D. R. Devries. 1985. Predation and the distribution and abundance of a pulmonate pond snail. Oecologia 66:93-99. Brown, K. M. and D. M. Lodge. 1993. Gastropod abundance in vegetated habitats: the importance of specifying null models. Limnology and Oceanography 38:217-225. 100 AMER. MALAC. BULL. 14(2) (1998) Brown, K. M. and B. H. Strouse. 1988. Relative vulnerability of six fresh- water gastropods to the leech Nephelopsis obscura (Verrill). Freshwater Biology 9:157-166. Browne, R. A. 1978. Growth, mortality, fecundity, and productivity of four lake populations of the prosobranch snail, Viviparus geor- gianus. Ecology 59:742-750. Browne, R. A. and W. D. Russell-Hunter. 1978. Reproductive effort in molluscs. Oecologia 27:23-27. Burch, J. B. 1982. Freshwater Snails (Mollusca: Gastropoda) of North America. United States Environmental Protection Agency Publication 600/3-82-026. 365 pp. Connell, J. H. 1983. On the prevalence and relative importance of inter- specific competition: evidence from field experiments. American Naturalist 122:661-696. Covich, A. P., T. A. Crowl, J. E. Alexander, Jr., and C. C. Vaughn. 1994. Predator-avoidance responses in freshwater decapod-gastropod interactions mediated by chemical stimuli. Journal of the North American Benthological Society 13:283-290. Crowl, T. A. 1990. Life history strategies of freshwater snails in response to stream permanence and predation: balancing conflicting demands. Oecologia 84:238-243. Crowl, T. A. and A. P. Covich. 1990. Predator-induced life history shifts in a freshwater snail. Science 247:949-95 1. Davis, G. M. 1982. Historical and ecological factors in the evolution, adaptive radiation, and biogeography of freshwater molluscs. American Zoologist 22:375-395. Dillon, R. T. 1981. Patterns in the morphology and distribution of gas- tropods in Oneida Lake, New York, detected using computer-gen- erated null hypotheses. American Naturalist 118:83-101. Dillon, R. T. 1987. A new Monte Carlo method for assessing taxonomic similarity within faunal samples: re-analysis of the gastropod community of Oneida Lake, New York. American Malacological Bulletin 5:101-104. Dillon, R. T. and E. F. Benfield. 1982. Distribution of pulmonate snails in the New River of Virginia and North Carolina USA: interaction between alkalinity and stream drainage area. Freshwater Biology 12:179-186. Dussart, G. B. J. 1987. Effects of water flow on the detachment of some aquatic pulmonate gastropods. American Malacological Bulletin 5:66-72. Foin, T. C. and A. E. Stiven. 1970. The relationship of environment size and population parameters in Oxytrema proxima (Say) (Gastropoda: Pleuroceridae). Oecologia 5:74-84. Gurevitch, J., L. L. Morrow, A. Wallace, and J. S. Walsh. 1992. A meta- analysis of competition field experiments. American Naturalist 140:539-572. Hairston, N. G., F. E. Smith, and L. B. Slobodkin. 1960. Community structure, population control, and competition. American Naturalist 44:421-425. Harman, W. N. 1968. Interspecific competition between Bithynia and pleurocerids. The Nautilus 82:72. Harman, W. N. 1972. Benthic substrates: their effect on freshwater Mollusca. Ecology 53:271-277. Harman, W. N. and C. O. Berg. 1971. The freshwater snails of central New York. Search 1:1-68. Hawkins, M. J. and G. R. Ultsch. 1979. Oxygen consumption in two species of freshwater snails (Goniobasis): effects of temperature and ambient oxygen tension. Comparative Biochemistry and Physiology 63A:369-372. Hill, W. R., H. L. Boston, and A.D. Steinman. 1992. Grazers and nutrients simultaneously limit lotic primary productivity. Canadian Journal of Fisheries and Aquatic Sciences 49:504-512. Horwitz, R. J. 1978. Temporal variability patterns and the distribution pat- terns of stream fishes. Ecological Monographs 48:307-321. Jacoby, G. Z. and S. J. Cary. 1996. Winter stoneflies (Plecoptera) in sea- sonal habitats in New Mexico, USA. Journal of the North American Benthological Society 15:690-699. Jenkinson, J. J. and R. M. Todd. 1998. Management of native mollusk resources. In: Aquatic Fauna in Peril: the Southeastern Perspective, G. W. Benz and D. E. Collins, eds. pp. 283-305. Aquatic Research Institute Special Publication 1, Chattanooga, Tennessee. Johnson, P. D., K. M. Brown, and C. V. Covell. 1994. A comparison of the macroinvertebrate assemblage in Doe Run Creek, Kentucky: 1960 and 1990. Journal of the North American Benthological Society 13:496-510. Kesler, D. H. and W. R. Munns, Jr. 1989. Predation by Belastoma flu- mineum (Hemiptera): an important cause of mortality in freshwa- ter snails. Journal of the North American Benthological Society 8:342-350. Klosiewski, S. P. 1991. The Role of Pumpkinseed Sunfish in Structuring Snail Assemblages in Northern Wisconsin Lakes. Ph. D. disserta- tion, Ohio State University, Columbus, Ohio. 147 pp. Lodge, D. M., K. M. Brown, S. P. Klosiewski, R. A. Stein, A. P. Covich, B. K. Leathers, and C. Bronmark. 1987. Distribution of freshwa- ter snails: spatial scale and the relative importance of physico- chemical and biotic factors. American Malacological Bulletin 5:73-84. Lodge, D. M., D. W. Kershner, J. E. Aloi, and A. P. Covich. 1994. Effects of an omnivorous crayfish (Orconectes rusticus) on a freshwater littoral food web. Ecology 75:1265-1281. Lotrich, V. A. 1973. Growth, production, and community composition of fishes inhabiting first, second and third order streams of eastern Kentucky. Ecological Monographs 43:377-397. Mangum C. and W. Van Winkle. 1973. Responses of aquatic invertebrates in declining oxygen conditions. American Zoologist 13:529-541. Matthews, W. J. 1988. North American prairie streams as systems for eco- logical study. Journal of the North American Benthological Society 7:387-409. McMahon, R. F. 1983. Physiological ecology of freshwater pulmonates. In: The Mollusca, Vol. 6, Ecology, W. D. Russell-Hunter, ed. pp. 359-430. Academic Press, Orlando, Florida. McMahon, R. F. 1985. Seasonal respiratory variation and metabolic com- pensation for temperature and hypoxia in the freshwater snail, Physella virgata virgata (Gould), with special reference to the effects of heated effluents. American Malacological Bulletin 3:243-265. Miller, A. M. and S. W. Golladay. 1996. Effects of spates and drying on macroinvertebrate assemblages of an intermittent and a perennial prairie stream. Journal of the North American Benthological Society 15:670-689. Moyle, P. B. 1988. Fishes: an Introduction to Ichthyology. Prentice Hall, Englewood Cliffs, New Jersey. 559 pp. Neves, R. J., A. E. Bogan, J. D. Williams, S. A. Ahlstedt, and P. W. Harfield. 1998. Status of aquatic mollusks in the southeastern United States: a downward spiral of diversity. In: Aquatic Fauna in Peril: the Southeastern Perspective, G. W. Benz and D. E. Collins, eds. pp. 41-83. Aquatic Research Institute Special Publication 1, Chattanooga, Tennessee. Osenberg, C. W. 1989. Resource limitation, competition and the influence of life history in a freshwater snail community. Oecologia 79:512- 519. Palmieri, M. D., J. R. Palmieri, and J. T. Sullivan. 1980. A chemical analysis of the habitat of nine commonly occuring Malaysian freshwater snails. Malayan Naturalist Journal 34:39-45. Pearson, W. D. and L. A. Krumholz. 1984. Distribution and Status of BROWN ET AL.: LOTIC GASTROPOD ECOLOGY 101 Ohio River Fishes. United States Environmental Protection Agency Report ORNL/Sub/79-783 1/1. 439 pp. Peterson, C. H. and R. E. Renaud. 1989. Analysis of feeding preference experiments. Oecologia 80:82-86. Richardson, T. D., J. F. Scheiring, and K. M. Brown 1988. Secondary pro- duction of two long lived lotic gastropods (Pleuroceridae: Elimia). Journal of the North American Benthological Society 7:234-241. Rosemond, A. D., P. J. Mullholland, and J. W. Ellwood. 1993. Top-down and bottom-up control of stream periphyton: effects of nutrients and herbivores. Ecology 74:1264-1280. Russell-Hunter, W. D. 1978. Ecology of freshwater pulmonates. Jn: The Pulmonates, Vol. 2A, Systematics, Evolution and Ecology, V. Fretter and J. Peak, eds. pp. 335-383. Academic Press. Orlando, Florida. Schoener, T. W. 1983. Field experiments on interspecific competition. American Naturalist 122:240-285. Sheldon, S. P. 1987. The effects of herbivorous snails on submerged macrophyte communities in Minnesota lakes. Ecology 68:1920- 1931. Snaydon, R. W. 1991. Replacement or additive designs for competition studies? Journal of Applied Ecology 28:930-936. Sokal, R. R. and F. J. Rohlf. 1981. Biometry, 2nd ed. W. H. Freeman and Company, San Francisco. 859 pp. Stein, R. A., C. G. Goodman, and E. A. Marschall. 1984. Using time and energetic measures of cost in estimating prey value for fish preda- tors. Ecology 68:1920-1931. Strong, D. R., Jr., D. Simberloff, L. G. Abele, and A. B. Thistle. 1984. Ecological Communities: Conceptual Issues and the Evidence. Princeton University Press, Princeton, New Jersey. 613 pp. Van der Schalie, H. and E. G. Berry. 1973. The effects of temperature on growth and reproduction of aquatic snails. Sterkiana 50:1-92. Vermeij, G. 1987. Evolution and Escalation. Academic Press, Orlando, Florida. 527 pp. Vermeij, G. J. and A. P. Covich. 1978. Coevolution of freshwater gas- tropods and their predators. American Naturalist 112:833-843. Von Brand, T., H. D. Baernstein, and B. Mehlman. 1950. Studies on the anaerobic metabolism and the aerobic carbohydrate consumption of some fresh water snails. Biological Bulletin 98:266-276. Williams, D. D. 1996. Environmental constraints in temporary fresh waters and their consequences for the insect fauna. Journal of the North American Benthological Society 15:634-650. Date of manuscript acceptance: 23 March 1998 na tettne | ok ae i An analysis of the community structure of subtidal and intertidal benthic mollusks of the Inlet of Bano (Ria de Ferrol) (northwestern Spain) Celia Olabarria!, Victoriano Urgorri2, and Jesus S. Troncoso3 1Laboratorio de Invertebrados y Ecologia del bentos, Facultad de Ciencias del Mar, Universidad Autonoma de Sinaloa, Paseo Claussen, s/n, 82000 Mazatlan (Sinaloa), Mexico, chispita@ola.icmyl.unam.mx 2Departamento de Bioloxia Animal, Facultade de Bioloxfa, Universidade de Santiago, E-15706 Santiago (A Coruna), Spain 3Departamento de Medio Ambiente e Recursos Naturais, Facultade de Ciencias del Mar, Universidade de Vigo, Vigo (Pontevedra), Spain Abstract: A study of the benthic communities of the Inlet of Bafio (Ria de Ferrol) (northwestern Spain) was carried out, encompassing the examination of 75 sampling stations (35 subtidal and 40 intertidal). Data were subjected to classification and ordination techniques. Two major assemblages were identified, and divided into five subgroups in terms of dominance of the species, constancy, and fidelity. The subtidal zone was characterized by the Abra alba (Wood, 1802) community. It was structured in facies as follows: The Nucula nitida Sowerby, 1833-Thyasira flexuosa (Montagu, 1803) facies, was the innermost, in the eastern area of the inlet; another facies existed in the transition toward the Clausinella fasciata (da Costa, 1778) community in the outer inlet area. In the intertidal zone the Cerastoderma edule (Linné, 1758)-Scrobicularia plana (da Costa, 1778) community was defined, which was also structured in facies: one facies, located at the mouth of the river, was dominated by Hydrobia ulvae (Pennant, 1777); another facies, found in the inner inlet area associated with a meadow of the seagrass Zostera noltii Hornem., 1832, characterized by Bittium reticulatum (da Costa, 1778), Loripes lacteus (Linné, 1758), C. edule, Venerupis senegalensis (Gmelin, 1791), and Rissostomia membranacea (Adams, 1800), and a third facies, situated at the border of the inlet, included Gibbula umbilicalis (da Costa, 1778), Littorina littorea (Linné, 1758), L. obtusata (Linné, 1758), V. senegalensis, etc., as dominant species. The analyses showed that sediment parameters (mainly grain size), organic matter in the subtidal stations, and grain size and depth are the most important factors govern- ing the distribution and abundance of the communities. Key words: community structure, intertidal, subtidal, mollusks, Ria de Ferrol, Spain Research on the macrobenthic communities of the Quintino et al., 1986; Sauriau et al., 1989). northern coast of the Iberian Peninsula has been carried out However, despite the plethora of faunistic studies by several authors, who have taken both a faunistic and an involving Galician mollusks, there is a need for research on ecological-descriptive approach. Their sources of informa- their sinecology. An exception to the lack of sinecological tion have been primarily articles on the ecology of the mac- studies is the work by Cadée (1968) in the Ria de Arousa. robenthos, and their ultimate goals have been to carry out Therefore the purpose of this work was to study the ecolo- analyses of the faunistic community as a whole. This 1s gy of the benthic populations of mollusks living on soft reflected by the great number of biocenotic studies on the substrata, both subtidal and intertidal, of the Inlet of Bano benthos living on soft substrata in recent years on the (Ria de Ferrol, northwestern Spain). The aim was to define Galician coasts (Viéitez, 1976, 1981; Anadon, 1980; Mora, the different communities living in the inlet and how they 1980; Penas and Gonzalez, 1983; Rodriguez Castelo and relate to physicochemical factors (granulometry, sorting Mora, 1984; Planas and Mora, 1984 a, b; Laborda, 1986; coefficient, organic matter, carbonates, and nitrogen). Planas, 1986; Lopez Serrano and Viéitez, 1987; Viéitez and Baz, 1988; Junoy and Viéitez, 1989, 1990; Mazé er al., 1990; Palacio et al., 1991, 1993; Currds and Mora, 1991, MATERIAL AND METHODS 1992; Pérez Edrosa and Junoy, 1993). Frequent studies have also been made on the mala- STUDY AREA cological communities of different geographical areas of The Inlet of Bano (Fig. 1) is located on the south- the European Atlantic coasts (Lande, 1975; Evans and ern border of the central channel in the Ria de Ferrol Tallmark, 1976; Petersen, 1977; Glémarec, 1978; Tunberg, between the Punta do Faro da Palma (43°27°52”N; 1981; Dewarumez, 1983; Gentil et al., 1986; Cornet, 1985; 08°16’°49”"W) and Punta Piteira (43°27’°57"N; American Malacological Bulletin, Vol. 14(2) (1998): 103-120 103 104 Iniet of Bano Fig. 1. Location of the 75 sampling stations in the Inlet of Bafio (Ria de Ferrol), Bano. (I, intertidal station; S, subtidal station). 08°15°37"W), with an area of 0.5 km? and a maximum depth of 18 m (Fig. 2). The inlet is oriented in a NNE-SSW direction; the prevailing winds are southwesterly for most of the year, except in summer when northeasterlies become dominant. The mean tidal range in the ria is 2.7 m, and tidal effects give rise to strong currents (up to 1.5 m/s in the ria’s central channel). Outward movement of water from the ria provokes flow to the southeast within the inlet, while movement into the ria provokes flow to the SSW within the AMER. MALAC. BULL. 14(2) (1998) S14 WwW & maeérl bed Z. noltii meadow 9 145m Esteiro river northwestern Spain. A. Spain. B. Galicia. C. Ria de Ferrol. D. Inlet of inlet. These currents, which are stronger at the mouth of the inlet than at more distal points, are the dominant factor affecting sediment distribution within the inlet (Fig. 3). The inlet is characterized by soft bottom bordered by a rocky strip. A maérl bed is present in four of the west- ern stations, between 4.5 and 10.5 m depth. A seagrass meadow of Zostera noltii Hornem., 1832, is present in the central intertidal zone. The distribution of different grain sizes suggests an OLABARRIA ET AL.: BENTHIC MOLLUSKS OF INLET OF BANO 105 -11.10 -10.25 9.00 -16.92 -12.91 -5.00 -420 -720 -13.20 -2.00 4.00 145m Fig. 2. Depth (m) of the subtidal (-) and intertidal (+) stations. enrichment gradient for fine fractions on the eastern part of the Inlet of Baflo. Sandy muds prevail in this area, while the coarsest sediment fractions are prevalent in the outer part. On the other hand, the percentages of organic matter are low, with enrichment toward the eastern part (Table 1). Nitrogen follows a variation pattern similar to that of organic matter. Calcium carbonate does not present a high percentage, with highest values found in the subtidal zone (Olabarria et al., 1996). The area has a temperate humid Atlantic climate and the hydrographic conditions show thermal variations of more than 5°C in the first 10 m. This variation decreases with depth. Salinity fluctuates with the tides and seasons, due to the fact that the ria has difficulty in draining because of the narrow central channel, which creates water masses having different salinities. The pH values are consistently between 7 and 9, and there are no anomalous data that would indicate a source of excessive acidity or alkalinity (COTOP, 1987). SAMPLE COLLECTION The sampling program, which was designed to pro- vide sufficient information on the distribution of the different species of mollusks, consisted of 35 subtidal and 40 intertidal Stations sampled from July 1991 to June 1992. Sampling points were selected along 12 parallel transects across the inlet at 100 m intervals, taking samples at the points that were judged by visual examination to show a change in nature, tex- ture, or substrate covering. Intertidal samples were addition- ally collected at the ends of each transect and, in the inner intertidal zone, samples were also taken every 100 m along each transect (Olabarria et al., 1996). In the subtidal zone, samples were collected by scuba diving. At each point, a 0.5 m2 square sample was taken, to a depth of approximate- ly 20 cm, using a rectangular shovel. All samples were sub- sequently wet-sieved through a series of sieves with 10, 2, and 0.5 mm mesh. Finally the sieved samples were trans- ported to the laboratory, and the living specimens were sort- ed by the remounting technique (Ros, 1975). Sediment samples were obtained for granulometric study. This con- sisted of an analysis of grain size, organic matter, nitrogen, and carbonates (Guitian and Carballas, 1976). DATA ANALYSES Data were organized into station by species matri- ces. The Shannon-Wiener and Pielou’s evenness indices (Washington, 1984) were used to assess species diversity and evenness. The data for the population studies were processed in two ways: (1) Analyses of qualitative data (presence- absence) of the species in the sampling stations were per- formed by applying the point correlation coefficient (®; Daget, 1976), whereby a correlation matrix 1s created, fol- lowed by classification into clusters through the clustering algorithm UPGMA, by means of the NTSYS-pe program & >2 mm @ 1.0-0.5 mm e@ 0.5-0.20 mm @ 0.20-0.05 mm @ <0.05 mm Fig. 3. Sediment characteristics of the Inlet of Bafio (Ria de Ferrol). Mean grain size in mm. 106 AMER. MALAC. BULL. 14(2) (1998) Table 1. Organic matter (MO), carbonates (CA), nitrogen (N) and sort coefficient (So; Trask, 1932) in subtidal and intertidal stations. Station MO CA N So Station MO CA N So $101 0.42 30.22 0.05 2.42 1103 0.59 19.64 0.08 > 2.50 $102 0.51 31.11 0.04 2.69 1104 0.43 20.57 0.04 1.80 $103 0.42 33.33 0.03 1.76 1105 0.67 9.43 0.06 1.58 $104 1.03 18.22 0.06 6.06 1106 0.39 12.86 0.03 2.58 $105 0.70 34.67 0.08 4.80 1107 0.55 DES 7, 0.04 2.55 $106 1.04 8.89 0.01 8.36 1108 0.54 2.10 0.06 1.95 $107 1.99 32.44 0.13 6.44 1109 0.66 1.71 0.06 > 1.65 $108 1.16 34.67 0.10 2.18 1110 2.00 8.57 0.06 3.71 S109 1.53 29.33 0.10 9.60 1114 0.23 1.71 0.02 > 1.29 $110 0.69 32.89 0.05 2.75 1115 0.47 1.03 0.03 1.79 S111 1.69 32.00 0.13 0.18 1116 1.41 0.34 0.10 5.07 $112 1.13 13.78 0.09 21.70 1117 0.63 26.67 0.08 Balt $113 0.97 13.78 0.08 2.54 1118 0.34 16.29 0.05 2.92 $114 1.37 21.33 0.07 20.99 1119 0.59 0.86 0.01 1.63 $115 0.44 34.67 0.04 2.00 1120 0.94 0.34 0.02 1.49 $116 0.58 34.67 0.04 1.78 1121 0.91 0.86 0.07 > 2.50 $117 0.32 4.89 0.03 2.08 1122 1.37 1.71 0.09 1.54 S118 1.55 24.89 0.09 9.57 1123 0.64 0.26 0.03 1.89 $119 0.44 17.33 0.03 2.63 1124 0.83 0.17 0.04 1.45 $120 0.59 18.40 0.04 3.39 1125 0.83 0.26 0.04 1.37 $121 0.87 27.56 0.08 2.54 1126 0.43 1.03 0.02 1.27 $122 0.60 30.22 0.07 7.50 1127 0.51 0.86 0.04 2.29 $123 0.45 16.44 0.04 2.35 1128 0.82 0.34 0.02 2.23 $124 0.69 30.22 0.06 >3.30 1129 1.04 0.86 0.09 1.38 $125 0.97 35.56 0.14 10.04 1130 1.23 0.26 0.05 0.32 S126 1.31 22.58 0.05 >3.90 1131 0.48 0.86 0.04 4.31 $127 1.83 2.22 0.12 >1.84 1132 0.72 0.17 0.03 1.41 $128 0.26 6.29 0.04 >3.50 1133 0.91 0.34 0.02 1.43 $129 1.80 17.33 0.12 8.48 1134 0.91 0.26 0.03 1.32 $130 1.51 25.78 0.11 9.00 1135 0.49 1.03 0.02 1.50 $131 2 16.27 0.10 8.50 1136 0.44 0.17 0.35 1.77 $132 0.61 1.33 0.04 1.61 1137 0.95 0.17 0.04 1.46 $133 2.04 6.22 0.11 5.21 1138 0.80 0.17 0.05 3.77 $134 0.93 19.02 0.03 >3.53 1139 0.76 0.86 0.07 3.92 $135 0.42 0.89 0.03 1.70 1140 0.63 0.86 0.05 7A7 1101 0.17 12.00 0.02 1.13 1141 0.69 4.29 0.06 3.08 1102 0.38 1.42 0.01 1.78 1142 0.38 24.89 0.03 1.58 1143 0.78 14.23 0.06 3.65 (ver. 1.60; Rohlf, 1990). ad - bc V(at+ b)(a+c)(b + d)(c +d) where a = number of species in both stations; b = number of species in the first station; c = number of species in the second station; and d = species absent in both stations. ® values ranged from -1 to+1; p=a+b+c+d,x2=p ®2. (2) Analyses of quantitative data of the species pre- sent in at least 10% of the stations were performed, omit- ting those which could create “noise” because of their nar- row distribution and scarcity. The Detrending Correspondence Analysis (DCA) was applied according to ® = the polynomial method, by means of the CANOCO-pc pro- gram (ver. 3.10; Ter Braak, 1988). The data were trans- formed according to the formula, x = logjo (x + 1). This technique plots the stations along the axes according to similarity in species composition. The axes are often inter- preted as environmental gradients whose identity can be analyzed by means of statistical correlations between the location of the stations on the axes and their environmental characteristics (Eleftheriou and Basford, 1989; Junoy and Viéitez, 1990). The environmental variables measured were: depth (positive values for intertidal stations; negative values for subtidal stations), percentages of the different granulometric fractions, organic matter, silt-clay ratio, car- bonates, nitrogen, and sort coefficient. These environmental OLABARRIA ET AL.: BENTHIC MOLLUSKS OF INLET OF BANO 107 Table 2. Species classifications according to constancy and fidelity indices. (Caii, addition of constancies of species A within each popula- tion; Na;, number of stations within population where species A exists; Nj, total number of stations within population). Constancy index Fidelity index Cai =(Na1/ Nj) x 100 (Cay / Caii) x 100 accidental < 10% rare < 12% occasional 11-33% not very common 13-25% accessory 34-50% common 26-50% preferential 51-66% very common 51-75% elective 67-90% constant 76-100% exclusive 91-100% factors were correlated with the axes through Spearman’s correlation analysis. Based on the group of stations defined using the point correlation coefficient, the species were classified according to the criteria of constancy and fidelity, based on number of species (Dajoz, 1971; Table 2), and fidelity- dominance product (FxD; Glémarec, 1964), including only species that appeared in at least two stations (Glémarec, 1964; Cabioch, 1968; Lastra et al., 1990; Junoy and Viéitez, 1990; Curras and Mora, 1991). RESULTS FAUNISTIC ANALYSIS, SPECIES DIVERSITY, AND ABUNDANCE The 75 samples analyzed yielded a total of 20,647 individuals belonging to 148 species, in which gastropods were the most abundant (62.4%; Table 3), followed by bivalves (36.7%); chitons and scaphopods were present in small numbers. The gastropods numerically dominated the intertidal zone, outnumbering the bivalves by more than twofold. This is probably caused by the large number of individuals belonging to abundant species such as Hydrobia ulvae and Rissoa parva. The latter was also present in the subtidal zone in a high number of samples along with Hinia reticulata, H. incrassata, and Onoba semicostata. There, bivalves were slightly more dominant than gastropods, due to the high number of individuals of Mysella bidentata, in addition to Papillicardium papillosum, and several species of Anomiidae. The bivalves Mytilus edulis and Venerupis senegalensis were the most abundant species in the inter- tidal zone. [Note: The species Littorina littorea, L. obtusa- ta, L. mariae, Lepidochitona cinerea, and M. edulis are nor- mally associated with hard bottoms. In some cases during this study, small stones were included in the soft sediment samples collected by shovel from the Inlet of Bano. These small stones provided hard substratum for the chitons, and also for attached algae, thus providing food for the periwin- kle species. ] The values of the Shannon-Wiener diversity index (Table 4) fluctuated between 0.00 (station 1134) and 4.35 (station S112), and evenness ranged from 0.06 (station 1135) to 0.94 (station S124). These two parameters were positively correlated (r = 0.57; p < 0.001). Diversity correlated positively with carbonate con- tent, percentage of coarse sand, and sorting coefficient, and negatively with depth and percentage of fine sand. Both rel- ative diversity and evenness showed positive correlations with the percentage of coarse sand, percentage of gravel, and the sort coefficient, and a slightly negative correlation with the percentage of fine sand (Table 5). The lowest population density was recorded at sta- tion 1135 with 32 ind/m2, and the highest density at station S111 (6,040 ind/m2). The mean density in the subtidal zone (1,294.6 ind/m2) was higher than that in the intertidal zone (931.9 ind/m2). Species richness varied widely, with a greater num- ber of species in the subtidal than in the intertidal zone. The lowest values for species richness were found at the inter- tidal stations near the mouth of the river, due to high abun- dance of the dominant species Hydrobia ulvae. Species richness correlated negatively with the percentage of fine sand and positively with carbonate content, and percentages of coarse sand, gravel, and silt-clay, which explains the increase in species richness at the outer subtidal stations with coarser grain size and higher silt-clay percentages. Table 3. Number of individuals (ind) and species (sp), and total percentage (%) of individuals in the subtidal and intertidal zones and in both. TOTAL ind sp % ind Polyplacophora 177 6 0.85 110 Gastropoda 1,288 89 62.39 5,206 Scaphopoda 9 1 0.04 9 Bivalvia 7,579 52 36.70 6,003 TOTAL 9,053 148 100.00 11,328 SUBTIDAL INTERTIDAL sp % ind sp % 6 0.99 67 4 0.72 82 45.95 7,676 41 82.37 1 0.07 0 0 0.00 52 52.99 1,576 32 16.91 141 100.00 9,319 dil 100.00 108 AMER. MALAC. BULL. 14(2) (1998) Table 4. Faunistic parameters for each subtidal and intertidal station: abundance, species richness (R), Shannon- Wiener diversity index (H’), Pielou’s evenness index (J’), and density (ind/m2). Station Abundance R H’ JS ind/m2 $101 273 31 4.25 0.86 1,092 $102 194 29 3.61 0.74 7716 $103 158 20 3.09 0.72 632 $104 112 19 3.43 0.81 448 $105 204 27 (3.85 0.81 816 S106 924 37 3.74 0.72 3,696 $107 461 36 «63.82 0.74 1,844 $108 1,188 39 «3.26 =©0.62 4,752 $109 658 38 863.15 0.60 2,632 $110 260 32 407 081 1,040 $111 1,510 45 2.81 0.51 6,040 $112 365 42 435 0.80 1,460 $113 802 49 419 0.75 3,208 $114 553 43 4.24 0.78 2,212 $115 218 33 4.11 0.81 872 S116 523 29 2.87 0.59 2,092 $117 264 19 2.95 0.69 1,056 $118 231 21 2.24 0.51 924 $119 127 24 3.39 0.74 508 $120 48 15 3.38 0.86 192 $121 85 24 3.00 0.65 340 $122 369 33. 2.87 0.57 1,476 $123 112 22 3.51 0.78 448 $124 29 15 3.69 0.94 116 $125 96 2 2.46 0.56 384 $126 337 29, 3.58 0.37 1,348 $127 40 8 2.39 0.79 160 $128 70 16 3.25 0.81 280 $129 82 21 3.84 0.87 328 $130 379 33 2.82 0.56 1,516 $131 126 21 3.57 0.81 504 $132 110 10 1.58 0.47 440 $133 205 26 3.28 0.69 820 $134 131 18 2.50 0.59 524 $135 84 8 1.94 0.65 336 T1101 142 24 3.99 0.87 568 1102 333 24 3.41 0.74 1,332 COMMUNITY STRUCTURE The classification (Fig. 4) that was obtained after applying the point correlation coefficient for 99% signifi- cance (® = 0.29) showed two large assemblages (A and B) which belong to the intertidal and subtidal stations, respec- tively. These assemblages were divided into five sub- groups: Al, A2, A3, B1, and B2. Subgroup Al (five sta- tions) located in the inner part of the inlet near the mouth of the river; subgroup A2 (18 stations) included a group of intertidal stations located along the border of the inlet; sub- group A3 (17 stations) included the intertidal stations of the central zone, mostly covered by Zostera noltii, which also comprised two subtidal stations; subgroup B1 (ten stations) located in the middle-eastern subtidal zone; subgroup B2 (23 stations) consisted of the outermost subtidal stations Station Abundance R H’ J ind/m2 1103 530 29 3.17 +0.62 2,120 1104 249 24 346 0.75 996 1105 162 12 2.28 0.63 648 1106 165 13. 2.60 0.70 660 1107 481 23 3.11 0.69 1924 1108 154 11 2.37 0.68 616 1109 387 17 3.48 0.85 1,548 1110 56 8 2.11 0.70 224 1114 152 17 2.68 0.66 608 1115 12 5 2.12 0.91 48 1116 56 16 3.08 0.77 224 1117 274 25 3.21 0.69 1,096 1118 79 18 3.24 0.77 316 1119 10 4 185 0.92 40 1120 24 9 243 0.76 96 1121 140 8 2.14 0.71 560 1122 271 17 2.43 0.59 1,084 1123 143 11 3.06 0.88 572 1124 343 14 2.11 0.55 1,372 1125 363 13. 1.78 0.48 1,452 1126 442 13. 161 0.43 1,768 1127 62 12 2.92 0.81 248 1128 53 9 2.59 0.82 212 1129 734 9 044 £0.14 2,936 1130 204 13. 1.72 = 0.46 816 1131 199 9 1.38 0.44 796 1132 388 4 0.20 0.10 1,552 1133 802 8 030 0.09 3,208 1134 45 1 0.00 — 180 1135 8 3 1.06 0.07 32 1136 171 1 0.00 — 684 1137 746 2 0.06 0.06 2,984 1138 400 11 2.49 0.72 1,600 1139 56 9 283 0.89 224 1140 20 6 1.94 0.75 80 1141 146 13° 2.21 0.59 584 1142 131 20 3.41 0.79 524 1143 186 21 3.48 0.79 744 which are subjected to stronger hydrodynamism. A non-parametric Kruskal-Wallis test proved that there were significant differences between assemblages A and B in terms of the percentages of fine sand, carbonates, and silt-clay, the sort coefficient, and diversity (p < 0.001), as well as the percentages of coarse sand, gravel, and organic matter (p < 0.05). Assemblage B stations had high- er mean percentages of gravel, coarse sand, silt-clay, and carbonates than assemblage A. It also surpassed assem- blage A in the mean percentage of organic matter, mean diversity value, and sort coefficient. However, assemblage A stations exhibited higher percentages of fine sand than those of assemblage B. The same analysis was later applied to the different subgroups and a multiple comparison test (Conover, 1971) was carried out a posteriori, to reveal the OLABARRIA ET AL.: BENTHIC MOLLUSKS OF INLET OF BANO 109 Table 5. Spearman’s rank correlations between ordination axes and community index, and environmental variables in Inlet of Bano. Depth (measured in relation to zero tidal level): subtidal stations (-m) and inter- tidal stations (+m). (H’, Shannon-Wiener diversity index; J’, Pielou’s evenness index; R, species richness; * p <0.05; **, p< 0.01; NS, not significant.). Environmental variables R J Gravel 0.44** 0.33** Coarse sand 0.43** 0.35* Medium sand -0.15NS 0.04NS Fine sand -0.69** -0.24* Carbonates 0.78** 0.13NS Organic matter 0.14NS -0.19NS Nitrogen 0.41** -0.03NS Silt-clay 0.45** -0.04NS Depth -0.68** -0.13NS Sort coefficient 0.44** 0.30* groups that significantly differed from one another in the different parameters (Table 6). The ordination analysis (DCA) was carried out on the subtidal and intertidal stations as a whole. However, because of the excessive concentration of points in the upper and lower quadrant of the negative part of axis I, a separate analysis for the subtidal and intertidal samples was performed. These analyses were based on the groups derived from the application of the point correlation coeffi- cient for a 99% level of significance. In the analysis of the intertidal stations, the first two axes accounted for 36.4% of the total variance (Fig. 5). Axis I had a slightly negative correlation with the depth parameter, and axis II correlated positively with the per- centage of fine sand and negatively with the percentage of carbonates and coarse sand (Table 5). In this analysis there are three groups of stations; on the positive side of axis I, both in the upper and lower quadrants, there is a group of stations that in the similarity (®) and constancy-fidelity analyses distinguished subgroup A3. In the upper left-hand quadrant, there was a group of closely-linked stations for which the most characteristic species was Hydrobia ulvae. This species reached densities of 2,960 ind/m2 (station 1137). This group defined subgroup A in the similarity analysis. Below, in the upper left-hand quadrant which con- tinued onto the lower quadrant, there was a group of sta- tions and species that characterized what has been defined as subgroup A2 using the point correlation and the constan- cy-fidelity indices. In the analysis of the subtidal stations, the coordi- nates of the first two axes accounted for 33.4% of the total variance (Fig. 6). Axis I had a high positive correlation with the percentage of fine sand and carbonates, and axis II displayed a correlation with the percentage of coarse sand and organic matter, and to a lesser extent with the percent- Subtidal Intertidal Axis I Axis II Axis I Axis II 0.51** -0.29NS_— -0.14NS_—-0.20NS_-0.52** 0.527" -0.39* 0.62** = -0.13NS_— -0.64** -0.09NS -0.05NS O.10NS — -0.08NS_—-0.04NS -0.69** 0.66** — -0.25NS 0.27NS_-0.67** 0.66** -0.59** = -0.40** — -0.05NS_—-0.67** -0.01NS 0.32NS — -0.52** — -0.21NS_— 0.1 2NS 0.22NS_ -0.47** — -0.04NS__-0.22NS 0.18NS -0.17NS.-0.29NS_—(.O5NS -0.56** 0.0SNS_—-0.06NS_—-0.39* 0.32* 0.50** -0.41* -O0.16NS — -0.09NS_—-0.41 ** age of carbonates and nitrogen (Table 5). In this analysis, a group of stations and species defining subgroup B1 can be seen in the upper and lower right-hand quadrants, located toward the positive end of axis I. On the negative side of axis I, in both the upper and lower quadrants, there is a group of stations and species that has been defined by the similarity index (®) and the constancy and fidelity indices as subgroup B2. In the vicinity of this group there is a group of stations (S126, $122, S112, S118, S130, S134) which are transition stations, grouped in the same cluster in the point correlation analysis (subgroup B2). These are sta- tions with modified bottoms, although originally they most likely had bottoms similar to the stations in subgroup BI. However, due to dredging operations, sediments from other Table 6. Multiple comparison test (Conover, 1971) between the sub- groups derived from the point correlation analysis (Al, A2, A3, BI, B2) indicating groups that showed no significant differences (p < 0.01). Gravel percentage Coarse sand percentage Al Abstract: The effect of blocking the release of the first polar body (PB1) or second polar body (PB2) with 3°C cold shock treatment on the development and ploidy of embryos was studied in the small abalone, Haliotis diversicolor (Lischke 1846); Groups of PB1 eggs were treated for 5, 10, 20, or 40 min, and resulted in 21 + 3.0%, 17 + 3.5%, 7 + 2.0%, and 0% normal gastrulae, respectively. The same treatment of PB2 eggs produced 12 + 2.6%, 15 + 2.6%, 3 + 2.5%, and 0% normal gastrulae, respectively. The percentage of normal embryos in the control group was 68 + 2.3%. Significant differences in percent nor- mal embryos (by ANOVA and Duncan’s analysis) were found between the treatments and the control (F316 = 19.28, p < 0.001) but not between equivalent treatments in the PB1 and PB2 groups. In the control groups, 41.8% of the embryos were diploid, while spontaneous haploids (9.0%), triploids (7.5%), and aneuploids (41.7%) also occurred. When the percentages of diploids, triploids, and aneuploids in the treated groups were compared with the control, there was a significant difference with respect to the diploid group (F4,4 = 6.88, p < 0.05) but not the triploid (F4,4 = 1.60, p > 0.05) or aneuploid (F44 = 3.05, p > 0.05) groups. Comparison (using Duncan’s analysis) of the treatments and the control of the factors of time and polar body in relation to diploidy showed a significant difference only in the 10 min group of PB1 and PB2 (p < 0.05) but no variation between PB1 and PB2 (p > 0.05). The results show a positive relationship between the length of immersion time at 3°C and the number of abnormal embryos, and no significant relationship between 10 min immersion and the development of tripoidy. This study illustrates that the formation of various levels of ploidy during chromosome segregation is random. No differ- ences in forming various ploidy levels were detected between the PB1 and PB2 groups. Key words: polar bodies, ploidy, development, aquaculture, Haliotidae Recently, much discussion in the aquacultural sci- The small abalone, Haliotis diversicolor (Lischke, ences has focused on the potential of chromosome manipu- 1846), is currently raised in commercial hatcheries in lation techniques, most specifically the artificial induction Taiwan (Chen and Yang, 1979). Triploid abalone could be of triploidy (three sets of chromosomes in each cell). Two an important part of the abalone culture industry if they can features of triploids are advantageous: (1) The nuclei of be shown to have a faster growth rate than diploid individu- triploid cells are larger to accommodate increased DNA in als. Abalone growers wish to produce the maximum the nucleus, and this leads to a concomitant increase in amount of flesh so preventing the onset of sexual maturity overall cell size. (2) Triploids have poorly developed using triploidy is desirable. To be commercially beneficial, gonads and produce far fewer gametes than diploids, result- the induction technique must result in a high percentage of ing in overall larger individuals. This is in contrast to triploids and a high percent survival. Possible factors diploid organisms which grow until sexual maturity is affecting these two parameters include the induction reached when growth slows down or stops as a large pro- method, the polar body that is blocked, the animal species, portion of energy becomes devoted to the production of conditions during induction, start time and duration, ambi- eggs or sperm. ent temperature, and the quality of the eggs and sperm. Previous investigations have demonstrated the In the Pacific abalone, Haliotis discus hannai (Ino, application of chemical agents, temperature (heat or cold), 1980), Arai et al. (1986) reported that 3°C cold shock pro- or hydrostatic pressure to induce triploidy (Stanley et al., duced high percentages of triploids. Kudo er al. (1991) 1981, 1984; Chaiton and Allen, 1985; Arai et al., 1986; used the same inducing conditions with H. diversicolor Quillet and Panelay, 1986; Dowing and Allen, 1987; diversicolor, but did not produce the same results. The lat- Yamamoto et al., 1988; Allen er al., 1989; Stephens, 1989; ter authors provided a more detailed analysis of early Kudo et al., 1991). Triploids have been induced in a vari- embryos which could partly account for this discrepancy. ety of fish and mollusks (Longo, 1972; Arai et al., 1986; Stanley et al. (1984) found that meiosis I triploid oysters Beaumont and Fairbrother, 1991). (inhibiting the first polar body; PB1) grew better than American Malacological Bulletin, Vol. 14(2) (1998): 139-147 139 140 AMER. MALAC. BULL. 14(2) (1998) meiosis II triploids (inhibiting the second polar body; PB2), however, their observation continues to be controversial (Beaumont and Fairbrother, 1991) and the induction of PB1 and PB2 triploids continues to be of interest. In this study, karyological analyses were used to determine ploidy of gas- trula-stage embryos of H. diversicolor following blocking of PB1 and PB2 with different exposure times to 3°C cold shock treatment. These findings should help clarify the optimum conditions for the manipulation of triploidization in this species. METHODOLOGY Gamete preparation Parental abalone were cultured from the seed of artificial propagation and broodstock in a cement pond (1.5 x 4.x 1.0 m) at the Tainan Station of the Taiwan Fishery Research Institute (Yang and Ting, 1987, 1988). 20 male and 20 female abalone (shell length 6-7 cm; 25-30 g) were placed separately into 24 | plastic containers (40 x 30 x 20 cm) filled with 10 1 of UV-irradiated, filtered seawater. Gametes were obtained following the methods of Chen and Yang (1979) by artificially inducing spawning with ultravi- olet (UV) light and thermal violence. Water temperature was gradually increased with a heater from 23° to 28°C over 5 hr (mean rate of increase 1°C/hr) and then gradually decreased with 4°C seawater from 28° to 23°C over 2.5 hr (mean rate of decrease 2°C/hr). Two or three such cycles were sufficient to induce male and female spawning. The water temperature was maintained at 24° + 1°C during spawning (Yang and Chen, 1979). Investigation of polar body release time Eggs (ca. 200 pm in diameter) were collected on a 60-um mesh screen. The unfertilized eggs were divided into groups of 104 individuals and each group was suspend- ed in a | liter beaker of seawater at 19°, 21°, 24°, 27°, or 30°C, respectively. Sperm solution (2 x 105 per ml) was also divided into five parts (10 ml/part) and each was placed in a 50 ml beaker at one of the five above-men- tioned temperatures. For insemination, | ml of the stock sperm solution was added to the egg solution at the same corresponding temperature. After insemination, a sample of eggs (102) was drawn every 2-3 min and fixed in 5% for- malin. The time of the release of the first and second polar bodies, as well as the first cleavage of zygotes, was defined as the time when 50% of the eggs in a sample had reached that stage. Cold shock treatment 3 min after insemination at 24°C (to inhibit the release of PB1), four samples of inseminated eggs (104) were collected on a 60-um mesh screen and placed into 3°C cold shock treatment for 5, 10, 20, and 40 min, respec- tively. Four additional samples were collected at 10 min post-insemination (to inhibit the release of PB2), and treat- ed in the same cold shock series. After treatment, each egg sample was placed into a 2 | beaker of seawater at 24°C at a density of 5 eggs/ml. The egg suspensions were not aerated and were maintained in fresh seawater by decanting and replacing the upper two-thirds of water every hour until the eggs had settled to the bottom of the beaker. Determination of percent normal embryos Morphologically normal embryos have been shown to have a higher survival rate in the subsequent floating stages. Following the technique of Yang and Chen (1979) for calculating the percentage of normal gastrulae to esti- mate the effect of the length of time in cold shock treat- ment, samples of embryos (102) were collected 6 h after insemination and fixed in 5% formalin. Normal embryos at this stage have a smooth surface with a circular shape. In abnormal embryos, the cells are damaged, are of different sizes, and their shape is not circular. Determination of ploidy Chromosome squashes were prepared from gastru- lae in each treatment using the methods of Allen et al. (1989). Ploidy was determined by counting chromosomes. Categories of ploidy were: haploid (1n = 15-16), diploid (2n = 31-32), triploid (3n = 47-48), tetraploid (4n = 63-64), and pentaploid (5n = 79-80). All others were considered aneuploid. Photographs were taken with Kodak Technical Pan Film 2415. Karyotype analyses were performed as described by Arai et al. (1986). Data analysis Analysis of Variance (ANOVA) and Duncanis analysis (Sokal and Rohlf, 1981; Statsoft, 1993) were used to determine the significance of differences between the control and treatments in terms of percent normal embryos and percent diploidy, triploidy, and aneuploidy. RESULTS Normal meiosis after insemination releases the first polar body (PB1) and second polar body (PB2), after which the egg undergoes the first cleavage (Fig. 1). In this study, temperature influenced the time of polar body release and cleavage of the egg. The higher the temperature, the short- er the time before polar body release and first cleavage (Fig. 2). At 24°C, PB1 was released at ca. 8 min and PB2 at YANG ET AL.: DEVELOPMENT OF HALIOTIS DIVERSICOLOR 14] ca. 19 min after insemination; the first egg cleavage occurred at ca. 43 min. These cytological events took less than half as long to occur in eggs maintained at 19°C than at 30°C. Cold shock effectively inhibited polar body release. Following insemination, treated eggs released only one polar body after anaphase II while untreated eggs undergo- ing normal meiosis released both PB! and PB2 (Fig. 3). From microscopic examination of the embryos, 68 + 2.3% of the control group exhibited normal gastrulae. In the PB1 groups, 21 + 3.0%, 17 + 3.5%, and 7 + 2.0% of the gastrulae were normal after treatment times of 5, 10, and 20 min, respectively, while for the PB2 groups, 12 + 2.6%, 15 + 2.6%, and 3 + 2.5% were normal. All eggs subjected to 40 min cold shock at the PB1 or PB2 stage produced deformed embryos by the gastrula stage. When comparing percent normal embryos in the treatment and control groups, all treatments of PB1 and PB2 groups differed sig- nificantly (p < 0.001) from the control. Between PB1 and PB2 groups, however, there were no significant differences (p > 0.05) for each of the treatment times. Once an egg completes meiosis I and II, the chro- mosomes of the male nucleus combine with those of the female nucleus. This is followed by mitosis and subsequent development. In our observations of the chromosomes in cells at the gastrula stage, the control and treatment groups exhibited six types of ploidy (Fig. 4; Table 1). In the con- trol group, 41.8% of the embryos were diploid, 9.0% hap- loid, 7.5% triploid, and 41.7% aneuploid. No tetraploids or pentaploids were found in the control. Variations in ploidy in relation to treatment time and in polar body between treatment and control groups were Ambient temperature (°c) ae 10 20 30 40 50 60 ; ; Time of polar body elapsed and egg first cleavage (min) Fig. 1. Inseminated egg of Haliotis diversicolor undergoing normal meio- sis. A. First polar body (PB1) released after anaphase I of meiosis I. B. Fig. 2. Time of first polar body (PB1) and second polar body (PB2) Second polar body (PB2) released after anaphase II of meiosis II. C. First release and first cleavage of the egg (EFC) in Haliotis diversicolor, as cleavage of the egg. Scale bar = 20 um. observed after insemination at five water temperatures. 142 AMER. MALAC. BULL. 14(2) (1998) Fig. 3. Inseminated egg of Haliotis diversicolor after anaphase II of meio- sis II. A. Without cold shock treatment; two polar bodies observed (arrows). B. After cold shock treatment; only one polar body (arrow) observed. Scale bar = 20 um. analyzed by ANOVA. There were significant differences in the diploid category for immersion time (F4.4 = 6.88, p < 0.05) and polar body (F2.6 = 8.27, p < 0.05). Duncan’s analysis of the factors of time and polar body in relation to diploidy showed that 10 min of immersion time differed significantly (p < 0.05) from 5, 20, and 40 min, and that there were significant differences between PBI and PBI with the control, but no significance (p > 0.05) between PB1 and PB2. Studies of the karyotypes of diploid, triploid, and tetraploid small abalone revealed that eight chromosomes are metacentric, seven are submetacentric, and one is subte- locentric/acrocentric (Fig. 5). The distribution of chromo- some numbers in control and treated groups is shown in Fig. 6. In the control group, as expected, most embryos were normal diploids with 32 chromosomes, while in all the treated groups, most embryos were aneuploid, i. e. 2n # 32. In the treated groups, only a small percentage of triploid and tetraploid embryos were found, most embryos being aneuploid. DISCUSSION Species and the environmental nature of the habitat affect the release time of the first and second polar bodies. In the Pacific oyster, Crassostrea gigas (Thunberg, 1793), ten maternal tetrad chromosomes undergo meiosis I and II, and release two polar bodies (Arai et al., 1986; Quillet and Panelay, 1986; Guo et al., 1992; Longo et al., 1993). The release times of PB1 and PB2 are 17 and 35 min, respec- tively, at 24-25°C (Allen et al., 1989); these and the time of the first egg cleavage are influenced by water tempera- ture (Arai et al., 1986; Kudo et al., 1991). In the small abalone studied here, the release times of PB1 and PB2 were 8 and 19 min, respectively. Haliotis discus hannai from northern Japan has release times of 15 and 38 min, respectively, at 25°C (Arai et al., 1986). In this report, vari- ation in temperature was a major factor affecting polar body release and cell cleavage time. In this study, the percentage of morphologically normal embryos was used as a measure of larval survival. Although “percent normal” is admittedly not equal to “per- cent survival,” normal morphological development does have a positive relationship with survival. This study of the effect of cold shock on the percentage of normal gastrulae showed that temperature was the major factor causing dam- age to eggs; the longer the immersion time, the greater the percentage of deformed embryos. All four PB! and PB2 groups showed a low percentage of normal gastrulae, but there were no differences between PB1 and PB2 groups (p > 0.05); confirmation of corresponding larval survival requires further studies. These results differ from those of Guo et al. (1992) who found that in the Pacific oyster, blocking PB1 always resulted in higher mortalities than did blocking PB2. The results published by Kudo et al. (1991) on Haliotis diversicolor diversicolor (i. e. survival rates of 57 and 16%, respectively, in PB1 and PB2 groups subjected to 10 min cold shock) are likewise contrary. Treated groups exhibited lower percentages of diploids than the control. This could indicate that the cold shock treatment effectively inhibited the polar body of the egg. However, the treatments did not produce a higher per- centage of triploids than diploids in the control, suggesting that chromosome segregation does not always lead to triploidization. The occurrence of haploidy and triploidy in the con- trol groups was not due to cold shock, suggesting that hap- loids and triploids occur spontaneously in small abalone. Spontaneous haploids could be caused by the failure of sperm pronucleus incorporation, while triploids have been YANG ET AL.: DEVELOPMENT OF HALIOTIS DIVERSICOLOR 143 Fig. 4. Categories of ploidization of chromosomes in Haliotis diversicolor embryos treated by cold shock after insemination. A. Haploid. B. Diploid. C. Triploid. D. Tetraploid. E. Pentaploid. F. Aneuploid. Scale bar = 20 um. 144 AMER. MALAC. BULL. 14(2) (1998) 2N pn cee © eS ee Se ate 1 2 3 4 5 6 7 8 KX m& ha Yu RA ah a4 nA i=} 10 11 42 13 14 15 16 ST/A — Spun 3N BRK wwe KAD REE Rum BKK RXK xek -B 4 2 3 4 5 S 7 B FYw~ Ga K EN Awr ARIK AX» Bah O48 3 10 11 12 13 14 15 16 SM ST/A 4N -—C ve Kad Ke XM QaAX MK! Reat FAxn) AKA X ytar 4 2 3 4 Ss 6 7 8 AK KR KKRAKR ARE KASS ARABS RAme UZAF Féhe 9 10 11 12 13 14 15 16 SM ST/A Fig. 5. Karyotypes of diploid (A), triploid (B), and tetraploid (C) Haliotis diversicolor. Pairs 1-8 are metacentric (M), 9-15 are submetacentric (SM), and pair 16 is subtelocentric/acrocentric (ST/A). Scale bar = 5 um. Table 1. Percentages of morphologically normal Haliotis diversicolor embryos (+ standard error), and of the various ploidy levels (with number of observations in parentheses) in the control and various treatments (immersion times), following 3°C cold shock blocking the first polar body (PB1) or second polar body (PB2). (1n, haploid; 2n, diploid; 3n, triploid; 4n, tetraploid; Sn, pentaploid; An, aneuploid; *, p < 0.05; **, p < 0.001). % Ploidization Group Polar % Normal N In 2n 3n 4n 5n An body Control 68 + 2.3 67 9.0 41.8 75 0.0 0.0 41.7 (153) (6) (28) (S) (28) 5 min PBI 21 + 3.0** 151 21.9 10.6 4.0 1.3 0.7 61.5 (147) (33) (16) (6) (2) (1) (93) 5 min PB2 12 + 2.6** 19 26.3 15.8 5.3 0.0 0.0 52.6 (104) (5) (3) (1) (10) 10 min PB1 17=23:5"* 240 75 16.7* 10.0 3.8 1.7 60.3 (187) (18) (40) (24) (9) (4) (145) 10min PB2 15 + 2.6** 91 8.8 20.9* 5.5 1.1 0.0 63.7 (147) (8) (19) (5) (1) (58) 20 min PB1 7 + 2.0** 283 7.1 16.3 11.0 1.8 0.4 63.4 (164) (20) (46) (31) (5) (1) (180) 20 min PB2 Sp a ea 15 0.0 13.3 6.7 6.7 0.0 73.3 (164) (2) () (1) (11) 40 min PBI 0 100 26.0 14.0 5.0 4.0 2.0 49.0 (179) (26) (14) (5) (4) (2) (49) 40 min PB2 0 13 38.5 0.0 0.0 0.0 0.0 61.5 (168) (5) (8) YANG ET AL.: DEVELOPMENT OF HALIOTIS DIVERSICOLOR 145 30 | A: Control 25) (n=67) | 20 > .*) c @ 15 =! o ov ic 10 5 2 10-2030 40 50 60 70 80 90 100 Chromosome number 30 3 ; B: PB1, 5 minutes F: PB2, 5 minutes 25 (n=151) (n=49) 20 2 > | e g eo eo i 10 “1 ‘| | | | 0! pt be ee oO ii oes 10 20 30 40 50 60 70 80 90 100 10 20 30 40 50 60 70 80 90 = 100 Chromosome number Chromosome number 20 ; 8 C7 PEA, 10 minutes G: PB2, 10 minutes (n=240) 7 (n=91) 15 6 > > 5 2 2 @ 10 wo 4 a o g 2 3 ir rg 5 2 1 0 7 | on o- 2! i ih ener 70 80 90 100 10 20 30 40 50 60 70 80 90 100 Chromosome number Chromosome number 201) 7 H: PB2, 20 minutes | O: PB1, 20 minutes . ” | (n=283) {n=t5) 15 2 a a c ec o 10 o s J > co @ 2 ir o1 5 o! 1 ee coy 0 see 10 20 30 40 50 60 70 80 90 100 10 20 30 40 50 60 70 80 90 100 Chromosome number Chromosome number 6 a8 E: PB1, 40 minutes |: PB2, 40 minutes a6 (n=400) 5 (n=13) 44 4 on, 12 > 2 2 § 10 o 3 3 | o 8 > 2 < uw 6 u 2 x 1 2 0 ToT ot oo —_ 0 ae eee ty 10 20 30 40 50 60 70 80 90 = 100 10 20 30 40 50 60 70 80 90 100 Chromosome number Chromosome number Fig. 6. Distributions of chromosome counts in the control and treatment groups of Haliotis diversicolor embryos treated with 3°C cold shock blocking the first polar body (PB1) or second polar body (PB2). Untreated control (A). Treatment of PB1 groups: 5 min (B), 10 min (C), 20 min (D), 40 min (E). Treatment of PB2 groups: 5 min (F), 10 min (G), 20 min (H), 40 min (I). (n, numbers of embryos). 146 AMER. MALAC. BULL. 14(2) (1998) reported in amphibians and fish and are probably caused by fertilization of unreduced ova (Fankhauser, 1945; Thorgaard and Gall, 1979). In this study of small abalone, 41.7% aneuploids were observed in the control. Aneuploids have commonly been assumed to be artifacts caused by poor karyological techniques such as chromosome loss or overestimation. In this study, a statistical significance between the control and treated groups (p > 0.05) demonstrated that aneuploids were not a consequence of cold shock treatment. The for- mation of aneuploid gametes in mollusks can be common, particularly under conditions of environmental stress (Dixon, 1982). Guo et al. (1992) used the chemical cytochalasin B to inhibit PB1, and the resulting aneuploids in the various treatments were distributed around two peaks. The high level in small abalone is in contrast to a low proportion (9%) in the Pacific oyster (Guo ef al., 1992), perhaps indicating that the abalone is more sensitive to the environment. Aneuploidy often leads to abnormality and death in higher animals (Dixon, 1982). In studies where ploidy was determined at juvenile or adult stages (Stanley et al., 1981; Downing and Allen, 1987), aneu- ploids could have been absent because they had already died during early embryonic development. The high per- centage of aneuploidy in small abalone suggests that block- ing the polar bodies caused disruption of normal chromo- some segregation, and could explain the higher mortality rate at the settling stage (90%) experienced during artificial propagation (Yang and Ting, 1988). The results of this study regarding triploids and tetraploids inhibiting PB1 or PB2 agree with those for the Pacific oyster (Crassostrea gigas; Guo et al., 1992), the American oyster [C. virginica (Gmelin, 1791); Stanley et al., 1981], the mussel [Mytilus edulis (Linné, 1758); Yamamoto and Sugawara, 1988], and the Pacific abalone (Haliotis discus hannai,; Arai et al., 1986). On the other hand, results of this study differ from those of other studies in which only triploid embryos (Quillet and Panilay, 1986) or only tetraploid embryos (Stephens, 1989) were observed to inhibit PB1. This difference could be due to different experimental methods and conditions. Comparison of treat- ed groups and controls in the diploid category were a con- sequence of cold shock at 10 min. Variation in these envi- ronmental factors was probably also responsible for the dif- ferent proportions of ploidy observed in the immersion times and between the two polar bodies in this study. The percentage of normal gastrulae and production of triploids showed that the optimal cold shock immersion time is between 5 and 10 min. The production of various levels of ploidy by inhibiting PB1 and PB2 demonstrated the complex nature of chromosome segregation. More rigid inducing conditions and the mechanism of chromosome segregation will be investigated in future studies. ACKNOWLEDGMENTS This research was partially supported by the National Science Council, Republic of China. The authors are grateful to Dr. I-Chiu Liao for supporting this research and also express their sincere thanks to Prof. Bruce L. Robertson for his valuable comments. LITERATURE CITED Ahmed, M. 1973. Cytogenetics of oysters. Cytologia 38:337-347. Allen, S. K., Jr., L. Downing, and K. K. Chew. 1989. Hatchery Manual for Producing Triploid Oysters. Washington Sea Grant Program, University of Washington Press, Seattle. 27 pp. Arai, K. F., F. Naito, and K. Fujino. 1986. Triploidization of the Pacific abalone, Haliotis discus hannai (Ino) with temperature and pres- sure treatments. Bulletin of the Japanese Society of Scientific Fisheries 52(3):417-422. Beaumont, A. R. and J. E. Fairbrother. 1991. Ploidy manipulation in mol- luscan shellfish: a review. Journal of Shellfish Research 10(1):1- 18. Chaiton, J. A. and S. K. Allen. 1985. Early detection of triploidy in the lar- vae of Pacific oysters, Crassostrea gigas, by flow cytometry. Aquaculture 48:35-43. Chen, H. C. and H. S. Yang. 1979. Artificial propagation of abalone, Haliotis diversicolor Lischke. China Fisheries Monthly 314:3-9. Dixon, D. R. 1982. Aneuploidy in mussel embryos (Mytilus edulis L.) originating from a polluted dock. Marine Biology Letters 3:155- 161. Downing, S. L. and S. K. Allen. 1987. Induced triploidy in the Pacific oys- ter, Crassostrea gigas: optimal treatments with cytochalasin B depend on temperature. Aquaculture 61:1-15. Fankhauser, G. 1945. The effects of chromosome number on amphibian development. Quarterly Review of Biology 20:20-78. Guo, X., K. Cooper, W. K. Hershberger, and K. K. Chew. 1992. Genetic consequences of blocking polar body I with cytochalasin B in fer- tilized eggs of the Pacific oyster, Crassostrea gigas: 1. Ploidy of resultant embryos. Biological Bulletin 183:381-386. Ino, T. 1980. Fisheries in Japan, Abalone and Oyster, Japan Marine Products Photo Materials. Tokyo, 165. Kudo, M., K. Arai, and K. Fujino. 1991. Triploidization of Haliotis diver- sicolor diversicolor by cold shock. Nippon Suisan Gakkaishi 57(7):1263-1267. Longo, F. J. 1972. The effects of cytochalasin B on the events of fer- tilization in the surf clam, Spisula solidissima. 1. Polar body for- mation. Journal of Experimental Zoology 182:321-344. Longo, F J., L. Mathews, and D. Hedgecock. 1993. Morphogenesis of maternal and paternal genomes in fertilized oyster eggs (Crassostrea gigas): effects of cytochalasin B at different periods during meiotic maturation. Biological Bulletin 185:197-214. Lu, J. K. 1986. The Combined Effects of Salinity and Temperature on Meiosis and Early Mitosis of the Pacific Oyster (Crassostrea gigas) Oocytes. Masters Thesis, University of Washington, Seattle. 110 pp. Quillet, E. and P. J. Panelay. 1986. Triploidy induction by thermal shocks in the Pacific oyster, Crassostrea gigas. Aquaculture 57:271-279. Sokal, R. R. and F. J. Rohlf. 1981. Biometry, 2nd ed. W.H. Freeman, YANG ET AL.: DEVELOPMENT OF HALIOTIS DIVERSICOLOR 147 San Francisco. 859 pp. Statsoft, Inc. 1993. Statistica for Windows, release 4.5B. Tulsa, Oklahoma. Stanley, J. G., S. K. Allen, Jr., and H. Hidu. 1981. Polyploidy induced in the American oyster, Crassostrea virginica, with cytochalasin B. Aquaculture 23:1-10. Stanley, J. G., H. Hidu, and S. K. Allen, Jr. 1984. Growth of American oysters increased by polyploidy induced by blocking meiosis I but not meiosis II. Aquaculture 37:147-155. Stephens, L. B. 1989. Inhibition of the First Polar Body Formation in Crassostrea gigas Produces Tetraploids, not Meiotic I Triploids. Masters Thesis, University of Washington, Seattle. 96 pp. Strathmann, M. F. 1987. Reproduction and Development of Marine Invertebrates of the Northern Pacific Coast. University of Washington Press, Seattle. 126 pp. Thorgaard, G. H. and G. A. E. Gall. 1979. Adult triploids in a rainbow trout family. Genetics 93:961-973. Yamamoto, S. and Y. Sugawara. 1988. Induced triploidy in the mussel, Mytilus edulis, by temperature shock. Aquaculture 72:21-29. Yamamoto, S., Y. Sugawara, T. Nomaru, and A. Oshino 1988. Induced triploidy in the Pacific oyster, Crassostrea gigas, and pertor- mance of triploid larvae. Tohoku Journal of Agriculture Research 39(1):47-59. Yang, H. S. and H. C. Chen. 1979. Effects of temperature and salinity on the embryonic development of abalone, Haliotis diversicolor supertexta. Journal of Marine Science 21:78-84. Yang, H. S. and Y. Y. Ting. 1987. Experiments on the husbandry and maintenance of abalone broodstock. Bulletin of the Taiwan Fisheries Research Institute 43:171-177. Yang, H. S. and Y. Y. Ting. 1988. Artificial propagation of abalone: broodstock development of abalone, Haliotis diversicolor super- texta, in ponds. Bulletin of the Taiwan Fisheries Research Institute 44:199-202. Date of manuscript acceptance: | February 1998 The gametogenic cycle of Brachidontes exustus (Linné, 1758) (Bivalvia: Mytilidae) at Wassaw Island, Georgia Mary L. Sweeney and Randal L. Walker Shellfish Aquaculture Laboratory, University of Georgia Marine Extension Service, 20 Ocean Science Circle, Savannah, Georgia 31411-1011 U.S.A. Abstract: The gametogenic cycle of a population of Brachidontes exustus (Linné, 1758) was studied from January to December 1993 at the northern beach of Wassaw Island, Georgia. Staging values were assigned for Gonadal Indexing (GI) of histologically prepared gonadal tissue: 1 = spent; 2 = partially spawned; 3 = early active; 4 = late active; 5 = ripe. From January to March and September to December, males had significantly higher (ANOVA, p = 0.0249) GI values when compared to females. Because male and female GI showed no significant difference (ANOVA, p = 0.1244) over the major repro- ductive period (April to August), they were averaged per sample period. Rapid gonadal development occurred from March to mid-May and follicles were ripe by mid-July. A minor spawn might have occurred in late July. A major spawn occurred from September to November. Of 261 animals sampled, the sex ratio was significantly different (Chi-square = 7.48; p < 0.05) from parity (1.00 female:1.43 male) and 9.6% were sexually undifferentiated. The greatest per- centage of sexually undifferentiated individuals occurred in March (66.7%) with fewer from January to April. No hermaphrodites were observed. Key words: Brachidontes, gametogenesis, mussel, reproduction, sex ratio Brachidontes exustus (Linné, 1758), commonly inhabits rock pilings, sea walls, and wharf pilings in the intertidal zone and is most abundant in lower intertidal regions (Seed, 1980). The mussel ranges in the United States from North Carolina to Texas (Abbott, 1974). Information concerning species of Brachidontes is limited. Studies include life habitats (Seed, 1980), surface and shell morphology (Fuller and Lutz, 1988, 1989), and larval development (Campos and Ramorino, 1980; Fields and Moore, 1983). In addition, Morton (1988) studied the reproductive cycle of B. variabilis (Krauss, 1848) in a Hong Kong mangrove and showed that juveniles grew rapidly, are sexually mature at one-year of age, and had a short life span (< 3 years). Gametogenesis in B. variabilis was initiated in winter with individuals spawning in summer and being spent by October-November (Morton, 1988). Gametogenic cycles have been studied in other marine mussels including Geukensia demissa (Dillwyn, 1817) (Heffernan and Walker, 1989; Borrero, 1987; Brousseau, 1982); Modiolus capax (Conrad, 1937) (Aguirre and Ramirez, 1989); Modiolus metcalfei (Hanley, 1843) (Lopez and Gomez, 1982); Modiolus modiolus (Linné, 1758) (Seed and Brown, 1977; Jasim and Brand, 1989); Modiolus philippinarum (Hanley, 1843) (Walter and Cruz, 1980); Mytilus edulis (Linné, 1758) (Seed and Brown, 1977; Brousseau, 1983); and Perna picta (Born, 1778) (Shafee, 1989). Gametogenic studies of species in the southeastern U. S. include comparisons of G. demissa pop- ulations in South Carolina (Borrero, 1987), an energy flow study of a G. demissa population at Sapelo Island, Georgia (Kuenzler, 1961), and a study of reproduction in G. demis- sa at Wassaw Sound, Georgia (Heffernan and Walker, 1989). There is little information on the life history of Brachidontes exustus and no gametogenic study. The pur- pose of this study was to determine the gametogenic cycle of a population of B. exustus at Wassaw Island, Georgia, through histological examination. METHODS Between 15 and 20 individuals of Brachidontes exustus (1-2 cm in length) were collected monthly from January to March 1993, biweekly from April to September 1993, and monthly from November to December 1993, from the northern beach of Wassaw Island, Georgia (Fig. 1). Inclement weather prevented an October 1993 sam- pling. Individuals occurred at the base of dead oak trees on the shoreward side within the intertidal surf zone. Shell length (distance anterior-to-posterior) was American Malacological Bulletin, Vol. 14(2) (1998):149-156 149 150 AMER. MALAC. BULL. 14(2) (1998) WASS AW $OUND WASSAW ISLAND Fig. 1. Location of Brachidontes exustus population occurring at the base of dead oak trees in the intertidal surf zone on Wassaw Island, Georgia. determined using Vernier calipers. The visceral mass was fixed in Davidson’s solution (Humason, 1967) for 48 hr, rinsed in 50% alcohol overnight, and stored in 70% alco- hol. Tissue samples were dehydrated in a graded alcohol series, cleared in toluene, and embedded in Paraplast®. Paraffin blocks were sectioned (7-8 um) with a Leica 820 IT Microtome. The tissue sections were deparaffinized with toluene, rehydrated to water, and then stained with Harris, Hematoxylin and counterstained with Eosin (Bancroft and Stevens, 1982). Slides were examined and photographed under a light microscope. Gonadal tissue was rated according to characteris- tics of gametes and stage of maturity, as described for other species (Brousseau, 1982, 1983; Heffernan and Walker, 1989). Gonadal tissue also occurred in the mantle of all individuals. The gametogenic cycle was divided into the following categories: Sexually Undeveloped = 0. No male or female gametes detected. Follicles small and shrunken with little or no lumen. Male Developmental Stages: Spent = |. Follicles empty with a large lumen space. A few spermatozoan cells scattered throughout the follicle. Gonads appear much smaller than those in ripe indi- viduals. Blood cells present in the follicles (Fig. 2a). Partially spawned = 2. Most follicles with the clear lumen surrounded by clusters of spermatozoa (Fig. 2c). Early active = 3. A narrow-to-wide band of large cells (spermatogonia and/or spermatocytes) located inside each testis follicle. No smaller spermatids or sperma- tozoa cells present. Follicle lumen large and devoid of cells (Fig. 2e). Late active = 4. A wide band of larger cells (spermatogo- nia and spermatocytes) present along the inside of the testis follicle. Lumen small to nonexistent. Some fol- licles with small pink streaks. Smaller spermatozoa present toward the center of the follicle (Fig. 2g). Ripe = 5. Gonads full of spermatocytes, spermatids, and spermatozoa, making it difficult to differentiate cell types. Follicles stained black to dark purple with pink streaks in the center (Fig. 21). Female Developmental Stages: Spent = 1. Follicles mostly empty with only a few scattered oocytes. Blood cells present; follicles smaller than when ripe (Fig. 2b). Partially spawned = 2. Most of the oocytes mature, each with a large nucleus; many eggs unattached to the follicle wall. Lumen space large. Many follicles with a few or no mature oocytes (Fig. 2d). Early active = 3. Follicle lumen devoid of cells. Young oocytes with small nuclei attached to the follicle wall (Fis; 2). Late active = 4. Oocytes large, elongate, some attached to the follicle wall. Follicles with several mature oocytes present but lumen incompletely filled with eggs (Fig. 2h). Ripe = 5. Many large mature oocytes present; in most spec- imens the lumen space is filled. A few oocytes SWEENEY AND WALKER: GAMETOGENIC CYCLE OF BRACHIDONTES EXUSTUS 151 os Te ©. Sea, . &° ) ow* wo a di Fig. 2. Reproductive stages in male and female Brachidontes exustus: a. Spent male. b. Spent female. c. Partially spawned male. d. Partially spawned female. e. Early active male. f. Early active female. g. Late active male. h. Late active female. I. Ripe male. j. Ripe female. Scale bars = 20 pm (a, b, c, d, e, g, 1) or 50 um (f, h, j). 152 AMER. MALAC. BULL. 14(2) (1998) attached to follicle wall (Fig. 2j). The sex of each individual was determined and the sex ratio of the population (biweekly sample size, N = 15) was tested for non-parity by Chi Square analyses (Elliott, 1977). Statistical analyses of male and female gonadal index values were analyzed by Analysis of Variance (ANOVA) (a = 0.05) using SAS for PC computer (SAS Institute, 1989). T-tests (& = 0.05) (SAS Institute, 1989) were performed on gonadal index values and shell length data to determine if significant differences occurred at any specific time period. RESULTS Individuals varied in shell length from 11.3 to 24.6 mm (mean = 17.3 + 0.15 mm SE; N = 261). A total of 261 mussels were sexed with 9.6% sexually undifferentiated, 37.2% female, and 53.2% male. Sex ratio was significantly different from parity (Chi-square = 7.48; p < 0.05) with a ratio of 1.00:1.43 female to male. Female and male gonadal indices are presented in Fig. 3. ANOVA revealed that males had significantly (p = 0.0249) higher GI values over time than females. Males had significantly higher GI values, as determined by T-tests from January to March and from September to December. No significant difference (ANOVA; p = 0.1244) in GI value occurred between males and females from April to August. Males had a longer gametogenic cycle than females. Females showed a more rapid drop in GI value (N = 7, mean = 4.86 + 0.14; N = 6, mean = 1.83 + 0.65) over males (N = 8, mean = 5.00 + 0.00; N = 9, mean = 3.67 + 0.53) during peak spawning (August-September). Gonadal development was evident throughout Mean Gonadal Index (GI) + SE most of the year, except March. The lowest mean GI value for males, females, and sexually undifferentiated individu- als combined occurred in March (N = 15, mean = 0.58 + 0.29; Fig. 4), during which the highest percentage of undif- ferentiated animals (66.7%) occurred. The mean GI for males and females combined increased from March (N = 15, mean = 0.58 + 0.29) to April (N = 15, mean = 2.27 + 0.38) and reached a peak in mid-July (N = 15, mean = 4.79 + 0.29). By mid-July, 92.9% of the population was ripe and 7.1% was partially spawned. Low percentages of individu- als in the partially spawned stage occurred from May to July (Fig. 4), which could be representative of dribble spawning, i. e. when males and females continually produce and release sperm and egg cells, respectively. Indications of spawning were not reflected in mean GI values (Figs. 3-4) during May to July. The combined mean GI value for males and females increased to 4.31 in mid-August, with 76.9% ripe individuals and 23.1% at the partially spawned stage. The highest GI value (N = 15, mean = 4.93 + 0.07) was reached in late August when 93.3% of the sampled popula- tion was ripe and 6.7% were spent. A major spawn com- menced in early September when there was a decrease in mature gametes (N = 15, mean GI = 4.25 + 0.39). Spawning continued until December (N = 21, mean GI = 1.95 + 0.43) when 40% was spent, 15% was partially spawned, 20% was undifferentiated, and 25% was at the ripe stage (Figs. 4-5). DISCUSSION The gametogenic cycle of Brachidontes exustus from a population in coastal Georgia showed a unimodal -@- Female Gl —o— Male Gl T T T_T T_T Glan | aT | Jan Feb Mar Apr May dune Jul Aug Sept Nov Dec Fig. 3. Mean monthly gonadal index values + SE for male and female Brachidontes exustus collected at Wassaw Island, Georgia. SWEENEY AND WALKER: GAMETOGENIC CYCLE OF BRACHIDONTES EXUSTUS 5.5 eS 3.5 Mean Gonadal Index (GI) + SE T T | ae | T T T Jan Feb Mar Apr May T June 153 T Nov Dec T Aug T Jul Sept Fig. 4. Mean monthly mean gonadal index values combined for male and female Brachidontes exustus collected at Wassaw Island, Georgia. Percent Frequency of Stages WU \ N 14-Jan [_] 21-Apr | V/A Qa + ee a = wT o & eo ®@ 5S 5 5 u = SS ae ee om” © OO es s s a - Oo Date 15-Jul F HRipe Ei Late Active Bi Early Active N NWSPartially Spawned fi Spent N OUndifferentiated | f Ss D DB AaaQ > re) 5> 3 3 9 09 D9 o oe wy | Oe oO oO w (o>) w+ (ap) oO - AN nN Fig. 5. Relative frequency of gonadal stages over time in Brachidontes exustus collected at Wassaw Island, Georgia. pattern (Figs. 3-4) during 1993, when gamete production increased from March until July, with a major spawn occur- ring in August and September. Dribble spawning occurred from May to August (Fig. 4) and this phenomenon was similar to the spawning pattern observed in Mytilus edulis (see Brousseau, 1983). Brousseau (1983) attributed dribble spawning as a response to variations in environmental con- ditions over a period of time. The spawning period (August-September 1993) of B. exustus at Wassaw Island showed similarities to a Geukensia demissa population in Georgia, which spawned from August to October 1984 and July to September 1985 (Heffernan and Walker, 1989). The results of the present study were also similar to Kuenzler’s (1961) results (spawning July-September) for a population of G. demissa on Sapelo Island, Georgia, and to those of a study of a population of G. demissa at North Inlet Estuary, South Carolina (Borrero, 1987). Borrero (1987) showed that individuals from low intertidal areas increased in gonadal development from May to early August, with a minor spawn occurring in June and a major spawn from August through September. However, a longer spawning period, resembling our results, occurred in a Hong Kong mangrove population of B. variabilis (see Morton, 1988). Morton (1988) found that the population of B. variabilis ripened in the summer and by November and December 1985, the population was spent. The lowest mean gonadal index value in March (0.58 + 0.29) for Brachidontes exustus was followed by an increase throughout the spring until peak values were attained in July (mean = 4.79 + 0.21). The increase in 154 AMER. MALAC. BULL. 14(2) (1998) Table 1. Habitat, sex ratio, and percent hermaphrodites of various mytilid mussels. (E, estuarine; D, deep sea; M, marine; ND, not determined). Species Habitat N Female/Male % Herm. Source Brachidontes exustus M 261 1.00:1.43* 0 This Study B. variabilis E 247 0.80: 1.00 ND Hulings, 1986 B. variabilis E 330 1.00:0.69* 0 Morton, 1988 Choromytilus meridionalis (Krauss, 1848) M 700 0.75:1.00* 0 Griffiths, 1977 Crenomytilus grayanus (Dunker, 1882) ND ND 1.00:1.00 ND Dolgov, 1985 Geukensia demissa E 354 1.00:1.00 ND Brousseau ,1982 G. demissa E 429 1.00:1.00 0 Borrero, 1987 G. demissa E 498 1.00:1.00 0 Heffernan and Walker, 1989 Idas argenteus (Jetfreys,1876) D 101 1.00:0.83 0 Dean, 1993 Lithophaga bisulcata (d’ Orbigny, 1842) M 157 1.00:1.00 ND Scott, 1988 96 1.00:1.00 ND L. nigra (d@’ Orbigny, 1842) ND 231 1.00:0.91 ND Barkati and Asif, 1984 Modiolus barbatus (Linné, 1758) ND 975 1.00:1.41 ND Cahour and Lucas, 1968 1060 1.00:1.66 ND M. metcalfei E 1280 1.00:1.00 0.16 Lopez and Gomez, 1982 M. modiolus M 5965 1.00:1.00 sats Jasim and Brand, 1989 M. philippinarum ND 848 1.00:1.00 ND Walter and Cruz, 1980 Mytella guyanensis (Lamarck, 1819) ND 388 3.47:1.00* ND Sibaja, 1986 Mytilus edulis ND ND 1.00:1.18 ND Coe, 1943 M. edulis M 14,661 1.00:1.13* ND Seed, 1969 M. edulis M 534 1.00:1.00 0.02 Brousseau, 1983 M. edulis ND 3000 1.00:1.00 ND Dolgov, 1985 M. edulis ND 450 1.00:0.90 0 Sprung, 1983 M. edulis M 346 1.00:0.79 0.3 Kautsky, 1982 M. galloprovincialis (Lamarck, 1819) ND ND 1.00:1.00 ND Kudinskii and Shurova, 1990 Perna perna (Linné, 1758) ND 275 1.00:1.00 0 Lasiak, 1986 P. picta ND 800 1.00:1.20* ND Shafee, 1989 P. viridis (Linné, 1758) M 1.00:1.00 0 Barkati and Ahmed, 1974 648 1.00:0.98 0 1.00:0.79 0 P. viridis ND 1000 1.00:1.00 0 Walter, 1982 P. viridis ND 1760 1.00:0.78 ND Vakily et al., 1988 P. viridis ND 526 1.00:1.29 <0.1 Lee, 1988 Perumytilus purpuratus (Lamarck, 1819) | ND 797 1.00:1.00 ND Lozada and Reyes, 1982 Septifer virgatus (Wiesmann, 1837) ND 382 1.00:0.78 0 Morton, 1995 Xenostrobus securis (Lamarck, 1819) E 354 1.00:0.91 ND Wilson, 1969 316 1.00:0.84 ND * Reported to be significantly different from a 1:1 ratio by a Chi-Square test. ** Percent hermaphrodite not determined, but stated in text that some hermaphrodites were observed. gonadal development during spring reflects the pattern of gametogenesis reported for Geukensia demissa populations in Georgia by Heffernan and Walker (1989). The highest percentage of sexually undifferentiated B. exustus (66.7%) occurred in March. Sexually undifferentiated mussels occurred in lower percentages during January (10.5%), February (23.1%), April (20.0%), November (26.7%), and December (20.0%) (Fig. 5). Heffernan and Walker (1989) found that during February 1984, 50.1% of the G. demissa sampled were sexually undifferentiated. Also, a higher per- centage of undifferentiated G. demissa occurred in the win- ter months (November 90% and December 60%), followed by decreasing proportions of undifferentiated animals, and an increase in sexually developing individuals throughout the spring (Heffernan and Walker, 1989). Within Mytilidae, sex ratios vary substantially (Table 1). In this study, Brachidontes exustus had signifi- cantly fewer females than males (1.00:1.43). For B. vari- abilis, Hulings (1986) found that significantly more males occurred in a population in the Gulf of Aqaba (Red Sea), while Morton (1988) found greater number of females than males in a Hong Kong estuarine population. As with most members of the Mytilidae, Brachidontes exustus is dioecious and exhibits stable gonochronism. No hermaphroditic B. exustus were found. Hulings (1986) and Morton (1988) observed no hermaphro- dites for B. variabilis. Within the Mytilidae, hermaphro- ditism is a rare event (Table 1). SWEENEY AND WALKER: GAMETOGENIC CYCLE OF BRACHIDONTES EXUSTUS 155 ACKNOWLEDGMENTS The authors wish to thank the U. S. Department of Interior, Fish and Wildlife, Wassaw Island National Wildlife Refuge personnel for allowing us to collect mussels for this study. Ms. A. Boyette is thanked for the graphics herein. Ms. D. Thompson is thanked for typing the manu- script. This work was funded by the University of Georgia Marine Extension Service with minor support by the Georgia Sea Grant College Program under grant number NA66RG0282. LITERATURE CITED Abbott, R. T. 1974. American Seashells, 2nd ed. Van Nostrand Rheinhold, New York. 663 pp. Aguirre, M. C. G. and L. F. B. Ramirez. 1989. Circlo reproductivo de mejillon, Modiolus capax (Conrad, 1937) (Bivalvia: Mytilidae: Anisomyaria) en la Bahia de Los Angeles, California, Mexico. Anales Centro de Ciencias del Mar y Limnologia 16:157-170. Bancroft, J. D. and A. Stevens. 1982. Theory and Practice of Histological Techniques, 2nd ed. Churchill Livingstone, New York. 662 pp. Barkati, S. and M. Ahmed. 1974. Reproductive cycle of the marine mussel Perna viridis. Pakistan Journal of Zoology 6:31-40. Barkati, S. and M. Asif. 1984. Reproductive cycle and sex ratio of the date mussel Lithophaga nigra (d’Orbigny). Pakistan Journal of Zoology 16:1-7. Borrero, F. J. 1987. Tidal height and gametogenesis: reproductive varia- tions among populations of Geukensia demissa. Biological Bulletin 173:160-168. Brousseau, D. J. 1982. Gametogenesis and spawning in a population of Geukensia demissa (Pelecypoda: Mytilidae) from Westport, Connecticut. The Veliger 24:247-251. Brousseau, D. J. 1983. Aspects of reproduction of the blue mussel, Mytilus edulis (Pelecypoda: Mytilidae) in Long Island Sound. Fishery Bulletin 81:733-739. Cahour, A. and A. Lucas. 1968. Sex ratio et hermaphrodisme accidentel chez Modiolus barbatus (L.) mollusque bivalve. Comptes Rendus Hebdomadaires des Séances de L’Academie des Sciences 162:221-224. Campos, B. M. and L. M. Ramorino. 1980. Larval and early benthic stages of Brachidontes granulata (Bivalvia: Mytilidae). The Veliger 22:277-281. Coe, W. R. 1943. Sexual differentiation in mollusks. I. Pelecypods. Quarterly Review of Biology 18:154-164. Dean, H. K. 1993. A population study of the bivalve /das argenteus Jeffreys, 1876, (Bivalvia: Mytilidae) recovered from a submerged wood block in the deep North Atlantic Ocean. Malacologia 35:21-41. Dolgov, L. V. 1985. Sex determination in populations of Crenomytilus grayanus and Mytilus edulis differing in size-age compositions. Soviet Journal of Marine Biology 11:82-89. Elliott, J. M. 1977. Some methods for the statistical analysis of samples of benthic invertebrates. Freshwater Biological Association, Scientific Publication 25, 160 pp. Fields, A. and E. Moore. 1983. The larval biology of Brachidontes modio- lus (Linné, 1767) (Bivalvia: Mytilidae). The Veliger 26:52-61. Fuller, S.C. and R. A. Lutz. 1988. Early shell mineralogy, microstructure, and surface sculpture in five mytilid species. Malacologia 29:363- 371. Fuller, S.C. and R. A. Lutz. 1989. Shell morphology of larval and post- larval mytilids from the north-western Atlantic. Journal of the Marine Biological Association of the United Kingdom 69:181- 218. Griffiths, R. J. 1977. Reproductive cycles in littoral populations of Choromytilus meridionalis (Kr.) and Aulacomya ater (Molina) with a quantitative assessment of gamete productions in the former. Journal of Experimental Marine Biology and Ecology 30:53-71. Heffernan, P. B. and R. L. Walker. 1989. Gametogenic cycles of three bivalves in Wassaw Sound, Georgia. III. Geukensia demissa (Dillwyn, 1817). Journal of Shellfish Research 8:327-334. Hulings, N. C. 1986. Aspects of the reproduction of rocky intertidal mol- lusks from the Jordan Gulf of Aqaba (Red Sea). The Veliger 28:318-327. Humason, G. L. 1967. Animal Tissue Techniques. W. H. Freeman Company, San Francisco. 569 pp. Jasim, A. K. and A. R. Brand. 1989. Observations on the reproduction of Modiolus modiolus in Isle of Mann waters. Journal of the Marine Biological Association of the United Kingdom 69:373-385. Kautsky, N. 1982. Quantitative studies on gonadal cycle, fecundity, repro- ductive output and recruitment in a Baltic Mytilus edulis popula- tion. Marine Biology 68:143-160. Kudinskii, O. Yu. and N. M. Shurova. 1990. Sex realization in the mussel Mytilus galloprovincialis in the northwestern part of the Black Sea. Soviet Journal of Marine Biology 16:278-282. Kuenzler, E. J. 1961. Structure and energy flow of a mussel population in a Georgia salt marsh. Limnology and Oceanography 27:75-90. Lasiak, T. 1986. The reproductive cycles of the intertidal bivalves Crassostrea cucullata (Born, 1778) and Perna perna (Linnaeus, 1758) from the Transkei coast, southern Africa. The Veliger 29:226-230. Lee, S. Y. 1988. The reproduction cycle and sexuality of the green mussel Perna viridis (L.) (Bivalvia: Mytilidae) in Victoria Harbor, Hong Kong. Journal of Molluscan Studies 54:317-325. Lopez, M. D. G. and E. D. Gomez. 1982. Reproductive cycle of the brown mussel (Modiolus metcalfei) in Calatagan, Batangas. Kalikasan, Philippine Journal of Biology 11:74-82. Lozada, E. and P. Reyes. 1982. Reproductive biology of a population of Perumytilus purpuratus at El Tabo, Chile (Mollusca: Bivalvia: Mytilidae). The Veliger 24:147-154. Morton, B. 1988. The population dynamics and reproductive cycle of Brachidontes variabilis (Bivalvia: Mytilidae) in a Hong Kong mangrove. Malacogical Review 21:109-117. Morton, B. 1995. The population dynamics and reproductive cycle of Septifer virgatus (Bivalvia: Mytilidae) on an exposed rocky shore in Hong Kong. Journal of Zoology, London 235:485-500. SAS Institute. 1989. User’s Guide. Statistics. Version 6. SAS Institute Inc., Cary, North Carolina. 486 pp. Scott, P. J. B. 1988. Initial settlement behavior and survivorship of Lithophaga bisulcata (d’Orbigny) (Mytilidae: Lithopaginae). Journal of Molluscan Studies 54:97-108. Seed, R. 1969. The ecology of Mytilus edulis L. (Lamellibranchiata) on exposed rocky shores. I. Breeding and settlement. Oecologia 3:277-316. Seed, R. 1980. A note on the relationship between shell shape and life habits in Geukensia demissa and Brachidontes exustus (Mollusca: Bivalvia). Journal of Molluscan Studies 46:293-299. Seed, R. and R. A. Brown. 1977. A comparison of the reproductive cycles of Modiolus modiolus (L.), Cerastoderma (Cardium) edule (L.), and Mytilus edulis L. in Strangford Lough, Northern Ireland. Oecologia 30:173-188. Shafee, M. S. 1989. Reproduction of Perna picta (Mollusca: Bivalvia) from the Atlantic coast of Morocco. Marine Ecology Progress Series 53:235-245. 156 AMER. MALAC. BULL. 14(2) (1998) Sibaja, W. G. 1986. Madurez sexual en el mejillon chona Mytella guya- nensis Lamarck, 1819 (Bivalvia: Mytilidae) del manzlar en Jicaral Puntarenas, Costa Rica. Revista de Biologia Tropical 34:151-155. Sprung, M. 1983. Reproduction and fecundity of the mussel Mytilus edulis at Helgoland (North Sea). Helgolander Meeresuntersuchungen 36: 243-255. Vakily, J. M., S. Tuaycharoem, and J. Nugranad. 1988. Analysis of length and weight characteristics of green mussel, Perna viridis from Thailand. Asian Fisheries Society 1:165-174. Walter, C. 1982. Reproduction and growth in the tropical mussel Perna viridis (Bivalvia: Mytilidae). Kalikasan, Philippine Journal of Biology 11:83-87. Walter, C. and R. D. Cruz. 1980. Studies on the reproductive cycle of Modiolus philippinarum (Bivalvia: Mytilidae). Kalikasan, Philippine Journal of Biology 9:121-136. Wilson, B. R. 1969. Survival and reproduction of the mussel Xenostrobus securis (Lamarck) (Mollusca: Bivalvia: Mytilidae) in a Western Australian estuary. Part II]. Reproduction, growth and longevity. Journal of Natural History 3:93-120. Date of manuscript acceptance: 18 July 1995 Aspects of gametogenesis of Diplodon rotundus gratus (Wagner, 1827) (Bivalvia: Hyriidae) in Brazil Wagner Eustaquio Paiva Avelar and Sonia Helena Santesso T. de Mendonca Department of Biology, Faculty of Philosophy, Sciences and Letters of Ribeirao Preto, University of Sao Paulo, 14049-901 Ribeirao Preto, SP, Brazil Abstract: The reproductive biology of the freshwater bivalve Diplodon rotundus gratus (Wagner, 1827) was studied in specimens collected from the Pardo River, in the northeastern region of the State of Sao Paulo, Brazil. Analysis of the gametogenic cycle revealed that the species is a functional her- maphrodite. The gametogenic cycle began at the end of spring (November) and continued until the beginning of winter (June). In spring and summer the population was found to consist of pure females with only a small proportion of hermaphroditic individuals. Active gametogenesis occurred between February and June. Mature oocytes and sperm occurred simultaneously in the fall. Spawning is probably affected by exogenous factors such as temperature and prolonged droughts associated with host biology. In winter, spring, and summer, no fish with glochidia were found, although glochidial release occurred from February to August. Key words: Mollusca, Bivalvia, Hyriidae, gametogenesis, reproduction Several reports of the reproductive processes in species of Unionoidea in North America and Europe are available (Arey, 1932; Ghosh and Ghose, 1972; Zumoff, 1973; Wood, 1974a, b; Heard, 1975; Nagabhushanam and Lohgaonker, 1978; Dartnall and Walkey, 1979; Chung, 1980; Kat, 1983). Particularly outstanding for the southern hemisphere are the studies by Peredo and Parada (1984, 1986) on Diplodon chilensis chilensis Gray, 1828, and by Jones et al. (1986) on Australian Hyriidae. However, few studies are available about the life cycle involving gameto- genesis, time of incubation, glochidial elimination, host fish, and duration of parasitism for most freshwater bivalve species in southern regions. Systematic data and some bio- logical data, especially concerning the larval forms of freshwater bivalves, have been reported by Ortmann (1921), Morretes (1949), Bonetto (1954, 1959, 1961a, b, 1962, 1964, 1965), Bonetto and Ezcurra (1963), Mansur (1970, 1973), Hebling and Penteado (1974), Hebling (1976), and Mansur and Anflor (1981). It is important to understand the freshwater bivalve fauna of Brazil from many biological aspects, including their physiology, ecology, and systematic position. This is even more important for limnic bivalves that are poorly known and have been studied by few investigators. One such species is Diplodon rotundus gratus (Wagner, 1827), a species of wide geographic distribution on the South American continent (Bonetto, 1965). It is endemic to rivers of the Ribeirao Preto region (21°07’S, 47°45’W) where individuals live buried in muddy or sandy-muddy substra- tum quite close to the surface. Morretes (1949), Bonetto (196la, 1965), Parodiz (1968), Mansur (1970), and Hebling and Penteado (1974) have provided data about its biology, systematics, larval forms, and functional anatomy. Like most unionid bivalves, D. rotundus gratus incubates embryos in its inner demibranchs where the embryos differ- entiate into glochidia after about one month, as observed in the South American fauna by Bonetto (1959, 1961a, b). Our objective was to investigate the gametogenic cycle (development of gametes) and some aspects of the reproductive cycle (spawning and development) of Diplodon rotundus gratus to expand knowledge of its biol- ogy and to provide background information for future studies. MATERIALS AND METHODS Over a period of one year (February 1987 to January 1988) approximately 50 adult specimens were col- lected monthly from the Pardo River for histological study of the gametogenic cycle. The animals were located by probing the bottom of the river with the aid of feet and hands. They were fixed in aqueous Bouin’s solution at the collection site, removed from their shells after 24 hr, pre- American Malacological Bulletin, Vol. 14(2) (1998): 157-163 157 158 AMER. MALAC. BULL. 14(2) (1998) served in 70% alcohol, then dehydrated, cleared, and embedded in paraffin for histology. Approximately 20 7- um-thick transverse histological sections were obtained from each specimen (especially in the posterior region near the stomach where gonadal acini are more developed; Hebling and Penteado, 1974), mounted on microscope slides, and stained with Harris’ hematoxylin and eosin, Mallory’s trichrome, or combined stains (periodic-acid Schiff’s + alcian blue + Harris’ hematoxylin). The sections were analyzed under a light microscope for gametogenic phase and the various stages of follicular and marsupial development throughout the year. From February 1991 to January 1992, approximate- ly 50 adults were collected monthly for a complementary study of the reproductive cycle, again by probing the river bottom with feet and hands. The animals were transported alive to the laboratory in insulated boxes. The valves were pried open with a spatula to examine the general condition, development, and color of the marsupium. These macro- scopic analyses were complemented by microscopic exami- nation of a smear of the left inner demibranch of each ani- mal. The observed stages of the marsupium were later com- pared to the microscopic analysis of the histological sec- tions made earlier. At the bivalve collection site, fish were captured with a net of 2-cm mesh. These were transported alive to the laboratory in insulated boxes, placed in aquaria, identi- fied, and examined externally on the same day under a stereoscopic microscope to determine the presence of glochidia and/or lasids on the gills, fins, tail fins, and body. RESULTS GAMETOGENESIS In general, gametogenesis in Diplodon rotundus gratus is similar to that described by Peredo and Parada (1984, 1986) for D. chilensis chilensis. D. rotundus gratus is a hermaphrodite, with the gonad located in the visceral mass between the loops of the alimentary canal (Hebling and Penteado, 1974). The female portion occupies the lower two-thirds of the visceral mass, extending from the region of the digestive diverticula in front of the stomach to the posterior end of the foot. The male portion occupies the upper third of the visceral mass, extending from the ventral region backward to that below the renopericardial complex. Both portions consist of irregular tubes or branches. These sac-like follicles or acini differ in size and shape and are always delimited and enveloped by a follicular membrane. Inside the follicles are the male and female repro- ductive cells (Fig. 1), present in various phases of develop- ment and recognizable by their shape, size, and other char- acteristics. The developmental stages observed in D. rotun- dus gratus were similar to those described by Jones et al. (1986), with some modifications. Follicular Stage I is the beginning of gametogenesis and the phase of follicle for- mation. Components include: (1) precursor cells (mother- cells; Fig. 2), large and irregular cells with weak chromatin in parallel to the follicular wall, which give rise to sper- matogonia or oogonia after the first division; (2) sperm- morulae in males (Figs. 2-3), dark multinucleate spheres usually at the periphery of the follicle, defined in D. chilen- sis chilensis by Peredo and Parada (1984); and (3) nutritive cells in females (Fig. 5). Follicular Stage II is the active phase of gamete production and maturation (Fig. 3). Follicular Stage III is the phase of gamete elimination (spawning) to the gonoducts, defined here in two subdivi- sions: IIIa, partial elimination, and IIIb, extensive elimina- tion and end of gametogenesis. Follicular Stage IV is the period of recovery following gametogenesis. In the male (Fig. 4), follicles are quite reduced in number and size or almost totally absent; a few cells occasionally remain inside the follicle (Fig. 4). In the female, Stage IV follicles exist in three conditions: (1) small gonad with fewer folli- cles of reduced size with abundant interfollicular tissue [remaining oocytes assumed to be eliminated or to suffer lysis (Fig. 6)]; (2) follicles initiating a new cycle where nutritive cells are abundant; some oocytes still present and some follicles undergoing lysis; (3) all follicles in the process of lysis (Fig. 7). MARSUPIUM AND GLOCHIDIA After fertilization, zygotes migrate to the inner demibranchs (marsupia) where they develop and transform into glochidia. Glochidia, in turn, require a host to com- plete their life cycle. Development is completed through parasitism of fish that occur in the region. Fish captured in the Pardo River included species of Astyanax, Curimata, Holochestes, Piabina, Steindachnerina, Pimelodella, Lebistes, Geophagos, Leporinus, and Eingenmannia, all of which could have been infested. The most intensive infesta- tions were observed in the species Steindachnerina inscul- pita (Fernandez-Yépez, 1948), Holochestes heterodon (Eigenmann, 1915), and Pimelodella sp. The highest infes- tation of these fishes occurred in April. When living bivalves were examined in the labora- tory, the demibranchs were observed to differ in color. Some were whitish, some dark brown, and others were lighter brown with small white spots. Microscopic exami- nation of white demibranchs revealed only developing embryos still within the vitelline membrane. Dark brown demibranchs were filled with fully differentiated glochidia, with larval shells, adductor muscles, and glochidial hooks. Lighter brown demibranchs contained a mixture of devel- AVELAR AND MENDONCA: REPRODUCTION OF DIPLODON ROTUNDUS GRATUS 159 20am Figs. 1-4. Diplodon rotundus gratus, histological sections. 1. General aspect of gonad showing male and female follicles. 2. Spermatogenesis. 3. Male folli- cle in Follicle Stage II. 4. Male follicle in Follicle Stage IV. (ff, female follicle; mc, mother-cell; mf, male follicle; s, spermatid; S, spermatogonia; s1, first-order spermatocyte; s2, second-order spermatocyte; sm, sperm-morulae; sp, spermatozoon). oping embryos as well as fully differentiated glochidia. Examination of histological sections allowed obser- vation of marsupia with empty demibranchs, with embryos, and with fully developed glochidia. All phases were histo- logically well-defined but when the marsupium was empty, two conditions were observed: a preincubation phase char- acterized by thickening of the inner demibranch wall, and a second, post-incubation phase in which the marsupium was small, with less-thickened walls, and occasionally present- ing bubble-like empty spaces. Three developmental stages of the marsupium were defined for analytical purposes. Marsupial Stage I included empty marsupia, i.e. the absence of embryos or glochidia, including those ready to receive fertilized oocytes (preincubation phase) or those that had already eliminated glochidia (post-incubation phase). Marsupial Stage II included incubating marsupia, containing embryos in various phases of development still surrounded by vitelline membranes even when almost fully mature. Marsupial Stage HI comprised marsupia containing glochidia free from the vitelline membrane, recognized by the presence of adductor muscles, glochidial hooks, and lar- val shells (Fig. 8). REPRODUCTIVE CYCLE Monthly analysis of histological sections revealed two periods in the reproductive cycle of Diplodon rotundus gratus, one from February to July, and the other from August to January. The first period was characterized by predominance of both male and female follicles simultane- ously manifesting active oogenesis and spermatogenesis, with large numbers of sperm-morulae and nutritive cells (Fig. 5) in their respective follicles at the beginning of this period. Beginning in June, the number of functional her- maphrodites decreased gradually, with an increase in the number of functionally female individuals (Fig. 9). According to follicular development in the histolog- ical sections, marsupia were in most cases initially empty (Marsupial Stage I), with gradual growth in terms of the presence of embryos and glochidia, peaking in Marsupial Stages Ht and HI during the months of May and August (Fig. 10). When marsupial smears were analyzed, syn- chrony was noted between the stages of incubation observed in histological sections and those in the smears (Fig. 11). In April there was clear evidence of extensive elimination of both male and female gametes (Follicular Stage IIIb) (Figs. 12-14). After intense and rapid production 160 AMER. MALAC. 1004m BULL. 14(2) (1998) Figs. 5-8. Diplodon rotundus gratus, histological sections. 5. Female follicles in Follicular Stage I. 6. Female gonad in Follicular Stage IV with interfollic- ular tissue. 7. Female follicles in Follicular Stage TV showing lysis of oocytes. 8. Marsupium with glochidia. (nc, nutritive cell; pvo, previtellogenic oocyte). of gametes in February and March, with abundant sperm- morulae and nutritive cells, considerable reduction in the latter two was observed beginning in April, with a conse- quent reduction in gametogenesis. In the second period of reproduction (August to January), more and more specimens were functionally female as determined by histological examination, with the highest percentages occurring in August, September, and October (Fig. 9). In contrast, beginning in November, there was an increase in the percentage of functional hermaphro- dites accompanied by a concomitant reduction of functional females. With respect to gonadal development, beginning in August, both male and female follicles of virtually all specimens were in Follicular Stage IV. Marsupial differentiation observed in the histologi- cal sections showed that many embryos and glochidia (Marsupial Stages II and II, respectively) were present in large numbers in August and that Marsupial Stage I (post- incubation phase) began to predominate in September, with a clear decrease in inner demibranch size. During this peri- od there was also synchrony between the stages of gonadal development and marsupial development, except in October 1991 when the percentage of individuals with mar- supia containing immature glochidia reached 54% (Figs. 10-11). The number of nutritive cells present in female fol- licles decreased gradually until August and began to gradu- ally increase again beginning in September, coinciding with the spring and summer seasons of more intense rainfall. Sperm-morulae in male follicles (Figs. 2-3) were abundant in February and March, markedly decreased from April to October, and increased beginning in November, leading to maximum values in February and March (Table 1). Only small percentages of the collected specimens had non-reproductive follicles: 2% in June, 7% in July, and 0.2% in September. DISCUSSION According to the classification of Coe (1943), Diplodon rotundus gratus is a functional hermaphrodite. In the population studied here, hermaphroditism was observed AVELAR AND MENDONCA: REPRODUCTION OF DIPLODON ROTUNDUS GRATUS 161 9 100 —— Stage I --—- Stage I === Stage- I 100 90 80 70 O/ 60 %o 50 40 30 20 10 — Stage I ---Stage I -- -Stage I FMAMJJASONOD J Months Figs. 9-11. Diplodon rotundus gratus, annual percentages of individuals collected over one year (February 1987 to January 1988, Figs. 9-10, and February 1991 to January 1992, Fig. 11). 9. Percentages of hermaphrodit- ic, male, and female individuals. 10. Percentages of individuals in Marsupial Stages I-III as observed in histological sections. 11. Percentages of individuals in Marsupial Stages I-III as observed in marsu- pial smears. in practically all months of the year, with higher percent- ages at the end of summer and in the fall. In winter and spring the population consisted of individuals with mostly female follicles, followed by a new gradual increase in her- maphrodites beginning in November. In Diplodon rotundus gratus, as also in Argopecten irradians (Lamarck, 1819) (fide Coe, 1943), spermatic tis- sue occupies the dorsal portion of the gonad with ovarian tissue in the ventral portion. The two portions are not mor- phologically demarcated. Follicles simultaneously present- ing male and female elements are rare and there is no dif- ference in color between the two follicle types. Two types of spermatogenesis occur in Diplodon rotundus gratus: in the first, the mother-cell undergoes nor- 100F m 80 70 t % «. ) 50 40 : 30 : 20 ‘ 10 FMA M J 1 fine Sa it T Te \ iat T Tor T ez) 2 a oe 100 ELL LDL LILLE CA USD DGD DLS LD IMD LD LID AD LG LADD al NA 14 80h N N N SANA 60 E¥Stage I % al : N N eee I aol N : C)Stage Mo 30 N : N WStage mp 20-F N NStage ¥ 10h : : Figs. 12-14. Diplodon rotundus gratus, annual percentages of individuals collected over one year (February 1987 to January 1988) in Follicular Stages I-IV as observed in histological sections. 12. Oogenesis in her- maphroditic specimens. 13. Spermatogenesis in hermaphroditic speci- mens. 14. Oogenesis in functionally female specimens. 162 AMER. MALAC. BULL. 14(2) (1998) Table 1. Diplodon rotundus gratus, percentages of hermaphrodites plus functional males (H + M) among total individuals (N) collected with sperm-morulae present or absent, over a period of one year. (February 1987 to January 1988). Month N H+M % present % absent February 44 35 75 3 March 48 40 48 38 April 46 34 4 72 May 49 43 6 82 June 50 2h 2 54 July 46 8 4 15S August 49 ») 4 8 September 50 5 6 4 October 38 6 3 13 November 50 15 26 2 December 50 11 8 14 January 50 11 16 6 mal cell division giving rise to first- and second-order sper- matocytes, spermatids, and spermatozoa. In the second type, the mother-cell divides mitotically and the resulting cells remain joined with an intrafollicular arrangement resembling that of a morula stage (from which the term sperm-morulae was derived); the sperm-morulae, in turn, undergo meiosis, giving rise to spermatocytes, spermatids, and spermatozoa. This second type results in larger num- bers of spermatozoa, perhaps as an adaptation in response to the short reproductive period of these bivalves. The developing oocytes are connected to the follic- ular wall by a peduncle and are nourished by nutritive cells until they mature. Mature oocytes that are not released and that remain in the follicles after the elimination phase are reabsorbed by lysis from August to February. Jones et al. (1986) described a similar case in Cucumerunio novahol- landiae (Gray, 1834). Analysis of the marsupia of Diplodon rotundus gra- tus revealed the absence of an incubation period in spring or summer, confirmed by the almost total absence of embryos and glochidia. During this period, only female fol- licles exist due to total regression of male follicles. However, the marsupial smears of animals collected in 1991 showed the existence of an incubation peak in October, with about 54% of the individuals presenting embryos in the marsupium, in contrast to the marsupia observed in histological sections. Temperature could be a factor determining spawn- ing activity. Several investigators have demonstrated that the spawning activity of marine bivalves is directly related to temperature (Chipperfield, 1953; Allen, 1962; Wilson and Hodgkin, 1967; Lunetta, 1969). In Diplodon rotundus gratus, the spawning period begins in March when water temperature decreases. At the end of July, when the lowest temperatures occur, spawning virtually ceases and the ani- mals enter a process of recovery and follicular rearrange- ment that extends throughout August, September, October, and the beginning of November. Gametogenesis begins at the end of spring and con- tinues during the summer, coinciding with the period of heavy rainfall when the water volume of the river increases considerably and temperatures reach high levels. Jones et al. (1986) described a similar cycle in Australian limnic bivalves. The duration of the period of incubation in Diplodon rotundus gratus is similar to that in several unionaceans from the temperate zone. Organogenesis in D. rotundus gratus lasted about 30 days. Yokley (1972), Yager and Neves (1986), and Bruenderman and Neves (1993) also observed short periods of incubation in limnic bivalves in the Northern Hemisphere. Only one reproductive peak was observed in Diplodon rotundus gratus, in contrast to D. chilensis chilensis, in which reproduction is a continuous process (Peredo and Parada, 1986). In D. rotundus gratus, there is synchrony in reproduction and in male and female gamete maturation, as also recorded by Jones et al. (1986) and Peredo and Parada (1984, 1986). Fish captured by this study and found to be infest- ed with glochidia coincidentally occurred at the end of summer and fall, demonstrating that the species of Hyriidae occurring in the Pardo River present cycles similar to that in Diplodon rotundus gratus. Unfortunately it was not pos- sible to identify the glochidia parasitizing these fish, but more detailed studies of the life cycle of these bivalves are currently being planned. ACKNOWLEDGMENTS This research was supported by CNPq (Conselho Nacional de Desenvolvimento Cientifico e Tecnologico). We are grateful to Mr. Anténio José Colusso and Alvaro da Silva Costa for help with the field work and to Dr. Ricardo Macedo Correia e Castro for identification of the fish. LITERATURE CITED Allen, J. F. 1962. Gonad development and spawning of Brachidontes recurvus in Chesapeake Bay. The Nautilus 75:149-156. Arey, L. B. 1932. The nutrition of glochidia during metamorphosis. Journal of Morphology 53(1):201-221. Bonetto, A. A. 1954. Nayades del Rio Parana. El género Diplodon en el biotopo islefio del Parana médio e inferior. Secretaria de Agricultura, Ganaderia e Industrias, Santa Fé 62:1-56. Bonetto, A. A. 1959. Contribucién al conocimiento de las glochidias del AVELAR AND MENDONCA: REPRODUCTION OF DIPLODON ROTUNDUS GRATUS 163 género Diplodon y su aplicacion a los estudos sistematicos. Actas del Primer Congreso Sudamericano de Zoologia, Argentina 2:43- 59. Bonetto, A. A. 1961a. Investigaciones acerca de las formas larvales en el género “Diplodon” y su aplicacién a los estudios sistematicos. Direcci6n General de Recursos Naturales, Santa Fé 1961:1-48. Bonetto, A. A. 1961b. Nuevas notas sobre formas larvales de nayades Sud y Centroamericanas. Physis 21(62):332-335. Bonetto, A. A. 1962. Especies nuevas y poco conocidas de nayades del sistema del Rio de la Plata y outras cuencas proximas. Direccién General de Recursos Naturales, Santa Fé 8:213-224. Bonetto, A. A. 1964. Las especies del género Diplodon (Mollusca: Unionacea) en los Rios de la Pendiente Atlantica del sur del Brasil. Physis 24(68):323-328. Bonetto, A. A. 1965. Las especies del género Diplodon en el sistema hidrografico del Rio de la Plata (Mollusca, Unionidae). Anais do Segundo Congresso Latino-Americano de Zoologia, Sado Paulo 2:37-54. Bonetto, A. A. and I. D. Ezcurra. 1963. Notas malacologicas. Physis 24(67):17-21. Bruenderman, S. A. and R. J. Neves. 1993. Life history of the endangered fine-rayed pigtoe Fusconaia cuneolus (Bivalvia: Unionidae) in the Clinch River, Virginia. American Malacological Bulletin 10(1):83-91. Chipperfield, P. N. J. 1953. Observation on the breeding and settlement of Mytilus edulis (L.) in British waters. Journal of the Marine Biological Association of the United Kingdom 32:449-476. Chung, E. Y. 1980. Reproductive cycle and breeding season of the fresh- water clam, Anodonta (Sinanodonta) woodiana (Lea). Bulletin of the Korean Fishery Society 13(4):135-144. Coe, W. R. 1943. Sexual differentiation in mollusks. I - Pelecypods. The Quarterly Review of Biology 197:154-164. Dartnall, H. J.G. and M. Walkey. 1979. The distribution of glochidia of the swan mussel, Anodonta cygnea (Mollusca) on the three-spined stickleback Gasterosteus aculeatus (Pisces). Journal of Zoology 189:31-37. Ghosh, C. and K. C. Ghose. 1972. Reproductive system and gonadal activ- ities in Lamellidens marginalis (Simpson, 1900). The Veliger 14(3):283-288. Heard, W. H. 1975. Sexuality and other aspects of reproduction in Anodonta (Pelecypoda: Unionidae). Malacologia 15:81-103. Hebling, N. J. 1976. The functional morphology of Anodontites trapezeus (Spix) and Anodontites trapesialis (Lamarck) - Bivalvia, Mycetopodidae. Boletim de Zoologia da Universidade de Sao Paulo 15:265-298. Hebling, N. J. and A. M. G. Penteado. 1974. Anatomia funcional de Diplodon rotundus gratus Wagner, 1827 (Mollusca, Bivalvia). Revista Brasileira de Biologia 34(1):67-80. Jones, H. A., R.D. Simpson, and C. L. Humphrey. 1986. The reproductive cycles and glochidia of fresh-water mussels (Bivalvia: Hyriidae) of the Macleay River, northern New South Wales, Australia. Malacologia 27(1):185-202. Kat, P. W. 1983. Sexual selection and simultaneous hermaphroditism among the Unionidae (Bivalvia: Mollusca). Journal of Zoology, London 201:395-416. Lunetta, J. E. 1969. Fisiologia da reproducao dos mexilhdes (Mytilus perna - Mollusca, Lamellibranchia). Boletim da Faculdade de Filosofia Ciencias da Universidade de Sado Paulo, Zoologia e Biologia Marinha, N. S. 26:33-111. Mansur, M. C. D. 1970. Lista dos moluscos bivalves das famllias Hyriidae e Mycetopodidae para o Estado do Rio Grande do Sul. lheringia, Zoology 39:33-95. Mansur, M. C. D. 1973. Morfologia do sistema digestivo das espécies do género Diplodon (Spix, 1827) do Rio Guaiba, Rio Grande do Sul (Unionacea - Hyrtidae). Iheringia, Zoology 43:75-90. Mansur, M. C. D. and L. M. Anflor. 1981. Diferengas morfoldgicas entre Diplodon charruanus (Orbigny, 1835) e Diplodon pilsbryi Marshall, 1928 (Bivalvia, Hyriidae). Iheringia, Zoology 60:101- 116. Morretes, F. L. 1949. Ensaio de catalogo dos moluscos do Brasil. Arquivos do Museu Paranaense, Curitiba 7:5-216. Nagabhushanam, R. and A. L. Lohgaonker. 1978. Seasonal reproductive cycle in the mussel Lamellidens corrianus. Hydrobiologia 61:9- 14. Ortmann, A. E. 1921. South American naiades; contribution to the knowl- edge of the fresh-water mussels of South America. Memoirs of the Carnegie Museum 8(3):451-622. Parodiz, J. J. 1968. Annotated catalogue of the genus Diplodon (Unionacea - Hyriidae). Sterkiana 30:1-22. Peredo, S. and E. Parada. 1984. Gonadal organization and gametogenesis in the fresh-water mussel Diplodon chilensis chilensis (Mollusca: Bivalvia). The Veliger 27:126-133. Peredo, S. and E. Parada. 1986. Reproductive cycle in the freshwater mus- sel Diplodon chilensis chilensis (Mollusca: Bivalvia). The Veliger 28:418-425. Wilson, B. R. and E. P. Hodgkin. 1967. A comparative account of the reproductive cycles of five species of marine mussels (Bivalvia - Mytilidae) in the vicinity of Fremantle, Western Australia. Australian Journal of Marine and Freshwater Research 18:175- 203. Wood, E. M. 1974a. Development and morphology of the glochidium larva of Anodonta cygnea (Mollusca: Bivalvia). Journal of Zoology, London 173:1-13. Wood, E. M. 1974b. Some mechanisms involved in host recognition and attachment of the glochidium larva of Anodonta cygnea (Mollusca: Bivalvia). Journal of Zoology, London 173:15-30. Yager, B. L. and R. J. Neves. 1986. Reproductive cycle and fish hosts of the rabbit’s foot mussel Quadrula cylindrica strigillata (Mollusca: Unionidae) in the upper Tennessee River drainage. American Midland Naturalist 116:329-340. Yokley, P., Jr. 1972. Life history of Pleurobema cordatum (Rafinesque, 1820) (Bivalvia: Unionacea). Malacologia 11:351-364. Zumoff, C. H. 1973. The reproductive cycle of Sphaerium simile. Biological Bulletin 144:212-228. Date of manuscript acceptance: 11 August 1997 Piscine shah fi eles. - 245 OL, i : t as ae i. GE fe wpe” 4 eS: Asin all E . : af ay es erate wag RD A BS Ea ate : 1 IE yc, * le dae ; he Fn eieiencill Reet ae ass Aa i > Mh Aah ahiet nh Spe ele . eo i 4 - wasene 7 : J - E ee are Peat rec ne m. fy (Awe a. Ro i ahant ; ‘oil : agi": =3 TP dp rar GT 7 > * PWR Te) es o> (REL TY rg cae a Dy Wee: Les NAD eo Ayal ha _ eh Ul Saiki 7 ; AE wis” - 1 9b OP Sau wish ay — : vel oF Mogradakt sehen VRigha 1, cena - hited Migtheal ie baw oth uarehd ty - ‘We meine Haat aah anh ae ..* shee ‘igs, adl TH 14 5 (ht lee as howatheclty Tay MP. hoy) Aa eit . Survival and growth of juvenile freshwater mussels (Unionidae) in a recirculating aquaculture system Francis X. O’Beirn*, Richard J. Neves, and Michelle B. Steg Virginia Cooperative Fish and Wildlife Research Unit**, Department of Fisheries and Wildlife Sciences, Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24061-0321, U. S. A., mussel@vt.edu [RJN] Abstract: An indoor recirculating aquaculture system was constructed to provide suitable conditions to culture juvenile freshwater mussels. In the first of three growth trials, Villosa iris (I. Lea, 1829) juveniles were cultured for 22 wk, and grew from an initial mean length of 0.4 mm to 2.7 mm. Survival was 26.8% overall. In the second trial, growth and survival were compared between juveniles of V. iris held in sediment and without sediment. The initial mean length of both groups was 2.7 mm, and this experiment ran for 17 wk. The juvenile mussels in sediment grew to a mean length of 5.7 mm with 85% survival, significantly greater (p < 0.01) than juveniles held without sediment (4.5 mm, 74% survival). In the third trial, two cohorts of juvenile Lampsilis fasciola Rafinesque, 1815, increased in length from 1.1 mm and 1.4 mm to 3.3 mm and 4.1 mm, respectively, with comparable survival (78.7% versus 64.5%). Results of these trials demonstrate that juvenile mussels can be reared successfully within recirculating systems. One of the factors deemed important in suc- cessful culture is continuous feeding of an appropriate food source. In this study, a unialgal culture of Neochloris oleabundans Chantanachat and Bold, 1962, was used throughout. Regular cleaning of the system and water replacement also was important. Finally, the culture of juveniles in sediment appears to be an important factor in ensuring good growth and survival. This phenomenon could be related to pedal feeding behavior, proper orientation of the mussels for filtering efficiency, or stability from physical disturbance. Key words: Unionidae, freshwater mussels, Vi/losa, Lampsilis, algae, recirculating aquaculture Major declines in freshwater mussel populations 1982; Zale and Neves, 1982; Neves and Widlak, 1987; (Unionidae) were recorded by the early twentieth century Dimock and Wright, 1993). (Smith, 1899; Coker et al., 1921) and were attributed pri- Historically, reports on the aquaculture of juvenile marily to overfishing for a burgeoning pearl button industry unionids have conveyed little detailed empirical informa- (Coker, 1919). More recently, declines in unionid numbers tion. Many of the data are anecdotal, with little or no moni- have been noted and ascribed to pollution, habitat destruc- toring of the progress of the juveniles throughout the cul- tion or alteration, competition from non-indigenous bivalve ture period (Lefevre and Curtis, 1912; Coker et al., 1921; species [e. g. Asian clam, Corbicula fluminea (Miller, Howard, 1922). Culture attempts on unionids using artifi- 1774), and the zebra mussel, Dreissena polymorpha cial media have resulted in some success (Isom and (Pallas, 1771)], as well as to harvest for shell to produce Hudson, 1982; Keller and Zam, 1990). However, the appli- nuclei for pearl culture. The realization of this dramatic cation of these results to the large-scale and long-term decline by resource managers and scientists within the last propagation of unionids remains speculative and untested. two decades has prompted an increase in much needed Artificial culture of endangered and threatened mussel studies on basic life history and ecology of numerous species has been strongly recommended in recovery plans unionid species. Such studies continue as more is learned of as a Strategy to enhance declining population numbers, as environmental and host-fish requirements for survival and well as for the reintroduction of species to sites within their reproduction in this group of bivalves (Isom and Hudson, historic ranges. With these propagation needs in mind, we initiated a project focusing on factors influencing the growth and aia survival of juvenile unionids in a captive environment. To Laboratory, Wachapreague, Virginia 23480, U.S. A. ; , : me sais Meee shat this end, a recirculating system was constructed to rear Jointly supported by the Biological Resources Division-USGS, Virginia ; : a Department of Game and Inland Fisheries, and Virginia Polytechnic juvenile mussels under controlled conditions of food Institute and State University. rations, water temperature and flow, and to monitor growth *Present address: Virginia Institute of Marine Science, Eastern Shore American Malacological Bulletin, Vol. 14(2) (1998):165-171 165 166 AMER. MALAC. BULL. 14(2) (1998) and survival. This report describes feeding and mainte- nance protocols for the recirculating culture system and presents results of growth and survival trials with juvenile mussels. MATERIALS AND METHODS The recirculating culture system was located within a greenhouse facility at the Virginia Tech Aquaculture Center, Blacksburg, Virginia. The greenhouse has the capa- bility of maintaining fairly constant temperature either by the use of an evaporative cooling system or propane heaters. However, water temperatures throughout the study, while differing from ambient temperatures especially in winter months, exhibited some seasonal fluctuation. The recirculating system used for culturing juvenile mussels consisted of a rectangular reservoir tank (225 1) from which water was pumped by a 1/25 hp magnetic drive pump through half-inch PVC line, to one end of an elongat- ed raceway-type tank, which served as the primary cultur- ing chamber. The raceways were 3 m in length and 66 cm in width. Water flow into the raceway was regulated and measured using an in-line flow meter. Water was intro- duced into the raceway by ten parallel jets across the width of the chamber, which provided adequate uniformity in flow within the system. The water in the raceway was grav- ity-fed back to the reservoir via a standpipe. Water capacity in the raceway could be varied by adjusting the height of the standpipe. Aeration in the system was provided by air pumped into the reservoir and supplemented by the pump- ing action. Throughout the growth trials, the water depth in the raceway was 20 cm, giving a volume of 1701. Municipal water was conditioned (dechlorinated) by aeration for a period greater than 24 hr prior to use in the culturing system. Because the municipal water was soft (< 55 mg/l CaCO3), hardness of the water was increased by the addition of well-water (hardness of 450 mg/] CaCO3) to maintain an overall hardness in the system at ca. 200 mg/1 CaCO3. Throughout the growth trials, juvenile mussels were fed the unicellular green alga Neochloris oleoabundans Chantanachat and Bold, 1962, which has been shown to support good growth in juvenile mussels (Gatenby, 1994). Mussels were fed algae at a density of 10,000 cells/ml either once or twice daily. Ultraviolet-sterilized and dechlorinated water was used for algal cultures and the recirculating system. Living algal cultures were maintained semi-continuously in 250 | Kalwall clear plastic tubes (Kalwalls Aquacenter, Leland, Mississippi). The unialgal cultures were not axenic. Enrichment of the algal culture was achieved by the addition of appropriate quantities of nutrient media (Ukeles, 1971). We attempted to harvest algae at the late exponential growth phase, although this might not have been achieved for all feeding events. Mussel Species The growth trials described herein were conducted on common species of mussels to determine the potential success of such a system on endangered species. One surro- gate for these rare species is the rainbow mussel, Villosa iris (I. Lea, 1829), common to streams and rivers of the upper Tennessee River system. This species was selected because of its relative abundance, and the success with which large quantities of juveniles can be acquired from host fishes. Also, the habitat requirements of V. iris (riffles and shoals) are similar to those of many endangered species. In addition, two cohorts of the wavy-rayed lamp- mussel, Lampsilis fasciola Rafinesque, 1820, were utilized in growth trials. This species also is common to the upper Tennessee River system. Fish infestation and collection of juveniles were accomplished by standard techniques (Zale and Neves, 1982). Rock bass (Ambloplites rupestris Rafinesque, 1817) were infested with glochidia of V. iris, and largemouth bass (Micropterus salmoides Lacépéde, 1802), were used as the host fish for L. fasciola. Growth Trial 1 Young rainbow mussels with a mean shell length of 0.4 mm were placed in three petri dishes (18 cm x | cm) with 600-700 mussels per dish. Fine sediment (< 120 pm), collected from the Clinch River at Carbo, Virginia, and aer- ated during storage, was placed in each dish to a depth of 1- 1.5 mm. The introduction of feed to the mussels in this first trial was not regimented, as the primary purpose was to fine-tune the system, in terms of adjusting flow and ensur- ing good circulation. The flow rate of water through the raceway was 6 I/min. Growth (shell length) was assessed at 2, 4, 8, 16, and 22 wk of the trial from a sample of juveniles seived from the sediment. Fresh sediment was placed in the culture dishes after each sampling event. Water was changed in the system every 10-12 d (with one exception). This pilot study was initiated in August and terminated in December 1996. Growth Trial 2 A second trial followed the first, using those mus- sels surviving the first trial supplemented with mussels of the same age from another study. Given the relatively small number of animals held in the system, it was anticipated that water quality within the system would not be seriously compromised. However, water was changed on a weekly basis and water quality variables were more closely moni- tored in this study. On a weekly basis, temperature, pH, hardness, un-ionized ammonia, and dissolved oxygen were measured. Also, older juveniles require a greater rate of | | | | BEIRN ET AL.: UNIONIDAE GROWTH IN RECIRCULATING AQUACULTURE 167 water change (flow) than younger individuals (Jiang Li- Fan, Freshwater Fisheries Research Center, Wuxi, P. R. China, pers. comm.), therefore, flow rate across the dishes was increased to 9 I/min. Two culture techniques were evaluated within the system during this trial. The first culture method employed a sediment substratum (< 350 pm). Three replicate petri dishes contained sediment to a depth of 0.5-0.7 cm. The growth of juveniles in these dishes was compared to that of juveniles in dishes (N = 3) without substratum. The ratio- nale for this comparison was to determine whether sedi- ment was a necessary component in the survival and growth of juvenile mussels at this particular size (age) in their development. If not, then sediments could be discarded from the culture protocols, thus making maintenance and cleaning of the culture system much easier. Upon initiation of the second growth trial, 100 juveniles were placed within each petri dish. Growth (length of 30 juveniles per dish) and survival (total living count per dish) were monitored every 2 wk for the duration of the study (17 wk). Growth Trial 3 Two cohorts of juvenile Lampsilis fasciola were placed in the recirculating system to evaluate their growth and survival with the rearing protocols previously described. At the start of this trial, cohort | had a mean size of 1.1 mm. Growth and survival were estimated four times over the subsequent 12 wk. Cohort 2 had a mean size of 1.4 mm upon initiation of this trial, and these mussels were sampled only twice during the 12 wk period. RESULTS Growth Trial 1 Juvenile mussels in the three dishes, at an initial mean shell length of 0.4 mm, exhibited a steady rate of growth for the 22 wk period, achieving a mean length of 2.7 mm (Fig. 1). Survival after the first two sampling periods (4 wk) was high (approx. 95%) in all dishes. However, between the second and third sampling interval (8 wk), high mortality (60-65%) was observed in all dishes. Thereafter, mortality was low, concluding with an overall survival of 26.8%. Water temperatures in the dishes from October to December ranged from 24.5-15.4°C, with a steady decline as winter progressed. Growth Trial 2 The mean lengths of juvenile mussels at the start of growth trial were identical (mean = 2.7) for the sediment and no-sediment treatments (Fig. 2). Growth for the first two sampling periods (4 wk) was slow for both treatments; 4 i} 3 = D c 25 ® par rT) <= ” 1 4 0 T f T T T 0 5 10 15 20 25 Weeks Fig. 1. Mean shell length (mm + 2 SE) of juvenile rainbow mussels, Villosa iris, over 22 wk (Trial 1). —@— Sediment —#— No Sediment Shell Length Weeks Fig. 2. Mean shell length (mm + 2 SE) of juvenile rainbow mussels, Villosa iris, in sediment and no-sediment treatments over 17 wk (Trial 2). however, after 6 wk, a discernible difference was apparent between the two treatments. The mean size of juveniles in the sediment treatment was 3.5 mm, whereas that of juve- niles in the no-sediment treatment was 3.1 mm. This differ- ence was further exaggerated after 9 wk, when the mussels in the sediment and no-sediment treatments had mean sizes 168 AMER. MALAC. BULL. 14(2) (1998) 100 + —@®— Sediment 60 - —m— No Sediment Survival (%) Weeks Fig. 3. Mean survival of juvenile rainbow mussels, Villosa iris, in sedi- ment and no-sediment treatments over 17 wk (Trial 2). of 4.4 mm and 3.8 mm, respectively. Upon completion of the study at 17 wk, the mussels held in sediment were 5.7 mm; no-sediment mussels had attained a mean size of only 4.5 mm. Repeated measures Analysis of Variance on these data, after termination of the study, indicated that the mus- sels held in sediment were significantly larger than those held without sediment (p < 0.01). Overall, survival also dif- fered between the two treatments (Fig. 3). Those juveniles held in sediment had an overall survival of 85%, whereas the no-sediment treatment mussels exhibited significantly lower survival (74%). It was noted throughout this study that many juveniles held in sediment attached to the con- tainer surface and each other by byssal threads. This trait made separation for measuring and enumeration difficult. Conversely, those mussels cultured in dishes without sedi- ment produced no byssal threads at any period in the study. Growth Trial 3 Cohort | had grown to a mean size of 2.3 mm after 4 wk and 3.3 mm after 12 wk, when the trial was ter- minated (Fig. 4). Cohort 2 had grown to 3.9 mm after 10 wk and to 4.1 mm by 12 wk. Survival of cohort 1 and cohort 2 after 12 wk was 78.7% and 64.5%, respectively. Water Quality Water temperatures throughout the study fluctuat- ed with air temperatures and ranged between 15.4°C and 26.4°C. The overall mean water temperature in the culture system was 20.6°C. Dissolved oxygen (mg/l) and pH val- ues remained relatively stable throughout the study; both had mean values of 8.3. Un-ionized ammonia showed some variability, ranging from 0 to 0.053 mg/1, with a mean value of 0.013 mg/l. Water hardness had a mean value of 213 mg/l CaCO3, close to the targeted value of 200 mg/1 CaCO3. However, there was considerable variation in hard- ness values between samplings (range 138-277 mg/l CaCOQ3). DISCUSSION The growth of Villosa iris in Trial 1 was notable; 2.7 mm after 22 wk. This rate of growth exceeds that of 2.9 mm after 39 wk (Gatenby et al., 1996), and 1.8 mm after 20 wk under different culture conditions for the same species (Gatenby et al., 1997). The rainbow mussel is a small, slow-growing species in the Tennessee River system. When compared to other estimates of growth rates observed in the wild, the growth rate and sizes recorded in this study were considerably greater than those previously reported. Neves and Widlak (1987) estimated the mean size of juvenile freshwater mussels from a variety of species (including V. iris) as 2.7 mm after 1 yr (52 wk). Survival of our juvenile mussels also was comparable to those reported in previous studies. Gatenby ef al. (1996) reported 8% sur- vival of V. iris after 165 days, and 30-38% survival after 140 days under their best culture conditions (Gatenby et al., 1997). After 154 days, the mean survival in growth trial 1 Shell Length Weeks Fig. 4. Mean shell length (mm + 2 SE) of two cohorts of the wavy-rayed lampmussel, Lampsilis fasciola, over 12 wk (Trial 3). BEIRN ET AL.: UNIONIDAE GROWTH IN RECIRCULATING AQUACULTURE 169 was 26.8%, which included the large single mortality event that occurred between the second and third sampling peri- ods. The exact cause of this high mortality event is unknown; however, it is our opinion that the cause was poor water quality due to an oversight. Because the sched- uled water change did not occur between the second and third sampling periods, the sediments were fouled with fila- mentous algae, previously shown to result in high juvenile mussel mortality (Yang, 1996). The water quality condition in the interstitial sediments, while unknown, was also assumed to be less than optimum. Juveniles of even eury- topic species such as Utterbackia imbecillis (Say, 1829) and Pyganodon cataracta (Say, 1817) have been shown to be very sensitive to conditions of hypoxia and low pH, as well as elevated temperatures (Dimock and Wright, 1993). The performance of juvenile mussels in growth Trial 2 also proved informative. The significantly greater growth of mussels held in sediment as compared with those with no sediment was somewhat surprising. We compared growth and survival of juvenile rainbow mussels cultured with and without sediment in this trial. If comparable or better growth was achieved without the sediment, then the use of sediment could be dispensed with, which would make the maintenance and cleaning of the culture systems easier. However, this was not the case. Those animals held in sediment had greater growth and survival than those juveniles with no substratum at all. The reason for this phe- nomenon is unclear; however, we offer some possible rea- sons for the differences observed. Juvenile mussels could obtain nutrition from the sediment, to include organic mat- ter or micro-organisms. The substratum also allows juve- nile mussels to orient themselves properly for feeding. By doing so, they can siphon from higher in the water column and avoid the slower water in the boundary layer near the sediment-water interface (Vogel, 1981; Mann and Lazier, 1991). The mussels in the sediment were clearly seen to orient themselves in this manner, whereas the mussels with- out sediment lay on one valve with no vertical orientation. The sediment offers a certain amount of stability, which would buffer juvenile mussels from disturbance by turbulent water flow or vibrations originating outside the culture vessels. Those juveniles held without sediment were subjected to greater disturbance, as witnessed by the buffet- ing (from both flow and vibrations) that they received in the dishes. When disturbed in such a manner, the mussels had a tendency to cease feeding and close up. This cessa- tion of feeding could result in their slower growth and reduced survival. The response of the wavy-rayed lamp- mussel cohorts in growth trial 3 seems to support this dis- turbance hypothesis (Fig. 4). Both cohorts appeared to have similar growth rates until the first sampling of cohort 1. Thereafter, the growth of this cohort appeared to decrease dramatically with further samplings. The first sampling of cohort 2 did not occur until week 10, after which the growth rate decreased appreciably. Sampling involved removal of the mussels from water and much handling for periods in excess of 20 min. Stress associated with the han- dling of juvenile mussels could have contributed to a reduc- tion in growth. We speculate that disturbance associated with the sampling of these juveniles (measuring and count- ing) had a detrimental effect on the mussels. The juvenile mussels held without sediment would undoubtedly be exposed to greater disturbance, and this could be reflected in reduced growth rate, as we observed. Vanderploeg et al. (1995) determined that filtration rate in Lampsilis radiata siliquoidea (Barnes, 1823) was comparable between mus- sels in sediment and those lying on their sides. However, they used adult mussels and warned that such a phenome- non might not apply to all unionids. It has long been recog- nized that filtration rate in bivalves is adversely affected by mechanical or physical disturbance (Jorgensen,1960; Mohlenberg and Riisgard, 1979). The degree of stress that young mussels experi- enced in our study can also be evidenced by the fact that the mussels held in sediment produced copious amounts of byssal threads to secure themselves to the bottom of the culture dish. Conversely, the mussels held without sediment produced no byssal threads. After the study was completed, the mussels held without sediment were placed in sediment, and within | wk they produced byssal threads. We hypothe- size that the sediment provided appropriate habitat for the mussels, which consequently produced byssal threads to secure themselves within the substratum. Those held with- out sediment were in unstable habitat and did not produce byssal threads. A conclusion of our study is that young freshwater mussels, without suitable substratum, do not exhibit high growth and survival. This is supported by evi- dence from marine mussels, Mytilus spp., where production of byssal threads is mediated by the favorability of the habitat encountered (Morton, 1992), and is a result of net energy surplus (Hawkins and Bayne, 1992). The water quality parameters recorded throughout the study were not detrimental to juvenile mussels (Table 1). Temperature fluctuations, despite their range, were gradual and well within the range that juvenile mussels experience in the rivers and streams of southwestern Virginia. Hardness levels also fluctuated considerably, yet the values were such that the water was considered moder- ately hard (Landau, 1992). Dissolved oxygen and pH val- ues remained relatively stable and were not deemed stress- ful to the juveniles mussels. Un-ionized ammonia levels were considerably lower than levels reported to cause mor- tality in the marine northern quahog, Mercenaria mercenar- ia (Linné, 1758) (see Stevens, 1982). Despite the differences observed in growth Trial 2, the growth and survival in both treatments with Villosa iris 170 AMER. MALAC. BULL. 14(2) (1998) Table 1. Summary of water chemistry during juvenile mussel culture experiments. Date Temperature Dissolved Oxygen (°C) (mg/l) 10/25/96 21.0 7.8 10/31/96 24.6 7.4 11/5/96 21.7 7.9 11/8/96 22.5 79 11/14/96 15.4 9.1 11/18/96 20.0 7.6 11/22/96 20.8 8.6 11/30/96 17.4 9.3 12/14/96 19.7 9.3 12/22/96 15.7 9.6 12/28/96 23.1 8.6 1/4/97 26.4 Tes) 1/11/97 18.9 9.4 1/18/97 15.9 10.9 1/24/97 17.9 8.9 2/2/97 23.0 8.4 2/8/97 19.5, 8.7 2/16/97 17.2 9.9 2/21/97 24.6 TD 3/7/97 19.5 - 3/16/97 17.9 8.8 3/21/97 24.0 5.8 3/31/97 18.3 9.0 4/4/97 22.5 6.1 4/14/97 20.9 6.5 4/22/97 20.9 7.9 Mean + SD 20.4 + 2.9 8.3 41.2 were excellent, and matched or exceeded the rates observed in previous culture attempts (Yang, 1996; Gatenby ef al., 1997). The commercial culture of marine bivalves is con- sidered successful if a survival rate of 1-5% is realized from egg to planting size (M. Castagna, Virginia Institute of Marine Science, Eastern Shore Laboratory, pers. comm.; M. Pierson, Cherrystone Aquafarms, Cheriton, Virginia, pers. comm.). Our culture attempts compare favorably with these estimates. However, it must be noted that brooding organisms (e. g. freshwater mussels) tend to exhibit higher survival than non-brooding planktotrophic species (e. g. many marine bivalves). The growth and survival observed with the two cohorts of Lampsilis fasciola was encourag- ing, and provided evidence that the system and techniques employed in these studies can be applied successfully to other species of freshwater mussels. No single factor can be identified as causative in the success of our culture attempts. It is our opinion that the combination of the sys- tem (providing unidirectional consistent water flow with adequate dissolved oxygen levels) and the husbandry tech- niques (daily feeding and regular cleanings) were extreme- ly important for the successful culture of the juvenile unionids. It has long been appreciated that successful cul- ture of marine bivalves can only be achieved by vigilant monitoring, cleaning, and constant feeding through the lar- pH Un-ionized Hardness Ammonia (mg/1) (mg/l CaCO3) 8.5 0.013 255 8.4 0.008 138 8.4 - 160 8.2 - - 8.2 - 250 8.6 - - 8.7 0.037 235 8.6 0.022 235 8.5 0.007 205 8.8 0.000 193 8.6 0.043 243 8.9 0.053 249 8.6 0.040 206 8.4 0.000 220 8.1 0.002 201 8.4 0.001 190 8.0 0.001 201 8.5 0.008 217: 8.4 0.003 264 = _ 200 8.3 0.025 190 8.3 0.001 167 V3 0.000 144 8.0 0.001 230 7.8 0.001 261 Tea 0.000 215 8.3+0.4 0.013 + 0.0016 213 +37.9 val and juvenile phases (Castagna and Kraeuter, 1981). The ability to consistently rear large numbers of freshwater mussels, from the juvenile stage to a size suitable for relo- cation to natural environments, is a much-needed step in the conservation of rare unionids or the production of com- mercially valuable species. Results of our study demon- strate that this goal is achievable with further research into the factors influencing the growth and survival of freshwa- ter mussels within indoor culture systems. ACKNOWLEDGMENTS The authors wish to thank Braven Beaty for his extensive assis- tance in producing juvenile mussels, and Bruce Parker and Catherine Gatenby for providing the stock culture of algae. Funding for this research was provided by the States of Virginia and Tennessee, and by the U. S. Fish and Wildlife Service. LITERATURE CITED Castagna, M. and J. N. Kraeuter. 1981. Manual for Growing the Hard Clam Mercenaria. Virginia Institute of Marine Science Special BEIRN ET AL.: UNIONIDAE GROWTH IN RECIRCULATING AQUACULTURE 171 Report in Applied Marine Science And Ocean Engineering 249, Gloucester Point, Virginia. 110 pp. Coker, R. E. 1919. Fresh-water mussels and mussel industries of the United States. Bulletin of the United States Bureau of Fisheries 36: 13-89. Coker, R. E., A. F. Shira, H. W. Clark, and A. D. Howard. 1921. Natural history and propagation of freshwater mussels. Bulletin of the United States Bureau of Fisheries 37:75-181. Dimock, R. V. and A. H. Wright. 1993. Sensitivity of juvenile freshwater mussels to hypoxic, thermal and acid stress. The Journal of the Elisha Mitchell Scientific Society 109:183-192. Gatenby, C. M. 1994. Development of a Diet for Rearing Juvenile Freshwater Mussels. Masters Thesis, Virginia Polytechnic Institute and State University, Blacksburg, Virginia. 121 pp. Gatenby, C. M., R. J. Neves, and B. C. Parker. 1996. Influence of sedi- ment and algal food on cultured juvenile freshwater mussels. Journal of the North American Benthological Society 15:597-609. Gatenby, C. M., B. C. Parker, and R. J. Neves. 1997. Growth and sur- vival of juvenile rainbow mussels, Villosa iris (Lea, 1829) (Bivalvia: Unionidae) reared on algal diets and sediment. American Malacological Bulletin | 4:57-66. Hawkins, A. J. S. and B. L. Bayne. 1992. Physiological interrelations, and the regulation of production. Jn: The Mussel Mytilus: Ecology, Physiology, Genetics and Culture. Developments in Aquaculture and Fisheries Science, Vol. 25, E. Gosling, ed. pp. 171-222. Elsevier, Amsterdam. Howard, A. D. 1922. Experiments in the culture of fresh-water mussels. Bulletin of the United States Bureau of Fisheries 38:63-89. Isom, B. G. and R. G. Hudson. 1982. /n vitro culture of parasitic freshwa- ter mussel glochidia. The Nautilus 96:147-151. Jorgensen, C. B. 1960. Efficiency of particle retention and rate of water retention in undisturbed lamellibranchs. Journal du Conseils Internationale Exploration de Mer 26:94-116. Keller, A. E. and S. G. Zam. 1990. Simplification of in vitro culture tech- niques for freshwater mussels. Environmental Toxicology and Chemistry 9:1291-1296. Landau, M. 1992. Introduction to Aquaculture. John Wiley and Sons, New York. 440 pp. Lefevre, G. and W. C. Curtis. 1912. Studies on the reproduction and artifi- cial propagation of freshwater mussels. Bulletin of the United States Bureau of Fisheries 30: 105-201. Mann, K. H. and J. R. N. Lazier. 1991. Dynamics of Marine Ecosystems: Biological-Physical Interactions in the Ocean. Blackwell Scientific Publications, Cambridge. 466 pp. Mohlenberg, F. and H. U. Riisgard. 1979. Filtration rate, using a new indi- rect technique, in thirteen species of suspension-feeding bivalves. Marine Biology 54:1 43-147. Morton, B. 1992. The evolution and success of the heteromyarian form in the Mytiloidae. Jn: The Mussel Mytilus: Ecology, Physiology, Genetics and Culture, Developments in Aquaculture and Fisheries Science, Vol. 25, E. Gosling, ed. pp. 21-52. Elsevier, Amsterdam. Neves, R. J. and J. C. Widlak. 1987. Habitat ecology of juvenile freshwa- ter mussels (Bivalvia: Unionidae) in a headwater stream in Virginia. American Malacological Bulletin 5:1-7. Smith, H. M. 1899. The mussel fishery and pearl button industry of the Mississippi River. Bulletin of the United States Fish Commission 18:289-314. Stevens, F. S. 1982. Sensitivity of juvenile hard clams (Mercenaria mer- cenaria) to ammonia. Journal of Shellfish Research 2:107. Ukeles, R. 1971. Nutritional requirements in shellfish culture. /n: Artificial Propagation of Commercially Valuable Shellfish, K. S. Price and D. L. Maurer, eds. pp 43-64. University of Delaware, Newark. Vanderploeg, H. A., J. R. Liebig, and T. F. Nalepa. 1995. From picoplankton to microplankton: temperature-driven filtration by the unionid bivalve Lampsilis radiata siliquoidea in Lake St. Clair. Canadian Journal of Fisheries and Aquatic Sciences 52:63-74. Vogel, S. 1981. Life in Moving Fluids: the Physical Biology of Flow. Princeton University Press, Princeton, New Jersey. 352 pp. Yang, C. 1996. Experiments to Culture Juvenile Freshwater Mussels in Small Tanks, Floating Containers and Sediment Beds. Masters Thesis, Virginia Polytechnic Institute and State University, Blacksburg, Virginia. 97 pp. Zale, A. V. and R. J. Neves. 1982. Fish hosts of four species of lampsiline mussels (Mollusca: Unionidae) in Big Moccasin Creek, Virginia. Canadian Journal of Zoology 60:2535-2542. Date of manuscript acceptance: 16 April 1998 Fa é a “ , t “6 = a = hs = 1 aa] a yo SSS - 1 - 3 \ - tac | | = ™ i t > { = 7 a Asi eae, 8 a oat : \s) tsau@ evesad = ‘ce VA geo i a id det bal @ ~ = es ae | eee Slankyes. =a 7 = * nn ainaine ft < ; . Wh erug io § q 9 ive eat : ‘ Fa ee Silas Ss ‘ =. 2 Tye Lae ie: : © estat & ne Fy iV "Kcdadlvott 7 € olltigye rT = ibaa. ; BYS, ie j Veesitt a 1S pee lah uu s e i : : Spam let Hegel Tea! 7 ran ie Said | oe aguntiea ag Feeding interactions between native freshwater mussels (Bivalvia: Unionidae) and zebra mussels (Dreissena polymorpha) in the Ohio River Bruce C. Parker!, Matthew A. Patterson!, and Richard J. Neves2 ! Department of Biology, Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24061, U.S. A. 2 Virginia Cooperative Fish and Wildlife Research Unit, Department of Fisheries and Wildlife Sciences, Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24061, U.S. A. Abstract: The effects of zebra mussel infestation on the feeding of native unionids in the Ohio River were evaluated through gut contents and available food in the water column. In 1996, heavily infested Amblema plicata (Say, 1817) and Quadrula pustulosa (I. Lea, 1831) had significantly less (p < 0.01) organic matter in their guts (1.4 and 0.6 mg ash-free dry weight [AFDW], respectively) than lightly infested specimens (4.6 and 1.8 mg AFDW, respective- ly), and heavily infested Q. pustulosa had a significantly lower (p < 0.05) mean algal cell number (1.8 x 104) in the guts than lightly infested specimens (3.9 x 105). However, mean algal cell numbers in the guts of heavily infested and lightly infested A. plicata (5.7 x 105 versus 9.1 x 105, respectively) were not significantly different (p = 0.17). In 1997, significant reductions (p < 0.05) in total algal cells and organic matter in gut samples again occurred for heavily versus lightly infested individuals of both species. In addition, gut contents of individual A. plicata from one of two sites contained significantly less (p < 0.05) organic matter (0.92 versus 4.55 mg AFDW) and fewer algal cells (9.4 x 104 versus 2.3 x 105) than the combined gut contents of all zebra mussels (18- 33 mm in length) attached to their shells. Gut analyses also revealed significant diet overlap between native unionids and infesting zebra mussels. Water samples collected from just above the mussel beds in 1997 showed that algal densities and total suspended solids at the heavily infested site (> 360 zebra mussels/m2) were reduced by more than 50%, when compared to samples collected from the surface. Thus, competitive interactions or interference by zebra mussels likely reduced the availability of algal and detrital food resources for consumption by unionids. Key words: algae, zebra mussels, unionids, Ohio River, competition Since its introduction into Lake St. Clair, the zebra Mackie, 1994; Nalepa, 1994; Schloesser and Nalepa, mussel, Dreissena polymorpha (Pallas, 1771), has greatly 1994). By attaching to the shells of unionids, zebra mussels reduced phytoplankton and bacterioplankton levels in the can directly affect unionid survival by interfering with Great Lakes (Wu and Culver, 1991; Maclsaac et al., 1992; feeding, respiration, balance, burrowing, and locomotion Cotner et al., 1995; Fanslow et al., 1995; Heath et al., (Mackie, 1991; reviewed by Schloesser et al., 1996). Large 1995). Phytoplankton levels in Lake Erie, for example, densities of zebra mussels, however, also can affect unionid dropped 62-92% (Leach, 1993), and planktonic diatoms survival indirectly by reducing or removing food resources decreased 85% despite sufficient nutrients for growth from the water column (Lewandowski, 1976; Hebert ef al., (Holland, 1993). Consequently, Secchi disk transparencies 1991; Mackie, 1991; Haag er al., 1993). A large gill-area to in Lake Erie have increased 85% (Leach, 1993). body-dry-weight ratio, and a large number of gill cirri in Phytoplankton grazing by zebra mussels also can alter the individual zebra mussels, allow for increased filtration effi- composition of the phytoplankton community. In Lake ciency and filtration rate relative to those of native unionids Huron, for example, zebra mussel feeding has shifted domi- (Silverman et al., 1995). Filtration rates of the freshwater nance from diatoms to filamentous green algae (Lowe and mussel, Lampsilis siliquoidea (Barnes, 1823), for example, Pillsbury, 1995), and recent studies show selective rejection were found to be only one-tenth the filtration rate of indi- of the nuisance bluegreen alga Microcystis by zebra mus- vidual zebra mussels (Heath er al., 1995). In laboratory sels, such that Microcystis becomes dominant in the plank- experiments, Baker and Hornbach (1997) reported that ton (H. Vanderploeg, NOAA, pers. comm.) Amblema plicata (Say, 1817) filtered 74 ml/hr, while the 28 Zebra mussel colonization of the Great Lakes also infesting zebra mussels filtered 130 ml/hr as a group. Thus, has caused dramatic declines in the survival and fitness of relatively small populations of zebra mussels can affect the native freshwater mussel populations (Hebert ef al., 1991; feeding of unionids. Hunter and Bailey, 1992; Haag er al., 1993; Gillis and Zebra mussel populations in the lower Ohio River American Malacological Bulletin, Vol. 14(2) (1998): 173-179 173 174 AMER. MALAC. BULL. 14(2) (1998) have achieved densities comparable to those in the Great Lakes (350,000/m2; A. Miller, USACOE, pers. comm.). Because of documented impacts to the phytoplankton com- munities and native mussel populations of the Great Lakes, large populations of zebra mussels in the Ohio River could have similar consequences for native mussel populations. Strayer and Smith (1996) found that low zebra mussel infestation rates in the Hudson River were associated with high unionid mortality and hypothesized that reduced food resources might be the cause. No studies, however, have directly confirmed whether zebra mussels affect the feeding of unionids in a riverine environment where organic materi- als are continually supplied from upstream. Thus, the objective of this study was to determine whether zebra mussels reduce unionid ingestion of phytoplankton and organic matter by (1) ingesting similar food resources, and (2) reducing food resources at the sediment-water interface. METHODOLOGY On 23 July 1996, ten specimens each of the three- ridge, Amblema plicata, and the pimpleback, Quadrula pus- tulosa (1. Lea, 1831), were collected from a lightly infested site on the Ohio River near Parkersburg, West Virginia, which had a mean density of 0.3 zebra mussels/m2, and a maximum of one zebra mussel/unionid (P. Morrison, USFWS, pers. comm.). On 16 August 1996, ten specimens of A. plicata were collected from a heavily infested site near Paducah, Kentucky, which had 3,600 zebra mussels/m2 (A. Miller, USACOE, pers. comm.). Specimens of Q. pustu- losa were difficult to find at this site, soon 21 August 1996, ten specimens were collected from another heavily infested site near Maysville, Kentucky, which had 360 zebra mus- sels/m2 and a maximum of 92 zebra mussels/unionid (P. Morrison, USFWS, pers. comm.). In the field, mussel bod- ies were removed from shells, weighed, preserved in 95% ethanol, and transported to the laboratory for analysis. In 1997, ten specimens each of A. plicata and Q. pustulosa were collected from a highly infested (370 zebra mussels/m2) and a lightly infested (< 1 zebra mussel/m2) site on the Ohio River for gut content analysis. In addition, all zebra mussels, 18-33 mm in length, attached to the shells of A. plicata were removed and preserved in 95% ethanol for gut content analysis. Zebra mussels 18-33 mm in length were chosen, because it is difficult to remove the entire gut contents of smaller individuals. At each collec- tion site, water samples with algae were collected from the the surface and overlying the mussel bed, fixed with acid Lugol’s solution (Saraceni and Ruggiu, 1969), and placed in settling chambers to compare the density and relative abundance of algal genera using inverted microscopes. Aliquots of 100 ml were then filtered through pre-ashed Whatman GF‘F filters, dried (100°C), and ashed (500°C) to determine the ash-free dry weight (AFDW) of seston. Gut contents of unionids and of zebra mussels attached to Amblema plicata were individually removed from each specimen, pooled, then suspended in 3 ml of water, and fixed with 50 pl of acid Lugol’s solution for analysis. A 50 pl aliquot of the gut contents was placed on a microscope slide. Ocular grids divided the field of view into 59 transects, and algal cells were counted and identi- fied to genus from two transects using an Ausjena/Nomarsky microscope at 400X. The variability of this semi-quantitative method (+ 20%, a = 0.05) was deter- mined using ten counts from the same sample. The remain- ing gut contents were collected on pre-ashed Whatman GF/F filters, dried (100°C), and ashed in a muffle furnace (500°C) overnight to determine AFDW. Mean algal cell numbers and mean AFDW values in the gut samples of each species were compared by ANOVA. RESULTS In 1996, significant differences in total algal cells and organic matter were observed in guts of lightly and heavily infested unionids (Table 1). While mean algal cell numbers in guts of lightly and heavily infested Amblema plicata (5.7 x 105 versus 9.1 x 105 cells) were not signifi- cantly different (p = 0.17), the gut contents of lightly infest- ed A. plicata had significantly more (p < 0.01) organic mat- ter (4.6 mg AFDW) than heavily infested specimens (1.4 mg AFDW). Heavily infested Quadrula pustulosa showed significantly lower (p < 0.05) organic matter and mean algal cell number (0.6 mg AFDW and 1.8 x 104 cells, respectively) than lightly infested specimens (1.8 mg AFDW and 3.9 x 105 cells, respectively). In 1997, significant reductions in organic matter content and total algal cells also were observed in guts of heavily infested unionids (Table 2). Organic matter content and total algal cells were significantly lower (p < 0.05) in the guts of heavily infested Amblema plicata (0.9 mg Table 1. Mean algal cell number and ash-free dry weight (AFDW; + SD) in guts of Amblema plicata and Quadrula pustulosa heavily infested (H) and lightly infested (L) with zebra mussels, July-August 1996. Species N Algal Cell Number AFDW (mg) A. plicata (L) 11 9.1 x 105 +6.0 x 105 4.6+0.9 A. plicata (H) 10 5.7 x 105+ 4.9 x 105 1.4+0.7 Q. pustulosa (L) 10 3.9 x 105+2.8 x 105 1.8+1.0 Q. pustulosa (H) 10 1.8 x 104+9.2 x 103 0.6 + 0.3 PARKER ET AL.: FEEDING INTERACTIONS 175 Table 2. Mean algal cell number and ash-free dry weight (AFDW, + SD) in guts of Dreissena polymorpha and of Amblema plicata and Quadrula pustulosa heavily infested (H) and lightly infested (L) with zebra mussels, July-August 1997. Species N Algal Cell Number AFDW (mg) A. plicata (L) 10 5.5 x 106+2.4x 106 5.1 41.7 A. plicata (H) 10 9.4x 104+7.6x 104 0.9+0.8 Q. pustulosa (L) 10 1.9x 106+ 1.3 x 106 2.0+1.2 Q. pustulosa (H) 9 6.9 x 104+7.3 x 104 0.3 +0.2 D. polymorpha 9 2.3 x 105+ 1.3 x 105 46+3.6 AFDW and 9.4 x 104 cells, respectively) than lightly infest- ed specimens (5.1 mg AFDW and 5.5 x 106 cells, respec- tively). Significant reductions in both organic matter con- tent and total algal cells also were observed for heavily infested (0.3 mg AFDW and 6.9 x 104 cells, respectively) versus lightly infested (2.0 mg AFDW and 1.9 x 106 cells, respectively) Quadrula pustulosa. Examination of unionid guts in 1996 and 1997 indi- cated that Amblema plicata and Quadrula pustulosa readily ingested a significant amount of detritus (ca. 90%) along with algal cells between 4 and 80 pm in length. Diatoms (Bacillariophyta) and green algae (Chlorophyta) dominated gut samples, and the dominant algal genera in 1996 (Chlorella, Cyclotella, Navicula, Melosira, and Scenedesmus; Table 3) and 1997 (Chlorella, Cyclotella, Mougeotia, Melosira, and Scenedesmus; Table 4) were sim- ilar. Usually, the relative abundance of algal genera within unionid guts was very similar to relative abundances of algae in water samples collected from the river bottom (Table 5). Interestingly, the pennate diatom Synedra domi- nated water samples at the lightly infested site (43%), but few if any of these > 100 pm-long cells were found in the unionid guts. In 1997, algal cell densities and seston AFDW at the water surface (8.37 x 104 cells/ml and 9.0 mg/I, respectively) and above the mussel bed (8.42 x 104 cells/ml and 9.0 mg/l, respectively) were nearly identical at the lightly infested site (Table 5). Seston AFDW (5.0 mg/l) and algal cell densities (2.2 x 104 cells/ml) overlying the mussel bed at the heavily infested site, however, were greatly reduced compared to samples taken from the surface (7.2 mg/l and 8.3 x 104 cells/ml, respectively). In 1997, individual Amblema plicata collected from the heavily infested site were infested with > 100 zebra mussels/unionid, averaging 50 zebra mussels between 18 and 33 mm in length. Gut contents of zebra mussels con- tained large amounts of detritus. Pooled samples of zebra mussel guts contained five times more (p < 0.05) organic matter (4.55 mg versus 0.92 mg, respectively) and twice as many algal cells (2.3 x 105 cells versus 9.4 x 104 cells, respectively) than the specimen of A. plicata to which they were attached (Table 3). In addition, the dominant algal genera — Chlorella, Cyclotella, Mougeotia, Melosira, and Scenedesmus — and range of cell sizes (4-70 ppm) in zebra mussel guts were nearly identical to those in infested A. pli- cata (Table 4). Relative abundances of algal genera within zebra mussel guts were similar to relative abundances in water samples collected from the river bottom (Table 5). DISCUSSION Reduced food resources at the sediment-water inter- face can cause decreased growth rates in bivalves despite adequate food resources at the water surface (Frechette and Bourget, 1985). This phenomenon could be particularly important for unionids in the lower Ohio River because zebra mussels occur in large densities at the sediment-water interface and often attach directly to the shells of unionids. In fact, water samples collected from just above the mussel beds in 1997 showed that algal densities and AFDW at the heavily infested site (> 360 zebra mussels/m2) were reduced by more than 50% when compared to samples col- lected from the surface. Thus, it appears that zebra mussel densities in heavily infested sections of the Ohio River con- tribute to reduced food resources at the sediment-water interface. Regardless of reductions in total algal cell densities, the effects of zebra mussel infestation on unionid ingestion can be reduced if zebra mussels and unionids selectively feed on different food types. Currently, studies on selective feeding in bivalves appear to be inconclusive. Some authors have concluded that bivalves select food particles of high quality (Allen, 1914; Loosanoff and Engel, 1947; Shumway et al., 1985), while others have concluded that feeding is non-selective (Churchill and Lewis, 1924; Gale and Lowe, 1971; Bayne et al., 1976). Thus, selective feeding could be species dependent. Zebra mussels have been reported to ingest a wide range of food particles between 0.7 and 450 pm (Mikheyev, 1967; Jorgensen et al., 1984). However, Sprung and Rose (1988) indicated that retention efficiency in zebra mussels is maximized for food particles between 5- 35 pm; Ten Winkle and Davids (1982) also concluded from gut content analysis that zebra mussels select particles between 15-50 pm. In our study, zebra mussel gut samples contained food particles between 4-70 pm in maximum dimension. In comparison, Miura and Yamashiro (1990) indicated that the unionid Anodonta calipygos (Kobelt, 1879) ingests food particles between 0.5 and 100 pm. Like for the zebra mussel, maximum retention efficiencies were 176 AMER. MALAC. BULL. 14(2) (1998) Table 3. Percent relative abundances (SD) of algae in guts of lightly (L) and heavily (H) infested Amblema plicata and Quadrula pustulosa collected from the Ohio River, July-August 1996. (+, presence [< 2%]; -, absence). Algae A. plicata (L) Chlorophyta Ankistrodesmus - Chlamydomonas + Chlorella 15.3 (8.7) Chlorococcum + Closterium - Coelastrum _ Cosmarium - Gonium - Mougeotia + Oedogonium - Oocystis + Pediastrum - Scenedesmus 6.2 (5.9) Schroederia + Selenastrum - Spirogyra + Staurastrum + Bacillariophyta Achnanthes + Cocconeis + Coscinodiscus + Cyclotella 11.8 ( Cymbella + Diatoma + Fragilaria - Gomphonema + Melosira 5.7 (4.0) Navicula 26.5 (7.4) Nitzschia = Pinnularia + Pleurosigma = Stephanodiscus - Surirella Synedra Tabellaria +++ Cyanoprokaryota Chroococcus + Merismopedia - Oscillatoria + Spirulina ~ Other Chromulina - Dinobryon + Peridinium ~ Chroomonas + found for intermediate-sized particles (5-30 pm). Maximum filtration rates for the unionid Elliptio com- planata (Lightfoot, 1786) also were found for particles between 4 and 5 pm (Paterson, 1984). In our study, unionid gut analyses indicate that unionids ingest food particles between 4 and 80 pm in maximum dimension. Thus, zebra mussels and unionids in the Ohio River ingest food A. plicata Q. pustulosa Q. pustulosa (H) (L) (H) = + = = + = 15.0 (11.2) 42.3 (18.8) 63.0 (29.4) + + = = + = = + = = + = + —_ — + + = + — — + _ — 5.2 (4.7) 6.6 (5.8) z = + = = + = = + = + + = 45.1 (11.1) 7.0 (3.8) 31.9 (30.8) + + 7 = + = = + = — + = 16.0 (11.0) 8.8 (8.2) 2.8 (6.4) 2.2 (0.6) 10.6 (6.2) = + a + + = ‘one + — + = — + + = + + = + — = —t + pay iow. + = + ie a + + = resources of similar size. Relative abundances of algal genera ingested by zebra mussels and the two unionid species were very simi- lar to those in water samples collected immediately above the mussel bed, giving no evidence of selective feeding. The only exception was the pennate diatom Synedra which was not readily ingested by unionids or zebra mussels, PARKER ET AL.: FEEDING INTERACTIONS 177 Table 4. Percent relative abundances (SD) of algae in guts of Dreissena polymorpha and lightly (L) and heavily (H) infested Amblema plicata and Quadrula pustulosa collected from the Ohio River, July-August 1997. (+, presence [S$ 2%]; —, absence). Q. pustulosa Q. pustulosa D. polymorpha (L) (H) Algae A. plicata A. plicata (L) (H) Chlorophyta Ankistrodesmus = - Chlamydomonas - + Chlorella 1.7 (1.4) 13.8 (10.1) Chlorococcum + + Cosmarium - - Crucigenia - - Gonium ~ - Mougeotia 40.6 (7.5) 5.1 (7.0) Oedogonium + - Pandorina + - Pediastrum + + Scenedesmus 9.5 (3.6) 5.7 (2.0) Trebouxia + = Bacillariophyta Coscinodiscus - - Cocconeis - = Cyclotella 21.3 (11.9) 38.5 (18.6) Diatoma + - Gomphonema + = Melosira 13.7 (9.2) 25.9 (16.7) Navicula + + Stephanodiscus + ~ Synedra - - Tabellaria - = Cyanoprokaryota Aphanocapsa - - Chroococcus + - Oscillatoria - - Other Peridinium - = probably because of its cell length of > 100 pm. In unionid gut samples, Mougeotia, Cyclotella, and Melosira did appear in slightly greater abundance than in water samples. Regardless of selective feeding, unionid and zebra mussel gut contents contained a nearly identical assemblage of dominant algal genera. Thus, by ingesting food particles of similar size and type, zebra mussels in the Ohio River can compete directly with native unionids for food resources. However, competition can only be confirmed if the food items measured to assess diet overlap constitute a signifi- cant portion of the total diet of one of the potential com- petitors (Buss and Jackson, 1981). Unfortunately, there are no studies on the dietary or nutritional requirements of freshwater mussels. Significant reductions in the mean AFDW and total algal cell number from gut samples of heavily infested versus lightly infested native freshwater mussels indicate that interference competition for food resources is occur- ring in the lower Ohio River. Since the arrival of the zebra + = = + + + 9.6 (8.8) 30.8 (24.7) 37.5 (16.8) + + + + _— — + = 4 + — _ 21.2 (11.7) 5.3 (9.3) 14.2 (12.2) + — + + 8.2 (5.9) 10.9 (9.6) 13.9 (10.7) + = = Be = = 25.9 (6.9) 16.2 (12.7) 5.4 (4.5) Bs = = 17.6 (9.4) 28.5 (18.3) 13.7 (10.4) + - + + = = + mae _— + = + + rs = + = = + — _ + a = mussel, unionid mortality thus far at various sites in the lower Ohio River has been estimated at 20-40% (P. Morrison, USFWS, pers. comm.). Interference with unionid feeding by zebra mussels, however, also could have long term consequences for the overall fitness of native mussels. Recent studies have shown that native unionids from the heavily infested Ohio River have significantly reduced glycogen levels relative to unionids from lightly infested areas (Patterson et al., 1997). Glycogen is an important energy reserve for animals, especially bivalves (de Zwann and Zandee, 1972; Barber and Blake, 1981; Bayne and Newell, 1983; Haag et al., 1993), and significant reductions can lead to chronic mortality or declines in reproductive success. Gonad development in marine bivalves, for exam- ple, has been shown to continue despite reduced energy reserves (Gabbott and Bayne, 1973; Bayne, 1975), but sub- sequent growth rates and energy reserves of the developing larvae decreased (Bayne, 1972; Helm et al., 1973; Bayne et al., 1975). Thus, by reducing food resources and interfering 178 AMER. MALAC. BULL. 14(2) (1998) Table 5. Percent relative abundances of algae at the surface and just above the mussel bed in the heavily (H) and lightly infested (L) Ohio River, July- August 1997. (+, presence [< 2%]; —, absence). Algae Surface Bottom Surface Bottom (L) (L) (H) (H) Chlorophyta Ankistrodesmus - + = = Chlamydomonas 6 5 16 11 Chlorella 10 8 25 23 Chlorococcum + 2 + 7 Gonium — + _ = Mougeotia 8 2 6 5 Oedogonium + + - + Pediastrum + + = + Scenedesmus 12 16 13 12 Staurastrum + + + - Bacilariophyta Coscinodiscus - + + + Cyclotella 4 5 3 13 Diatoma + + = + Melosira + 2 4 =) Navicula - + - + Stephanodiscus + + - + Synedra 12 43 11 10 Cyanoprokaryota Chroococcus - 2 + a Merismopedia 8 7 - = Oscillatoria + 6 3 3 Other Mallomonas — - - + with normal feeding, zebra mussels could have their great- est effect on the long-term persistence of unionid beds in the lower Ohio River through reduction in reproductive success and recruitment. ACKNOWLEDGMENTS We thank Patty Morrison, Mitch Ellis, and others at the Ohio River Islands National Wildlife Refuge, as well as Dr. Andrew Miller and the United States Army Corps of Engineers for help in collecting mussels from the Ohio River. Thanks also go to Catherine Gatenby and the many volunteers who assisted with collection and processing of mussels in the field. Additionally, we thank Ashleigh Funk, Lana Shurts, and Doug Smith for laboratory assistance. This study was funded by Quick Response Funds of the Biological Resources Division, U. S. Geological Survey. LITERATURE CITED Allen, W. R. 1914. The food and feeding habits of freshwater mussels. Biological Bulletin 27: 127-147. Baker, S. M. and D. J. Hornbach. 1997. Acute physiological effects of zebra mussel (Dreissena polymorpha) infestation on two unionid mussels, Actinonaias ligamentina and Amblema plicata. Canadian Journal of Fisheries and Aquatic Sciences 54:512-519. Barber, B. J. and N. B. Blake. 1981. Energy storage and utilization in rela- tion to gametogenesis in Argopecten irridians concentricus (Say). Journal of Experimental Marine Biology and Ecology 52:121- 134. Bayne, B. L. 1972. Some effects of stress in the adult on the larval devel- opment of Mytilus edulis. Nature 237:459. Bayne, B. L. 1975. Reproduction in bivalve mollusks under stress. In: The Physiological Ecology of Estuarine Organisms, J. Vernberg, ed. pp. 259-277. University of South Carolina Press, Columbia, South Carolina. Bayne, B. L. and R. C. Newell. 1983. Physiological energetics of marine molluscs. In: The Mollusca, Vol. 4, Physiology, Part 1, A. S. M. Saleuddin and K. W. Wilbur, eds. pp. 407-515. Academic Press, New York. Bayne, B. L., P. A. Gabbott, and J. Widdows. 1975. Some effects of stress in the adult on the eggs and larvae of Mytilus edulis. Journal of the Marine Biological Association of the United Kingdom 55:675- 689. Bayne, B. L., R. J. Thompson, and J. Widdows. 1976. Physiology I. In: Marine Mussels, Their Ecology and Physiology, B. L. Bayne, ed. pp. 121-206. Cambridge University Press, Cambridge. Buss, L. W. and J. B. C. Jackson. 1981. Planktonic food availability and suspension-feeder abundance: evidence of in situ depletion. Journal of Experimental Marine Biology and Ecology 49:151- 161. Churchill, E. P. and S. I. Lewis. 1924. Food and feeding in freshwater mussels. Bulletin of the United States Bureau of Fisheries 39:439- 471. Cotner, J. B., W. S. Gardner, J. R. Johnson, R. H. Sada, J. F. Cavaletto, and R. T. Heath. 1995. Effects of zebra mussels (Dreissena poly- morpha) on bacterioplankton: evidence for both size-selective consumption and growth stimulation. Journal of Great Lakes Research 21(4):517-528. de Zwann, A. and D. I. Zandee. 1972. Body distribution and seasonal changes in the glycogen content of the common sea mussel Mytilus edulis. Comparative Biochemistry and Physiology 43A:53-58. Fanslow, D. L., T. F. Nalepa, and G. A. Lang. 1995. Filtration rates of the zebra mussel (Dreissena polymorpha) on natural seston from Saginaw Bay, Lake Huron. Journal of Great Lakes Research 21(4):489-500. Frechette, M. and E. Bourget. 1985. Food limited growth of Mytilus edulis L. in relation to the benthic boundary layer. Canadian Journal of Fisheries and Aquatic Sciences 42:1 166-1170. Gabbott, P. A. and B. L. Bayne. 1973. Biochemical effects of temperature and nutritive stress on Mytilus edulis. Journal of the Marine Biological Association of the United Kingdom 53:269-286. Gale, W. F. and R. L. Lowe. 1971. Phytoplankton ingestion by the finger- nail clam, Sphaerium transversum (Say), in pool 19, Mississippi River. Ecology 52(3):507-513. Gillis, P. L. and G. L. Mackie. 1994. Impact of the zebra mussel, Dreissena polymorpha, on populations of Unionidae (Bivalvia) in Lake St. Clair. Canadian Journal of Zoology 72:1260-1271. Haag, W. R., D. L. Berg, D. W. Garton, and J. L. Farris. 1993. Reduced survival and fitness in native bivalves in response to fouling by the introduced zebra mussel (Dreissena polymorpha) in western Lake Erie. Canadian Journal of Fisheries and Aquatic Sciences 50:13-19. Heath, R. T., G. L. Fahnenstiel, W. S. Gardner, J. F. Cavaletto, and S. Hwang. 1995. Ecosystem-level effects of zebra mussels PARKER ET AL.: FEEDING INTERACTIONS Ie) (Dreissena polymorpha): an enclosure experiment in Saginaw Bay, Lake Huron. Journal of Great Lakes Research 21(4):501- 516. Hebert, P. D., C. C. Wilson, M. H. Murdoch, and R. Lazar. 1991. Demography and ecological impacts of the invading mollusc Dreissena polymorpha. Canadian Journal of Zoology 69:405-409. Helm, M. M., D. L. Holland, and R. R. Stephenson. 1973. The effect of supplementary algal feeding of a hatchery breeding stock of Ostrea edulis on larval vigour. Journal of the Marine Biological Association of the United Kingdom 53:673-684. Holland, R. E. 1993. Changes in planktonic diatoms and water transparen- cy in Hatchery Bay, Bass Island area, western Lake Erie, since the establishment of the zebra mussel. Journal of Great Lakes Research 19(3):617-624. Hunter, R. D. and J. F. Bailey. 1992. Dreissena polymorpha (zebra mus- sel): colonization of soft substrata and some effects on unionid bivalves. The Nautilus 106(2):60-67. Jorgensen, C. B., T. Kiorboe, F. Mohlenberg, and H. U. Riisgard. 1984. Ciliary and mucus-net filter feeding, with special reference to fluid mechanical characteristics. Marine Ecology Progress Series 15:283-292. Leach, J. H. 1993. Impacts of the zebra mussel (Dreissena polymorpha) on water quality and fish spawning reefs in western Lake Erie. /n: Zebra mussels: Biology, Impacts, and Control, T. F. Nalepa and D. W. Schloesser, eds. pp. 381-397. Lewis Publishers, Boca Raton, Florida. Lewandowski, K. 1976. Unionidae as a substratum for Dreissena poly- morpha Pall. Polskie Archiwum Hydrobiologie 23(3):409-420. Loosanoff, V. L. and J. B. Engel. 1947. Effect of different concentrations of micro-organisms on feeding of oysters (O. virginica). United States Fish and Wildlife Service Fisheries Bulletin 51:31-57. Lowe, R. L. and R. W. Pillsbury. 1995. Shifts in the benthic algal commu- nity structure and function following the appearance of zebra mussels (Dreissena polymorpha) in Saginaw Bay, Lake Huron. Journal of Great Lakes Research 21(4):558-566. Maclsaac, H. J., W. G. Sprules, O. E. Johannsson, and J. H. Leach. 1992. Filtering impacts of larval and sessile zebra mussels (Dreissena polymorpha) in western Lake Erie. Oecologia 92:30-39. Mackie, G. L. 1991. Biology of the exotic zebra mussel, Dreissena poly- morpha, in relation to native bivalves and its potential impact in Lake St. Clair. Hydrobiologia 219:25 1-268. Mikheyev, V. P. 1967. [Filtration nutrition of the Dreissena.] (In Russian). Trudy vsesoyuzogo Nauchno-issledovatelnogo Instituta Prudovogo Rybnogo Khoziaistva 15:117-129. [not seen] Miura, T. and T. Yamashiro. 1990. Size selective feeding of Anodonta calipygos, a phytoplanktivorous freshwater bivalve, and viability of egested algae. Japanese Journal of Limnology 51(2):73-78. Nalepa, T. F. 1994. Decline of native unionid bivalves in Lake St. Clair after infestation by the zebra mussel, Dreissena polymorpha. Canadian Journal of Fisheries and Aquatic Sciences 5\:2227- 2233. Paterson, C. G. 1984. A technique for determining apparent selective fil- tration in the fresh-water bivalve Elliptio complanata (Lightfoot). The Veliger 27(2):238-241. Patterson, M. A., B. C. Parker, and R. J. Neves. 1997. Effects of quaran- tine times on glycogen levels of native freshwater mussels (Bivalvia: Unionidae) previously infested with zebra mussels. American Malacological Bulletin 14(1):75-79. Saraceni, C. and D. Ruggiu. 1969. Techniques for sampling water and phytoplankton. In: A Manual on Methods for Measuring Primary Production in Aquatic Environments, R. A. Vollenweider, ed. pp. 5-7. Blackwell Scientific Publications, Oxford. Schloesser, D. W. and T. F. Nalepa. 1994. Dramatic decline of unionid bivalves in offshore waters of western Lake Erie after infestation by the zebra mussel, Dreissena polymorpha. Canadian Journal of Fisheries and Aquatic Sciences 51:2234-2242. Schloesser, D. W., T. F. Nalepa, and G. L. Mackie. 1996. Zebra mussel infestation of unionid bivalves (Unionidae) in North America. American Zoologist 36:300-310. Shumway, S. E., T. L. Cucci, R. C. Newell, and C. M. Yentsch. 1985. Particle selection, ingestion, and absorption in filter-feeding bivalves. Journal of Experimental Marine Biology and Ecology 91:77-92. Silverman, H., E. C. Achberger, J. W. Lynn, and T. H. Dietz. 1995. Filtration and utilization of laboratory-cultured bacteria by Dreissena polymorpha, Corbicula fluminea, and Carunculina tex- asensis. Biological Bulletin 189:308-319. Sprung, M. and U. Rose. 1988. Influence of food size and food quantity on the feeding of the mussel Dreissena polymorpha. Oecologia 77:526-532. Strayer, D. L. and L. C. Smith. 1996. Relationships between zebra mussels (Dreissena polymorpha) and unionid clams during early stages of the zebra mussel invasion of the Hudson River. Freshwater Biology 36(3):771-779. Ten Winkle, E. H. and C. Davids. 1982. Food selection by Dreissena poly- morpha Pallas (Mollusca: Bivalvia). Freshwater Biology 12:553- 558. Wu, L. and D. A. Culver. 1991. Zooplankton grazing and phytoplankton abundance: an assessment before and after invasion of Dreissena polymorpha. Journal of Great Lakes Research \7 (4):425-436. Date of manuscript acceptance: 6 March 1998 7 > Aa i mt De we aa | ‘ — \ 7 *. t “nd eas : ae f os eg US 0) Kye Whee (Ue i Sow aa a - — Voie aio Ce he } 2 at ; ee : : oh Aone ee perry. Gor! a f — H ae oe ‘ - ce ¥ 2 w - & Gio Wierik: jae y : as , 7 a 1 { | " Research Note: A nondestructive method for cleaning gastropod radulae from frozen, alcohol-fixed, or dried material Wallace E. Holznagel Aquatic Biology Program, Department of Biological Sciences, University of Alabama, Box 870345, Tuscaloosa, Alabama 35487-0345 U.S.A. Abstract: A new method for cleaning tissue from gastropod radulae is presented that is non-destructive compared to the methods presently in use. Present methods are time sensitive because the caustic solutions used can damage radula. This new method uses a detergent and Proteinase K enzyme which are not time sensitive. This method will work with frozen, alcohol-fixed, or dried material but not with formalin-fixed material. In addition, the tissue super- natant can be used for subsequent standard DNA-extraction techniques. Key words: gastropod, snail, radula, cleaning method, preparation The present methods for cleaning gastropod radulae for viewing use either sodium hypochlorite (NaOCl) or potassium hydroxide (KOH). Because of the caustic nature of these chemicals, careful monitoring of the specimen is required to avoid damaged or dissolved radulae. An alter- native method routinely used to digest tissue for DNA extraction (Sambrook ef al., 1989) has proven to be as effective in cleaning radulae and avoids the likelihood of unwanted erosion of the specimens. This method is effec- tive for preparing radulae from frozen, alcohol-fixed, newly dried, and from dried specimens collected over 30 years ago. This method has not proven to be useful for cleaning radulae from specimens preserved in formalin because of the protein cross-linking that occurs with this method of fixation. MATERIALS The necessary equipment and stock solutions required are listed in Table 1. In addition, the following lab equipment is needed: dissecting microscope, microforceps, scalpel, and pipettes. Additional useful tools include: insect pins (very good probes) and sharpened applicator sticks (excellent tools for manipulating the radula during mount- ing). Cleaned radulae for scanning electron microscopy (Fig. 1) were sputter-coated with gold-palladium and scanned on a Hitachi S-2500 scanning electron microscope at the University of Alabama. THE METHOD With freshly frozen material the gastropod is gener- ally partially extended so that it is a simple matter to extract the animal from its shell by grasping the foot and body behind the head and gently pulling. Typically in alcohol- fixed and/or dried material, the animal has withdrawn into the shell. Removal of the snail body from the shell can be accomplished by carefully cracking the shell. After separat- ing the head from the rest of the animal, rinse the head with water to remove any extraneous material that might have adhered to it (i. e. shell fragments). Place each specimen in an individual 1.5 ml microcentrifuge tube and add 500 pl NET buffer and 10 pl Proteinase K (Table 1). Close the tube and tape it to the mixer platform, and place the mixer an incubator at 37°C. With fresh or alcohol-fixed material, cleaning should take about 2-3 hr. Dried tissue specimens can require 2-4 d, depending on the amount of starting material. For dry material it may be necessary to replace the NET buffer and Proteinase K after 24-48 hr. Once cleaning is complete, as determined by examination under a dissecting microscope, remove the radula and rinse in sev- eral changes of deionized water to removed all traces of the NET buffer. The remaining supernatant can be used to obtain DNA using DNA standard phenol/chloroform extraction techniques. The radula is then stored in 25% ethanol until ready for mounting. At this point the clean radula can be mounted by a procedure appropriate for the type of visualization to be used. Some cleaned radulae have American Malacological Bulletin, Vol. 14(2) (1998): 181-183 18] AMER. MALAC. BULL. 14(2) (1998) j “ ' a f Ne q Fig. 1. Examples of radulae cleaned by the described method. A-B. Jo fluvialis (Say, 1825). C-D. Leptoxis (Athearnia) crassa anthonyi (Redfield. 1854). E-F. Terebia granifera (Lamarck, 1822). G-H. Gyrotoma pyramidatum Shuttleworth, 1845. The radulae were prepared from the following types of tissue: fresh frozen (A-B); alcohol fixed (C-D); alcohol fixed and dried; (E-F); 30-year old unfixed dried material (G-H). Figs. A, C, E, and G show the central and left lateral tooth; B, D, and H show the left pair of inner and outer marginal teeth; and F shows the right pair of inner and outer marginal teeth. Scale bars = 50 um. HOLZNAGEL: METHOD FOR CLEANING RADULAE 183 Table 1. Equipment and stock solutions. Incubator Platform mixer 1.5 ml microcentrifuge tubes NET buffer 1 ml Tris pH 8.0 2 ml 0.5 M ethylene diamine tetraacetic acid (EDTA) Iml 5 M NaCl 20 ml 10% sodium dodecyl] sulfate (SDS) (Sigma-Aldrich Chemical Company no. L4522) 76 ml deionized water Proteinase K stock* 20 mg Proteinase K* (Sigma-Aldrich Chemical Company no. P4914) 1 ml deionized water * store at -20 °C been stored in 25% ethanol for up to 6 mo before mounting and no degradation was observed. The radula and the radular ribbon do not appear to be affected by the enzymatic action of Proteinase K as evi- denced by scanning electron micrographs (Fig. 1). Therefore in this cleaning method careful monitoring is not necessary. An additional benefit of this procedure is that the radular ribbon from dried material is rehydrated and becomes flexi- ble enough to undergo the rigors of mounting. LITERATURE CITED Sambrook, J., E. F. Fritch, and T. Maniatis. 1989. Molecular Cloning, a Laboratory Manual, 2nd ed. Cold Spring Harbor Laboratory Press, New York. 3 vols. Date of manuscript acceptance: | February 1998 Research Note: In search of Rossia pacifica diegensis S. S. Berry, 1912 K. M. Mangold], R. E. Young?, and Craig R. Smith2 1Basel, Switzerland 2Department of Oceanography, University of Hawaii, Honolulu, Hawaii 96822 U. S. A. Abstract: We describe the eggs of Rossia pacifica S. S. Berry, 1911, from a depth of 1,200 m off of southern California. The large size of the eggs and the great depth and low temperature from which they were retrieved contrasts strongly with smaller eggs of this species taken in shallower, warmer water along the western coast of the contiguous United States. This raises questions as to whether more than one taxon in the genus Rossia is present in this area. Key words: Cephalopoda, Rossiidae, larvae, eggs, subspecies S. S. Berry first described the sepiolid cephalopod Rossia pacifica from Alaskan waters in 1911. In 1912, he provided a more complete description based on specimens taken from southern Alaska to southern California. His specimens were trawled from depths of at least 30 to 310 m, possibly as deep as 550 m. In recent years, the known habitat of the species has been extended through the Bering Sea to the Tsushima Straits between Korea and southern Japan (Nesis, 1987), and southward from off of the south- ern California coast to about 28°N off of Baja California (F. G. Hochberg, pers. comm.); the known depth range remains approximately the same. Considerable information now exists on the biolo- gy of Rossia pacifica (Brocco, 1971; Summers, 1985; Anderson, 1987, 1991; Summers and Colvin, 1989; Anderson and Vanderwerff, 1989; Anderson and Shimek 1994). Off of the northwestern coast of the continental United States, it spawns clusters of white eggs with each egg about | cm in diameter and ovoid in shape, except for a flattened area where the eggs are cemented to the substra- tum (Anderson and Shimek, 1994). The eggs have a tough outer coating and a papilla that lies opposite the flat attach- ment site (Anderson, 1991). Off Washington state, scuba divers have observed egg clusters attached to rocks at ocean depths of 18 to 30 m (Anderson and Shimek, 1994), and eggs have been spawned and reared in local laborato- ries and public aquaria (Summers, 1985; Anderson, 1991). Eggs have developed at 10°C in a temperature-controlled aquarium and from 6-12°C (monthly means) in a running seawater system that reflected ambient ocean temperatures. Under both regimes, the minimum hatching times were four and three-quarters months with spontaneous hatching continuing for an additional two months (Summers, 1985). At hatching, the young measured 6-8 mm mantle length (ML) (Summers, 1985; Anderson, 1991). Recently, Shevtsov and Radchenko (1997) reported R. pacifica eggs with developing embryos from the Bering Sea at a depth of 250 m and a temperature of 1.6°C. In the northern Japan Sea, the distribution and spawning of R. pacifica has been discussed by Shevtsov and Mokrin (1997). In his 1912 paper, Berry noted that specimens of Rossia pacifica caught off of San Diego, California, were captured at greater depths than in other localities and dif- fered morphologically. Specifically, they (1) were smaller and more slender and delicate, (2) had relatively larger fins, and (3) carried suckers on the arms in predominantly two, rather than four, rows. Berry (1912) designated these spec- imens as a new subspecies, R. p. diegensis, while apparent- ly feeling that the morphological differences were sufficient for specific status. His type specimens are presently deposited in the U. S. National Museum of Natural History (Washington, DC; (USNM 214376). However, as these specimens were geographically separated from other Californian specimens found in Monterey Bay, he believed intergrades might eventually be found that would negate specific status. The possible validity of this taxon has been virtually ignored since his description. MATERIAL AND METHODS The material examined was collected by the R/V NEW HORIZON, 13 December 1995, with a 25 ft (7.6 m) semiballoon trawl at 33°10.25°N, 118°25.2’W (tow mid- American Malacological Bulletin, Vol. 14(2) (1998):185-187 185 186 AMER. MALAC. BULL. 14(2) (1998) point) in the Santa Catalina Basin, California, at a depth of 1,204-1,222 m. The bottom temperature at this locality was 4.1°C. The trawl had an outer mesh of 3.8 cm (stretch) and an inner mesh of 1.3 cm (stretch). The bottom in this area is a silty clay with a disaggregated median grain size of 4.0 um and the benthic macrofaunal community is heavily dominated by deposit-feeding polychaetes (Kukert and Smith, 1992). The specimens are deposited in the Santa Barbara Museum of Natural History. RESULTS Three large cephalopod eggs were retrieved from the trawl. Two measured 18 x 16 mm, one of which is fig- ured (Fig. 1A). The third was 14 mm in one dimension; damage prevented measurement in the other axis. The two larger eggs had well-developed embryos; the smaller egg had a badly malformed embryo. The surface of the egg cap- sules had been somewhat abraded by the trawl, preventing description other than to note that a papilla was present and that the capsule wall was about 0.5 mm thick and transpar- ent (Fig. 1B). No obvious flat attachment site was found on the capsule. The developing embryos appeared to be at roughly the same stage of development. One was dissected out. It has a mantle length of 2 mm at about Naef’s stage XVII or XVIII (Naef, 1928) and the size of the external yolk sac, not measured due to fracturing, dwarfs the size of the embryo (Fig. 1B). Large fins arise from the sides of the mantle rather than from the dorsal side (Fig. 1C). Each fin is nearly semicircular in outline and subterminal in posi- Fig. 1. Rossia eggs and embryos. A. Single egg in outer capsule. Photographed with reflected light. B. Single egg with embryo (arrow) and portion of outer capsule separated from egg. Distance between two lines on label paper is 1 cm. Photographed with transmitted light. C. Dorsal view of embryo showing fin shape, free dorsal mantle margin, and terminal spine. Photographed with reflected light. D. Posterior view of stained (methylene blue) embryo. White arrow indicates terminal spine; black arrow indicates hatching organ. Photographed with reflected light. E. Posterior view of unstained embryo. Photographed with reflected light. Scale bars = 1 cm (A, B), 1 mm (C, D, E). (I-IV, left arms; T = tentacle). MANGOLD ET AL.: ROSSIA PACIFICA DIEGENSIS 187 tion, the dorsal mantle margin is not fused with the head (Fig. 1C), and the hatching organ and terminal spine are easily seen (Figs. 1C, D). The embryo extends at right angles to the yolk sac with the arms spread laterally over the surface of the yolk (Figs. 1D, E). It has eight arms and two tentacles (Figs. 1D, E), with two rows of developing suckers on arms II and III; the armature of the other arms was not examined. DISCUSSION Although the embryonic development of Rossia pacifica has not been described, our embryos can be identi- fied on the following basis. They can be placed in the Decapodiformes by the presence of eight arms and two ten- tacles, and in Sepiolidae by the (1) stubby, blunt shape of the mantle, (2) position and shape of the fins, and (3) pres- ence of a terminal spine (Boletzky, 1991). Within the Sepiolidae, its identity is further narrowed by the lack of a dorsal head-mantle fusion. Within the North Pacific Ocean, the only sepiolids lacking head-mantle fusion are Rossia and Heteroteuthis, but only Rossia is presently known from Californian waters. In addition, the small, pelagic Heteroteuthis hatches at less than 2 mm ML (Boletzky, 1978) and can be eliminated by the large size of our eggs. By process of elimination, therefore, the eggs belong to a species of Rossia. To our knowledge, the only other record of rossi- ian eggs from this depth is that of Neorossia caroli (Joubin, 1902) which was taken in a trawl at 1,234 m in the Mediterranean Sea (Villaneueva, 1992). This species, known to reach depths of 1,744 m (Villaneueva, 1992), could be the deepest living of all rossiian species. The eggs described here are also very large for a member of Rossiinae as this group typically has eggs of 5-7 mm in diameter (Lu er al., 1992). Judging from the known ratio of egg size to hatchling size in shallow-water Rossia pacifica, an embryo hatching from an egg of 18 mm should be more than 10 mm ML. The only species of Rossia known from Californian waters is R. pacifica. The eggs described here differ from the R. pacifica eggs taken off of the northwest- ern coast of the continental United States in that they are much larger and were found at greater depths and at lower temperatures. Either R. pacifica exhibits great variability in these parameters or another Rossia, possibly R. p. diegensis, is present. On the other hand, R. pacifica could be a cold-water, large-egged form and the second taxon could be an unnamed warmer-water, small-egged form found at rather shallow depths along the western coast of the contiguous United States. In the latter case, R. p. diegensis would be invalid. We strongly urge researchers to reexamine the possibility that a second taxon exists. LITERATURE CITED Anderson, R. C. 1987. Field aspects of the sepiolid squid Rossia pacifica (Berry, 1911). Western Society of Malacologists, Annual Report 20:30-32. Anderson, R. C. 1991. Aquarium husbandry of the sepiolid squid Rossia pacifica. American Association of Zoological Parks and Aquariums, 1991 Annual Conference, Proceedings:206-211. Anderson, R. C. and R. L. Shimek. 1994. Field observations of Rossia pacifica (Berry, 1911) egg masses. The Veliger 37:117-123. Anderson, R. C. and J. E. Vanderwerff. 1989. In pursuit of the suburban squid. Sea Frontiers 35:165-169. Berry, S. S. 1911. Preliminary notices of some new Pacific cephalopods. Proceedings of the United States National Museum 40:589-590. Berry, S. S. 1912. A review of the cephalopods of western North America. Bulletin of the Bureau of Fisheries 30:267-336 + pls. Boletzky, S. von. 1978. Premiéres données sur le développement embry- onaire du sepiolidé pélagique Heteroteuthis. Haliotis 9:81-84. Boletzky, S. von. 1991. The terminal spine of sepiolid hatchlings: its development and functional morphology (Mollusca, Cephalopoda). Bulletin of Marine Science 49:107-112. Brocco, S. L. 1971. Aspects of the Biology of the Sepiolid Squid Rossia pacifica Berry. Masters thesis, University of Victoria, British Columbia, Canada. 151 pp. Kukert, H. and C. R. Smith. 1992. Disturbance, colonization and succes- sion in a deep-sea sediment community: artificial-mound experi- ments. Deep-Sea Research 39:1349-1371. Lu, C. C., A. Guerra, F. Palumbo, and W. C. Summers. 1992. Order Sepioidea Naef, 1916. In: “Larval” and Juvenile Cephalopods: A Manual for Their Identification, M. J. Sweeney, C. F. E. Roper, K. M. Mangold, M. R. Clarke, and S. v. Boletzky, eds. Smithsonian Contributions to Zoology 513:1-282. Naef, A. 1928. Die Cephalopoden. Fauna und Flora des Golfes von Neapel 35, monogr., I-2, 357 pp. Nesis, K. 1987. Cephalopods of the World. Translated from Russian by B. S. Levitov. Edited by L. A. Burgess. T. F. H. Publications, Neptune City, New Jersey. 351 pp. Shevtsov, G. A. and N. M. Mokrin. 1997. Distribution and biology of Rossia pacifica (Cephalopoda, Sepiolidae) in the Russian exclu- sive zone of the Japan Sea [abstract]. Program and Abstracts, 63rd Annual Meeting, American Malacological Union, and 30th Annual Meeting, Western Society of Malacologists:57. Shevtsov, G. A. and V. I. Radchenko. 1997. Discovery of an egg mass with embryos of Rossia pacifica (Cephalopoda, Sepiolidae) in the Okhotsk Sea [abstract]. Program and Abstracts, 63rd Annual Meeting, American Malacological Union, and 30th Annual Meeting, Western Society of Malacologists:57. Summers, W. C. 1985. Ecological implications of life stage timing deter- mined from the cultivation of Rossia pacifica (Mollusa, Cephalopoda). Vie et Milieu 35:249-254. Summers, W. C. and L. J. Colvin. 1989. On the cultivation of Rossia paci- fica (Berry, 1911). Journal of Cephalopod Biology \(1):21-31. Villanueva, R. 1992. Deep-sea cephalopods of the northwestern Mediterranean: indications of up-slope ontogenetic migration in two bathybenthic species. Journal of Zoology, London 227:267- 276. Date of manuscript acceptance: 25 June 1998 waren 4 A =e VGN SYMPOSIUM: TRADITIONAL VERSUS PHYLOGENETIC SYSTEMATICS OF MOLLUSKS Organized by GARY ROSENBERG ACADEMY OF NATURAL SCIENCES OF PHILADELPHIA AMERICAN MALACOLOGICAL SOCIETY SANTA BARBARA, CALIFORNIA 25 JUNE 1997 189 Reconciling observed patterns of temporal occurrence with cladistic hypotheses of phylogenetic relationship Helena Fortunato Center for Tropical Paleoecology and Archaeology, Smithsonian Tropical Research Institute, Box 2072, Balboa, Republic of Panama, stri03.ancon.fortunae @ic.si.edu Abstract: Cladistic analyses of well-sampled groups with a complete and abundant fossil record can yield phylogenetic hypotheses that conflict with stratigraphic data, even to the extent of supporting phylogenies that appear to invert the stratigraphy. This is most probably due to the convergent evolution of similar morphologies (7. e. homoplasy), rather than the inadequacy of the fossil record. Several ways have been proposed to resolve this problem (stratophenetics, stratocladistics, character refinement, etc.). This paper proposes an iteractive technique to construct a tree more consistent with the observed fossil record by building separate phylogenies for different stratigraphic intervals which can then be assembled into a composite phylogeny. Columbellid gastropods of the genus Strombina Morch, 1852, were used to test this approach. Strombina originated and diversified in the Caribbean during the Miocene and Pliocene. During the Pliocene, they became nearly extinct in the Caribbean, but diversified in the eastern Pacific. Phylogenies of 42 species based only on shell morphology (49 characters, 186 states) yielded trees with high stratigraphic inconsistency and ghost lineages that require the presence of descen- dants 10 million years or more before the first appearance of their hypothesized ancestors. Removal of species that originated after the Pliocene resolved most of these stratigraphic inconsistencies although some ghost lineages remained. This Miocene/Pliocene tree was then used to root the trees illustrating the relationships among Pleistocene/Holocene species. This final composite tree is more consistent with the hypothesized fossil record for the group than the original tree. Key words: phylogeny, fossil record, consistency, morphology, gastropods, Strombina Most analyses of extant taxa exclude fossil data component to calibrate phylogenetic hypotheses. However, with the justification that the fossil record is too incomplete the fossil record is often incongruent with cladistic topolo- and too poorly dated to be useful (Hennig, 1966, 1981; gies based on morphology alone. There are two sources of Nelson, 1989; Patterson, 1981; Goodman, 1989). Although disagreement: (a) when the cladistic results require the this is true for many taxa without durable skeletons, many presence of long ghost lineages that conflict with a rich fos- marine invertebrate taxa such as mollusks, bryozoans, and sil record for the time interval of interest (Cheetham and corals have a rich fossil record. Moreover, fossil collections Hayek, 1988; Jackson and Cheetham, 1990, 1994; Wagner, of these taxa can greatly outnumber collections of extant 1995; Wagner and Erwin, 1995); or (b) when ancestor- species. Also, the ages of many fossil collections are descendant relationships are turned upside-down relative to known within a range of 1-2 million years, and often better, their stratigraphic sequence (Jackson and Cheetham, 1994). which is adequate to order occurrences over the typically This disagreement can be by as much as 5-10 million years. 20 million years of history of most genera (Cheetham, Such strong conflict between a good fossil record and a 1987; Budd, 1989; Budd er al., 1994, 1996). This can given cladistic hypothesis suggests that morphological con- depend on the density of sampling (7. e. number of samples vergence (i. e. homoplasy) could be responsible for putative and sampling interval), but even in studies of taxa with a relationships within the branches of the cladograms sparse fossil record the inclusion of fossils can help dis- (Campbell and Barwick, 1988, 1990; Allmon, 1989; criminate morphospecies and increase the consistency of Bodenbender, 1994). cladistic hypotheses, thus providing a more effective means The fossil record can be employed in different ways of rooting trees than comparison with a living outgroup to try to resolve this disagreement. For example, extinct (Cheetham and Hayek, 1988; Jackson and Cheetham, 1994; taxa can be used to infer character polarity (Donoghue et Cheetham and Jackson, 1995; Marshall, 1995). al., 1989; Novacek, 1992b). The fit between stratigraphic Fossils can provide reliable estimates of the order in information and trees can be used to choose among most which lineages appear in the history of life (Gardiner, 1982; parsimonious trees based on morphological data alone Gauthier er al., 1988; Norell and Novacek, 1992a, b; (Norell, 1993; Huelsenbeck, 1994; Suter, 1994; Benton and Novacek, 1992a; Foote, 1996), thus providing a temporal Storrs, 1994, 1996). Stratocladistics (Fisher, 1980, 1988, American Malacological Bulletin, Vol. 14(2) (1998):191-200 19] 192 AMER. MALAC. BULL. 14(2) (1998) 1991, 1992, 1994; Clyde, 1994; Clyde and Fisher, 1997) uses stratigraphic information to calculate the “stratigraphic parsimony debt” - a measure of the discrepancies between expected and observed order of stratigraphic occurrences. Although in this case both types of data are allowed to interact at the same level, morphological data are still assigned more weight, thus biasing the output. Common to many of these studies is the tendency to test the adequacy and completeness of the fossil data by quantifying the con- gruence between stratigraphic information and phylogenet- ic hypothesis based on morphological data (Marshall, 1990, 1994, 1997; Huelsenbeck, 1994; Benton and Hitchin, 1996; Huelsenbeck and Rannala, 1997). Several authors (Gingerich, 1979; 1994; Cheetham, 1987; Budd, 1988; Stanley et al., 1988; Geary, 1990; Wei, 1994) have advocat- ed the use of stratophenetics, a clustering technique, to reconstruct relationships using both Recent and fossil species, using overall morphologic similarity to connect populations through time. Moreover, stratophenograms give excellent results for groups with a very complete, well-preserved fossil record, like foraminiferans, cheilostome bryozoans, some gastropod groups, and small vertebrates. Another way of using the fossil record is to con- struct separate analyses of subsets of a clade between major pulses of evolution (Budd and Coates, 1992). The justifica- tion for this method is the assumption that convergent taxa tend to originate during such pulses or turnovers. These separate analyses can then be used to assemble a composite tree across the pulse by building on the tree for the previous interval, such that later taxa cannot totally change hypothe- sized relationships among much older taxa. This method assumes that observed stratigraphic ranges correspond closely to “true” stratigraphic ranges and that members of different but contemporary clades had approximately the same probabilities of preservation and recovery. Consequently, it considers the possibility that the phyloge- netic hypothesis could be incorrect, rather than blaming the fossil record for its deficiency. The aim of this paper it is to present a case study where this methodology was used to choose among competing hypotheses of relationships. The final accepted tree is one that better reconciles observed and expected patterns of temporal occurrences for the group in question. This methodology applies only to species-level analyses, differing in this from stratocladis- tics, which can be used in cases of uncertain specific affini- ties and also with supraspecific taxa. MATERIALS AND METHODS TAXA AND CHARACTERS The Isthmus of Panama provides a well-document- ed example of a turnover event that occurred approximately 3-2 Ma (Allmon et al., 1993, 1996; Jackson er al., 1993, 1996; Budd et al., 1996; Cheetham and Jackson, 1996; Fortunato and Jackson, 1996). The large molluscan fossil collections now available for the Neogene to Recent of tropical America, with greatly improved geochronological data (Coates et al., 1992; Collins and Coates, 1993; Coates and Obando, 1996; A. G. Coates, unpub.) and from tens to hundreds of localities, contain thousands of specimens belonging to several highly diverse clades (J. B. C. Jackson et al., unpub.). They permit a look at evolution at the species level and test the hypothesis that homoplasy arises during mass extinction/origination events. The taxon used in this study is a group of small buc- cinoidean gastropods, mostly 1-5 cm long, the so-called Strombina-group (Jung, 1986, 1989). This group comprises approximately 110 species ranging from the early Miocene (about 20 Ma) through the Holocene, with an excellent fos- sil record and several thousand specimens collected. Their alpha taxonomy has been revised recently (Jung, 1989), and is assumed to be correct (or correctable with little effort) for this study. The wide distribution of the group, both in space and time, is an important aspect for the technique used here. Species diversity slowly increased over time, mainly in the Caribbean. In the eastern Pacific there were always many fewer species until approximately 1.8 Ma, at the Plio-Pleistocene boundary, when there was a sudden geographic shift in diversity from the Caribbean to the east- ern Pacific (Jackson et al., 1993). The fact that there are fewer collections (and specimens) from the eastern Pacific than from the Caribbean sea could affect the observed diversity pattern through time. Nevertheless, in spite of being very abundant, most taxa of the Strombina-group occur only in one region and only during one time interval, which, given the extensive records overall, cannot be mere- ly a sampling artifact. Of the five genera that comprise the Strombina- group, only the genus Strombina Morch, 1852, was used for this study. This is the most abundant and diverse genus in the group, with 42 species, of which 20 originated during the last three million years. Previous results based in both shell morphology and anatomy strongly support the mono- phyly of this genus (Fortunato and Jung, 1995). The material analyzed here (Table 1) consisted of all available fossil and Holocene collections from more than 20 institutions and private collections used by Jung (1989) for his taxonomic revision of the genus Strombina. These collections were supplemented by the fossil collec- tions from the Panama Paleontological Project (PPP), main- tained in the Naturhistorisches Museum, Basel, Switzerland. The characters used in this study were selected from a larger suite of characters, including both measurements FORTUNATO: RECONCILING PHYLOGENIES WITH THE FOSSIL RECORD [93 Table 1. Number of lots and specimens used in this study. Subgenus Lots Specimens Strombina 362 7085 Spiralta 459 2646 Lirastrombina 291 1363 Arayina 1] 117 Recurvina 310 2969 Costangula 98 603 (continuous) and qualitative (discrete) characters of shell morphology and anatomy, as defined in a previous work to test the monophyly of the Strombina-group (Fortunato and Jung, 1995). Only shell morphology was considered here to allow equal evaluation of both extant and fossil species. Of these 49 characters, six are measurements of overall shell and aperture dimensions; 38 are qualitative (discrete) char- acters related to various aspects of shell ornamentation and armor, as well as aperture armor; and five are composite characters referring to overall shell size, shape, and spire, as well as apertural size and shape. All characters were unordered and equally weighted. Unknown states were coded as missing. Continuous characters were coded using the Duncan’s procedure for multiple-comparison tests of species means (Winer, 1971). Detailed descriptions of these characters, as well as coding procedures and the data matrix, will be published elsewhere. ANALYTIC PROCEDURES Species-level cladistic phylogenies for 42 Strombina species were generated using Hennig86 version 1.5 (Farris, 1988). Trees were rooted on the three oldest species of the genus Sincola Olsson & Harbison, 1953, which is the second oldest genus of the Strombina-group. Both Sincola and Strombina first occur in the proto- Caribbean region approximately 18 million years ago. Heuristic searches were performed using the ‘“mhennig” followed by “bb” options to generate trees. Exhaustive searches were impossible to carry out because of the exces- sive amounts of time required. No consensus trees were requested; all most-parsimonious trees obtained in each analysis were evaluated and only the trees that were con- gruent with the time of origination and extinction hypothe- sized for the group in question were retained. The single tree that was most consistent with the observed fossil record was then used for the reconstructions. Table 2 summarizes the sequence of procedures and results for all four sets of cladistic analyses. All assump- tions were the same for the four sets of analyses, only the species included varied. For Analysis 1, all 42 species root- ed on three Sincola species were used regardless of strati- graphic level. Evaluation of the most parsimonious trees yielded by this analysis (the “Total” tree; Fig. 1) showed that taxa from different stratigraphic levels were evenly scattered across the trees, indicating extremely high strati- graphic inconsistency. The taxa were therefore subdivided into two groups, pre- and post-turnover, depending on their time of origination. Analysis 2 was run using only the 22 species that originated before the turnover event that took place in tropical America (“‘pre-turnover” species; origina- tion times older than three million years). The trees yielded by this analysis (the “Mio-Pliocene”’ tree; Fig. 2) were eval- uated and taken as a fixed point for further work. Analysis 3 was subdivided into two phases, each consisting of a Table 2. Sequence of procedures used in each analysis and the results obtained. Analysis 1 Input: 42 ingroup taxa 3 outgroup taxa 49 characters Procedures: Morphologic parsimony “Total tree” 2 main clades Result: Stratigraphic inconsistency Analysis 2 22 ingroup taxa older than 3 Ma 3 outgroup taxa 49 characters Morphologic parsimony Removal of taxa younger than 3 Ma “Mio-Pliocene tree” 2 main clades Stratigraphic consistency Analysis 3A 11 Mio-Pliocene ingroup taxa (left clade) 3 outgroup taxa 49 characters Morphologic parsimony Break between taxa 28/27 Include oldest Strombina (taxon 4) Add younger taxa to each lineage Final tree for Strombina/ Recurvina clade Analysis 3B 9 Mio-Pliocene group taxa (right clade) 3 outgroup taxa 49 characters Morphologic parsimony Break between taxa 28/27 Include oldest Strombina (taxon 4) Add younger taxa to each lineage Final tree for Lirastombina/ Spiralta/Costangula clade Analysis 4 All trees Connect each tree sequentially following stratigraphic order Retain main clades delineated in Analysis 1 Composite tree Stratigraphic consistency 194 series of separate cladistic analyses including the younger species which originated after the turnover event (the “‘post- turnover” species; origination times younger than three mil- lion years). In each of these phases, the extant species were added to each subclade as defined by the “Mio-Pliocene” tree and supported by the “Total” tree. The results of the third set of analyses were then used to build a composite tree (Analysis 4) based on the “Mio-Pliocene” tree and with all of the extant species added. Ancestor-descendant rela- tionships were interpreted for each of these subclades based on the results of Analyses 2 and 3. The taxon of the same subgenus which lay at the tip of the “Mio-Pliocene” tree was considered to be the ancestor of the subsequent Holocene species. The trees yielded by Analysis 3 were attached to these ancestor taxa. In another words, all analyses were performed as though a systematist decided to test for homoplasy three million years ago, before the major turnover event took place in Central America. One of the assumptions is that the relationships that existed then could not be changed by any subsequent events. The removal of younger species from the “Mio-Pliocene” tree minimized most of the possi- ble effects of the subsequent radiation during this period which could obscure older relationships. co onan @pwp = OO Oo Millions of Years 16 © Lirastrombina @ Sincola O Strombina ? @ Strombina AMER. MALAC. BULL. 14(2) (1998) RESULTS AND DISCUSSION Analysis 1 (“‘Total’” tree). Four equally most parsi- monious trees, with length 703, CI = 0.21, and RI = 0.52, were obtained. The main differences in tree topology were related to the relative positions of taxa 17 and 18. In one case taxon 17 was hypothesized to be sister of the group formed by taxa 10 and 18, whereas in the other case taxon 18 was hypothesized to be sister of the group formed by taxa 10 and 17. The former tree was accepted because it was more consistent with the observed fossil record of the group. Fig. | shows a reconstruction of the accepted tree plotted against the stratigraphic column (the “Total” tree). Fig. | shows striking stratigraphic inconsistency. One-third of the branches are ghost lineages that extend for more than ten million years without a single fossil occur- rence or with fossils that appear several million years after their origination times as hypothesized by the cladogram (i. e. taxa 9, 12, 40/41). The tree also hypothesizes that sev- eral extant species are ancestors of species that became extinct 10-15 Ma. This is the case for taxon 12, a living species reported from a region that yielded many other fos- sil species, but of which there are no fossil occurrences. According to the relationships postulated in Fig. 1, this 41 40 3 42 43 Mh OB @& 37 & O Arayina m RecuMna V Splratta A Costangula Fig. 1. Reconstructed 42-species cladogram (‘‘Total” tree) showing the geologic time scale for calibration. Numbers on terminal branches represent species; symbols on terminal branches represent subgenera; thick black lines are real fossil data through time; gray lines represent ranges with uncertain age assign- ment; thin vertical lines represent ghost lineages hypothesized by the cladogram. FORTUNATO: RECONCILING PHYLOGENIES WITH THE FOSSIL RECORD taxon is the ancestor of clades 13/18 and 5/8, as well as of taxon 28; however, most of these hypothesized descendent species became extinct several million years before taxon 12 originated. The same situation occurs with several other taxa (i. e. taxa 36, 38, etc.). These long ghost lineages also push back the radiation of the whole group to a time, approximately 15 Ma, when, although there is an excellent fossil record for many other molluscan groups in the same region, there is no sign of such a radiation in the Strombina- group. Despite these stratigraphic inconsistencies, the sub- generic assignments of the species are mostly conserved, and several major clades are delineated, thus the problem is mainly one of stratigraphic inconsistency between clades and subclades. Such stratigraphic inconsistency is probably due to an extensive morphologic convergence caused by the evolution of similar forms during the turnover as exist- ed in the earlier history of the group. Similar problems of homoplasy are common among corals, bryozoans, etc. (Budd and Coates, 1992; Jackson and Cheetham, 1994; Cheetham and Jackson, 1995, 1996). Analysis 2 (““Mio-Pliocene” tree). This analysis yielded eight most-parsimonious trees, of 419 steps, CI = 0.31, and RI = 0.48. The major differences among these trees were in the basal taxa. In several trees the node that joined taxa 30 and 31 was basal to the rest of the tree, 195 whereas in other trees the node joining taxa 28 and 35 was basal. The tree in which taxa 28 and 35 were basal was accepted because these taxa have older origination times. The “Mio-Pliocene” accepted tree is plotted strati- graphically in Fig. 2. Most of the long ghost lineages have disappeared. None of the younger species included in the analysis are hypothesized to be ancestors of much older taxa. Once again, the subgeneric taxonomic assignments hold together and two major clades are clearly defined. Analysis 3A (18 species, “Strombina/Recurvina” clade). The right subclade of the “Mio-Pliocene” tree (Fig. 2) comprised all species of the subgenus Recurvina Jung, 1989 (closed squares), and nearly all species of Strombina s. 5. (closed circles). Three Strombina s. s. species (taxa 25, 26, 27) grouped with the left clade. This result supports the initial questionable subgeneric assignments of taxa 26 and 27 (Jung, 1989) and suggests that taxon 25 could also belong to another subgenus. The analysis of the 11 Mio- Pliocene species of these subgenera, plus the oldest Strombina s. s. (taxon 4) (which consistently appeared in a basal position in all previous analyses) rooted on Sincola, yielded only one tree with exactly the same topology, thus confirming the stability of this clade. When the two younger species of Strombina, three of Recurvina, and two of Costangula Jung, 1989 (which had consistently grouped with the other two subgenera in (*) 1 ) 2 3 1 18 y-) 9 4 A Fe 20 ms iS) a “4 ® 6 re) e Y uu a o 7 Y g 8 3 4 on 37 9 re 10 eu . f Ln S 1s r) 4 ‘ 18 16 7 18 19 2» © Urastrombina © Stombina? © Arayina V Splratta @ SIncola @ Strombina m Recurina Fig. 2. Reconstructed 22-species cladogram (“Mio-Pliocene” tree) with geologic time scale for calibration. All symbols as in Fig. 1. This tree was the fixed point for Analyses 3A and 3B. 196 AMER. MALAC. BULL. 14(2) (1998) Tree 1 pe caribaea K4 T quirosana prstsp32 L=35 Fl isi [2 —=STsp33 soy pstse35 ==STsp36 30-= L STsp37 2of REsp38 zef 7=REsp39 Lar pe Sospso 2ssicospa1 2 of peer 24 REsp43 bof E ( RBsp44 2Cagspas, Tree: 2 S. caribaea a. -quirosana S.pigea sof STsp4 39 STsp29 i po7Esrsp28 3a] pretsP39 36) p-Stsp31 igs. Pad | ic seergp a3 po stepsa Sethe STsp37 =32 r=STsp35 294+=sTsp36 Lol f-REsSp38 re[ REsp39 bof COsp40 2slcospai Leg fee see 24) prRBsp43 =23 [ C REsp44 224 REsp45 Fig. 3. Two most parsimonious trees yielded by Analysis 3A, including 19 species: 11 Mio-Pliocene species of the subgenera Strombina (ST) and Recurvina (RE) (= right clade of the “Mio-Pliocene” tree) plus seven extant species: two Strombina s. s., three Recurvina, and two Costangula (CO) rooted on three species of Sincola (S. caribaea Gabb, 1873; S. quirosana H. K. Hodson,1931; S. pigea Olsson, 1964). The oldest Strombina 5s. s. in the analysis (taxon 4) was also used here. the “Total tree” analysis), were added to the analysis and rooted on Sincola, the resulting two most-parsimonious trees (Fig. 3) differed only in the relative position of taxon 37 versus 35 and 36. The tree in which taxon 35, the oldest, branched first (Fig. 3, Tree 1) was accepted. Fig. 4 shows the reconstructed “Mio-Pliocene” tree (Fig. 2) with the “Strombina/Recurvina” clade (Fig. 3) added to the right side. Analysis 3B (23 species, ‘“Lirastrombina/Spiralta/ Arayina” clade). The left clade of the ““Mio-Pliocene” tree comprised all species of three subgenera: Lirastrombina Jung, 1989 (open diamonds), Spiralta Jung, 1989 (open inverted triangles), and Arayina Jung, 1989 (open squares), as well as the three Strombina s. s. (taxa 25/27) mentioned above. Analysis of the ten Mio-Pliocene species of these subgenera plus the oldest Strombina species (which had been basal in all prior analyses), rooted on Sincola, yielded five trees. The major differences among these trees were in the relative positions of taxa 25 and 26. The tree in which the node that joined taxa 25 and 26 was basal was accepted based on the older origination times of taxon 26. This topology also retained taxa 26 and 27 close together. In spite of these changes, all topologies maintained all species belonging to the left clade in the same relative positions they had in the “Mio-Pliocene” tree, thus confirming the stability of this clade. The “Lirastrombina/Arayina/Spiralta” clade involved three different and large subgenera, and it was analyzed in two separated phases. In phase 1, when the ten extant species of Lirastrombina were added to the three Mio-Pliocene species plus the Strombina s. s. (taxon 4) and rooted on Sincola, the resulting two most-parsimonious trees (Fig. 5) differed only in the relative positions of taxa 13 and 14. In one of these trees, taxon 14 was hypothesized as a sister taxon of taxon 13, whereas in the second tree, taxon 13 was hypothesized as branching earlier. Because taxon 14 has an earlier origination time than taxon 13, the tree in which it is hypothesized to arise at least at the same time (as a sister taxon of 13 and not later) was accepted (Fig. 5, Tree 1). Due to extreme homoplasy, the extant taxon 10 always nested very deep within the Lirastrombina tree, resulting in a decrease of overall stratigraphic consis- tency. This problem was solved by the elimination of this species from this phase of the analysis; this species was later added to the “Total tree” based on its position in the first analysis (see Fig. 1). In phase 2, when the two extant species of Spiralta were added to the Mio-Pliocene species of Spiralta and Arayina plus taxon 4 and rooted on Sincola, the result was a single tree (Fig. 6). Analysis 4 (composite tree). The trees resulting from Analyses 2 and 3 were then added to the left side of the “Mio-Pliocene”’ tree. Fig. 7 shows the final reconstruct- ed tree. This tree has a high stratigraphic consistency even though ages were not used directly as characters. It has two major clades: one composed of members of the Strombina, Recurvina, and Costangula subgenera; the other is com- posed mainly of members of the Lirastrombina, Arayina, and Spiralta subgenera. These same major clades were delineated in the original 42-species tree (Fig. 1), but some of their relationships were originally obscured by morpho- logic convergence which then confused the stratigraphic relationships. A few ghost lineages, presumably due to missing fossil data, still remain. CONCLUSIONS The observed stratigraphic order of species occur- rences (first and last appearances) should be used as one more source of data in phylogenetic inferences. The tech- nique proposed here is a simple and independent way of assessing rival phylogenetic hypotheses obtained through parsimony methods. The final tree is more consistent with FORTUNATO: RECONCILING PHYLOGENIES WITH THE FOSSIL RECORD 197 co oan aga on -_ © Millions of Years anonz2za 16 © Urastrombina @ Sincola ®@ Strombina O Stombina? CO Arayina @ Recunvina % 39 40 4142 43 44 V Splratta A Costangula Fig. 4. Reconstructed cladogram of the 22 Mio-Pliocene species plus the seven extant species of the right clade (“Strombina/Recurvina” clade), with geo- logic time scale for calibration. All symbols as in Fig. 1. Tree zs caribaea La a pigea L iene -quirosana 261 prstse4 Lest i eae Leal (pee leestErtepi7 LIsp11 rtsps LIsp8s { pretse? liya L, serene erisps Tree 2 S. caribaea Lc ]/:27;=S.pigea S.quirosana r 26 Worse 25] leu eoe 24] LIsp18 eg sfEgtsp17 Lisp12 a (eee 16—= Ltisps Lor ico lis i Wer rses 14 LIsp6 3Crtsps Fig. 5. Two most-parsimonious trees yielded by the analysis of 13 species (Mio-Pliocene plus extant) of the subgenus Lirastrombina (LI) rooted on Sincola. This was the first phase of Analysis 3B. Other taxon labels as in Fig. 3. f=19 Bee S.quirosana Lael STsp4 lia 7 [et esPed 16 (eee aw L, ale as eaRsp20 SPsp23 “Lua (een aie Ex [pete Soe 11-=—SPsp21 Fig. 6. Single most-parsimonious tree yielded by the analysis of eight species (Mio-Pliocene plus extant) of the subgenera Spiralta (SP) and Arayina (AR) rooted on Sincola. This was the second phase Analysis 3B. Other taxon labels as in Fig. 3. the fossil data and uses fewer statements about relative probabilities of preservation, thus conforming to the princi- ple of parsimony. Several main aspects of the methodology used and its implications for the study of phylogenetic relationships should be noted here. Others will be considered in a later work dealing mainly with further refinement of this tech- nique, as well as with its comparison with similar methods. The quality of geochronological data used is very important for the correct interpretation of the results obtained with AMER. MALAC. BULL. 14(2) (1998) % 39 40 4142 43 44 198 6 6 7 8 910 1112 21 0 1 2 3 4 3 nm § re) 6 @ 7 wt sy : 4 ve) 9 10 Cou On = 18 4: | 16 16 7 18 19 20 © Urasttombina O Stombina ? @ Sincola @ Stombina O Arayina @ Recunina V Splratta 4 Costangula Fig. 7. Reconstructed cladogram including the 42 initial species. This is the final composite tree based on the “Mio-Pliocene” tree plus the trees yielded by Analyses 3A and 3B. All symbols as in Fig. 1. this methodology. Uncertain age assignments and low stratigraphic resolution will make difficult the decision about time intervals into which to divide the analyses, thus increasing the uncertainty in recovering the true history of the clades under study. Sampling can also be a problem: uneven geographic and stratigraphic sampling will give an incomplete picture of the true abundance and diversity of the groups studied, and its effects can be confounded with homoplasy. Another important point is the question of how well the species-level taxonomy is employed. The “Total tree” preserved most of the subgeneric units (with only a few exceptions, some of which were already doubtful). This result matches earlier work (Fortunato and Jung, 1995) using both shell morphology and anatomy, thus adding confidence to the methodology employed to recover the true history of the clade. In spite of this agreement with the structural component of the analysis, the temporal aspect was violated both by the extremely long-lasting ghost lineages and by the hypothesized order of origina- tions of several ancestor-descendant pairs. The removal of all taxa that originated during the last three million years resolved many of these incongruences with the known order of fossil occurrences. Most of the hypothesized cladistic relationships (i. e. ancestor-descendant pairs and sister taxa) among the Mio-Pliocene taxa have a high level of agreement with their origination time as shown by the fossil record. This also helped to resolve most of the long- lasting ghost lineages. These results all agree with the assumption that most of the stratigraphic inconsistencies present in the first tree (Fig. 1) were due to high levels of homoplasy after and before the turnover event rather than to the incompleteness of the fossil record, incorrect taxo- nomic assignments, or sampling bias. It is important to notice that the evolutionary pat- terns hypothesized by the trees obtained before and after the removal of the young species are very different. These evolutionary scenarios and its implications for the history of the group will be analyzed elsewhere. It is enough to mention here that whereas the tree represented in Fig. | implies an extensive radiation of the whole group approxi- mately 15 Ma, that represented in Fig. 6 suggests that there were two radiations (or three), one approximately 10 Ma, and the second probably just before the final closure of the isthmus of Panama, and extending for some time thereafter in the eastern Pacific. This scenario of several possible radiations agrees with the fossil record. The technique outlined in this paper is still in its preliminary stage. As such it suffers from several biases which are currently being resolved. One of these is the lack of stratigraphic confidence intervals for the samples. The approach used here assumes that the sampling bias is not large enough to significantly override the effects of homo- plasy. As part of the development of this technique, strati- graphic intervals will be calculated following Wagner’s (1995) methodology. This will separate the effects of homoplasy from those of inadequate sampling, thus testing the consistency of the method. Also as part of the work in progress, the same data set will be employed using Fisher’s methodology (Fisher, 1992; Clyde and Fisher, 1997) and the resulting trees compared with the composite tree FORTUNATO: RECONCILING PHYLOGENIES WITH THE FOSSIL RECORD 199 obtained through the technique presented here. This will provide a test of the power of both methodologies and eval- uate their capability of recovering the true history of this group. Regardless of the need for further development of this technique, it is clear that conventional parsimony analyses are only the first step towards recovering the true history of a clade. Cladistic hypotheses should be tested against the stratigraphic data if at all possible. Failure to do this can distort the evolutionary histories and falsify our predictions about evolutionary patterns. ACKNOWLEDGMENTS This paper was first presented in at the 63rd annual meeting of the American Malacological Union as part of the symposium, “Traditional versus phylogenetic systematics” organized by G. Rosenberg. J. Jackson, A. Budd, A. Cheetham, K. Johnson, A. Coates, and two anonymous reviewers criticized the manuscript and made many helpful suggestions. This work was supported by the Smithsonian Tropical Research Institute and by a grant from the Swiss National Science Foundation to P. Jung, J. Jackson, and H. Fortunato. LITERATURE CITED Allmon, W. D. 1989. Paleontological completeness of the record of lower Tertiary mollusks, U. S. Gulf and Atlantic coastal plains: implica- tions for phylogenetic studies. Historical Biology 3:141-158. Allmon, W. D., G. Rosenberg, R. W. Portell, and K. S. Schindler. 1993. Diversity of Atlantic coastal plain mollusks since the Pliocene. Science 260:1626-1629. Allmon, W. D., G. Rosenberg, R. W. Portell, and K. S. Schindler. 1996. Diversity of Pliocene-Recent mollusks in the western Atlantic: extinction, origination, and environmental change. Jn: Evolution and Environment in Tropical America, J. B. C. Jackson, A. G. Coates, and A. F. Budd, eds. pp. 271-302. University of Chicago Press, Chicago. Benton, M. J. and R. Hitchin. 1996. Testing the quality of the fossil record by groups and by major habitats. Historical Biology 12:111-157. Benton, M. J. and G. W. Storrs. 1994. Testing the quality of the fossil record: paleontological knowledge is improving. Geology 22:111- 114. Benton, M. J. and G. W. Storrs. 1996. Diversity in the past: comparing cladistic phylogenies and stratigraphy. In: Aspects of the Genesis and Maintenance of Biological Diversity, M. E. Hochberg, J. Clobert, and R. Barbault, eds. pp. 19-40. Oxford University Press, Oxford. Bodenbender, B. E. 1994. Stratocladistic reconstruction of blastoid evolu- tionary history. Geological Society of America, Abstracts with Programs 26:A427. Budd, A. F. 1988. Large-scale evolutionary patterns in the reef-coral Montastrea: the role of phenotypic plasticity. Proceedings of the 6th International Coral Reef Symposium 3:393-398. Budd, A. F. 1989. Biogeography of Neogene Caribbean reef corals and its implications for the ancestry of eastern Pacific reef corals. Memoires of the Association of Australasian Paleontologists 8:219-130. Budd, A. F. and A. G. Coates. 1992. Nonprogressive evolution in a clade of Cretaceous Montastrea-like corals. Paleobiology 18:425-446. Budd, A. F., K. G. Johnson, and J. B. C. Jackson. 1994. Patterns of replacement in Late Cenozoic Caribbean coral communities. Geological Society of America, Abstracts with Programs 25:454. Budd, A. F., K. G. Johnson, and T. A. Stemann. 1996. Plio-Pleistocene turnover and extinctions in the Caribbean reef-coral fauna. In: Evolution and Environment in Tropical America, J. B. C. Jackson, A. G. Coates, and A. F. Budd, eds. pp. 168-204. University of Chicago Press, Chicago. Campbell, K. S. W. and R. E. Barwick. 1988. Geological and paleontolog- ical information and phylogenetic hypotheses. Geological Magazine 125:207-227. Campbell, K. S. W. and R. E. Barwick. 1990. Paleozoic dipnoan phyloge- ny: functional complexes and evolution without parsimony. Paleobiology 16:143-169. Cheetham, A. H. 1987. Tempo of evolution in a Neogene bryozoan: are trends in single morphologic characters misleading? Paleobiology 13:286-296. Cheetham, A. H. and L.-A. C. Hayek. 1988. Phylogeny reconstruction in the Neogene bryozoan Metrarabdotos: a paleontologic evaluation of methodology. Historical Biology 1:65-83. Cheetham, A. H. and J. B. C. Jackson. 1995. Process from pattern: tests for selection versus random change in punctuated bryozoan speci- ation. In: New Approaches to Speciation in the Fossil Record, D. H. Erwin and R. L. Anstey, eds. pp. 184-207. Columbia University Press, New York. Cheetham, A. H. and J. B. C. Jackson. 1996. Speciation, extinction, and the decline of arborescent growth in Neogene and Quaternary cheilostome Bryozoa of tropical America. In: Evolution and Environment in Tropical America, J. B. C. Jackson, A. G. Coates, and A. F. Budd, eds. pp. 205-233. University of Chicago Press, Chicago. Clyde, W. C. 1994. Stratigraphic congruence vs. morphological congru- ence in phylogenetic analysis. Geological Society of America, Abstracts with Programs 26(7):A122. Clyde, W. C. and D. C. Fisher. 1997. Comparing the fit of stratigraphic and morphologic data in phylogenetic analysis. Paleobiology 23(1):1-19. Coates, A. G., J. B. C. Jackson, L. S. Collins, T. M. Cronin, H. J. Dowsett, L. M. Bybell, P. Jung, and J. A. Obando. 1992. Closure of the Isthmus of Panama: the near-shore marine record of Costa Rica and western Panama. Geological Society of America Bulletin 104:814-828. Coates, A. G. and J. A. Obando. 1996. The geologic evolution of the Central American isthmus. /n: Evolution and Environment in Tropical America, J. B. C. Jackson, A. G. Coates, and A. F. Budd, eds. pp. 21-56. University of Chicago Press, Chicago. Collins, L. S. and A. G. Coates. 1993. Marine paleobiogeography of Caribbean Panama: last Pacific influences before closure of the Tropical American Seaway. Geological Society of America, Abstracts with Programs 25:A428-429. Donoghue, M., J. Doyle, J. Gauthier, A. Kluge, and T. Rowe. 1989. The importance of fossils in phylogeny reconstruction. Annual Review of Ecology and Systematics 20:431-460. Farris, J. S. 1988. Hennig86, version 1.5, Program and Documentation. Port Jefferson Station, New York. Fisher, D. C. 1980. The role of stratigraphic data in phylogenetic infer- ence. Geological Society of America, Abstracts with Programs 12(7):426. Fisher, D. C. 1988. Stratocladistics: integrating stratigraphic and morpho- logic data in phylogenetic inference. Geological Society of 200 America, Abstracts with Programs 20(7): A186. Fisher, D. C. 1991. Phylogenetic analysis and its application in evolution- ary paleobiology. In: Analytic Paleobiology, Short Courses in Paleontology, No. 4, N. L. Gilinsky and P. W. Signor, eds. pp. 103-122. Paleontological Society, University of Tennessee, Knoxville. Fisher, D. C. 1992. Stratigraphic parsimony. In: MacClade: Analysis of Phylogeny and Character Evolution, version 3.0, W. P. Maddison and D. R. Maddison, eds. pp. 124-129. Sinauer Associates, Sunderland, Massachusetts. Fisher, D. C. 1994. Stratocladistics: morphological and temporal patterns and their relation to phylogenetic process. /n: Interpreting the Hierarchy of Nature, L. Grande and O. Rieppel, eds. pp.133-171. Academic Press, San Diego. Foote, M. 1996. On the probability of ancestors in the fossil record. Paleobiology 22(2):141-151. Fortunato, H. and J. B. C. Jackson. 1996. Neogene speciation in (one or two) small oceans. 6th American Paleontological Convention Abstracts, Paleontological Society Special Publications 8:128. Fortunato, H. and P. Jung. 1995. The Strombina-group (Neogastropoda: Columbellidae): a case study of evolution in the neotropics. Geological Society of America, Abstracts with Programs 27:A52. Gardiner, B. 1982. Tetrapod classification. Zoological Journal of the Linnean Society 74:207-232. Gauthier, J., A. Kluge, and T. Rowe. 1988. Amniote phylogeny and the importance of fossils. Cladistics 4:105-209. Geary, D. H. 1990. Patterns of evolutionary tempo and mode in the radia- tion of Melanopsis (Gastropoda; Melanopsidae). Paleobiology 16:492-511. Gingerich, P. D. 1979. The stratophenetic approach to phylogeny recon- struction in vertebrate paleontology. In: Phylogenetic Analysis and Paleontology, J. Cracraft and N. Eldredge, eds. pp. 41-77. Columbia University Press, New York. Gingerich, P. D. 1994. Confidence limits on stratigraphic ranges: estimat- ing times of origination and extinction from the fossil record. Geological Society of America, Abstracts with Programs 26:A122. Goodman, M. 1989. Emerging alliance of phylogenetic systematics and molecular biology: a new age of exploration. Jn: The Hierarchy of Life, B. Fernholm, K. Bremer, and H. Jornvall, eds. pp. 43-61. Nobel Symposium 70, Elsevier Science Publishers, Amsterdam. Hennig, W. 1966. Phylogenetic Systematics. University of Illinois Press, Urbana. 263 pp. Hennig, W. 1981. Insect Phylogeny. John Wiley, New York. 528 pp. Huelsenbeck, J. P. 1994. Comparing the stratigraphic record to estimates of phylogeny. Paleobiology 20(4):470-483. Huelsenbeck, J. P. and B. Rannala. 1997. Maximum likelihood estimation of phylogeny using stratigraphic data. Paleobiology 23(2):174- 180. Jackson, J. B. C. and A. H. Cheetham. 1990. Evolutionary significance of morphospecies: a test with cheilostome bryozoan. Science 248:579-583. Jackson, J. B. C. and A. H. Cheetham. 1994. Phylogeny reconstruction and the tempo of speciation in cheilostome Bryozoa. Paleobiology 20(4):407-423. Jackson, J. B. C., P. Jung, A. G. Coates, and L. S. Collins. 1993. Diversity and extinction of tropical American mollusks and the emergence of the Isthmus of Panama. Science 60:1624-1626. Jackson, J. B. C., P. Jung, and H. Fortunato. 1996. Paciphilia revisited: transisthmian evolution of the Strombina-group (Gastropoda: Columbellidae). In: Evolution and Environment in Tropical America, J. B. C. Jackson, A. G. Coates, and A. F. Budd, eds. pp. AMER. MALAC. BULL. 14(2) (1998) 234-270. University of Chicago Press, Chicago. Jung, P. 1986. Neogene paleontology in the northern Dominican Republic. 2. The genus Strombina (Gastropoda: Columbellidae). Bulletins of American Paleontology 90(324): 1-42. Jung, P. 1989. Revision of the Strombina-group (Gastropoda: Columbellidae), fossil and living. Distribution, biostratigraphy and systematics. Mémoires Suisses de Paleontologie 111:1-298. Marshall, C. R. 1990. Confidence intervals on stratigraphic ranges. Paleobiology 16(1):1-10. Marshall, C. R. 1994. Confidence intervals on stratigraphic ranges: partial relaxation of the assumption of randomly distributed fossil hori- zons. Paleobiology 20(4):459-469. Marshall, C. R. 1995. Stratigraphy, the true order of species’s originations and extinctions, and testing ancestor-descendant hypotheses among Caribbean Neogene bryozoans. Jn: New Approaches to Speciation in the Fossil Record, D. H. Erwin and R. L. Anstey, eds. pp. 207-235. Columbia University Press, New York. Marshall, C. R. 1997. Confidence intervals on stratigraphic ranges with nonrandom distributions of fossil horizons. Paleobiology 23(2):165-173. Nelson, G. J. 1989. Origin and diversification of teleostean fishes. New York Academy of Sciences Annals 167:18-30. Norell, M. A. 1993. Tree based approaches to understanding history: com- ments on ranks, rules, and the quality of the fossil record. American Journal of Science 293-A:407-417. Norell, M. A. and M. J. Novacek. 1992a. The fossil record and evolution: comparing cladistic and paleontologic evidence for vertebrate his- tory. Science 255:1690-1693. Norell, M. A. and M. J. Novacek. 1992b. Congruence between superposi- tional and phylogenetic patterns: comparing cladistic patterns with the fossil record. Cladistics 8:319-337. Novacek, M. J. 1992a. Fossils, topologies, missing data, and the higher level of phylogeny in eutherian mammals. Systematic Biology 41(1):58-73. Novacek, M. J. 1992b. Fossils as critical data for phylogeny. Jn: Extinction and Phylogeny, M. J. Novacek and Q. D. Wheeler, eds. pp. 46-88. Columbia University Press, New York. Patterson, C. 1981. Significance of fossils in determining evolutionary relationships. Annual Review of Ecology and Systematics 12:195- 223. Stanley, S. M., K. L. Wetmore, and J. P. Kennet. 1988. Macroevolutionary differences between two major clades of Neogene planktonic Foraminifera. Paleobiology 14:235-249. Suter, S. 1994. Cladistic analysis of cassiduloid echinoids: trying to see the phylogeny for the trees. Biological Journal of the Linnean Society 53:31-72. Wagner, P. J. 1995. Stratigraphic tests of cladistic hypothesis. Paleobiology 21(2):153-178. Wagner, P. J. and D. H. Erwin. 1995. Phylogenetic tests of speciation hypotheses. In: New Approaches to Speciation in the Fossil Record, D. H. Erwin and R. L. Anstey, eds. pp. 87-122. Columbia University Press, New York. Wei, K.-Y. 1994. Stratophenetic tracing of phylogeny using SIMCA pat- tern recognition technique: a case study of the late Neogene planktic Foraminifera Globoconella clade. Paleobiology 20:52- 65. Winer, B. J. 1971. Statistical Principles in Experimental Design, 2nd ed. McGraw-Hill, New York. 907 pp. Date of manuscript acceptance: 20 July 1998 Review of shell reduction and loss in traditional and phylogenetic molluscan systematics, with experimental manipulation of a negative gain character Paula M. Mikkelsen Department of Invertebrates, American Museum of Natural History, Central Park West at 79th Street, New York, New York 10024-5192, U.S. A., mikkel @amnh.org Abstract: Modern phylogenetic methods are improving our basis for molluscan systematics and our understanding of evolutionary processes. The use of traditional characters in a phylogenetic analysis helps us directly contrast their accustomed “taxonomic value” with synapomorphies suggested by a clado- gram. While most cladistic characters are structurally complex, opisthobranch and pulmonate gastropods exhibit numerous characters which are losses - in shell, operculum, radula, etc. - some as presumed synapomorphies for higher-level taxa. These losses, called negative gains, can be complete (absence) or partial (reduction). To describe and code such characters, we are forced to assess morphology that is not observable. How we do so can affect tree topology, and thus the final hypothesis. This in turn determines what sequences of character evolution are supported, what monophyletic clades are recognized, and if translated into a hierarchical classification, what established taxa are confirmed or rejected. Negative gain characters must be explicitly defined to ensure repeatability; in particular, “reduced” must consider the type of reduction (in size or composition), the level of reduction (expressed qualitatively or numeri- cally), and presumed homologies of an unobservable feature. Inclusion of negative gain characters in an analysis can document the extent of putative paral- lelism or “trends” for character loss in a lineage; still, subjective coding decisions have profound effects. These points are illustrated here by three datasets derived from recent literature, on sacoglossan and notaspidean opisthobranchs and on sigmurethran pulmonates, in which the shell exists in fully present, reduced, and fully absent states. By manipulating only shell characters in these multi-system datasets, through different a priori assumptions and coding alternatives (binary, multistate unordered, multistate ordered, uncoded/mapped), changes in the resulting cladogram(s) were induced, ranging from extreme, to slight, to unchanged. In the absence of methodological preference, the use of multiple methods is advised, with conclusions based on all results and on confidence in carefully coded characters. Key words: cladistics, classification, Mollusca, Opisthobranchia, Pulmonata In the last decade of systematic malacology, phylo- marily discarded in favor of phylogenetic ones. genetic methodology (= cladistics) has allowed the re-eval- Whenever cladistic revision of a taxonomic group uation of traditional molluscan systematics in ways that are is attempted, two questions must be addressed: (1) are the more objective and repeatable than previously possible. traditional taxa monophyletic clades? and (2) are the tradi- Although the advantages of this technique are now well- tional taxon-defining characters synapomorphies? The most accepted, cladistic analyses often suggest dramatically dif- effective method of answering both of these questions is to ferent evolutionary relationships, and thus hierarchical tax- actually code and use traditional characters as part of the onomic classifications, than those long-considered as phylogenetic dataset. Only by doing so can the accustomed dogma. Because ranked classifications are necessary means “taxonomic value” of a character be directly compared with of convenience and communication, radical changes from the pattern of character evolution revealed by a cladogram. familiar taxonomic arrangements can place the phyloge- Cladistics depends on the inheritance of derived neticist at odds with workers in applied systematics (e. g. characters, and performs optimally when these characters education, ecology, collection management, biopolitics) as are structurally complex. But in many traditional characters well as other disciplines (e. g. neurophysiology), needing of many taxa, the derived condition is absence (= presumed proper labels for their research subjects. The systematic secondary loss of a primitively present feature) and is research community is likewise impacted when newly- called a negative gain. In cladistic terms, instead of ple- named taxa, created in a fervor to label clades, are rapidly siomorphic 0 = absent, and derived | = present, one has the Overturned, often by the same author(s) (e. g. reverse: 0 = present, | = absent. Difficulties are compound- Triganglionata Haszprunar, 1985b = Allogastropoda ed when an intermediate stage is involved, i. e. when the Haszprunar, 1985a). It is therefore prudent to be confident feature is still present but appears reduced or simplified in of cladistic results before traditional arrangements are sum- some fashion. The intermediate state changes the character American Malacological Bulletin, Vol. 14(2) (1998):201-218 201 202 from discrete (present/absent) to continuous (coded as 0 = present, | = reduced, 2 = absent), wherein the character state boundaries (especially between present and reduced) can be ambiguous. Traditional systematics is replete with examples of taxa defined (at least in part) by derived absences, e. g. rep- tiles without limbs (Serpentes; snakes). Among mollusks, Ponder and Lindberg (1997: 205) suggested that “loss of plesiomorphic structures, rather than their structural modi- fication, accounts for much of the homoplasy seen in gas- tropods.”” Examples of traditional taxonomic losses across the phylum include the entire jaw/radula complex in Bivalvia, ctenidia in Scaphopoda and Heterobranchia, the operculum in Marginellidae and most Olividae, the radula in Pyramidellidae and Retusidae, and jaws in Neogastropoda (see Boss, 1982; Willan, 1987; South, 1992; Salvini-Plawen and Steiner, 1996). Nearly half of the tradi- tional taxonomic characters of Cephalaspidea (Opisthobranchia; “bubble snails”) involve reduction or loss (= total reduction) of a feature (Mikkelsen, 1993). Reduction of the shell is probably the most-cited example, and has presumably occurred many times throughout the course of molluscan evolution. Reasons proposed for such losses include: ecophenotypic (streamlining for burrowing, flexibility for slithering into crevices), physiological (scarcity of calcium in the environment, constraints of para- sitic life), and developmental (miniaturization, paedomor- phosis). [Fong et al. (1995: 251) suggested that loss might have less to do with adaptation than with “indirect selec- tion” when there is “relaxation of ... stabilizing selection” on a character; they viewed the reductive process as evolu- tionarily polarized, from nonfunctionality, through atrophi- cation, to complete loss.] Regardless of the actual underly- ing cause, when shell reduction is considered ‘tan evolu- tionary trend” or characteristic of a taxon, it becomes a negative gain character in cladistics. Although negative gain characters are not impossi- ble to accommodate in cladistics, they can be “especially problematic” (Bieler, 1992: 315; Mikkelsen, 1993). First, because homology centers on structure for recognizing homologous states of a character (by at least one definition; see reviews by Hall, 1994), negative gains are more diffi- cult to code because there is nothing (in the case of absent features) to interpret. Second, because the term “reduced” can encompass a broad range of factors (reduced in size, thickness, sculpture, complexity, function, etc.; Pogue and Mickevich, 1990; Proctor, 1996), one must assure that the reduced state of one taxon is the same (homologous?) with that of the other taxa being investigated; thus precise defin- ition of kind and amount of reduction is required. Third, how we code absence is controversial: although unknown or inapplicable character states are most often treated as question marks in datasets, Pimentel and Riggins (1987) Ps AMER. MALAC. BULL. 14(2) (1998) advocated that absence is a valid character state when it is apomorphic. Finally, because reduction or loss of a feature can conceivably occur more than once in a lineage, nega- tive gains carry the threat of increased homoplasy (= extra steps in an analysis, reflected in a character consistency index of < 1.0). So, negative gain characters raise recurrent uncertainties about: (1) homology (is the present feature homologous with the absent feature?), (2) definition (how do we describe something we can’t see?), (3) procedure (how do we code absence and/or reduction?), and (4) rela- tive usefulness (should we omit such characters from the analysis if they are inherently homoplastic?). Focussing on the molluscan shell, the goals of this study are: (1) to review shell reduction and loss as an example of a traditional qualitative character in mollusks; (2) to discuss how shell reduction and loss can be used as a negative gain character for cladistic analysis, in terms of implied homologies, requisite definition, and inherent diffi- culties; (3) using published datasets, to show how results and conclusions can change with experimental manipula- tion of shell reduction and loss characters; (4) to consider the pros and cons of coding and analytical alternatives; and (5) to emphasize the cladistic utility of negative gain char- acters, as means of hypothesizing the occurrence of homo- plasy and testing accustomed taxonomic value. DEFINITION AND REVIEW OF SHELL REDUCTION A large external shell composed of calcium carbon- ate and secreted by the mantle is synapomorphic for, and thus plesiomorphic within, conchiferan mollusks (Brusca and Brusca, 1990; Lindberg and Ponder, 1996). Absence (= loss) of the shell is therefore apomorphic within the group (Ponder and Lindberg, 1997). Ontogenetic evidence sup- ports this interpretation; a shell is present in the larval form of most shell-less mollusks. Loss of the adult shell has pre- sumably occurred in parallel in octopod Cephalopoda and a number of times in two major gastropod lineages: Opisthobranchia (including some Anaspidea, Notaspidea, and Sacoglossa, and all Nudibranchia and Gymnosomata, see Gosliner and Ghiselin, 1984; Gosliner, 1991), and Pulmonata (including especially Soleolifera and Philomycidae; see South, 1992). Opisthobranchs are most renowned in this regard; as noted by Solem (1974: 117), “The one clear trend in evolution [of opisthobranchs] is toward loss of the shell.” Several lineages of mollusks show “the most clearly defined trend in shell variation” (Solem, 1974: 16) where the derived loss is less than complete, i. e. where the shell is present but reduced. A reduced (= vestigial, rudimentary) shell is listed as a traditional character for higher taxa from MIKKELSEN: SHELL REDUCTION IN PHYLOGENETICS 203 the subclass (e. g. Opisthobranchia) to family (e. g. Teredinidae) level (Table 1). However, shell reduction can encompass: (1) reduced in size relative to the overall size of the body (also generally meaning that the mollusk cannot fully retract into its shell), and/or (2) reduced in thickness, i. e. thin-walled, fragile, and/or weakly calcified. While not qualifying as shell reduction alone, other recurring attri- butes of reduced shells include: (1) auriculiform or plate- like shape, with an oversized body whorl (= rapidly expanding whorls), and enlarged aperture; (2) streamlined with regard to reduction in spines, ribs, and other sculpture; and (3) completely internal or incompletely internalized by hypertrophied, overlying mantle folds. Evolutionary modification from shell present to reduced, then, can represent different evolutionary path- ways, and caution must be used that one is coding only one kind of reduction within a single transformation series. For example, the large, paper-thin shell of a Haminoea (Opisthobranchia: Cephalaspidea: Haminoeidae) and the small, thick one of a Fissurellidea (Vetigastropoda: Fissurellidae) can both be called reduced, therefore theoreti- cally they could both be identically coded. Yet the character state change is not identical - one is reduced by becoming Table 1. Supraspecific taxa listing “shell reduced” [but not meaning “absent’] as a traditional character. Based on descriptions by Boss (1982) unless otherwise noted. (S, reduced in size; T, reduced in thickness). “Prosobranchs” Fissurellidae: Fissurellidea group (McLean, 1984) S Naticidae: Sininae ST Lamellariidae ST Carinariidae ST Opisthobranchs Cephalaspidea ST or T Runcinoidea ST Philinoglossacea ST Sacoglossa: Oxynoidae ST Anaspidea: Notarchidae ST Notaspidea S Pulmonates Amphibulimidae ST Limacidae S Testacellidae S Bivalves Galeommatoidea ST or T Teredinidae S Cephalopods Coleoidea (also Brusca and Brusca, 1990) ST Teuthoidea (also Brusca and Brusca, 1990) ST Loliginidae ST Octopoda (also Brusca and Brusca, 1990) ST Cirrata ST Opisthoteuthidae ST Incirrata ST Boltaenidae ST thinner, the other by becoming smaller. In most taxa, the shell is actually reduced in both size and thickness (Table 1). Thus the transformation is a mixture of two different evolutionary pathways that can occur together or indepen- dently. Although the shell itself is homologous within Mollusca, the reduced state of one shell might not be homologous with the reduced state of another. Likewise the absent state of one shell (e. g. Bursatella, an anaspid) is not necessarily homologous with the absent state of another (e. g. a nudibranch). Proctor (1996: 144) recognized this problem by stating that “losses of a character state may be falsely homologized, since although there may be many independent losses of a character state, seldom are there structural or behavioral clues to this independence.” In- depth examination of molluscan larval shells, which are present in nearly all shell-less groups (Thompson, 1976), could provide ontogenetic evidence for decision-making in this area. In addition to defining the pathway of reduction, the limits of reductive character states must be clearly defined. How much smaller or thinner does the shell have to be, to be coded as reduced rather than as fully present? Will levels of reduction be coded as separate character states, and if so, are these meaningful levels, or arbitrary cut-offs within a continuous morphocline? Several solutions have been uti- lized in the past: capacity of the animal to withdraw into the shell (Boss, 1982); and relative sizes of the shell and man- tle, expressed qualitatively or as a numerical ratio (McLean, 1984; Willan, 1987; Bieler and Mikkelsen, 1992). In summary, for coding and analyses to be repeat- able, defining the shell condition (i. e. the kind and degree of reduction) is a fundamental step. To minimize a priori reasoning, shell condition must be as explicitly defined as possible, ideally without the use of imprecise terms such as vestigial or rudimentary. CODING ALTERNATIVES AND METHODS In this study, published datasets were used to illus- trate if and how a change in coding adult shell reduction can produce different results. For these purposes, I am assuming that reduction has been in each case rigorously defined as required above, limiting the reduced state to one homologous pathway and one level. In each case dataset, four alternative coding choices are used, reflecting different a priori assumptions and phylogenetic philosophies. Binary Coding (two separate binary characters: (1) 0 = present, I = absent; and (2) 0 = fully present, I = reduced): In this method, taxa with shells absent (character 1 = 1) are coded “?” (= unknown or inapplicable) for char- acter 2 because the appearance of something that is not pre- sent cannot be determined (Maddison, 1993). [This is not 204 AMER. MALAC. BULL. 14(2) (1998) the same as additive binary coding, which divides a multi- state character into subcharacters and produces the same cladogram as additive (= ordered) multistate character (Hauser and Presch, 1991; Wiley et al., 1991).] Multistate Unordered Coding (single multistate character: 0 = fully present, I = reduced, 2 = fully absent): This coding avoids the **?”’s necessary with Binary Coding (above). All characters were treated as fully unordered (= non-additive, minimally connected), that is, considering any character state change (0 to 1, | to 2, 0 to 2, plus reversals) as possible in a single step. Multistate Ordered Coding: Same coding as the previous, but this character (only) was treated as a linearly ordered character (= additive, maximally connected). This assumes that reduced is a requisite intermediate stage between fully present and fully absent, therefore, e. g. a change from 0 to 2 requires two steps instead of one. All other multistate characters in each dataset were left unordered. Uncoded/Mapped: Here shell condition was not coded or used in the analysis, but was subsequently mapped onto the resultant tree(s). This method has been used by authors when data are missing in a large number of taxa (Mikkelsen, 1996), or when negative gain characters are thought to be overly homoplastic (Ponder and Lindberg, 1996, 1997; although one survey [Proctor, 1996] found that potential homoplasy has seldom been cited as a reason to exclude characters). This reflects a decision, a priori, not to allow the character to play a role in tree construction. Each trial or test dataset, then, consisted of four analyses: Binary Coding, Multistate Unordered Coding, Multistate Ordered Coding, and Uncoded/Mapped. Each used the parsimony-based algorithms of Hennig86 (Farris, 1988), using in each trial an algorithm which resolved in a reasonable amount of time and yielded a total number of trees which could be rapidly analyzed (i. e. no memory overflows). Because this is a demonstration, it was not important that these be rigorous analyses, only comparable ones. The same algorithm was used within each trial set of four analyses; all characters were given equal weight (= unweighted) in all analyses. Uncoded/Mapped analyses used the binary datasets, but with shell-reduction characters inactivated within Hennig86. Character analysis was assist- ed using Clados (Nixon, 1992). Cladograms were rendered for publication using Component (Page, 1993). Cladistic analyses most often result in more than one, often many, most-parsimonious trees (MPTs). Although “one would ideally examine the implications for character evolution on all equally acceptable phylogenies” (Maddison, 1991: 315), most authors condense the MPTs either through consensus trees or successive approxima- tions weighting (Carpenter, 1988). Because these decrease the amount of information revealed by the analysis, an alternative method of summarizing topologies was used here. When a series of taxa occur in a consistent region of all trees, but in varying arrangements within that region, repetitive regional topologies can be identified that are much smaller in number than the total number of trees. Each regional topology can be examined separately for implied patterns of character evolution. For this discussion, only those regions in which character state changes relevant to shell reduction or loss occurred are presented in full. Other regions are abbreviated here, and the arrangement of taxa comprising each such region (although admittedly not unimportant to tree length and construction) was generally disregarded. TEST DATASET RESULTS Although manipulation of hypothetical datasets in studies such as these are often instructive, one is inevitably left wondering how similar manipulations would affect real data. Therefore, the early choice was made here to conduct these trials on actual datasets. Unfortunately, very few pub- lished datasets have used shell reduction and loss as coded characters, perhaps for the reasons cited above. Some stud- ies have included a shell present/reduced character (e. g. Bieler and Mikkelsen [1992] coded 0 = subequal to mantle, 1 = significantly smaller than mantle; Jensen [1996a] used 0 = large, 1 = small), but to be most effective here, datasets using taxa with shells in at least three possible states (pre- sent/reduced/absent) were desired. Experimental results using three test datasets are here presented: sacoglossan (Jensen, 1996b) and notaspidean opisthobranchs (Willan, 1987), and sigmurethran pulmonates (Tillier, 1989). Each dataset included characters from a wide range of soft anato- my systems (e. g. mantle cavity, alimentary tract, nervous system, reproductive system) in addition to the shell char- acters. Summaries of the results in each section (below) emphasize: the number of MPTs, monophyletic clades sup- ported (including traditional taxa, for ease of discussion), and interpretation of shell reduction/loss evolution from the cladogram(s). NOTE: Although these experimental trials used real datasets, the object was neither to revise the results of the original published version nor to criticize the original analysis. These matrices were analyzed here as experimental datasets for demonstration purposes only, and these results should not be interpreted as rigorous phyloge- netic reanalyses of taxa. Support for named clades is men- tioned only to illustrate changes in result depending on cod- ing alternative used. SACOGLOSSAN OPISTHOBRANCHS The opisthobranch subgroup Sacoglossa (= Ascoglossa, “‘leaf-slugs’’) includes members with full-sized MIKKELSEN: SHELL REDUCTION IN PHYLOGENETICS 205 shells, others with shells reduced in size (both of these cate- gories are reduced in thickness), and many without shells as adults. Shell presence/absence has played a traditional role in sacoglossan classification, with two groups, Oxynoacea and Placobranchacea, containing shelled and unshelled forms, respectively. In 1996, Jensen published a phylogenetic analysis of the Sacoglossa, including 35 ingroup taxa (mostly at genus-level) and 52 characters from the shell, mantle cavi- ty, gross morphology, circulatory, digestive, reproductive, and nervous systems, and egg mass. Her dataset (Jensen, 1996b: table 4) included five taxa with fully present shells, three with reduced shells, and 27 without shells as adults. Her character list treated the shell using three binary char- acters: (1) 0 = present, 1 = absent; (2) 0 = univalved, | = bivalved; and, (3) 0 = covering whole body, | = reduced. Character 2 accomodated shell condition in the famous “bivalved gastropods” (Julia, Berthelinia) and was not manipulated here. Characters | and 3 reflect shell loss or reduction in size, and were those involved in manipulation. Because Jensen’s character list and data matrix (Jensen, 1996b: tables 3-4) were used here largely unchanged from the published version (altered only by adding an all-zero character O and correcting character 21 for Mourgona to 0; K. R. Jensen, pers. comm., 1996), they are not reproduced here. The Hennig86 algorithms mhennig* and bb* (multi- ple passes plus branch-swapping) were used for these analyses. Several monophyletic clades were consistent and are abbreviated here for discussion: (1) 27 unshelled taxa hereafter combined as Placobranchacea; (2) the bivalved gastropods, Julia + Berthelinia, hereafter abbreviated as Juliidae; and (3) Oxynoe + Lobiger + Roburnella (usually united as a clade), hereafter (when monophyletic) as Oxynoidae. Variation within monophyletic Oxynoidae was disregarded, as was one taxon, Cylindrobulla, which was consistently basal and unresolved with the outgroup (Jensen’s “‘Ancestor’’). Binary Coding. Analysis of the original dataset produced 56 MPTs of length 155 (CI 0.41, retention index [RI] 0.76). These fell into eight topologies (Figs. 1-8: 1, 14 MPTs; 2, 12 MPTs; 3-5, eight MPTs each; 6-8, two MPTs each). Shell reduction (character 3 = 1) occurred indepen- dently of shell loss (character | = 1) (i. e. a branched char- acter state tree) in six of the eight topologies (Figs. 3-8, 30 of 56 MPTs, or 54%). However, in spite of independent binary coding, shell reduction was prerequisite to shell loss (i. e. a linear character state tree) in the remaining two topologies (Figs. 1-2), which included the most frequently- occurring topologies (14 and 12 MPTs, respectively) and nearly half of the MPTs (26 of 56, or 46%). The traditional Oxynoacea was supported by four topologies (Figs. 5-8, 14 MPTs or 25%). Multistate Unordered Coding. This method com- bined binary characters | and 3 to form a single multistate shell character: 0 = present, | = reduced, 2 = absent. The combined character replaced character 1, and character 3 was thus eliminated. The algorithm likewise produced 56 MPTs, in the same topologies as Binary Coding, and of nearly identical statistics (length 155, CI 0.41, RI 0.75). Character state trees were therefore also unchanged from Binary Coding. Multistate Ordered Coding. Using the same dataset as Multistate Unordered Coding, this analysis was run with character | (only) ordered. 26 MPTs resulted, of length 155 (CI 0.41, RI 0.76). The result comprised only two topologies, identical in form (Figs. 1-2) and number (14 + 12 MPTs, respectively) to those requiring shell reduc- tion prerequisite to loss in the previous two analyses (i. e. only those with linear character state trees). Unlike the pre- vious two cases, traditional Oxynoacea was not supported by any of the resultant trees. Uncoded/Mapped. This analysis used the original (binary) dataset, but with characters | and 3 inactivated. The result was 60 MPTs of 153 steps (CI 0.40, RI 0.75), including 56 trees of the same eight topologies realized by the previous analyses (Figs. 1, 14 MPTs; 2, 12 MPTs; 3-5, eight MPTs each; 6-8, two MPTs each), plus two new topologies (Figs. 9-10, two MPTs each). In these two new topologies (Figs. 9-10), the Oxynoidae became unresolved (apparently in the absence of its single synapomorphy, shell reduced [it was also united by two homoplastic character state changes in the nervous system]). When mapped onto the trees, shell loss remained synapomorphic for the Placobranchacea, in evidence of support for the clade by three other synapomorphies (in vascular and reproductive characters) plus other homoplas- tic character state changes (in the nervous system). Shell reduction was independent of shell loss in 30 MPTs (Figs. 3-8, 50%); reduction was prerequisite to loss in the remain- ing 30 MPTs (50%), including the two new topologies (Figs. 9-10) and, again, the most frequently-occurring topologies (Figs. 1-2). Notably the two new topologies (Figs. 9-10) required reversals in shell reduction, from shell reduced back to fully present in Juliidae. Traditional Oxynoacea was supported by four topologies (Figs. 5-8, 14 MPTs or 23%). Summary. (1) Change in coding from binary to multistate produced no change in the results if the analysis was run unordered. Ordering the shell reduction/loss char- acter resulted in fewer MPTs, restricted to those which required shell reduction as an intermediate step before shell loss; in this case, this was a subset of the Binary or Multistate Unordered Coding results. Eliminating the shell reduction/loss characters produced more MPTs than any other coding alternative. (2) Regardless of coding alterna- 206 AMER. MALAC. BULL. 14(2) (1998) 3 4 oh NO Outgroup Outgroup Outgroup Outgroup Ascobulla Ascobulla Ascobulla Ascobulla Volvatella Volvatella Volvatella Volvatella JULIDAE JULIIDAE JULIDAE JULODAE e e A OXYNOIDAE 4 OXYNOIDAE OXYNOIDAE OXYNOIDAE @ e e A ® A S. : PLACOBRANCHACEA ry PLACOBRANCHACEA 7 PLACOBRANCHACEA > PLACOBRANCHACEA Ol © N 0 Outgroup Outgroup Outgroup Outgroup Ascobulla Ascobulla Ascobulla Ascobulla Vohatella VoWvatella Volvatella JULODAE e JULIDAE JULIIDAE JULIIDAE A OXYNOITDAR e e ; e OXYNOIDAE OXYNOIDAE OXYNOIDAE Volvatella e e e e PLACOBRANCHACEA PLACOBRANCHACEA PLACOBRANCHACEA PLACOBRANCHACEA © —_ Oo Outgroup Outgroup Ascobulla Ascobulla Figs. 1-2: Ui oe Volvatella Volvatella ie) JULIIDAE S suLaDAE Fi 3-8 a zs 1gZs. 3-8: Oxynoe Oxynoe & U ™ee Lobiger Lobiger - Roburnella : Rob iH . kare Figs. 9-10: U——-e-—~-$ = PLACOBRANCHACEA = PLACOBRANCHACEA Figs. 1-10. Sacoglossan opisthobranch test dataset results and character state trees. Figs. 1-8. Most-parsimonious tree topologies from Binary, Multistate Unordered, and Multistate Ordered Coding. Figs. 9-10. Same, from Uncoded/Mapped results. (*, shell reduction; ¢*, shell loss; 0, reversal to shell fully pre- sent; arrow, position of shell reduction if Juliidae shell is reduced before becoming bivalved [see text]). tive, identical traditional taxa were generally supported. Placobranchacea consistently formed a clade even if shell reduction/loss was eliminated from tree construction. Oxynoacea was supported by most MPTs, but was unre- solved in two topologies where shell reduction/loss charac- ters were eliminated. (3) Regardless of coding alternative, shell reduction/loss characters usually served as synapo- morphies for various clades: shell absent for Placobranchacea, and shell reduced for Oxynoidae (most cases). (4) Regardless of coding alternative, tree topology was generally sustained, even in Uncoded/Mapped analy- ses. (5) Shell reduction was prerequisite to shell loss in about 50% of the MPTs produced by each analysis, and in the most-frequently occurring topologies of each analysis. This was true even in unordered analyses. (4) Reversals occurred only in the Uncoded/Mapped analysis, involving only the Juliidae regaining a fully present shell from the reduced shell of Oxynoacea. This hypothesis could be envi- sioned if the bivalved gastropod shells of Juliidae could only be derived from a reduced shell; Kay (1968) in fact noted the close “approximation” of a single juliid valve to the reduced shell of Lobiger. It is interesting to note that the node at which shell reduction occurred can be easily shifted in each tree to accommodate the Juliidae (Figs. 1-8, arrows); if this is done, shell reduction becomes prerequi- site to shell loss in two more topologies (Figs. 3-4), with reflective shifts from branched to linear character state trees. NOTASPIDEAN OPISTHOBRANCHS Willan (1987) published one of the first phyloge- netic analyses involving opisthobranchs, in his investiga- tion of the Notaspidea (= Pleurobranchomorpha, “side- gilled slugs’). Here again, condition of the shell has played a traditional role in classification (Boss, 1982; Marcus, 1984, 1985), including forms with external cap-shaped shells (Umbraculacea), and those with reduced or absent MIKKELSEN: SHELL REDUCTION IN PHYLOGENETICS 207 11 12 13 14 Outgroup Outgroup Outgroup Outgroup 3 3 Tylodina Tylodina Umbraculum Umbraculum Anidolyta Anidolyta Tylodina Tylodina 3 Oo : Umbraculum Umbraculum 3 Anidolyta Anidolyta PLEUROBRANCHACEA PLEUROBRANCHACEA PLEUROBRANCHACEA PILEUROBRANCHIACEA Outgroup 1 6 Outgroup 1 Li Outgroup UMBRACULACEA UMBRACULACEA UMBRACULACEA @ PLEUROBRANCHAEINAE . PLEUROBRANCHAEINAE = PLEUROBRANCHAEINAE Bathyberthella Pleurobranchus Bathyberthella 2 2 Berthella Berthella Pleurobranchus Pleurobranchus Bathyberthella Berthella Berthellina Berthellina Berthellina e e Pleurehdera Pleurehdera Pleurehdera Figs. 11-17. Notaspidean opisthobranch test data results. Figs. 11-14. Most-parsimonious tree topologies, basal region. Figs. 15-17. Same, top region. (2, shell internalization; 3, shell decalcification; *, shell reduction; ¢*, shell loss; 0, reversal to shell calcified (character 3); arrow, position of shell internalization (character 2) if Pleurobranchaeinae are included; double slash, demarcation between basal and top tree regions). shells (Pleurobranchacea: Pleurobranchidae: Pleuro- branchinae [reduced in some] and Pleurobranchaeinae [lost in all]). Willan’s (1987: tables 3-5) original dataset included characters of the shell, mantle cavity, head-foot, digestive system, and reproductive system. The 57 characters in 11 genus-level taxa (Willan, 1987: table 5) included four bina- ry shell characters: (1) O = present, | = absent; (2) 0 = external, | = internal; (3) 0 = calcified, 1 = uncalcified; and (11) 0 = mantle and shell subequal in size, | = mantle larger than shell. While characters 2 and 3 are attributes common- ly noted in reduced shells (see above), characters | and 11 are expressions of shell loss and reduction, respectively, and were the only ones manipulated here. Following minor modifications to facilitate the parsimony-based Hennig86 analysis (see Notes in Table 2), the resulting list of 41 char- acters (Table 2) was subjected to the exhaustive Hennig86 algorithm, ie* (implicit enumeration). Each analysis produced the same number of trees (12 MPTs) of comparable length and tree statistics, with repetitive topology in two regions. The basal part of the tree, including the taxa of traditional Umbraculacea (Tylodina, Anidolyta, and Umbraculum), appeared in four topologies of three MPTs each (Figs. 11-14). Traditional Tylodinidae (Tylodina + Anidolyta) and Umbraculacea formed monophyletic clades in only one topology or three MPTs (25%) each (Figs. 13 and 11, respectively). The “top” of the tree, comprising the eight taxa of Pleurobranchacea (and Pleurobranchidae) consistently placed the two traditional subfamilies, Pleurobranchaeinae and Pleurobranchinae, as monophyletic sister-groups. The three taxa of Pleurobranchaeinae (Pleurobranchella + Pleurobranchaea + Euselenops) formed a consistent mono- phyletic clade, hereafter combined as Pleurobranchaeinae. The five taxa of Pleurobranchinae (Bathyberthella + Berthella + Pleurobranchus + Berthellina + Pleurehdera) appeared in three topologies of four MPTs each (Figs. 15- 17). Within this, Berthellina + Pleurehdera formed another consistent monophyletic clade (traditionally unnamed). Shell absent (character | = 1) was consistently 208 AMER. MALAC. BULL. 14(2) (1998) Table 2. Notaspidean taxon list (with shell condition), character list, and data matrix (based on Willan, 1987). TAXON LIST Gastropoda Heterobranchia Opisthobranchia Notaspidea Umbraculacea Fam. Tylodinidae Tylodina - present, external, uncalcified Anidolyta - present, external, uncalcified Fam. Umbraculidae : Umbraculum - present, external, calcified Pleurobranchacea Fam. Pleurobranchidae Subfam. Pleurobranchinae Pleurobranchus - present, internal, uncalcified Berthella - present, internal, uncalcified Bathyberthella - present, internal, uncalcified Pleurehdera - reduced, internal, uncalcified Berthellina - reduced, internal, uncalcified Subfam. Pleurobranchaeinae Pleurobranchella - absent Pleurobranchaea - absent Euselenops - absent CHARACTER LIST . Dummy all-zero. . Shell: 0, present; 1, absent (coding reversed). . Shell: 0, external; 1, internal. . Shell: 0, calcified; 1, uncalcified. . Periostracum: 0, smooth; 1, rough or lamellate. . Muscle scar: 0, incomplete; |, intermediate suspensor present; 2, complete. . Shell shape: 0, circular; 1, rectangular. . Shell location: 0, central; 1, anterior; 2, posterior (Willan character 9, coding adjusted). . Shell size relative to body size: 0, large; 1, medium; 2, small (Willan character 10). . Shell size relative to mantle size: 0, subequal; 1, mantle larger than shell (Willan character 11). . Mantle: 0, smooth; 1, pustulose; 2, puckered (Willan character 12). . Mantle spicules: 0, absent; 1, present (Willan character 13). . Mantle border, anteriorly: 0, entire; 1, weakly emarginate; 2, deeply cleft (Willan character 14). . Mantle margin: 0, entire; 1, slightly crenulate; 2, deeply serrate (Willan character 16). . Mantle and oral veil: 0, separate; 1, fused (Willan character 18). . Oral tentacles: 0, separate; 1, joined by oral veil (Willan character 21). . Oral veil width relative to body: 0, very narrow; 1, narrow; 2, moderately broad; 3, very broad (Willan character 22). . Oral veil papillae: 0, absent; 1, present along anterior edge (Willan character 23). . Rhinophores: 0, separated; 1, together, unfused; 2, together, fused (Willan character 24). . Pedal gland: 0, absent; 1, present (Willan character 27). . Gill location: 0, well back, posterior right; 1, posterior right; 2, from left corner to posterior midline (Willan character 31). . Gill attachment, extent: 0, half length; 1, less than half length; 2, almost entire length (Willan character 32). . Gill rachis: 0, smooth; 1, pustulose (Willan character 33). . Anus relative to gill basement membrane: 0, at middle; 1, in front of hind end; 2, above hind end; 3, well behind gill (Willan character 34). . Median buccal gland: 0, absent; 1, present (Willan character 38). . Radular rachidian teeth: 0, present; 1, absent (Willan character 39; coding reversed). . Radular lateral teeth, denticle at base: 0, absent; 1, present (Willan character 40). . Radular lateral teeth: 0, not lamellate; 1, lamellate (Willan character 42). . Labial cuticle: 0, two separate thickenings (jaws); 1, continuous thickened ring (Willan character 43; coding reversed). . Mandibular elements: 0, oval or polygonal; 1, cruciform (Willan character 44; coding reversed). . Mandibular elements, blades: 0, denticulate; 1, smooth (Willan character 45; coding reversed). . Reproductive condition: 0, monaulic; 1, diaulic; 2, triaulic (Willan character 46). . Penial autospermal groove: 0, present; 1, absent (Willan character 48; coding reversed). . Penis location: 0, at base of right oral tentacle; 1, anterior midline; 2, on right side in front of gill (Willan character 49; coding adjusted). . Penis: 0, non-protrusible; 1, protrusible (Willan character 50). (Continued) ontinue MIKKELSEN: SHELL REDUCTION IN PHYLOGENETICS 209 Table 2. (continued) 35. Allosperm receptacles: 0, two; 1, one (Willan character 52; coding reversed). 36. Receptaculum seminis, origin: 0, low; 1, high (Willan character 53; coding reversed). 37. Prostate gland: 0, surrounding autosperm canal; 1, absent; 2, present as distinct organ (Willan character 54; coding adjusted). 38. Penial gland: 0, absent; 1, present (Willan character 55). 39. Penial sack: 0, absent; 1, present (Willan character 56). 40. Vas deferens, coiling within penial sack: 0, absent; 1, present (Willan character 57). Notes: Changes to Willan’s original coding were: 1. Autapomorphies were removed (characters 7, 15, 19, 20, 25, 26, 28, 30, 35, 36, 41, 47, and 51). 2. Inexplicably, some character states were originally numbered 0-1-2-etc., while others were numbered 1-2-3-etc. The latter characters (3, 5, 8-12, 14, 16, 17, 22, 24, 31-34, 39, 40, 45, 46, 49, 52, 54, and 55) were renumbered here so that the most plesiomorphic state was 0. With this accomplished, character 17 became all 0’s, so was eliminated here. 3. In comparing the original coding scheme (Willan, 1987: table 4) with the character list (including plesiomorphic and apomorphic states), some of Willan’s apomorphic character states were found inappropriately coded as 0. Coding was here adjusted (usually simply reversed) for characters 1, 9, 37, 39, 43, 44, 45, 48, 49, 52, 53, and 54. Following this recoding, character 37 became an autapomorphy and was eliminated. 4. Some taxa were originally listed as having multiple character states. These were coded here with the most plesiomorphic state. When this was done, char- acters 8 and 29 became all 1’s and dashes (inapplicable), so were eliminated. 5. A probable error was noticed in one of the original shell characters (character 3, shell calcified/uncalcified), where the text claimed “only Umbraculum calcifies its shell to any degree” (Willan, 1987: 220) while Umbraculum and all other taxa except Pleurobranchus and Bathyberthella were originally coded as uncalcified. This character was recoded here according to Willan’s statement, with the outgroup and Umbraculum as calcified (0) and all remaining taxa as uncalcified (1; although Marcus, 1985, noted that the shell of Tylodina includes a calcified part); the last three taxa in the matrix were coded here as “inapplicable” (-). 6. An all-zero outgroup was used, and an all-zero character 0 was added. DATA MATRIX Outgroup ONO0DDD0DD00D000000000000000000000000000000000 Tylodina 000 101000000000 10000000200101— 00201-0000 Anidolyta 0001 100000000101000000020101 1— 10201-0000 Umbraculum 00001 20010000200— 10220301001—-01—- 2000 Berthella 001100100001 10011021110011000112101111100 Pleurobranchus 00110010002120011021111211000101101111000 Berthellina 001 1001121010001 10201 10211010112101000100 Pleurehdera 00110011110-00011021110211110102101010100 Bathyberthella 0011001000000001 1021 110211000002101010000 Pleurobranchella Ol-------- 10- 01121001101110000011011- 2011 Pleurobranchaea Ol-------- 10- 01121011101100000011011- 2011 Euselenops Ol-------- 00- 0113101111111000011101101000 synapomorphic for Pleurobranchaeinae (Figs. 15-17). Shell reduced (character 9 = Willan’s character 11 = 1) was a consistent synapomorphy for the Berthellina + Pleurehdera clade, and was never prerequisite to shell loss (consistently a branched character state tree as in that given for Figs. 3- 8). Shell internalization (character 2 = 1) was synapomor- phic for Pleurobranchinae, but the sister clade, Pleurobranchaeinae, was coded ‘*?” for this character, so the character state change could be a synapomorphy for Pleurobranchacea (Figs. 15-17, arrows). Shell decalcifica- tion (character 3 = 1) was less definitive, occurring below Tylodina in all top-region topologies (and synapomorphic for a large clade above Umbraculum in two topologies, Figs. 13-14), but requiring reversals in Umbraculum in two topologies (Figs. 11-12). Binary Coding and Multistate Unordered Coding. These two analyses produced 12 MPTs of 76 steps (CI 0.72, RI 0.73). The Multistate Unordered Analysis com- bined characters | and 9 into one character: 0 = present (and subequal in size relative to mantle), 1 = reduced (in size relative to mantle), and 2 = absent. This replaced char- acter 1, and character 9 was eliminated from the data matrix. Multistate Ordered Coding. This analysis used the same data matrix as Multistate Unordered Coding, but the analysis was run with character | (only) ordered. The resulting 12 MPTs were of 77 steps (CI 0.71, RI 0.73). Uncoded/Mapped. This analysis used the binary dataset in Table 2 but with characters 1 and 9 deactivated within Hennig86. The resulting 12 MPTs were of 74 steps (CI 0.71, RI.0.72). Summary. Regardless of coding alternative, there was no change in resultant topologies using the notaspidean dataset (although the tree length and statistics did vary slightly). Interpretation of monophyletic clades and evolu- tion of shell reduction/loss therefore did not change either. 210 AMER. MALAC. BULL. 14(2) (1998) SIGMURETHRAN PULMONATES Like opisthobranchs, pulmonates include represen- tatives with full, reduced, or absent shells; those with reduced shells are called either semislugs or slugs accord- ing to degree of reduction of the visceral mass (slugs hav- ing completely internalized shells, and semislugs having external shells that are too small to retract into; T. Pearce, pers. comm.; those lacking shells are also called slugs; Solem, 1974). Unlike in opisthobranchs, however, the degree of shell reduction does not play a traditional role in classification. The shells of land slugs have been judged of limited taxonomic value because of their variability (Reuse, 1983), and because of the convergence “inherent in” reduced shells (Solem, 1978). Terrestrial gastropods there- fore present a similar-but-different case for this analysis. Unfortunately, no robust phylogenetic analysis including a dataset involving all shell morphotypes (pre- sent/reduced/absent) suitable for this manipulation has been published. The best available data come from Tillier’s (1989) massive anatomical monograph of Stylom- matophora, which presented character coding for taxa that were translatable into a data matrix for this demonstration. Tillier’s work included nearly 200 taxa (Tillier, 1989: appendix A), but only 17 characters (Tillier, 1989: appen- dix E) from relatively few organ systems (digestive, excre- tory, and nervous). To reduce the data matrix to more-nor- mal proportions of characters and taxa, a subset of taxa was chosen for this analysis. Solem (1974) noted that only the Sigmurethra includes fully shelled forms, semislugs, and slugs, so for this analysis, representatives of the sig- murethran subgroup Aulacopoda were selected from Tillier’s data. The final dataset included two suprafamilial groups, six families, and 22 genus-level ingroup taxa plus an all-zero outgroup (Table 3). Of the six families, one con- tains snails only, three contain snails plus either slugs or semislugs (or both), and two contain slugs or semislugs only. Tillier’s 17 original characters (Tillier, 1989: appendix E; as clarified by Emberton and Tillier, 1995) were trans- formed into cladistic characters as noted in Table 3 (see Notes). No shell reduction character was originally coded by Tillier, although taxa were indicated as “semislug” or “slug” where applicable (Tillier, 1989: appendix B). For this analysis, four character states were used to code the shell as: fully present, reduced (semislug), reduced (slug), and absent (determined in part from familial descriptions in Boss, 1982). Two reduced categories were used to preserve the distinction between the shells of semislugs and slugs, assuming (perhaps in oversimplification) that the shell’s condition in slugs is further reduced from that in semislugs. The lists of taxa (22) and characters (20, with shell reduc- tion in binary form) and the data matrix appear in Table 3. Hennig86’s algorithm, mhennig* (multiple passes without branch-swapping), was used for all analyses (more robust algorithms completed, but produced extraordinarily large numbers of trees). Because of the extreme variability in the results of this analysis, repeating topologies could not be identified. In general, suprafamilial and familial groups were not supported, and because the focus of this demon- stration lies in the pattern(s) of character evolution implied by the various results, only examples and character state trees will be presented. Binary Coding. Analysis of the dataset produced a single MPT (Fig. 18) of 76 steps (CI 0.46, RI 0.57). Shell reduction (character 2 = | or 2) occurred in parallel three times. Two of these paths were direct unmodified changes to slug- (character 2 = 2) or semislug-type (character 2 = 1) reduction. The third was a path through semi-slug type reduction to two changes to slug-type reduction and one to shell loss (character | = 1). Multistate Unordered Coding. This analysis com- bined binary characters | and 2 to form a single multistate shell character: 0 = present, 1 = reduced (semislug), 2 = reduced (slug), 3 = absent. The combined character replaced character 1; character 2 was eliminated. The algo- rithm produced a single MPT of exactly the same topology, length, and statistics as that produced in Binary Coding (Fig. 18). Multistate Ordered Coding. This analysis used the same modified dataset as Multistate Unordered Coding, but the analysis was run with character | (only) ordered. This scenario thus presupposed that shell reduction occurs in this lineage in the following linear order: semislug (reduced) to slug (reduced) to shell absent. The analysis produced seven MPTs of 77 steps (CI 0.45, RI 0.57). Six of the seven trees produced a character state tree and topology similar to that in Fig. 19. The character state tree reflected the linear ordering, but included one reversal from semislug-type reduction back to unreduced. The seventh MPT was similar except lacked the reversal, through relocating the shelled snails to the basal region of the tree. Uncoded/Mapped. With characters | and 2 deacti- vated from the data matrix in Table 3, a single MPT was produced of 70 steps (CI 0.45, RI 0.55). When mapped on the tree, shell reduction characters produced a generally more complex pattern than in the previous methods (Fig. 20). Reduction occurred three times in parallel, two of these being direct unmodified changes to slug-type reduc- tion. The third pathway was semislug-type reduction lead- ing to slug-type reduction once, shell loss once, and rever- sals to unreduced shells twice. Summary. (1) Although Binary and Multistate Unordered Coding produced the same result, Multistate Ordered Coding produced more trees (all different from the previous); Uncoded/Mapped again produced one tree, but again of a unique topology. (2) Suprageneric groups (i. e. MIKKELSEN: SHELL REDUCTION IN PHYLOGENETICS 24 traditional superfamilies and families) were inconsistent, therefore monophyletic clades differed drastically depend- ing on coding alternative used. (3) Shell reduction charac- ters formed synapomorphies for at least one multi-taxon clade in all trees, but the supported clades were again dependent upon coding alternative. (4) The hypothesis of evolution of shell reduction and loss (expressed in character state trees) was highly dependent upon coding alternative. Ordering of the character states (unreduced to semislug- type reduction to slug-type reduction to loss) never occurred except when forced by Multistate Ordered Coding. DISCUSSION WHAT IS THE “CORRECT” METHOD? Each of the above trials illustrated four ways of handling one traditional character in a cladistic dataset. The results ranged from differing greatly among methods (pul- monates), to differing slightly (sacoglossans), to not differ- ing at all (notaspideans). The degree of difference corre- sponded here to what qualitatively appears to be relative strength of the total dataset (especially of non-shell charac- ters therein), expressed by (in pulmonate, sacoglossan, and notaspidean datasets, respectively) (1) the Hennig86 algo- rithm that would readily resolve (m*, m*bb*, and ie*), (2) range of CI (0.36-0.37, 0.40-0.41, and 0.71-0.72), (3) the amount of overlap among the trial results (very little, some, and total), and (4) the amount of change caused by deacti- vation of the shell reduction/loss characters (much shorter, but more MPTs; slightly shorter, slightly more MPTs; and no change). Interestingly, one other dataset manipulated in this same manner reacted like the notaspidean dataset - ie* algorithm, total overlap of results, no change caused by deactivating the character in question. This was the trun- catellid gastropod dataset published by Rosenberg (1996), manipulating a gill character originally coded as O = pre- sent, 1 = reduced, and 2 = absent; this result can also be attributed to the strength of the original dataset. Binary versus Multistate Coding. Hauser and Presch (1991) theorized that an unordered analysis contain- ing multistate characters should produce exactly the same results as a binary analysis. This was universally true here for all three datasets. Binary coding of shell reduction/loss characters requires the use of missing character states (= unknown or inapplicable question marks) in the dataset. According to the literature, these can increase the number of MPTs (Wilkinson, 1995) or can cause other missing-data prob- lems attributed to long-distance influence (Maddison, 1993). However in these test datasets, the presence of ‘““?”’s did not appear to induce problems, perhaps because these were “safe circumstances” as described by Maddison (1993: 579) where the “inapplicable” regions of the trees were confined to a single taxon (pulmonates) or clade (sacoglossans and notaspideans). Other arguments against Binary Coding include: (1) binary characters could be redundant or not completely independent (Pimentel and Riggens, 1987; Maddison, 1993), and (2) if there is in fact evidence for an ordered transformational character, this evidence will be forfeited through binary coding and can be lost to the result (Pogue and Mickevich, 1990; Lipscomb, 1994). Ordered versus Unordered. Hauser and Presch (1991) reanalyzed 27 published datasets to test the results of ordered versus unordered characters; their ordered analy- ses affected all multistate characters in each dataset, rather than only one as in the test cases here. In agreement with their results, ordering did not demonstrably improve clade resolution, and the number of trees was not necessarily reduced by ordering. Slowinski (1993) reanalyzed 21 pub- lished matrices and found similar results: unordered analy- ses resulted in overall greater resolution and greater con- gruence. Hauser and Presch (1991: 253) noted that “the effect of ordered characters ... is, in part, based on their interaction with other characters in the datamatrix” so that each individual change to a dataset affects multiple levels. Nevertheless, achieving the best resolved and smallest number of trees out of a dataset is not acceptable rationale for ordering characters. Ordering restricts possible charac- ter state transformations and requires independent, corrobo- rating evidence that such a pattern occurred (or that others did not) (i. e. ontogeny; Hauser and Presch, 1991). Slowinski (1993) noted that linear ordering (the most common method) does not always convey the best possible assumption of transformation for multistate char- acters with four or more states. Here, linear ordering of the shell reduction/loss multistate character in the pulmonate dataset (Fig. 19) presupposed that semislug-type shell reduction preceded slug-type reduction, which in turn pre- ceded shell loss. In all other pulmonate results (Figs. 18, 20), this was not the case as evidenced by branched charac- ter state trees; shell loss was most often preceded only by semislug-type reduction. Choosing an unordered analysis, or “letting the algorithm decide,” clearly makes the fewest a priori assumptions, and will in fact reveal an ordered transforma- tion series if it is part of the MPT(s). Mickevich and Weller (1990) agreed in part (advocating ordering in general), not- ing that the pattern of character state change revealed by an unordered analysis can test the validity of a hypothetical ordered transformation series. In the case of the sacoglos- san analysis here, the 26 MPTs produced by ordering were a subset of those (56) produced by unordering; the ordered transformation was present in the two most frequently- 212 AMER. MALAC. BULL. 14(2) (1998) Table 3. Sigmurethran pulmonate taxon list (with shell condition), character list, and data matrix. Generic abbreviations (also used in Figs. 18-20) derived from superfamilial group, family, and genus names. Families and generic contents of families are as according to Vaught (1989); suprafamilial groupings are as according to Boss (1982) and South (1992) TAXON LIST Gastropoda Pulmonata Sigmurethra Aulacopoda Arionoidea Fam. Arionidae Hemphillia (ArArHm) - reduced, semislug Aphallarion (ArArAp) - reduced, slug Oopelta (ArArOo) - reduced, slug Fam. Philomycidae Philomycus (ArPhPh) - absent, slug Fam. Endodontidae Thaumatodon (ArEnTh) - present, snail Phrixgnathus (ArEnPh) - present, snail Limacoidea Fam. Helicarionidae Hemiplecta (LiHeHe) - present, snail Microparmarion (LiHeM1) - reduced, semislug Cystopelta (LiHeCy) - reduced, semislug Mariaella (LiHeMa) - reduced, slug Fam. Urocyclidae Trochozonites (LiUrTr) - present, snail Trochonanina (LiUrTc) - present, snail Acantharion (LiUrAc) - reduced, semislug Mesafricarion (LiUrMe) - reduced, semislug Atoxon (LiUrAt) - reduced, slug Elisolimax (LiUrE]) - reduced, slug Fam. Zonitidae Trochomorpha (LiZoTr) - present, snail Zonites (LiZoZo) - present, snail Phenacolimax (LiZoPh) - reduced, semislug Vitrinopsis (LiZoVi) - reduced, semislug Daudebardia (LiZoDa) - reduced, slug Plutonia (LiZoP1) - reduced, slug CHARACTER LIST 0. Dummy all-zero. 1. Shell: 0, present; 1, absent. 2. Shell: 0, fully present; 1, reduced (semislug); 2, reduced (slug). 3. Buccal mass (BM): 0, spheroidal to ovoidal tending toward cylindrical (BM1); 1, clearly cylindrical (BM2). 4. Esophageal crop (OC): 0, absent (OC1); 1, separated from gastric crop by distinct portion of esophagus (OC2); 2, separated from gastric crop by simple constriction (OC3); 3, as in OC3 but extending forward to nerve ring (OC4). 5. Gastric crop (SC): 0, cylindrical (SC1); 1, median portion inflated (SC2); 2, funnelform, widening from esophagus to stomach (SC2’); 3, anterior region inflated (SC3). 6. Gastric pouch (PS): 0, joining gastric crop without any constriction, distinctly wider than crop (PS1); 1, joining gastric crop without any constriction, slightly wider or no wider than crop (PS2); 2, separated from gastric crop by constriction, distinctly wider than crop (PS2’). 7. Intestine length (IL): 0, intestinal loops long, reaching level between distal limit of gastric pouch and middle of gastric crop (IL1); 1, intestine shorter, but loops distinct (IL2); 2, intestinal loops long, reaching proximally at least to level of distal limit of gastric pouch (IL2’); 3, intestinal loops reduced to almost flat sigmoid (IL3). 8. Ratio of kidney length:lung length (LR): 0, very short kidney, 0.45-0.7 (LR1); 1, 0.36-0.45 (LR2); 2, 0.7-1.0 (LR2’); 3, 0.25-0.36 (LR3); 4, very long kid- ney, 0.0-0.25 (LR4) [slugs and semislugs not scored, fide Emberton & Tillier, 1995: 203]. 9. Degree of closure of ureter (UR): 0, no closed retrograde ureter = complete mesurethry (UR1); 1, closed ureter reaching at most lung top (UR2); 2, ureteric tube reaching point between lung top and pneumostome (UR3); 3, ureteric tube reaching pneumostome = fully closed = complete sigmurethry (UR4). 10. Internal morphology of kidney (RR): 0, either two distinct regions (distal one usually lacking lamellae) or three distinct regions (median one either lack- ing lamellae or with lamellae different in appearance from those in proximal) (RR1); 1, kidney homogenous in internal morphology, with lamellae reaching distal region and level of kidney pore (RR2). 11. Cerebral commissure length (CC): 0, greater than 1.1 times right cerebral ganglion width (CC1); 1, 0.9-1.1 times right cerebral ganglion width (CC2); 2, less than 0.9 times right cerebral ganglion width (CC3). (Continued) MIKKELSEN: SHELL REDUCTION IN PHYLOGENETICS 213 Table 3. (continued) 12. Right cerebropedal connective length (CPD): 0, longer than twice right cerebral ganglion width (CPD1); 1, 1-2 times right cerebral ganglion width (CPD2); 2, shorter than right cerebral ganglion width (CPD3). 13. Ratio of lengths of left:right cerebropedal connectives (CPR): 0, < 0.9 (CPR1); 1, 0.9-1.1 (CPR2); 2, 1.1-1.5 (CPR3); 3, 1.5-2.5 (CPR4). 14. Right pleural ganglion position (PLD): 0, closer to pedal than cerebral ganglion = hypoathroid (PLD1); 1, closer to cerebral than pedal ganglion = epiathroid (PLD2). 15. Left pleural ganglion position (PLG): 0, closer to pedal than cerebral ganglion = hypoathroid (PLG1); 1, closer to cerebral than pedal ganglion = epiathroid (PLG2). 16. Visceral ganglion position relative to median plane of pedal ganglia (VG): 0, on right side (VG1); 1, in middle (VG2); 2, on left side (VG3). 17. Right parietal and pleural ganglia (PAD): 0, separate (PAD1); 1, in contact or fused (PAD2). 18. Left parietal ganglion position (PAG): 0, in contact with left pleural, or closer to it than to visceral ganglion, and separated from both by distinct connec- tive (PAG1); 1, closer to visceral than to left pleural ganglion, and separated from both by distinct connective (PAG2); 2, in contact with visceral ganglion only, and separated from left pleural by distinct connective (PAG3); 3, in contact or fused with both left pleural and visceral ganglia (PAG4). 19. Fusion of visceral ganglion (FG): 0, none (FG1); 1, with right parietal ganglion (FG2); 2, with left parietal ganglion (FG2’); 3, with both parietal ganglia (FG3). Notes: Tillier’s characters (Tillier, 1989: appendix E; as clarified by Emberton and Tillier, 1995) were transformed: 1. by changing original 1-2-2’-3 character codes (reflecting ordered character state changes, with primed states indicating branched character state trees) into linear 0-1-2-3 cladistic character states. 2. by coding any O-states in the original matrix (unexplained by Tillier and called “eliminated” by Emberton and Tillier, 1995) as “-.” 3. where more than one representative per genus was coded by Tillier, by combining them and choosing the most plesiomorphic state (but never “-”) for the cladistic data matrix. DATA MATRIX Outgrp —00000000000000000000 ArArHm § 00100311-31121000123 ArArAp 00200002-31- - - - - - - - ArArOo —00200000-31010000122 ArPhPh —01-00000-31221- - -1123 ArEnTh 00000311-10110110001 ArEnPh 00000301330212100101 LiHeHe 00000001230221001120 LiHeMi 00102-01231222 - 00133 LiHeCy 00100001 -30220 - - 0133 LiHeMa 00202111231- - - - -- - - LiUrTr 00000001 130222000120 LiUrTc 00000001030110001130 LiUrAc —00102300230211000131 LiUrMe 00102300 - - -211001130 LiUrAt 00203 -12 - 31221001130 LiUrE] 00203- 12231- - - - - - - - LiZoTr 00000201230221111110 LiZoZo 00000201031201001130 LiZoPh 00100001- 31222001131 LiZoVi 00100001230 - - - - - - - - LiZoDa_ 00210111- - -000001011 LiZoP! 00210111 - 31212002122 occurring topologies and 46% of the MPTs. Ordering only eliminated possible topologies from consideration. Uncoded/Mapped Characters. A posteriori map- ping of characters that one deems unusable is also a priori reasoning - these characters have been judged beforehand to play no role, or a conflicting role, in evolution of the group. If total evidence (usually combining molecular plus morphological characters in a single dataset) is the best accepted method of determining phylogenetic relationships (Kluge, 1989; de Queiroz et al., 1995), then the same should also be preferred at the morphological level, i. e. incorporating as many morphological characters (including negative gain characters) as possible. TRANSFORMING PHYLOGENETIC ANALYSES INTO CLASSIFICATIONS Revising taxonomic classification is usually not the sole (nor primary) question being approached using cladis- tics. Nevertheless, it is always tempting to translate resul- tant tree(s) following an analysis. Much of the discussion 214 AMER. MALAC. BULL. 14(2) (1998) Outgroup ArArOo ArEnTh ArEnPh LiUrTe LiUrAc LiUrMe LiHeHe LiUrTr LiZoZo LiZoTr ArArAp LiUrEl LiUrAt LiHeMa LiZeDa LiZoPI LiZoPh LiHeMi ArArHm ArPhPh LiHeCy LiZoVi Outgroup LiZoDa LiUrTe LiZoZo ArArAp LiUrEl LiUrAt LiHeHe LiZoTr ArPhPh ArArHm ArArOo LiHeMa LiZoPI LiHeMi LiZoPh LiUrTr LiHeCy LiZoVi ArEnTh ArEnPh LiUrAc LiUrMe 19 Outgroup ArEnTh ArEnPh LiUrTc LiZoZo LiUrAc LiUrMe LiZoVi LiUrTr LiHeHe LiZoTr LiHeCy ArArHm LiHeMi LiZoPh LiUrEl LiUrAt ArArAp ArPhPh ArArOo LiHeMa LiZoDa LiZoP1 Figs. 18-20. Pulmonate test dataset results, sample most-parsimonious cladograms and character state trees. Fig. 18. Binary and Multistate Unordered Coding. Fig. 19. Multistate Ordered Coding. Fig. 20. Uncoded/Mapped. See Table 3 for taxon abbreviations. (*, semislug-type reduction; ¢*, slug-type reduc- tion; 0, reversal to unreduced shell; L, shell loss; U, unreduced shell). here might have centered on differences in, for example, the number of superfamilies required by one cladogram over another. But, abandoning taxonomic rank (as suggest- ed by de Queiroz and Gauthier, 1992; Ponder and Lindberg, 1997; Roth, 1997), at least above ICZN-regulated family- level (ICZN, 1985), renders this discussion irrelevant. Above family-level, it is possible to refer to a clade without worrying about whether it is, for example, a subclass or an order. Abandonment of rank furthermore eliminates the need to erect meaningless “redundant categories” (reviewed by de Queiroz and Gauthier, 1992), e. g. an order Cylindrobullacea and family Cylindrobullidae for the monotypic clade defined only by the genus-level synapo- morphies of Cylindrobulla (Jensen, 1996b). If one chooses to do so, the number of recognized taxa derived from a cladogram depends on monophyletic MIKKELSEN: SHELL REDUCTION IN PHYLOGENETICS ZS clades (not grades or paraphyletic taxa; Bieler, 1990). As shown by these experimental manipulations, the choice of coding alternative can affect this result. Deciding which monophyletic clades to recognize is still a subjective step, and depends on synapomorphies and other node-defining character state changes. For example, in the sacoglossan trials, Placobranchacea was consistently supported here. Some of the resultant topologies resulted in a large number of taxa at a level equivalent to Placobranchacea, e. g. from Fig. 2: Ascobulla, Volvatella, Juliidae, and Oxynoidae. But whether these are recognized as five monophyletic clades, or four (Ascobulla, Volvatella, Juliidae, Oxynoidae + Placobranchacea), or three (Ascobulla, Volvatella, Juliidae + Oxynoidae + Placobranchacea), or two (Ascobulla, Volvatella + Juliidae + Oxynoidae + Placobranchacea), or one (all five combined) depends on at which node(s) suffi- cient support is recognized. If a cladogram such as that in Fig. 2, suggesting a drastically new classification, is chosen as the preferred tree, the degree of confidence placed in the dataset and its coded characters should determine how to proceed. Is this a robust analysis, with strong corroborating support (e. g. high bootstrap or other statistical values; large number of characters), and where the cladogram shows little change with each added character or taxon? Or is this preliminary, with a cladogram that is likely to change topology dramati- cally as new data are obtained? If the former, then the result must be trusted, regardless of how closely the cladogram agrees with traditional classification or an initial hypothesis of the “true tree.” If the latter, then construction of a hierar- chial classification should be postponed awaiting further data. CONCLUSIONS 1. Choosing characters and how to code them is the fundamental step in cladistics - and one not without subjec- tivity. Precise definition of negative gain characters, such as shell reduction and loss, is critical, requiring more “dili- gen[ce] in the detail and consistency of terminology” (Proctor, 1996: 145) than in the case of positive gain char- acters. How to code such characters is equally critical, and as was shown, different coding alternatives can (although not always) produce different cladograms, which can be translated into different classifications - a point made earli- er by Pogue and Mickevich (1990). 2. How a data matrix is analyzed is less critical, and it is not the goal of this paper to recommend one method over another. Although Multistate Unordered Coding might be interpreted as having the most support (Pimentel and Riggens, 1987), in the absence of theoretical or procedural preference, the best alternative is to use multiple methods (agreed by Hauser and Presch, 1991; Kim, 1993; but not by Wilkinson, 1992) at least during the iterative stage of char- acter development. As noted by Kim (1993: 335), “agree- ment among trees estimated by different methods lends greater credibility to the estimates.” Ponder and Lindberg (1997) used this approach, including recoding of multistate characters in their dataset as binary characters. This is com- parable to the method commonly used as an alternative to total evidence - running separate analyses of molecular ver- sus morphological (e. g. Hillis et al., 1996; Shaffer et al., 1997) or behavioral versus morphological data (e. g. Prum, 1990), and comparing results generated by the different methods, perhaps also in combined format. 3. Homoplasy should be hypothesized from a phylo- genetic analysis, not initially assumed. Negative gain char- acters, such as shell reduction and loss, have a high likeli- hood for homoplasy and as such might be preconceived as less informative than positive gain characters in a cladistic analysis. However, Sanderson and Donoghue (1996) showed that homoplasy in a cladogram (resulting in a low CI) can co-occur with a high level of confidence (expressed in that case by high bootstrap support values), which means that potentially homoplastic characters need not automati- cally be omitted. Furthermore, homoplasy does not necessarily indi- cate error or noise. Numerous parsimony-based analyses present high levels of homoplasy (Sober, 1992; Foley, 1993), and many authors have considered homoplasy to “constitute the majority of evolutionary change during the course of evolution and diversification of lineages” (Armbruster, 1996: 227; see also Hennig, 1966, 1983; Gosliner and Ghiselin, 1984; Sluys, 1989b; Moore and Willmer, 1997). Parallelism is often invoked in interpreta- tion of cladograms (e. g. Tassy, 1988; Sluys, 1989a; Erséus, 1990; Griswold, 1993; Jensen, 1996b; Salvini-Plawen and Steiner, 1996), absences are presented as synapomorphies (Ax, 1987; Miiller and Wagner, 1991), and homoplasy occurs in nearly every cladistic study; in a reanalysis of 38 published data matrices, Maddison (1991: table 2) obtained consistency indices [CIs] ranging from 0.198-0.808 (mean = 0.515). Wilkinson (1991) opined that parsimony methods fail not when homoplasy is rampant, but when homoplasy is based on misleading evidence - therefore, careful defini- tion of characters and character states is the key. Cladistics is a tool - albeit a powerful one - in studying molluscan sys- tematics; as such, it can help us study homoplasy. Including a limited number of negative gain characters (in conjunc- tion with sufficient positive gain characters) can serve to highlight homoplasy against a background of homology on the final tree (Platnick, 1977), and is the strongest method of testing hypotheses of loss. A low Cl-value does not (nec- 216 AMER. MALAC. BULL. 14(2) (1998) essarily) mean that a final result is poor. In these cases, a homoplastic character is informative as long as its suggest- ed transformation is not biologically impossible. Armbruster (1996: 240) noted that “homoplasy is common in characters with ecological significance” and “characters of ecological importance are often of little sys- tematic utility.” These same statements can be said of shell reduction and loss in the Mollusca: “in virtually every major group of mollusks there are some species in which the shell has become reduced to a remnant” (Solem, 1974: 16). Inclusion of these characters wherever possible in analyses could lead to more rigorous documentation of the number and extent of “trends toward shell loss” throughout the phylum. 4. Revising taxonomic classification is only part of why we do cladistics, and might not be supportable from a given analysis. Cladistic analyses rarely generate fully resolved trees with all monophyletic taxa supported by strong synapomorphies. Several equally supported classifi- cations can often be inferred from a single cladogram (Tassy, 1988), depending on the choice of monophyletic groups. Although new cladistic analyses (or reanalyses) are being regularly produced, a phylogenetic classification of all Mollusca still eludes us. In the interim, we must provi- sionally accept untested and even paraphyletic taxa in mol- luscan taxonomy, and avoid proposing unstable classifica- tions that will change with the next analysis. ACKNOWLEDGMENTS I thank Gary Rosenberg (Academy of Natural Sciences of Philadelphia) for inviting me to express these views in the AMU-spon- sored symposium, “Traditional versus Phylogenetic Systematics of Mollusks,” and for encouraging me to turn my musings into manuscript. He and Tim Pearce (Delaware Museum of Natural History [DMNH], Wilmington) provided valuable inroads into the pulmonate literature. I thank Liz Shea (formerly DMNH, now Bryn Mawr College, Bryn Mawr, Pennsylvania) for planting the initial seed for this investigation during dis- cussions of a classroom assignment. Dan Geiger (University of Southern California, Los Angeles), Thierry Backeljau (Royal Belgian Institute of Natural Sciences, Brussels), Jim McLean (Natural History Museum of Los Angeles County, California), Chuck Lydeard (University of Alabama, Tuscaloosa), and Claus Hedegaard (University of Aarhus, Denmark) pro- vided valuable comments and suggestions during the symposium. Riidiger Bieler (Field Museum, Chicago) obligingly trashed a first draft, to its ulti- mate benefit. LITERATURE CITED Armbruster, W. S. 1996. Exaptation, adaptation, and homoplasy: evolution of ecological traits in Dalechampia vines. In: Homoplasy: The Recurrence of Similarity in Evolution, M. J. Sanderson and L. Hufford, eds. pp. 227-243. Academic Press, San Diego, etc. Ax, P. 1987. The Phylogenetic System: The Systematization of Organisms on the Basis of Their Phylogenesis. English version of Das phylo- genetische System (Systematisierung der lebenden Natur aufgrund threr Phylogenese), 1984, translated by R. P. S. Jeffreys. John Wiley & Sons, Chichester, United Kingdom, etc. 340 pp. Bieler, R. 1990. Letters to the Editor: Haszprunar’s “clado-evolutionary” classification of the Gastropoda - a critique. Malacologia 31(2):371-380. Bieler, R. 1992. Gastropod phylogeny and systematics. Annual Review of Ecology and Systematics 23:311-338. Bieler, R. and P. M. Mikkelsen. 1992. Preliminary phylogenetic analysis of the bivalve family Galeommatidae. American Malacological Bulletin 9(2):157-164. Boss, K. J. 1982. Mollusca. In: Synopsis and Classification of Living Organisms, S. P. Parker, ed. pp. 945-1166. McGraw-Hill, New York. Brusca, R. C. and G. J. Brusca. 1990. Invertebrates. Sinauer Associates, Sunderland, Massachusetts. 922 pp. Carpenter, J. M. 1988. Choosing among multiple equally parsimonious cladograms. Cladistics 4(3):291-296. de Queiroz, K., M. J. Donoghue, and J. Kim. 1995. Separate versus com- bined analysis of phylogenetic evidence. Annual Review of Ecology and Systematics 26:657-681. de Queiroz, K. and J. Gauthier. 1992. Phylogenetic taxonomy. Annual Review of Ecology and Systematics 23:449-480. Emberton, K. C. and S. Tillier. 1995. Letter to the editor: Clarification and evaluation of Tillier’s (1989) stylommatophoran monograph. Malacologia 36(1-2):203-208. Erséus, C. 1990. Cladistic analysis of the subfamilies within the Tubificidae (Oligochaeta). Zoologica Scripta 19:57-63. Farris, J. S. 1988. Hennig86, version 1.5. Port Jefferson Station, New York. Foley, R. 1993. Striking parallels in early hominid evolution. Trends in Ecology and Evolution 8(6):196-197. Fong, D. W., T. C. Kane, and D. C. Culver. 1995. Vestigialization and loss of nonfunctinal characters. Annual Review of Ecology and Systematics 26:249-268. Gosliner, T. M. 1991. Morphological parallelism in opisthobranch gas- tropods. Malacologia 32(2):313-327. Gosliner, T. M. and M. T. Ghiselin. 1984. Parallel evolution in opistho- branch gastropods and its implications for phylogenetic methodol- ogy. Systematic Zoology 33(3):255-74 Griswold, C. E. 1993. Investigations into the phylogeny of the lycosoid spiders and their kin (Arachnida: Araneae: Lycosoidea). Smithsonian Contributions to Zoology 539:1-39. Hall, B. K. 1994. Homology. the Hierarchical Basis of Comparative Biology. Academic Press, San Diego, etc. 483 pp. Haszprunar, G. 1985a. The fine morphology of the osphradial sense organs of the Mollusca. 2: Allogastropoda (Architectonicidae, Pyramidellidae). Philosophical Transactions of the Royal Society 307B:497-505. Haszprunar, G. 1985b. The Heterobranchia - a new concept of the phy- logeny of the higher Gastropoda. Zeitschrift fiir phylogenetische Systematik und Evolutionsforschung 23(1):15-37. Hauser, D. L. and W. Presch. 1991. The effect of ordered characters on phylogenetic reconstruction. Cladistics 7(3):243-265. Hennig, W. 1966. Phylogenetic Systematics. English translation of Grundzeuge einer Theorie der phylogenetischen Systematik, 1950, D. D. Davis and R. Zangerl, trans. University of Illinois Press, Urbana. 263 pp. Hennig, W. 1983. Stammesgeschite der Chordaten. Verlag Paul Parey, Hamburg. 208 pp. MIKKELSEN: SHELL REDUCTION IN PHYLOGENETICS 217 Hillis, D. M., B. K. Mable, and C. Moritz. 1996. Applications of molecu- lar systematics: the state of the field and a look to the future. /n- Molecuiar Systematics, 2nd ed., D. M. Hillis, C. Moritz, and B. K. Mable, eds. pp. 515-543. Sinauer Associates, Sunderland, Massachusetts. International Commission on Zoological Nomenalcture (ICZN). 1985. International Code of Zoological Nomenclature, 3rd ed. International Trust for Zoological Nomenclature, London, and University of California Press. Berkeley and Los Angeles. 338 pp Jensen, K. R. 1996a. The Diaphanidae as a possible sister group of the Sacoglossa (Gastropoda, Opisthobranchia). In: Origin and Evolutionary Radiation of the Mollusca, J. D. Taylor, ed. pp. 231- 247. Oxford University Press, Oxford, United Kingdom. Jensen, K. R. 1996b. Phylogenetic systematics and classification of the Sacoglossa (Mollusca, Gastropoda, Opisthobranchia). Philosophical Transactions of the Royal Society of London B 351:91-122. Kay, E. A. 1968. A review of the bivalved gastropods and a discussion of evolution within the Sacoglossa. In: Symposia of the Zoological Society of London, 22, Studies in the Structure, Physiology & Ecology of Molluscs, V. Fretter, ed. pp. 109-134. Academic Press, New York. Kim, J. 1993. Improving the accuracy of phylogenetic estimation by com- bining different methods. Systematic Biology 42(3):331-340. Kluge, A. G. 1989. A concern for evidence and a phylogenetic hypothesis of relationships among Epicrates (Boidae, Serpentes). Systematic Zoology 38(1):7-25 Lindberg, D. R. and W. F. Ponder. 1996. An evolutionary tree for the Mollusca: branches or roots? In: Origin and Evolutionary Radiation of the Mollusca, J. D. Taylor, ed. pp. 67-75. Oxford University Press, Oxford, United Kingdom. Lipscomb, D. 1994. Cladistic Analysis Using Hennig86. George Washington University, Washington, D. C. 122 pp. Maddison, D. R. 1991. The discovery and importance of multiple islands of most-parsimonious trees. Systematic Zoology 40(3):315-328. Maddison, D. R. 1993. Missing data versus missing characters in phyloge- netic analysis. Systematic Biology 42(4):576-581. Marcus, Ev. d. B.-R. 1984. The western Atlantic warm water Notaspidea (Gastropoda, Opisthobranchia), parte 2. Boletim de Zoologico, Universidad de Sao Paulo, Brazil 8:43-76. Marcus, Ev. d. B.-R. 1985. Catalogue of the western Atlantic warm water Opisthobranchia. 10. The western Atlantic warm water Notaspidea (Gastropoda, Opisthobranchia), part 3. Umbraculacea. Boletim de Zoologico, Universidad de Sao Paulo, Brazil 9:1-15. McLean, J. H. 1984. Shell reduction and loss in fissurellids: a review of genera and species in the Fissurellidea group. American Malacological Bulletin 2:21-34. Mickevich, M. F. and S. J. Weller. 1990. Evolutionary character analysis: tracing character change on a cladogram. Cladistics 6(2):137-170. Mikkelsen, P. M. 1993. Monophyly versus the Cephalaspidea (Gastropoda, Opisthobranchia) with an analysis of traditional cephalaspid characters. Bolletino Malacologico 29:115-138. Mikkelsen, P. M. 1996. The evolutionary relationships of Cephalaspidea s.1. (Gastropoda: Opisthobranchia): a phylogenetic analysis. Malacologia 37:375-442. Moore, J. and P. Willmer. 1997. Convergent evolution in invertebrates. Biological Reviews 72:1-60. Miiller, G. B. and G. P. Wagner. 1991. Novelty in evolution: restructuring the concept. Annual Review of Ecology and Systematics 22:229- 256. Nixon, K. C. 1992. Clados, version 1.2. L. H. Bailey Hortorium, Cornell University, Ithaca, New York. Page, R. D. M. 1993. Component, version 2.0. The Natural History Museum, London Pimentel, R. A. and R. Riggins. 1987. The nature of cladistic data. Cladistics 3(3):201-209. Platnick, N. I. 1977. Parallelism in phylogeny reconstruction. Systematic Zoology 26:93-96. Pogue, M. G. and M. F. Mickevich. 1990. Character definitions and char- acter state delineation: the bete noire of phylogenetic inference. Cladistics 6:319-361. Ponder, W. F. and D. R. Lindberg. 1996. Gastropod phylogeny - chal- lenges for the 90s. In: Origin and Evolutionary Radiation of the Mollusca, J. D. Taylor, ed. pp. 135-154. Oxford University Press, Oxford, United Kingdom Ponder, W. F. and D. R. Lindberg. 1997. Towards a phylogeny of gastro- pod molluscs: an analysis using morphological characters. Zoological Journal of the Linnean Society 119:83-26S. Proctor, H. C. 1996. Behavioral characters and homoplasy: perception ver- sus practice. In: Homoplasy: The Recurrence of Similarity in Evolution, M. J. Sanderson and L. Hufford, eds. pp. 131-149. Academic Press, San Diego, etc. Prum, R. O. 1990. Phylogenetic analysis of the evolution of display behavior in the neotropical manakins (Aves: Pipridae). Ethology 84:202-231 Reuse, C. 1983. On the taxonomic significance of the internal shell in the identification of European slugs of the families Limacidae and Milacidae (Gastropoda, Pulmonata). Biologisch Jaarboek Dodonaea 51:180-200. Rosenberg, G. 1996. Independent evolution of terrestriality in Atlantic truncatellid gastropods. Evolution 50(2):682-693. Roth, B. 1997. Popular delusions, phantom taxa, and the weirdness of ranks [abstract]. Program and Abstracts, Combined Annual Meeting, American Malacological Union and Western Society of Malacologists, 21-27 June 1997, Santa Barbara, California: 51 Salvini-Plawen, L. v. and G. Steiner. 1996. Synapomorphies and ple- siomorphies in higher classification of Mollusca. In: Origin and Evolutionary Radiation of the Mollusca, J. D. Taylor, ed. pp. 29- 51. Oxford University Press, Oxford, United Kingdom. Sanderson, M. J. and M. J. Donoghue. 1996. The relationship between homoplasy and confidence in a phylogenetic tree. In: Homoplasy: The Recurrence of Similarity in Evolution, M. J. Sanderson and L. Hufford, eds. pp. 67-89. Academic Press, San Diego, etc. Shaffer, H. B., P. Meylan, and M. L. McKnight. 1997. Tests of turtle phy- logeny: molecular, morphological, and paleontological approach- es. Systematic Biology 46(2):235-268. Slowinski, J. B. 1993. “Unordered” versus “ordered” characters. Systematic Biology 42(2):155-16S. Sluys, R. 1989a. Phylogenetic relationships of the triclads (Platyhelminthes, Seriata, Tricladida). Bijdragen tot de Dierkunde 59(1):3-25. Sluys, R. 1989b. Rampant parallelism: an appraisal of the use of nonuni- versal derived character states in phylogenetic reconstruction. Systematic Zoology 38(4):350-370. Sober, E. 1992. Parsimony. /n: Keywords in Evolutionary Biology, E. F. Keller and E. A. Lloyd, eds. pp. 249-254. Harvard University Press, Cambridge, Massachusetts. Solem, A. 1974. The Shell Makers: Introducing Mollusks. John Wiley & Sons, New York, etc. 289 pp. Solem, A. 1978. Classification of the land Mollusca. Jn: Pulmonates, Vol. 2A, Systematics, Evolution and Ecology, V. Fretter and J. Peake, eds. pp. 49-97. Academic Press, London, etc South, A. 1992. Terrestrial Slugs: Biology, Ecology and Control. Chapman & Hall, London, etc. 428 pp. Tassy, P. 1988. The classification of Proboscidea: how many cladistic classifications? Cladistics 4(1):43-57. 218 AMER. MALAC. BULL. 14(2) (1998) Tillier, S. 1989. Comparative morphology, phylogeny and classification of land snails and slugs (Gastropoda: Pulmonata: Stylommatophora). Malacologia 30(1-2):1-303. Thompson, T. E. 1976. Biology of Opisthobranch Molluscs, Vol. 1. The Ray Society, London. 207 pp., 21 pls. Vaught, K. C. 1989. A Classification of the Living Mollusca. American Malacologists, Melbourne, Florida. 195 pp. Wiley, E. O., D. Siegel-Causey, D. R. Brooks, and V. A. Funk. 1991. The Compleat Cladist: A Primer of Phylogenetic Procedures. University of Kansas, Museum of Natural History [Special Publication 19], Lawrence, Kansas. 158 pp. Wilkinson, M. 1991. Homoplasy and parsimony analysis. Systematic Zoology 40(1):105-109. Wilkinson, M. 1992. Ordered versus unordered characters. Cladistics 8:375-385. Wilkinson, M. 1995. Coping with abundant missing entries in phylogenet- ic inference using parsimony. Systematic Biology 44(4):501-514. Willan, R. C. 1987. Phylogenetic systematics of the Notaspidea (Opisthobranchia) with reappraisal of families and genera. American Malacological Bulletin 5(2):215-241. Date of manuscript acceptance: 26 May 1998 Reproducibility of results in phylogenetic analysis of mollusks: a reanalysis of the Taylor, Kantor, and Sysoev (1993) data set for conoidean gastropods Gary Rosenberg Academy of Natural Sciences, 1900 Benjamin Franklin Parkway, Philadelphia, Pennsylvania 19103 U. S. A., rosenberg @acnatsci.org Abstract: Reanalysis of the Taylor, Kantor, and Sysoev (1993) data set on conoidean gastropods failed to reproduce their results. Taylor et al. found more than 900 trees of length 189; reanalysis yielded 32,700 trees of length 187. The number of trees they found was limited by the memory available on the computer used for the analysis. The Taylor ef al. consensus tree omitted the stated outgroup Benthobia (Pseudolividae); reanalysis including the outgroup yielded 3,149 trees of length 193, in all of which Benthobia fell within the ingroup. Strict and majority-rule consensus trees differed considerably in topology from those with Benthobia excluded. Reanalysis excluding the hypothetical ancestor, whose character states Taylor ef al. determined in part by ingroup analysis, yielded additional topologies of consensus trees. Only eight of 38 clades in the Taylor er al. tree appeared in all three strict consensus trees; 17 clades were not supported by any of the majority rule consensus trees. All three majority-rule consensus trees did support the transfer to Conidae by Taylor et al. of the turrid subfamilies Clathurellinae, Conorbinae, Oenopotinae, Mangeliinae, Daphnellinae, and Taraninae. This clade, however, did not appear in two of the strict consensus trees, so support for it is equivocal. Additional problems with the analysis include incorrect character mappings, use of characters primarily from one organ system, conflicts between text and data matrix, choice of taxa, and inclusion of data from taxa not included in the cladistic analysis in formulating the classification. The Taylor er al. data set does not support strong inferences about conoidean phylogeny, and there is not yet convincing evidence for abandoning the traditional classification of the group. Nonetheless, their data are an immensely valuable contribution to be built on as information about conoidean taxa, characters, organ systems, and outgroups accumulates. Key words: cladistics, critique, Conoidea, anatomy, classification Taylor, Kantor, and Sysoev (1993) presented a enor- There is always a danger that such a reanalysis will mous body of anatomical data on conoidean gastropods, be viewed as a personal attack, although the debate along with a phylogenetic analysis of their data. The phylo- between Bieler (1990) and Haszprunar (1990) over genetic tree they presented is reproduced here (Fig. 1), and Haszprunar’s (1988) “clado-evolutionary” classification of will be referred to hereafter as the TKS tree, from the the Gastropoda shows that salutary results are possible. It authors’ initials. Based on their analysis, Taylor et al. intro- was therefore with some trepidation that I approached Dr. duced a revised classification of the superfamily Conoidea, John Taylor and Dr. Yuri Kantor, at a much later date than I which differs in a number of respects from the traditional should have (July 1998), to ask for clarification of their classification that recognizes three families, Conidae, methods. They asked me to keep in mind that the analysis Turridae, and Terebridae. In particular, the turrid subfami- published in 1993 was performed in 1991, and that comput- lies Clathurellinae, Conorbinae, Oenopotinae, Mangeliinae, ing power and the sophistication of cladistic methods have Daphnellinae, and Taraninae were transferred to Conidae, both advanced considerably in the intervening seven years. and Drilliidae, Pseudomelatomidae, and Strictispiridae, for- With their cooperation, I have been able to include their merly subfamilies of Turridae, were elevated to familial comments about specific methodological points. status. Kohn and McLean (1994) critiqued the classifica- The question of whether the results of a phylogenet- tion, but did not try to replicate the analysis. I attempted to ic analysis can be reproduced can be addressed at several reproduce the results of Taylor et al., by reanalyzing their levels. Given the same data and the same assumptions, can data, but was unable to do so. I found many more, and the results be replicated? How much do the results change shorter trees, than they did, but because they gave insuffi- with different assumptions? These are essentially questions cient detail about methods, it was difficult to tell why we of the mechanics of the analysis and its internal consisten- had gotten different results. cy. Equally important, but much harder to answer: would a American Malacological Bulletin, Vol. 14(2) (1998):219-228 219 N N —_) AMER. MALAC. BULL. 14(2) (1998) Ancestor Clavus U oa 1 Splendrillia C | Drilliidae Hastula B pervieeca T ] Terebridae 2 4 Ouplicaria C Pseudomelatoma Pseudomelatomidae 7 Toxiclionella T Clavatulinae 8 Strictispira P Strictispiridae 6 Clavatula D Clonsia Clavatulinae 9 Clavulata S 1 Turricula N 10 Funa L en of re VeIGiaRaC Crassispirinae 5 14 PilsbryspiraN —— Zonulispirinae Aforia A ——————— Cochlespirinae 16 Lophiotoma L 17 Polystira A | Turrinae Micantapex P 20 Borsonia O ] Clathurellinae 15 21 Ophiodermella 19 Oenopota L ————— Qenopotinae 22 Tropidoturris F PE Anarithma M | Clathurellinae 24 Glyphostoma C 18 Genota WN ; 26 Benthofascts i Conorbinae 27 Conus V ———————. Coninae 25 Tomopleura V Clathurellinae Eucithara S$ 28 0 Mangelia N | Mangeliinae 31 Mangelia P 29 Thatcheria M Philbertia P 32 34 —— Philbertia L 33 Daphnella R Daphnellinae 35 Gymnobela E 36 Teretiopsis L 37 Abyssobela A 38 Taranis M | ———— Taraninae Fig. 1. TKS tree (50% majority-rule consensus tree from Taylor ef al., 1993: fig. 27). Names at the right show the new classification proposed by Taylor et al. Taxa are, in alphabetical order [with abbreviations in brackets]: Abyssobela atoxica Kantor & Sysoev, 1986 [Abyssobela A], Aforia abyssalis Sysoev & Kantor, 1987 [Aforia A], Anarithma metula (Hinds, 1843) [Anarithma M], Ancestor (hypothetical) outgroup, Benthobia outgroup, Benthofascis biconica (Hedley, 1903) [Benthofascis], Borsonia ochraea Thiele, 1925 [Borsonia O], Clavatula caerulea (Weinkauff, 1875) [Clavatula S; Clavatula C in Figs. 2-7], C. diadema Kiener, 1840 [Clavatula D], Clavus unizonalis (Lamarck, 1822) [Clavus U], Clionella sinuata (Born, 1778) [Clionella C], Conus ventricosus Gmelin, 1791 [Conus V], Daphnella reeveeana (Deshayes, 1863) [Daphnella R], Duplicaria colorata Bratcher, 1988 [Duplicaria C], Eucithara stromboides (Reeve, 1846) [Eucithara S], Funa latisinuata (E. A. Smith, 1877) [Funa L], Genota nicklesi Knudsen, 1952 [Genota N], Glyphostoma candidum (Hinds, 1843) [Glyphostoma C], Gymnobela emertoni (Verrill and Smith, 1884) [Gymnobela E], Hastula bacillus (Deshayes, 1859) [Hastula B], Lophiotoma leu- cotropis (Adams and Reeve, 1850) [Lophiotoma L], Mangelia nebula (Montagu, 1803) [Mangelia N], M. powisiana (Dautzenberg, 1887) [Mangelia P], Micantapex parengonius Dell, 1956 [Micantapex P], Oenopota levidensis (Carpenter, 1864) [Oenopota L], Ophiodermella inermis (Hinds, 1843) [Ophiodermella], Pervicacia tristis (Deshayes, 1859) [Pervicacia T], Philbertia linearis (Montagu, 1803) [Philbertia L], P. purpurea (Montagu, 1803) [Philbertia P], Pilsbryspira nymphia (Pilsbry and Lowe, 1932) [Pilsbryspira N], Polystira albida (Perry, 1811) [Polystira A], Pseudomelatoma penicillata (Carpenter, 1864) [Pseudomelatoma], Splendrillia chathamensis Sysoev and Kantor, 1989 [Splendrillia C], Strictispira paxillus (Reeve, 1845) [Strictispira P], Taranis moerchii (Malm, 1863) [Taranis M], Teretiopsis levicarinata Kantor and Sysoev, 1989 [Teretiopsis L], Thatcheria mirabilis Angas, 1877 | Thatcheria M], Tomopleura reevii (C. B. Adams, 1850) [Tomopleura V], Toxiclionella tumida (Sowerby, 1870) [Toxiclionella T], Tropidoturris fossata notialis Kilburn, 1986 [Tropidoturris F), Turricula nelliae spurius (Hedley, 1922) [Turricula N], and Vexitomina garrardi (Laseron, 1954) [Vexitomina G). (Reproduced with permission from the Bulletin of the Natural History Museum (London).] ROSENBERG: REPRODUCIBILITY IN PHYLOGENETIC ANALYSIS 2 different data set, with different characters or different rep- resentative taxa, lead to similar phylogenetic inferences? In principle, it should, because there is only one true tree. If a phylogenetic analysis fails in either internal or external consistency, then it is not a good basis for classification. Because I am not an expert on conoidean anatomy, I have considered mainly the internal consistency of the Taylor et al. data set, rather than attempting to formulate a revised data set. METHODS Parsimony analysis was performed using PAUP 3.1.1 (Swofford and Begle, 1993) on an Apple Macintosh Centris 660AV computer. Taylor et al. used a Macintosh SE (pers. comm.) running PAUP 3.0. Three versions of the data set were studied: Analysis I, the original data matrix, exactly as presented by Taylor et al.; Analysis II, the origi- nal matrix excluding the outgroup Benthobia (Pseud- olividae); and Analysis III, the original matrix excluding Benthobia and the hypothetical ancestor. The heuristic search option of PAUP was used, with all characters unordered and equally weighted and search settings as follows: Random addition sequence (10 repli- cates); tree-bisection-reconnection branch-swapping per- formed; MULPARS in effect; steepest descent in effect; branches having maximum length zero collapsed to yield polytomies; topological constraints not enforced; trees unrooted; no trees in memory at start of run. Taylor et al. (1993: 152) also performed a heuristic search but did not state any of the settings. They were unclear about ordering of characters and did not mention weighting. Strict and majority rule consensus trees were com- puted for the minimal trees for each set of taxa analyzed, along with partition lists showing clades found in one or more trees and their frequency of occurrence. MacClade 3.0.1 (Maddison and Maddison, 1992) was used to deter- mine the length of particular trees, based on the original matrix, again with all characters unordered and equally weighted. The following tree statistics were calculated: consis- tency index (CI), homoplasy index (HI), rescaled consisten- cy index (RC), and retention index (RI). RESULTS Within each run, only one to four of ten replicates were completed before PAUP’s maximum trees limit of 32,700 was reached, terminating the run. Therefore each analysis was run four times, with different random starting seeds to give different input orders of taxa. In some cases i) — tree memory was filled before the shortest trees were found. In particular, for Analysis II, some runs terminated with trees of length 188 and others with length 187 and for Analysis III some terminated at length 182, others at length 183. The results reported here are for the runs that yielded the greatest number of shortest trees; further analysis may yield yet shorter trees. In PAUP, when a run terminates because of the maximum tree limit, any non-minimal trees in memory (awaiting branch-swapping) are automatically discarded, so the number of most parsimonious trees found may be less 32,700. Search histories, which indicate dis- carded trees, are given in the figure captions of the strict consensus trees. Analysis I Reanalysis of the original data matrix yielded 3,149 trees of length 193. Benthobia, one of the designated out- groups, fell in the ingroup in 100% of the 3,149 trees (as seen from the partition list generated by PAUP), although its position was unresolved in the strict consensus tree (Fig. 2). In the majority rule consensus tree (Fig. 3), Benthobia grouped with Duplicaria and Pervicacia (Pervicaciinae) 82% of the time. Partition analysis showed that Hastula (Terebridae) fell as outgroup to all other conoideans in 37% of the trees, and Conus was outgroup in 3%. Conus plus Hastula was outgroup to the other conoideans in 16% (3% as a Clade, 13% paraphyletic). Clades that these and the fol- lowing consensus trees shared with the TKS tree (Fig. 1) are indicated in Table 1. Analysis IT Reanalysis of the original matrix with Benthobia excluded yielded 32,700 trees of length 187. Figs. 4-5 show the strict and majority rule consensus trees. Partition analysis showed that Conus did not group within the turrids in 3% of these trees. Clavus plus Splendrillia fell as out- group 43% of the time and the three terebrids fell as out- group 15%. The TKS tree (Fig. 1), which also excluded Benthobia, is two steps longer, with length 189. Analysis III Reanalysis of the original matrix with Benthobia and the hypothetical ancestor excluded yielded 10,490 trees of length 182. Strict and majority rule consensus trees are shown in Figs. 6-7. Of interest here is that the strict consen- sus tree supports clade 18 of the TKS tree: Conidae includ- ing Clathurellinae, Conorbinae, Oenopotinae, Mangeliinae, Daphnellinae, and Taraninae. DISCUSSION Although the capability to make data and methods i) N i) Ancestor Benthobia Pseudome latoma Clavatula D Clavatula C Turricula Clionella Funa Vexitomina Pilsbryspira Toxiclionella — Aas Strictispira i ae Clavus Splendrillia Lophiotoma eee eee ee Polystira eee ee Micantapex - Borsonia Tomopleura Tropidoturris eae Oe Ophioderme lla — Anarithma Glyphostoma Eucithara Mangelia N Mangelia P Philbertia P Philbertia L Daphnella Gymno bela Teretiopsis Abyssobe la Taranis Thatcheria Benthofascis Genota ee = te —_ Pervicacia Hastula Duplicaria Fig. 2. Analysis I (including Benthobia); strict consensus of 3,149 trees, length 193, CI = 0.285, RI = 0.613, HI = 0.715, RC = 0.175. Search histo- ry: replicate 1 (seed = 357612083), 1,938 trees found (length = 193); repli- cate 2 (seed = 1222301227), 395 additional trees found (length = 193); replicate 3 (seed = 63292021), 816 additional trees found (length = 193); replicate 4 (seed = 1306497612), MAXTREES limit (32,700) hit while swapping on tree #11,591, 29,551 non-minimal trees (length = 194) dis- carded. Total number of rearrangements tried = 407682616, time used = 27:26:26. explicit is one of the advantages of cladistics over tradition- al phylogenetic methods, some practitioners of cladistic methods do not take full advantage of this capability. For example Bieler (1990: 371) said of Haszprunar’s (1988) “clado-evolutionary” classification of the Gastropoda: “The presentation of the data is incomplete and inconsistent. The analysis is not repeatable” and raised concerns about depar- ture from standard cladistic practices. Reynolds (1997: 20) raised similar issues of “parsimony application and repeata- bility” in his reanalysis of Steiner’s (1992) data set on Scaphopoda. My reanalysis of the conoidean data set fur- ther emphasizes that without full details of methods, assumptions, and limitations, results might not be replica- ble. AMER. MALAC. BULL. 14(2) (1998) Ancestor Benthobia o—_[ nee 98 Duplicaria Pseudome latoma Clavatula D 92 Clavatula C 10 100. 10 ges Turricula 6 ql ee Clionella 10 oJ Funa 10 6 — Vexitomina Pilsbryspira 79— Toxiclionella 61 Strictispira Clavus 1 aa Splendrillia Lophiotoma —_—_——_— 5 10 jl eens Polystira 87 Aforia Micantapex 5 1 Borsonia 58 Ophioderme lla Benthofascis Tropidoturris 57 Oenopota Anarithma 8 eas Glyphostoma Eucithara 57 100—| Mangelia N rt ae Mangelia P Philbertia P Philbertia L 10 5 5 Gymno bela 57 100 9 5| Teretiopsis 5 2f—— Abyssobe la 100 Taranis 56 Daphnella Thatcheria 79 Genota Tomopleura Conus Hastula 63 Fig. 3. Analysis I (including Benthobia); 50% majority-rule consensus of 3,149 trees of length 193. I will show first that my Analysis II used the same data and the same assumptions as the analysis presented by Taylor et al. (1993), but got different results because of dif- ferences in computing power. I will then consider how dif- ferent assumptions about outgroups, characters, taxa and classification might affect the reliability of the results. RECONSTRUCTING THE METHODS Data matrix and ordering of characters From their stated methods, it is not clear whether Taylor et al. intended all characters to be unordered. On page 152, they flagged 18 characters as unordered: (2, 7, 23, 25-31, 33-38, 41-42), but in their table 2, only 15 char- acters were marked as unordered (2, 7, 23, 26-28, 30-31, 34-38, 40-41). Discrepancies involve characters 25, 29, 33, 40, and 42. All multistate characters were included on one list or the other, so presumably the published analysis was run with all characters unordered. This assumption is sup- ported in that few of the multistate characters are appropri- ate for linear orderings, but no alternate orderings were sug- gested. Oddly, a few binary characters (28, 30, 33, 35-36, ROSENBERG: REPRODUCIBILITY IN PHYLOGENETIC ANALYSIS Ancestor Pseudome latoma Strictispira Clavatula D Clavatula C Turricula Clione lla Funa Vexitomina Pilsbryspira Toxiclionella Clavus Splendrillia = Lophiotoma Polystira Aforia Micantapex Borsonia Tomopleura Tropidoturris Ophioderme lla Anarithma Glyphostoma Eucithara Mangelia N Mangelia P Philbertia P Philbertia L Daphnella Gymno bela Teretiopsis Abyssobe la Taranis Thatche ria Oenopota Benthofascis Genota Conus Pervicacia Duplicaria Hastula Fig. 4. Analysis II (excluding Benthobia), strict consensus of 32,700 trees, length 187, CI = 0.294, RI = 0.617, HI = 0.706, RC = 0.182. Search histo- ty: replicate 1 (seed = 581470058), 31,351 trees found (length = 187); replicate 2 (seed = 426658209), MAXTREES limit (32,700) hit while swapping on tree #31538, 1,349 additional trees found (length = 187). Total number of rearrangements tried = 764211632, time used = 24:40:51. 41-42), which are by definition unordered, appeared on each list. The discrepancies in the lists of unordered charac- ters raise the possibility that through some error the pub- lished matrix is not the version of the matrix that was ana- lyzed to produce the TKS tree. Taylor er al. (1993: 152) referred to making “small adjustments to the data set” as they explored their data, and Taylor and Kantor (pers. comm.) stated that they tried various scorings of characters and experimented with both ordered and unordered charac- ters. Assuming that all the characters were unordered (and equally weighted), I used MacClade to determine the length of the TKS tree, which, although it is a consensus tree, is fully resolved. It does indeed have the stated length of 189 steps. Thus the published matrix is consistent with the pub- Nw NO Oo Ancestor Pseudome latoma Clavatula D Clavatula C 9 6| 96 10 glee Turricula Clione lla 96 Funa 1 00 9 | eee (eee Vexitomina 1 00 Pilsbryspira Toxiclionella Strictispira Clavus py Stencil Lophiotoma rool —— Polystira Aforia Micantapex Tropidoturris Oenopota Borsonia Ophioderme lla Tomopleura Anarithma Glyphostoma Eucithara Mangelia N Mangelia P Philbertia P Philbertia L Gymno bela Teretiopsis Abyssobe la Taranis Daphnella Thatcheria Benthofascis Genota Conus Pervicacia Duplicaria Hastula 72 Fig. 5. Analysis II (excluding Benthobia); 50% majority-rule consensus of 32,700 trees of length 187. lished tree, and with the assumption that all characters were unordered. Exclusion of the outgroup Analysis I (original data matrix) found 3,149 trees of length 193 (Fig. 2-3). Taylor et al. (1993: 152) reported “over 900” trees of length 189, four steps shorter, for this matrix. Their consensus tree (Fig. 1), however, omits Benthobia, which was selected as the outgroup being “the most primitive non-conoidean neogastropod known” (1993: 152). In all trees of length 193, Benthobia fell within the ingroup, usually grouping with the Pervicaciinae. Taylor and Kantor (pers. comm.) stated that they excluded Benthobia from the analysis because of this behavior. Thus the analysis presented by Taylor er al. corresponds to Analysis II herein. I tried forcing Benthobia to stay in outgroup posi- tion, using the topological constraints option of PAUP. A heuristic search as described above resulted in 32,700 trees 224 AMER. MALAC. BULL. 14(2) (1998) Pseudome latoma Strictispira Clavus Splendrillia Lophiotoma Polystira Aforia Micantapex Borsonia Tomopleura Tropidoturris Ophioderme lla Anarithma Glyphostoma Eucithara Mangelia N Mangelia P Philbertia P Philbertia L Daphnella Gymno bela Teretiopsis Abyssobe la Taranis Thatcheria Oenopota Benthofascis Genota Conus Pervicacia Duplicaria Hastula Clavatula D Clavatula C Turricula Clionella Funa Vexitomina Pilsbryspira Toxiclionella Fig. 6. Analysis III (excluding Benthobia and hypothetical ancestor); strict consensus of 10,490 trees, length 182, CI = 0.302, RI = 0.622, HI = 0.698, RC = 0.188. Search history: replicate | (seed = 1): 10,490 trees found (length = 182); replicate 2 (seed = 1459270432): MAXTREES limit (32,700) hit while swapping on tree #16,743, 22,210 non-minimal trees (length = 183) discarded. Total number of rearrangements tried = 403556096, time used = 16:25:50. of length 194, one step longer than without the constraint. The strict consensus and majority rule consensus trees were quite similar to those for Analysis III (Figs. 6-7), in which the outgroups were excluded, but slightly more resolved because of small differences in support for a few clades (e. g. 100% versus 96%, 53% versus 47%) and so are not shown here. Shortest trees Analysis II (with Benthobia excluded) yielded 32,700 trees of length 187, two steps shorter than those found by Taylor et al. I ran the matrix several times using different starting seeds. In my analyses, runs always termi- nated because PAUP’s maximum limit of 32,700 trees was reached. This generally took ca. 24 hr. In some cases, trees of length 187 were found in only a few minutes; in others, the shortest trees found before memory overflowed were length 188. In no replicates were trees of length 189 Pseudome latoma Strictispira Clavus 10 po See Splendrillia a. Lophiotoma 100 Polystira Aforia Micantapex — 80. a Tropidoturris +10 9 7 Oenopota Borsonia Tomopleura 75 (ae Anarithma 64 Glyphostoma Eucithara r65 10 —{ — Mangelia N 10 glee Mage ia P 10 Philbertia P Philbertia L 10 6 oJ Gymnobela 96 Teretiopsis Lf — Abyssobe la Taranis Daphnella Thatche ria Benthofascis r Genota Conus Ophioderme lla Pervicacia 100L—— Duplicaria 10 eae ee ee ee Hastula Clavatula D Clavatula C 10 1 ool ___ Turricula 6 jie Ss Clione lla 10 of Funa 6 7 | —— Vexitomina Pilsbryspira Toxiclionella 10 Fig. 7. Analysis III (excluding Benthobia and hypothetical ancestor); unrooted, 50% majority-rule consensus of 10,490 trees of length 182. retained as the shortest trees. Taylor and Kantor (pers. comm.) told me that their runs terminated when the com- puter they used ran out of memory. This explains why Taylor et al. found trees of length 189 rather than 187 or 188 and why they found only 900+ trees rather than many thousands of trees: they simply did not have enough com- puting power. This should have been stated in the article, however, which otherwise gives the impression that the heuristic search routine was completed. Because Taylor et al. did not state the starting conditions of their analysis (e. gz. MAXTREES or other search settings) or the ending conditions, their results could not be reproduced. HYPOTHETICAL ANCESTOR Because there is no obvious sister group for Conoidea, Taylor et al. (1993: 152) used a hypothetical ancestral taxon “consisting of the underived states, where known, of the characters used in the analysis.’’ Construction of such a hypothetical ancestor is not unusual. A common procedure is to consider a variety of potential outgroups. If only one state occurs in the possible outgroups, the polarity is clear (Donoghue and Cantino, 1984). Taylor et al. appar- ently applied this type of reasoning in some cases (e. g. for accessory salivary glands; 1993: 142), but in many cases did not state how they made decisions about polarity. In other cases, the reasons they gave are known to be invalid, ROSENBERG: REPRODUCIBILITY IN PHYLOGENETIC ANALYSIS 225 Table 1. Consensus of consensus trees: clades co-occurring in each of the six consensus trees (Figs. 2-7) and the TKS tree (Fig. 1). (Trees from Anallysis] I, I, If]; MRC, majority rule consensus tree; SC, strict consensus tree). Clade shared by tree in: Fig. 2 Fig. 4 Fig. 6 Fig. 3 Fig. 5 Fig. 7 TKS Anal I Anal II Anal lll Anal I Anal II Anal III Clade SC SC SC MRC MRC MRC 1 x - xX X X x 2 z 3 - Xx Xx x 4 - x x x x X 5 a 6 X X Xx = 4 a - 8 9 Xx Xx X X x x 10 X - X x x 11 - x X 12 X X X x X X 13 - X - Xx 14 - = 15 - X 16 - - X Xx 17 X x X x x X 18 x X X X 19 - - 20 - - 21 - 22 - - 23 - = 24 - x x x 25 S : 26 27 S 28 - g 29 x X x X X X 30 X x x x X X 3] X x x X Xx X 32 X X X Xx Xx Xx 33 Xx X X Xx X Xx 34 35 36 - - - Xx X X 37 5 - - x Z x 38 = = 2 é 7 Total shared clades 10 10 13 19 16 19 such as the three criteria they explicitly mentioned (1993: 132). These are occurrence of character states (1) in taxa with the least derived states of other characters, (2) in most of the subfamilies of Turridae (i. e. common equals primi- tive), and (3) in distant rather than immediate outgroups (some members of Archaeogastropoda or “Meso- gastropoda” rather than other Neogastropoda). The first criterion assumes that primitive character States are correlated, but as Stevens (1980: 334) noted, “Correlations do not necessarily indicate anything about the evolutionary polarity of the characters states involved ... character states may be associated in various ways: all primitive, all derived, or mixed ...”” Of the second criterion, Stevens (1980: 335) stated, “the frequency of any character state in a group depends on the subsequent evolutionary history of the lineage with that state and is independent of the evolutionary polarity of that state.” Thus a primitive character might be rare if it belongs to a lineage that has not diversified or has suffered much extinction. A derived char- acter might be common if it arose early in the history of a clade. The third criterion ignores that the character states of distant outgroups cannot override the states in closer out- groups; at best they can render polarity ambiguous with regard to the ingroup (Maddison et al., 1984: fig. 3). Also, at least in some cases, Taylor et al. applied the first two cri- teria within the ingroup, rather than doing outgroup com- parisons (e. g. buccal mass position; 1993: 132). Because of doubt about the scoring of character states for the hypothetical ancestor, I tried running the analysis without it (and without Benthobia) (Analysis HI), finding 10,490 trees of length 182. Curiously, the consen- sus trees for these (Figs. 6-7) have more clades in common with the TKS tree than the consensus trees when the hypo- thetical ancestor was included: 13 versus ten (strict); 19 versus 16 (majority rule) (Table 1). In particular, clade 18, which groups Conidae with a subset of Turridae, appears in the strict consensus. CONSENSUS OF THE CONSENSUS TREES The strict consensus trees (with and without the out- groups) each share only ten to 13 clades with the TKS tree (Table 1), which has a total of 38 clades. Only eight clades are shared by all three strict consensus trees and the TKS tree (node numbers in parentheses): Clavatulinae (part) + Crassispirinae + Zonulispirinae (9); Clavatula caerulea + Turricula (12); Turrinae (17); Mangeliinae + Daphnellinae + Taraninae (29); Mangeliinae (30); Mangelia (31); Daphnellinae + Taraninae (32); and Philbertia + Daphnella + Gymnobela + Teretiopsis + Abyssobela + Taranis (33). One clade appears in all three strict consensus trees that does not appear in the TKS tree: Clavatulinae (all) + Crassispirinae + Zonulispirinae. Only 13 of the 38 clades are shared by all three majority-rule consensus trees, those in the strict consensus trees, listed above, and: Drilliidae (1); Pervicaciinae (4); Conidae sensu Taylor et al. (18); Anarithma + Glyphostoma (24); and Gymnobela + Teretiopsis + Abyssobela + Taranis (36). Seventeen clades supported by the Taylor et al. analysis are not supported by any of the trees presented herein: 2, 5, 7, 8, 14, 19, 20, 21, 22, 23, 25, 26, 27, 28, 34, 35, and 38. Taylor et al. (1993: 152) did not state the level of 226 AMER. MALAC. BULL. 14(2) (1998) support for any of the nodes in their consensus tree, saying instead “‘most of the branches are supported in 75-100% of the trees.” Because of this omission, it is not possible to tell where disagreement is significant. Lack of support in my analysis for a node supported in only 55% of their trees would not be surprising; lack of support for a node support- ed in 100% of their trees would be more interesting. The 17 nodes contradicted by my analysis average only 1.8 charac- ters per node (Taylor et al., 1993: table 4), whereas the 12 nodes supported by all three of my consensus trees average 3.3 characters per node. The mapping of characters to nodes is often ambiguous, however, and many alternative mappings to the ones in their table 4 exist. For example, character 3(1) was mapped to node 3, with loss in Pervicacia. An alternate mapping is independent gain in Hastula and Duplicaria. CHARACTERS Character mappings In addition to some characters in table 4 of Taylor et al. having alternate mappings, some mappings are simply incorrect. Kohn and McLean (1994) noted that character state 39(0) at node 27 (Benthofascis + Conus) should be mapped at node 26 instead, because it is present in Genota. Actually, 39(0) is a primitive state inherited from the root of the tree. State 39(1) (teeth on the outer lip) evolved at nodes 24 and 33 and in Eucithara; it was lost at node 36. There is no need to map 39(0) elsewhere in the tree. Similarly, character states 20(1), 35(0), and 38(1), mapped to node | (Drilliidae), are actually basal to the conoideans. Also at node 1, Clavus has state 7(2) not 7(1). Thus four of the five synapomorphies attributed to Drilliidae in Taylor et al.’s table 4 do not map to that node (in the given topology). The remaining character state, 37(0), has an ambiguous mapping, as node 3 has the same state, which thus could be basal, changing to 37(1) at node 5. Parallelism The majority of characters (77%) that Taylor et al. used are from a single organ complex, the foregut. Any classification based primarily on characters from one sys- tem is likely to be mislead by parallelisms. Cladograms based on characters from many organ systems are less like- ly to be misleading (Davis, 1979: 9). Taylor et al. (1993: 151) noted: many of the morphological trends in the Terebrinae involve partial to total loss of the foregut organs ..thus many of the characters were recorded as missing. In our earlier attempts at cladistic analysis, terebrid species tended to appear in rather disparate positions on the cladograms. Consequently, we have used only three species to represent the Terebrinae and Pervicaciinae, the taxa being the least-derived known for each group. Inclusion of characters from the reproductive, circulatory, excretory, and respiratory systems might have overridden the problem caused by the supposed parallel losses in the Terebridae. Paradoxically, understanding of the evolution of the foregut organs requires study of the other organ sys- tems - they cannot be studied in isolation. The hope is that homoplasy in any given system will be more than balanced by the overall phylogenetic signal in combined analysis across systems. An excellent example of this is Benthobia. Because of homoplasous characters in foregut anatomy, it was pulled into the ingroup, but the overall anatomy and morphology place it securely with the Pseudolividae (Kantor, 1991; see also Bouchet and Warén, 1985). The single species of Conus included in the analysis might have been pulled into Turridae because of parallelisms, just as Benthobia was pulled into Conoidea. Choice and scoring of characters Some of the ten non-foregut characters analyzed were overly simplified. For example, the division of shell form into “fusiform,” “coniform,” “turreted,” “terebri- form,” and “rounded” is simply unworkable in the continu- um of shapes that occur among the more than 4,000 living species of conoideans, especially because quantitative defi- nitions of these terms were not given (Taylor et al., 1993: 162). Number of protoconch whorls (< 2, > 2) is also an arbitrary division of a continuum, although it tends to cor- relate with developmental type. Sculpture of the protoconch (absent or weak versus present) combines many disparate types of sculpture as a single character state. Kohn and McLean (1994) noted that potentially useful shell characters were excluded, such as the resorp- tion of the inner shells whorls found in Conus and Benthofascis, which Taylor et al. mentioned (1993: 155), but did not include in their matrix. Taylor (1994: 435) stat- ed, “We considered, but did not include this character in the cladistic analysis. Its inclusion would have made no differ- ence to the structure of the cladogram except to add another apomorphy at the node of Benthofascis and Conus.” This clade, however, is not supported by any of the consensus trees presented herein. Inclusion of the character might well have resulted in support for this clade. There are also some inconsistencies in the scoring of foregut characters. Character 15 (septum dividing anteri- or and posterior areas of the rhynchocoel) was listed as pre- sent in Philbertia purpurea, Daphnella, Pervicacia and Duplicaria in Taylor et al.’s table 3, but a “probably homol- ogous septum” reported in Thatcheria (1993: 129) was not scored in the matrix. For Character 26, Strictispira paxillus ROSENBERG: REPRODUCIBILITY IN PHYLOGENETIC ANALYSIS 224 (Strictispiridae) has state 1 (Taylor et al., 1993: table 3), which was stated to occur only in Pseudomelatomidae (1993: 136), and Hastula bacillus was scored ‘“?”’ but has state 2 (1993: table 2). For Character 33, Pseudomelatoma, Splendrillia, Toxiclionella, and Lophiotoma were scored as having unfused odontophoral cartilages in table 3, but stat- ed to have fused cartilages on p. 133. CHOICE OF TAXA Although one of the major conclusions of the study was that some subfamilies of Turridae should be transferred to Conidae, only a single species of Conus from among the more than 500 living species was included in the analysis (two species were studied [Taylor et al., 1993: table 1]). The study thus did not provide evidence that Conus 1s monophyletic. Similarly, six of the subfamilies included in the analysis were represented by only one species each. Again, no evidence was provided that these groups are monophyletic, yet two, Pseudomelatomidae and Strictispiridae, were elevated to familial rank. Other taxa studied were excluded from the cladistic analysis, in the case of some Terebridae explicitly because they did not group as expected. Many other species listed in Taylor er al.'s table 1 did not appear in their table 3, but no explana- tion was given for their exclusion. One hopes that they were excluded because of insufficient data, or because character scores did not differ from those of included species, rather than for cladistic misbehavior. CLASSIFICATION The classification that Taylor er al. introduced does not follow the structure of their cladogram (Fig. 1), as noted by Kohn and McLean (1994): Turridae, Clavatulinae, and Clathurellinae are polyphyletic, and Crassispirinae, Conorbinae, and Daphnellinae are paraphyletic. Taylor er al. stated (1993: 157), “Information from taxa not included in the cladistic analysis (mainly radular characters) has also been used in constructing the classification.” Presumably these taxa were not included because other data on their foregut anatomy was not available. Unfortunately, inclusion of such information, however relevant, adds another layer of irreproducibility to the analysis. Such data might be used to chose between equally parsimonious trees, but beyond that, there seems little justification for their use. Taylor et al.’s classification conflicts with their tree because it is in part conservative. An alternative classifica- tion more consistent with the tree would introduce new subfamily names for parts of Clavatulinae, Clathurellinae, and Crassispirinae, and synonymize Conorbinae with Coninae, and Taraninae with Daphnellinae. That Taylor et al. refrained from naming new taxa is appropriate, given the preliminary nature of the data. This conservatism is at odds, however, with the elevation of Strictispiridae and Pseudomelatomidae to family level, which renders Turridae polyphyletic instead of paraphyletic. The essential conflict of the tree is that Conidae falls within Turridae. The solution that Taylor et al. adopted (transferring some turrid subfamilies to Conidae) left Turridae paraphyletic (ignoring for the moment the rank of Strictispiridae and Pseudomelatomidae). This is no better than accepting the traditional classification, in which Turridae might also be paraphyletic. At several points Taylor et al. stressed the limitations of their analysis. They noted that “small adjustments to the data set produced rather large changes in tree topography [sic] and the num- ber of alternative trees generated” (1993: 152), uncertainty about outgroups (p. 152), and inadequate sampling of taxa (p. 150). In view of these uncertainties, it is not clear why they introduced a new classification - producing the data and cladogram would have been enough. CONCLUSIONS The results of Taylor et al. (1993) cannot be repro- duced from their stated methods. The trees they reported are not minimal trees because the analysis they performed terminated prematurely, being limited by memory of their computer. Further problems arise from omission of the stat- ed outgroup, Benthobia, from the cladistic analysis, incor- rect character mappings, and conflicts between text and data matrix. The classification that Taylor et al. introduced does not follow the structure of their cladogram, taking into account data from taxa not included in the cladistic analy- sis. This adds further layers of irreproducibility to the results. Taking the data matrix as given, fewer than half of the clades they found were supported by my reanalysis. The matrix, however, is probably not a reliable basis for a phy- logenetic analysis. Most of the characters used were from a single organ complex, the foregut, making it likely that the parsimony analysis was mislead by parallelism, as evi- denced by the pseudolivid outgroup, Benthobia, being pulled into the ingroup. Also, some of character states of the hypothetical ancestor were determined by unreliable methods such ingroup comparison, so the polarities of char- acters were not well established. Addition of outgroup and ingroup taxa, rescoring of some characters, and of the hypothetical ancestor, and addition of characters from other organ systems would undoubtedly further modify the results. Nonetheless, I did find support for some aspects of the Taylor et al. analysis, taking the data matrix as given. Mangeliinae + Daphnellinae + Taraninae (node 29), and subsets thereof are supported in the strict consensus trees, as is Turrinae, and Clavatulinae + Crassispirinae + 228 AMER. MALAC. BULL. 14(2) (1998) Zonulispirinae (9), with Toxiclionella pulled in also, mak- ing Clavatulinae paraphyletic rather than diphyletic. Most interesting is that node 18, Conidae sensu Taylor et al. (Conidae + Clathurellinae, Conorbinae, Oenopotinae, Mangeliinae, Daphnellinae, and Taraninae) appears in all three majority-rule consensus trees. Support for this clade is not unequivocal, however, as it does not appear in two of the strict consensus trees. Given the limitations of the char- acter data and taxon sampling, it is yet possible that Conidae and Turridae will prove to be two separate clades. The single Conus species included in the analysis could well have been pulled into Turridae because of paral- lelisms, just as Benthobia was pulled into Conoidea. Malacologists will be justified in retaining the traditional classification, until more robust evidence for transferring turrid subfamilies to Conidae and for elevating Drilliidae, Pseudomelatomidae, and Strictispiridae to familial status is produced. Given the authors’ cautions about the preliminary nature of their analysis, their introduction of a new classifi- cation was premature. The monumental achievement of presenting the anatomical descriptions, assembling the data matrix, and presenting a preliminary tree would have been sufficient. Certainly it 1s much better to have this data available now than to wait five or ten years while the authors tackled other taxa and other organ systems to pro- duce a better-supported phylogeny. One advantage of cladistic over traditional methods is that intermediate results can be presented without the need to introduce a new classification. Once a new classification is introduced, inevitably some will adopt it, even if it is likely soon to be superseded. For example Turgeon et al. (1998) adopted Taylor et al.'s new classification of Conoidea. I note in closing that it was not the introduction of the new classification itself that spurred this reanalysis but that other workers were adopting a classification that I thought likely to be incorrect. Thus, the new classification did draw attention to ideas that have stimulated scientific discourse. Perhaps it has served its purpose after all. ACKNOWLEDGMENTS I thank John D. Taylor (The Natural History Museum, London) and Yuri I. Kantor (A. N. Severtzov Institute of Animal Morphology and Ecology, Moscow) for their willingness to discuss their methods with me. Two anonymous reviewers provided valuable comments on the manu- script. LITERATURE CITED Bieler, R. 1990. Haszprunar’s “Clado-evolutionary” classification of the Gastropoda - a critique. Malacologia 31:371-380. Bouchet, P. and A. Warén. 1985. Revision of the northeast Atlantic bathyal and abyssal Neogastropoda excluding the Turridae (Mollusca, Gastropoda). Bollettino Malacologico suppl. 1:121- 296. Davis, G. M. 1979. The origin and evolution of the gastropod family Pomatiopsidae, with emphasis on the Mekong River Triculinae. Academy of Natural Sciences of Philadelphia, Monograph 20, vii + 120 pp. Donoghue, M. J. and P. D. Cantino. 1984. The logic and limitations of the outgroup substitution approach to cladistic analysis. Systematic Botany 9:192-202. Haszprunar, G. 1988. On the origin and evolution of the major gastropod groups, with special reference to the Streptoneura. Journal of Molluscan Studies 54:367-441. Haszprunar, G. 1990. Towards a phylogenetic system of Gastropoda Part 1: traditional methodology - a reply. Malacologia 32:195-202. Kantor, Y. I. 1991. On the morphology and relationships of some olivi- form gastropods. Ruthenica 1:17-52. Kohn, A. J. and J. H. McLean. 1994. [Review of] Foregut anatomy, feed- ing mechanisms, relationships and classification of the Conoidea (= Toxoglossa) (Gastropoda). The Veliger 37:432-434. Maddison, W. P., M. J. Donoghue, and D. R. Maddison. 1984. Outgroup analysis and parsimony. Systematic Zoology 33:83-103. Maddison, W. P. and D. R. Maddison. 1992. MacClade: Analysis of Phylogeny and Character Evolution, ver. 3.0. Sinauer, Sunderland, Massachusetts Reynolds, P. 1997. The phylogeny and classification of Scaphopoda (Mollusca): an assessment of current resolution and cladistic reanalysis. Zoologica Scripta 26:13-21. Steiner, G. 1992. Phylogeny and classification of Scaphopoda. Journal of Molluscan Studies 58:385-400. Stevens, P. F. 1980. Evolutionary polarity of character states. Annual Review of Ecology and Systematics 11:333-358. Swofford, D. L. and D. P. Begle. 1993. PAUP: Phylogenetic Analysis Using Parsimony, ver. 3.1. Laboratory of Molecular Systematics, Smithsonian Institution, Washington, D.C. Taylor, J. D. 1994. Reply [to Kohn and McLean]. The Veliger 37:434-435. Taylor, J. D., Y. I. Kantor, and A. V. Sysoev. 1993. Foregut anatomy, feeding mechanisms, relationships and classification of the Conoidea (= Toxoglossa) (Gastropoda). Bulletin of the Natural History Museum, London (Zoology) 59:125-170. Turgeon, D. D., J. F. Quinn, Jr., A. E. Bogan, E. V. Coan, F. G Hochberg, W. G. Lyons, P. M. Mikkelsen, R. J. Neves, C. F. E. Roper, G. Rosenberg, B. Roth, A. Scheltema, F. G. Thompson, M. Vecchione, and J. D. Williams. Common and Scientific Names of Aquatic Invertebrates of the United States and Canada: Mollusks, 2nd ed. American Fisheries Society Special Publication 26, ix + 526 pp. Date of manuscript acceptance: 9 June 1998 AMERICAN MALACOLOGICAL SOCIETY FINANCIAL REPORT General Accounts 1997 Income and Expenses Ie GINSPEIN Gy STE A NOE Gl SOG Vee cece tests re Mai eats cr cewnsandil peanmcusotsone tomtaaneeccburereptialt ualag Ornette Aa, HTN ON TS ewe re tes cere eee eon eye eat nete re eee eager oth oe eave dene eens tees ene pa nace aa ttse Salads Membership ues (1095 O96, 1997) res a salctanssyccede can vonievs tev aeiara ne hrant vavewenmeapensinetee 18,159.75 Membership ues) 99 8). Bescoiotet tic ecco cee cosa oti pteueda igacu casa ars biamiavenad siaGaiede Drusmatau boomed 856.00 NIMEETE SUM ATI ESI VICE MCS ayhsnreatas ere tentor ree pivies son ane meee Neer baa a, cos wreranc es sentence mance es 2,419.28 Money Market Account Imterest:xisitesecacecnetvorecttisitvinececsdinavattasstuettecatess 1,123.40 ite, Membership-Endowaient: Hund iiesatSieccoasviiatscnetccncacecinseadesatepsncattes nee 108.92 S VII POSH ENO WINE I WA S series Mises i ccdcoesyioerscnsnsseeredynauarsssohans manny 1,186.96 Publicationselmcomes tse: cresscesomieeetateccsnuelcactatce tema tec adeasiae ub pean biunvat lr banaeie ns au teeel thy 13,428.71 VEE OTE IPM SUBS CELL OMS esac eeaceteg ete Sete eee erestacetece Paveecauncesvereeaeeeee 1,824.00 ANEB DOMESTIC SUBS CIIPUGUS aq. ceeceaesrccgec aeeersau seve eeese tena sede iloseyseuuneeeperece 412.44 ANTS Unspeciticth SUbSCHIpt OS sakirvegiacs «ust daectceeewesct co teescuanonedeaeytesbeee: 2, 711299 AEB PAD LAT OCS sce cagenenevcescsseevaeeny sae ke a eetaame apis tuend ocesutoe a saiernonsesehns 5,243.00 AUER URE OTUAL Mat Bes cet ree ret ences ietig eet tis cxn tao ive adasetanersane meas enamensnniges 2,985.26 Sales omO@iher PUbNCANOMS .\cc.-ccocrettets cance sadenaseecaa trina Meoannmuesstuetnaitess 252.00 PA rain tT IVS ETA Sera cease etc cece deans ttre cc cere ecto tee. arte een ceed a uk eas ete saetee 9,782.13 BOO MENISCUS UTP Ui Stene ete care resi datayeser ian) cast cavnsaszaates wi incdcatanhs 2,042.09 PODS MICELIIG SUMP IAS ce erccnsscecg Shores oar tencezceeacasusee rca veeateaduanee cineens ssseeeesnente 7,740.04 HE) Co MAU LOT Saree ce cms aces nee ee eee ettien saree cium evan ean sedsc tar al eas be dace mene MnP tis earl yabitneuitionsinsa 761.00 SUES UGE AIT UAW AC atest eae eae een me ec cere ee amnse nee tne ee an ah 500.00 Symposiuin’ Endowment Fund). satscd.ctsts ceietes.cmutinteieeesecedoncanasoenters 211.00 Student Grant Endowment Fun ..............cccccecececeeeeeeeeeeeececeeeeeeeeeneeeeenseeeenea 50.00 Fe Te Seal Cl CHILO UTS CITI (RUS ce aces coeds oe wast ee tele dee ate de aa ice a stancne ae re anencbe atime dmsareeventeee 530.22 Miscellaneous Income escccccscaetcsustcssansncetstuccneesauuneesu saneceds sucsenesevavavaueres celneteceavaseausnierdeeasencaee 130.00 IER GT EWN SiS oe reas ce ar Serene Oaks Gta cas Pra cancer acuetta detente cgc Li tasa wuent da taae iG ui ageaton eater deay es casda meter amhataaet? PLSSIG SME KPENSES esas antes, rcucns soseienye erae ses nent tae ot Od ods deat tltinTaita ts yale 64.94 Po Shek ey O21 Sh 4 0) wh choi ape ee Br err eee CPR RE ey ere ree eer ee 401.30 Members harman, EXPENSES. s9.0.aveevcesac ce ssnussncass ccerswassuneshicanaresannceasvaveassaunaeeys 67.12 IGE ASWRG Ie XCM SCs tite enn eticcinc ster etiotats tush ani Aine calevadssuleaciabilsraetcesiatuasiin 988.32 Put (iltatey WLS MADER S MIPS arcennennnseeteucaene teee tenn dag oan asa vese stnastanissuanuaaytbesionesngnereesnn ines 275.00 BAN KIM CIR SES Meee eee aera eee sna cone TRUCE Tea aiiac Par SN NCA NE ARULUS Eo ray secee sue neeuEae 35.84 IV UAV Ny NSEC yee sees ee aea eet tengo armen ee eee ence a PR 2 Be ord erate 448.94 INCOFPOT ALON HCC 25 cisco ccesencssouseesveneecedecsacesh sous seed ecudeas sseaveesidteccevede-neranmiaguneteeeenss 20.00 HS MrATICe/ Ts ONG COS aeaee, te cgee yegeveut eectneaes se onteenecttcnts are eset era calianigauacasistabawenianeees 604.00 RS TAMAS sree Seon gee centre cue see uaere reer re eemuaal ei acays veteccet wiosianayoaresticenes Muuam tN nmnamuneeen 168.00 IITVESTITIEN IE CES eis errerertee tte te neste crc atcaees cute ee catactiie sulantultasladdeusss uopeee entre rats 2.50 SUN yA oe Ca GCSE Sei ee nce cheep ete Sts a ded verdana cea ahenuehetet ees 15,680.96 PAIN LBs eetira cgeaienr atti Phaya vac Pace e eee pa nena va oases aah’ cia aeaedaoncecdvan evan ereesteayae 14,002.36 PAUVISEN CW SLCLLC Teeter retro eee aI OR eae mteae innit mane ae cement 1,678.60 ID Yt fale 0) t Cater Peeper e rere et ere eee teacher erty et mentee gate ener net net rey eer one ap ara 606.00 Abstract Mailing (1996 Meeting) .............:ceeeecceeeeeeneeeeeeneeeeeseneeeesnaeeeseneeeesanees 1,254.25 Officers’ Travel Expenses (To Annual Meeting) o.c.ce2.0i. ces taliesitisssncgttnedens 6,047.34 LUCENE LAE AWALOS aatrcuatsecvastoecsgcvestever sauces cae couceqeinacatesereryeneheanueeasdensesanes 1,000.00 MeO MICELI LD CIOS UL ey rea tee nataneesteteresaraseetates canteen cance eey eet een eae ean set cece 400.00 INES ig Gr SUTIN SUN arose sea eats tu wea ance tetas ese ane ode tot dees ence nia nya cacaceoPaecamin anaes beacamciacaes tat rcea caey candies wabhevsae ENDING BALANCE $113,862.68 ..$46,067.09 ...$28,064.5 | ..5 18,002.58 $126,167.85 65th ANNUAL MEETING THE AMERICAN MALACOLOGICAL SOCIETY PITTSBURGH, PENNSYLVANIA, U.S. A. JULY 4-9, 1999 The 1999 American Malacological Society meeting will be held at the Sheraton Station Square in Pittsburgh, Pennsylvania. In addition to an exciting symposium and various contributed sessions and posters, the meeting will host various special sessions and workshops of unique nature but broad interest. Symposium: New Looks at Old Molluscs: Recent Perspectives of Molluscan Evolution Organizers: Bud Rollins, University of Pittsburgh (snail@vms.cis.pitt.edu) and Ellis Yochelson Smithsonian Institution (yochelson.ellis@simnh.si.edu). Special Sessions and Workshops —Molluscan Genetics organized by Laura Adamkewicz, George Mason University (ladamkew @ gmu.edu). —Mollusks and Education: New Ideas from Museums to the Classrooms organized by M. Patricia Morse, University of Washington (mpmorse @u.washington.edu). —Biomineralization in Mollusks (including workshop) organized by Joseph Carter, University of North Carolina (clams @email.unc.edu). —Women in Malacology organized by Louise Russert-Kraemer, University of Arkansas (rkraemer @comp.uark.edu). —Malacology Curation Workshop organized and presented by Charlie Sturm, Carnegie Museum of Natural History (csturmjr+ @pitt.edu), Gary Rosenberg, Dick Petit, Jose Leal, Tim Pearce, and Kevin Cummings. ~The Clench Tapes and AMU History organized by Harold Murray (hmurray @trinity.edu). Additional AMS 1999 events will include a July 4th fireworks reception, access to Carnegie Museum of Natural History mollusk collection, auction with special molluscan art available, banquet cruise and a special student reception, plus a keynote address by Rich Lutz, Rutgers University, on “Deep-Sea Hydrothermal Vents: Exciting New Discoveries.” Art Bogan is organizing a field trip to high diversity freshwater sites in western Pennsylvania while Albert Kollar will lead a trip to excellent fossil sites, both on the last day of the meeting. Pittsburgh is an exciting, compact city with remarkable museums, theaters, and sports. Come early and stay late to take advantage of this completely revitalized city. The venue of the meeting is a full facility confer- ence hotel, the Sheraton Station Square, with ready access to all that Pittsburgh has to offer. Be sure to par- ticipate. A call for papers and posters will soon appear; please get your abstracts and registration materials in as early as possible. Deadline is 5 March 1999. For more information contact: Robert S. Prezant, President, AMS Office of the Dean Division of Mathematics and Natural Sciences Queens College, CUNY Flushing, NY 11367-1597 E-mail: rprezant@qc.edu 230 IN MEMORIAM Harold Lewis Alexander, J. E. 14:91 Avelar, W. E. P. 14:157 Brown, K. M. 14:27, 91 Carriker, M. R. 14:121 Chen, H.-C. 14:139 Cuezzo, M.G. 14:1 Edds, D. R. 14:41 Emberton, K. C. 14:87 Folino, N.C. 14:17 Fortunato, H. 14:191 Gatenby, C.M. 14:57 Graf, D. L. 14:35 Haimovici, M. 14:81 Holznagel, W. E. 14:181 Ituarte, C. F. 14:9 INDEX TO VOLUME 14 AUTHOR INDEX Kovalak, W. P. 14:67 Longton, G. D. 14:67 Mangold, K. M. 14:185 Martinez, E. 14:133 Mendonga, S.H.S. T. de 14:157 Mikkelsen, P. M. 14:201 Miller, E. J. 14:41 Neves, R. J. 14:57, 75, 165, 173 O’ Bern, F. X. 14:165 Obermeyer, B. K. 14:41 Olabarria, C. 14:103 Ortea, J. 14:133 Parker, B.C. 14:57, 75, 173 Patterson, M. A. 14:75, 173 Perez, J. A. A. 14:81 PRIMARY MOLLUSCAN TAXA INDEX Prophet, C. W. 14:41 Rosenberg, G. 14:133, 219 Santos, R. A. d. 14:81 Schloesser, D. W. 14:67 Smith, C. R. 14:185 Smithee, R. D. 14:67 Steg, M. B. 14:165 Sweeney, M. L. 14:149 Thorp, J. H. 14:91 Ting, Y.-Y. 14:139 Troncoso, J. S. 14:103 Urgorni, V. 14:103 Walker, R. L. 14:149 Yang, H.-S. 14:139 Young, R. E. 14:185 [first occurrence in each paper recorded; new tax (including species) in bold face] Abra 14:103 Abyssobela 14:220 Abyssotrophon 14:128 Acantharion 14:212 Acanthochitona 14:114 Acmaea 14:116 Actinonaias 14:51 Aeolididae 14:17 Aeolidoidea 14:17 Aforia 14:220 Alasmidonta 14:41 Alicula 14:133 Allogastropoda 14:20] Alvania 14:114 Amblema 14:41, 69, 76, 173 Amnicola 14:27, 97 Amphibulima 14:6 Amphibulimidae 14:203 Anarithma 14:220 Anaspidea 14:202 Anidolyta 14:207 Anodonta 14:37, 175 Anomia 14:114 Anomiidae 14:107 Aphallarion 14:212 Apogastropoda 14:122 Arayina 14:193 Archaeochilina 14:9 Archaeogastropoda 14:224 Argopecten 14:161 Arionidae 14:212 Arionoidea 14:212 Ascobulla 14:206 Ascoglossa 14:204 Astarte 14:127 Athearnia 14:182 Atoxon 14:212 Atys 14:133 Aulacopoda 14:212 Basommatophora 14:9 Bathyberthella 14:207 Benthobia 14:219 Benthofascis 14:220 Berthelinia 14:205 Berthella 14:207 Berthellina 14:207 Bittium 14:103 Bivalvia 14:35, 41, 57, 67, 107, 121, 149, 157, 173, 202 Boltaenidae 14:203 Boreotrophon 14:129 Borsonia 14:220 Boucardicus 14:87 Brachidontes 14:149 Brachystomia 14:114 Buccinidae 14:121 Buccinoidea 14:192 Bulimulidae 14:6 Bulla 14:133 Bunnya 14:1 Bursatella 14:203 Caecum 14:116 Caenogastropoda 14:29, 88 Calliostoma 14:117 Callochiton 14:117 Calyptraea 14:113 Campeloma 14:29, 97 Capulidae 14:121 Carinartidae 14:203 Cassidae 14:121 Cephalaspidea 14:133, 202 Cephalopoda 14:121, 203 Cerastoderma 14:103 Cerithiopsis 14:114 Zoe Chaetopleura 14:117 Chamelea 14:114 Chicoreus 14:129 Chilina 14:9 Chilinidae 14:9 Chione 14:129 Choromytilus 14:154 Chrysallida 14:112 Cirrata 14:203 Clathurellinae 14:219 Clausinella 14:103 Clavatula 14:220 Clavatulinae 14:220 Clavus 14:220 Clionella 14:220 Cochlespirinae 14:220 Coleoidea 14:203 Conidae 14:219 Coninae 14:220 Conoidea 14:219 Conorbinae 14:219 Conus 14:220 Corbicula 14:44, 165 Corbiculidae 14:44 corcovadensis, Cryptostrakon 14:1 Costangula 14:193 Crassispirinae 14:220 Crassostrea 14:64, 129, 142 Crenomytilus 14:154 Cryptostrakon 14:1 Cucumerunio 14:162 Cuspidaria 14:123 Cuthona 14:17 Cyclophoridae 14:88 Cyclophoroidea 14:88 Cylindrobulla 14:205 Cylindrobullacea 14:214 Cylindrobullidae 14:214 Cyprogenia 14:41 Cystopelta 14:212 Cytharella 14:113 Daphnella 14:220 Daphnellinae 14:219 Daudebardia 14:212 Dentalium 14:117 Digitaria 14:116 Diodora 14:117 Diplodon 14:157 Dosinia 14:114 Dreissena 14:52, 67,75, 165, 173 Dreissenidae 14:67 Drilliidae 14:219 Duplicaria 14:220 Eledone 14:81 Elimia 14:91 Elisolimax 14:212 Ellipsaria 14:42 Elliptio 14:37, 44, 70 Elysia 14:114 Endodontidae 14:212 Eucithara 14:220 Eupleura 14:129 Euselenops 14:207 Euspira 14:123 Fissurellidea 14:203 Fissurellidae 14:203 Fossaria 14:29 Funa 14:220 Fusconaia 14:42, 70, 76 Gaeotis 14:6 Galeommatoidea 14:203 Gastropoda 14:17, 27, 88, 107, 121, 133, 181, 208, 219 Genota 14:220 Geukensia 14:149 Gibbula 14:103 Glyphostoma 14:220 Gouldia 14:115 Gymnobela 14:220 Gymnosomata 14:202 Gyraulus 14:29, 97 Gyrotoma 14:182 Haedropleura 14:116 Haliotis 14:139 Haminoea 14:203 Haminoeidae 14:133, 203 Hastula 14:220 Helicarionidae 14:6, 212 Helicoidea 14:2 Helisoma 14:29, 91 Hemiplecta 14:212 Hemphillia 14:212 Heterobranchia 14:29, 202 Heteroteuthis 14:187 Hiatella 14:114 Hinia 14:107 Hyalopecten 14:123 Hydrobia 14:103 Hyriidae 14:157 Idas 14:154 Incirrata 14:203 To 14:182 Janthina 14-123 Jujubinus 14:116 Julia 14:205 Juliidae 14:205 Kellia 14:114 Laevapex 14:29 Lamellariidae 14:203 Lampsilis 14:36, 41, 68, 76, 165, 173 Lasmigonia 14:45 Lepidochitona 14:107 Leptochiton 14:113 Leptodea 14:45, 70 Leptoxis 14:182 Ligumia 14:45, 70 Lima 14:123 Limacidae 14:203 Limacoidea 14:212 Limopsis 14:123 Lirastrombina 14:193 Lithasia 14:91 Lithophaga 14:154 Littorina 14:103 Lobiger 14:205 Loliginidae 14:203 Lophiotoma 14:220 Loripes 14:103 Lucinoma 14:116 Lunatia 14:113 Lymnaea 14:29 Macoma 14:115, 123 Mactridae 14:123 Mangelia 14:113, 220 Mangeliinae 14:219 Manzonia 14:113 Margaritifera 14:36 Marginellidae 14:121, 202 Mariaella 14:212 Mariella 14:6 Megalonaias 14:42 Melanella 14:116 Mercenaria 14:169 Mesafricarion 14:212 Mesochilina 14:9 Mesogastropoda 14:224 Metostracon 14:1 Micantapex 14:220 Microparmarion 14:212 Modiolus 14:149 Moerella 14:114 Monia 14:116 Monodonta 14:112 Mourgona 14:205 Muricanthus 14:122 Muricidae 14:121 Muricoidea 14:122 Musculus 14:114 Myrtea 14:116 Mysella 14:107 Mytella 14:154 Mytilidae 14:149 233 Mytilus 14: 107, 146, 149, 169 Natica 14:128 Naticidae 14:121, 203 Naticoidea 14:122 Neochilina 14:9 Neogastropoda 14:202, 224 Neorossia 14:187 Notarchidae 14:203 Notaspidea 14:202 Nucula 14: 103, 127 Nudibranchia 14:17, 202 Obliquaria 14:45, 70 Obtusella 14:116 Octopoda 14:81, 203 Octopodidae 14:121 Odostomia 14:113 Oenopota 14:220 Oenopotinae 14:219 Olividae 14:202 Onchidoris 14:22 Onoba 14:107 Oopelta 14:212 Ophiodermella 14:220 Opisthobranchia 14:17, 202 Opisthotheuthidae 14:203 Ostrea 14:117 Oxynoacea 14:205 Oxynoe 14:205 Oxynoidae 14:203 Papillicardium 14:107 Parmella 14:6 Parvicardium 14:112 Patella 14:112 Pectinidae 14:123 Pellicula 14:6 Peltella 14:6 Perna 14:149 Perumytilus 14:154 Pervicacia 14:220 Pervicaciinae 14:221 Phenacolimax 14:212 Phestilla 14:21 Philbertia 14:220 Philomycidae 14:202 Philomycus 14:212 Philinoglossacea 14:203 Phrixgnathus 14:212 Physa 14:29, 92 Physella 14:92 Pilsbryspira 14:220 Pisidium 14:114 Placobranchacea 14:205 Planorbella 14:92 Pleurehdera 14:207 Pleurobema 14:45, 70 Pleurobranchacea 14:207 Pleurobranchaea 14:207 Pleurobranchaeinae 14:207 Pleurobranchella 14:207 Pleurobranchidae 14:207 Pleurobranchinae 14:207 Pleurobranchomorpha 14:206 Pleurobranchus 14:207 Pleurocera 14:94 Plutonia 14:212 Pododesmus 14:115 Polyplacophora 14:107 Polystira 14:220 Potamilus 14:45, 69 Promenetus 14:29 Pseudolividae 14:219 Pseudomelatoma 14:220 Pseudomelatomidae 14:219 Ptychobranchus 14:41 Pulmonata 14:2, 9, 29, 91, 202 Pyganodon 14:43, 70, 169 Pyramidellidae 14:202 Quadrula 14:41, 69, 76, 173 Ranellidae 14:121 Rapanidae 14:121 Raphitoma 14:113 Recurvina 14:193 Retusa 14:112 Retusidae 14:202 Rissoa 14:107 Rissostomia 14:103 Roburnella 14:205 Rossia 14:185 Runcinoidea 14:203 Sacoglossa 14:202 Scaphopoda 14: 107, 123, 202 Scrobicularia 14:103 Septifer 14:154 Sigmurethra 14:210 Sincola 14:193 Sininae 14:203 Soleolifera 14:202 Spiralta 14:193 Splendrillia 14:220 Dates of Publication Strictispira 14:220 Strictispiridae 14:219 Strombina 14:192 Strophitus 14:45 Stylommatophora 14:2, 210 Syndosmia 14:115 Syndosmya 14:123 Taraninae 14:219 Taranis 14:220 Tellina 14:117 Terebia 14:182 Terebridae 14:219 Terebrinae 14:226 Teredinidae 14:203 Teretiopsis 14:220 Testacellidae 14:203 Teuthoidea 14:203 Thatcheria 14:220 Thaumatodon 14:212 Thyasira 14:103 Tomopleura 14:220 Tonnidae 14:121 Toxiclionella 14:220 Toxolasma 14:41 Triganglionata 14:201 Tritogonia 14:45 Trochomorpha 14:212 Trochonania 14:212 Trochozonites 14:212 Trophonidae 14:12] Tropidoturris 14:220 Truncilla 14:42 Turbonilla 14:112 Turricula 14:220 Turridae 14:219 Turrinae 14:220 Turritella 14:116 Tylodina 14:207 Tylodinidae 14:207 Umbraculacea 14:206 Umbraculum 14:207 Uniomerus 14:45 Unionidae 14:57, 67, 75, 165, 173 Unionoidea 14:157 Urocyclidae 14:212 Urosalpinx 14:122 Utterbackia 14:43, 169 Valvata 14:29 Vayssiereidae 14:121 Venericardia 14:123 Veneridae 14:123 Venerupis 14:103 Venus 14:114 Venustaconcha 14:42 Vetigastropoda 14:129, 203 Vexitomina 14:220 victorhernandezi, Boucardicus 14:87 Villosa 14:57, 165 Vitreolina 14:116 Vitrinidae 14:6 Vitrinopsis 14:212 Viviparus 14:29 Volvatella 14:206 Xanthonychidae 14:1 Xanthonyx 14:1 Xenostrobus 14:154 Zonites 14:212 Zonitidae 14:212 Zonulispirinae 14:220 Volume 14(1), December, 1997 Volume 14(2), December, 1998 234 CONTRIBUTOR INFORMATION The American Malacological Bulletin serves as an outlet for reporting notable contributions in malacological research. Manuscripts concerning any aspect of original, unpublished research, important short reports, and detailed reviews dealing with molluscs will be considered for publi- cation. Each original manuscript and accompanying illustra- tions must be submitted with two additional copies for review purposes. Text must be typed on one side of 8-1/2 x 11 inch bond paper, double-spaced, and all pages num- bered consecutively with numbers appearing in the upper right hand corner of each page. Leave ample margins on all sides. Form of the manuscript should follow that outlined in the Council of Biology Editors Style Manual (sixth edition, 1994). This can be purchased from the CBE, 11 S. LaSalle Street, Suite 1400, Chicago, IL 60603, U.S.A. Text, when appropriate, should be arranged in sections as follows: 1. Cover page with title, author(s) and address(es), and suggested running title of no more than 50 characters and spaces. Authors should also supply five key words, placed at the base of this page, for indexing purposes. 2. Abstract (less than 5% of manuscript length) 3. Text of manuscript starting with a brief introduction followed by methodology, results, and discussion. Separate sections of text with centered subtitles in cap- ital letters. Acknowledgments 5. Literature cited 6. Figure captions All binomens must include the author and date attrib- uted to that taxon the first time the name appears in the manuscript [e. g. Crassostrea virginica (Gmelin, 1791)]. This includes nonmolluscan taxa. The full generic name along with specific epithet should be written out the first time that taxon is referred to in each paragragh. The gener- ic name can be abbreviated in the remainder of the para- graph as follows: C. virginica. References should be cited within text as follows: Hillis (1989) or (Hillis, 1989). Dual authorship should be cited as follows: Yonge and Thompson (1976) or (Yonge and Thompson, 1976). Multiple authors of a single article should be cited as follows: Beattie et al. (1980) or (Beattie et al., 1980). In the literature cited section of the manuscript refer- ences must also be typed double spaced. All authors must be fully identified, listed alphabetically and journal titles must be unabbreviated. Citations should appear as follows: Beattie, J. H, K. K. Chew, and W. K. Hershberger. 1980. Differential survival of selected strains of Pacific oys- ters (Crassostrea gigas) during summer mortality. Proceedings of the National Shellfisheries Association 70(2):184-189. Hillis, D. M. 1989. Genetic consequences of partial self fertilization on population of Liguus fasciatus (Mollusca: Pulmonata: Bulimulidae). American Malacological Bulletin 7(1):7-12. Seed, R. 1980. Shell growth and form in the Bivalvia. Jn: Skeletal Growth of Aquatic Organisms, D. C. Rhoads and R. A. Lutz, eds. pp. 23-67. Plenum Press, New York. Yonge, C. M. and T. E. Thompson. 1976. Living Marine Molluscs. William Collins Son and Co., Ltd., London. 288 pp. Illustrations should be clearly detailed and readily reproducible. All line drawings should be in black, high quality ink. Photographs must be on glossy, high contrast paper. All diagrams must be numbered in the lower right hand corners and adequately labeled with sufficiently large labels to prevent obscurance with reduction by one half. Magnification bars must appear on the figure, or the cap- tion must read Horizontal field width = x mm or x pum. All measurements must be in metric units. All illustrations submitted for publication must be fully cropped, mounted on a firm white backing ready for reproduction, and have author’s name, paper title, and figure number on the back. All figures in plates must be nearly contiguous. Additional figures submitted for review purposes must be of high quality reproduction. Xerographic reproduction of pho- tomicrographs or detailed photographs will not be accept- ablefor review. Abbreviations used in figures should occur in the figure caption. Indicate in text margins the appropri- ate location in which figures should appear. Color illustra- tions can be included at extra cost to the author. Original illustrations will be returned to author if requested. Any manuscript not conforming to AMB format will be returned to the author for revision. Final submission of accepted, revised manuscripts should include two typed copies of the text, tables, etc. and an additional copy in electronic form on 3.5” diskette. The electronic version should be readable as non-formatted ASCII files. New Taxa. The Bulletin welcomes complete descriptions of new molluscan taxa. Establishment of new taxa must conform with the International Code of Zoological Nomenclature (1985). Descriptions of new species-level taxa must include the following information in the order as given: higher taxon designation as needed for clarity; fami- ly name with author and date; generic name with author and date; Genus species author sp. nov. followed by numeration of all figures and tables; complete synonymy (if any); listing of type material with holotype and any other type material clearly designated along with complete museum catalogue or accession information; listing of all additional non-type material also with full museum deposi- tion information; type locality; diagnosis and full descrip- tion of material done in telegraphic style including measurements and zoogeographic distribution as neces- sary; discussion; etymology. Descriptions of new supraspe- cific taxa should include type species (for new genus) or type genus (for new family), diagnosis and full description done in telegraphic style, and list of included taxa. Proofs. Page proofs will be sent to the author and must be checked for printer’s errors and returned to the printer within a three day period. Significant changes in text, other than printer errors, will produce publishing charges that will be billed to the author. Mailing. All overseas mailing must be done via airmail. The American Malacological Society will not be responsi- ble for deferred publication of manuscripts delayed in sur- face mail. Charges. There are no mandatory page costs to authors lacking financial support. Authors with institutional, grant or other research support will be billed for page charges. The current rate is $35.00 per printed page. Acceptance and ultimate publication is in no way based on ability to pay page costs. Reprints. Order forms and reprint cost information will be sent with page proofs. The author receiving the order form is responsible for insuring that orders for any coau- thors are also placed at that time. Submission. Submit all manuscripts to Dr. Ronald B. Toll, Editor-in-Chief, American Malacological Bulletin, Depart- ment of Biology, Wesleyan College, 4760 Forsyth Road, Macon, Georgia 31297-4299 U.S.A. (e-mail: ron_toll @post.wesleyan-college.edu). Subscription Costs. Institutional subscriptions are avail- able at a cost of $48.00 per volume. Membership in the American Malacological Society, which includes personal subscriptions to the Bulletin, is available for $35.00 ($15.00 for students, $45 for affiliated clubs). All prices quoted are in U.S. funds. Outside the U.S. postal zones, add $5.00 seamail and $10.00 airmail per volume within North America, or $10.00 seamail and $20.00 airmail per volume in other locations. For subscription and back-issue information contact Dr. Paula M. Mikkelsen, Managing Editor, American Malacological Bulletin, Department of Invertebrates, American Museum of Natural History, Central Park West at 79th Street, New York, New York 10024-5192, U.S. A. (e-mail: mikkel@amnh.org). 0 Research Note: In search of Rossia pacifica diegensis S.S. Berry, 1912. K. M. MANGOLD, BOE SYOUNG, and CRAIG BR. SMU ovpacccscencstisessedcccrcstinuncntsnelvahie sasncoustatissouryeitecentsuedentinteraresis 185 SYMPOSIUM: TRADITIONAL VERSUS PHYLOGENETIC SYSTEMATICS OF MOLLUSKS Reconciling observed patterns of temporal occurrence with cladistic hypotheses of phylogenetic relationship. HELENA FORTUNATO... ccscesecseeeeseceecseeseeeeeaesaesetseseataeaes 19] —Review of shell reduction and loss in traditional and phylogenetic molluscan systematics, with experimental manipulation of a negative gain character. PAULA ML IEICE SEIN cists cas satay a acti plaaetcced teancsebeacigu nine oucaca cena tana ages vided nu bbcode eeyticeens 201 Reproducibility of results in phylogenetic analysis of mollusks: a reanalysis of the Taylor, Kantor, and Sysoev (1993) data set for conoidean gastropods. GARY ROSENBERG osc. cscccscccsiscncacessecssaceoesotedusssscscentocodencsancaneiuvddadenessdesnsiedessSennseassanecddeantcestecsesonedés 219 Fimantrall REPOM sssesseccsnwssescouvesancienzasndsanciecercteednobucssnssocaceadenchsotetsatbubuas cxenadiaa hunescessdueaenmisndedealaveudendaitedesuacnsis 229 PRAM U CE EINE cis os tcs nce cnc tccy etanastccccavieda csi esupuccaaeeaaelbcnntectanesrucencsuees iicsLoddassi oisassaencesdcaalenddleiaunevacuensapctuces: 230, Jet MO MON AM oiesesadeusesnsccsdiaiesesiedslevsnadsadeneinassiuasicveasystounsusuiuuuneiaioadecousnasssvtxonnndds bagaubaeaasdacnseduiedehspieaiabvonaseasn eae 231 Tne x:to: Volume 1:4 a22ooc sss a tsi acs obs Be diced cass ech eedeedl Avi de scande cen aebe's deus ge eddeusasdeaeeasdoseagedves oveassaives 232 AMERICAN MALACOLOGICAL BULLETIN VOLUME 15 1999 NUMBER 1 Journal of the American Malacological Society CONTENTS The utility of fossil data in phylogenetic analyses: a likelihood example using Ordovician-Silurian species of the Lophospiridae (Gastropoda: Murchisoniina). PETER J. WAGNER ....0........cccccccsseccesseccesssecessseeeesseeecesseueeeesssecessesssseccesesssesensuaapes 1 The Tonicella lineata (Wood, 1815) species complex (Polyplacophora: Tonicellidae), with descriptions of two new species. ROGER N. CLARK..............ccccccssesesesseeeseees 33 Glycogen concentration in the mantle tissue of freshwater mussels (Bivalvia: Unionidae) during starvation and controlled feeding. MATTHEW A. PATTERSON, BRUCE C. PARKER, and RICHARD J. NEVES 000... cceccsesseesseeeseeseeecseeeeeeseceeseeaeeseseeseeeesees 47 Variation in glycogen concentrations within mantle and foot tissue in Amblema plicata plicata: Implications for tissue biopsy sampling. TERESA J. NAIMO ath TEIVUY IVD. EOIN ROB aia anisms case nier sect eSasisas wer ukals exe chcdvace sc eon acdc dated cdesacatdn letewedsheeenieipaniadees 51 Recruitment in a freshwater unionid (Mollusca: Bivalvia) community downstream of Cave Run Lake in the Licking River, Kentucky. STEPHEN E. McCMURRAY, GUENTER A. SCHUSTER, and BARBARA A. RAMEY .00.0.0...ccccccccccccsssssesseseeseesecseeseesseneeeseeses 57 Recovery Status of Freshwater Mussels (Bivalvia: Unionidae) in the North Fork Holston River, Virginia. WILLIAM F. HENLEY and RICHARD J. NEVES ..............c:cccccccccsssssssssseseeeees 65 Historical and ontogenetic changes in shell width and shape of land snails on the island of Kikai. EMIKO HAYAKAZE and SATOSHI CHIBA ...0.0.......cceccccsesesesesesssesseseseeeseeeseseeeecaeeeees 75 New acavid land snails from Madagascar. KENNETH C. EMBERTON ....00..........cccccccccsssescesescescssceseseeees 83 Edentulina of Madagascar (Pulmonata: Streptaxidae). KENNETH C. EMBERTON...........0.0.00..ccccee 97 PRman tal Rep ONG oi desinscasctticnassatansdacecencrs sede esesetideteilebucwachadnesatvxqiedueis sence dvseissasmnuiasbanusiteedeauistanend ae 109 PMMAOUNOSMIOAE on casssss sas sisicazavsds nds we snsiieuobaddeadeasssepen otaasontslionancdensvinvéhehs$aduels hie oabedantasotinvesduasbinnaasaldidtessoeasonebialed 1 1d In Memoria .o.ccsscssssesssesssssessssssssesssesssssescssessssescssesessesccsuecesuee nants ba andesite ee 111 AMERICAN MALACOLOGICAL BULLETIN RONALD B. TOLL, Editor-in-Chief College of Natural Sciences and Mathematics BOARD OF EDITORS University of Central Arkansas Conway, Arkansas 72035 AMELIE SCHELTEMA Biology Department ASSOCIATE EDITORS Woods Hole Oceanographic Institution Woods Hole, Massachusetts 02543 ROBERT S. PREZANT Division of Math and Natural Sciences Queens College, CUNY Flushing, NY 11367-1597 JOHN A. ALLEN Millport, United Kingdom JOHN M. ARNOLD Honolulu, Hawaii, U.S.A. JOSEPH C. BRITTON Fort Worth, Texas, U.S.A. JOHN B. BURCH Ann Arbor, Michigan, U.S.A. EDWIN W. CAKE, JR. Ocean Springs, Mississippi, U.S.A. PETER CALOW Sheffield, United Kingdom JOSEPH G. CARTER Chapel Hill, North Carolina, U.S.A. ARTHUR L. CLARKE Portland, Texas, U.S.A. CLEMENT L. COUNTS, Ill Wallops Island, Virginia, U.S.A. THOMAS DIETZ Baton Rouge, Louisiana, U.S.A. WILLIAM K. EMERSON New York, New York, U.S.A. DOROTHEA FRANZEN Bloomington, Illinois, U.S.A. ROGER HANLON Woods Hole, Massachusetts, U.S.A. and Paleontology California Academy of Sciences San Francisco, California 94118-4599 BOARD OF REVIEWERS JOSEPH HELLER Jerusalem, Israel ROBERT E. HILLMAN Duxbury, Massachusetts, U.S.A. K. ELAINE HOAGLAND Washington, D.C., U.S.A. VICTOR S. KENNEDY Cambridge, Maryland, U.S.A. ALAN J. KOHN Seattle, Washington, U.S.A. LOUISE RUSSERT KRAEMER Fayetteville, Arkansas, U.S.A. JOHN N. KRAEUTER Baltimore, Maryland, U.S.A. ALAN M. KUZIRIAN Woods Hole, Massachusetts, U.S.A. RICHARD A. LUTZ Piscataway, New Jersey, U.S.A. GERALD L. MACKIE Guelph, Ontario, Canada EMILE A. MALEK New Orleans, Louisiana, U.S.A. MICHAEL MAZURKIEWICZ Portland, Maine, U.S.A. JAMES H. McLEAN Los Angeles, California, U.S.A. TIMOTHY A. PEARCE, Managing Editor Delaware Museum of Natural History Box 3937 Wilmington, Delaware 19807-0937 W. D. RUSSELL-HUNTER Department of Biology Syracuse University Syracuse, New York 13210 TERRENCE M. GOSLINER, Ex Officio Department of Invertebrate Zoology THOMAS R. WALLER Department of Paleobiology Smithsonian Institution Washington, D. C. 29560 ROBERT F. MCMAHON Arlington, Texas, U.S.A. ANDREW C. MILLER Vicksburg, Mississippi, U.S.A. BRIAN MORTON Hong Kong JAMES J. MURRAY, JR. Charlottesville, Virginia, U.S.A. RICHARD NEVES Blacksburg, Virginia, U.S.A. JAMES W. NYBAKKEN Moss Landing, California, U.S.A. A. RICHARD PALMER Edmonton, Canada WINSTON F. PONDER Sydney, Australia CLYDE F. E. ROPER Washington, D.C., U.S.A. NORMAN W. RUNHAM Bangor, United Kingdom DAVID H. STANSBERY Columbus, Ohio, U.S.A. FRED G. THOMPSON Gainesville, Florida, U.S.A. Cover: Io fluvialis (Say, 1825) is the logo of the American Malacological Society. THE AMERICAN MALACOLOGICAL BULLETIN is the official journal publication of the American Malacological Society. AMER. MALAC. BULL. 15(1) ISSN 0740-2783 AMERICAN MALACOLOGICAL BULLETIN VOLUME 15 1999 NUMBER | Journal of the American Malacological Society CONTENTS The utility of fossil data in phylogenetic analyses: a likelihood example using Ordovician-Silurian species of the Lophospiridae (Gastropoda: hy tk Murchisoniina). PETER J. WAGNER 0000... cece ceeceeeeteeeeeeeeeeeteteeeteeeeeeeeaeeee geese Ng ALAS Ps a é The Tonicella lineata (Wood, 1815) species complex (Polyplacophora: Tonicellidae), with descriptions of two new species. ROGER N. CLARK. .............ccccccccccseeseeeeens 33 Glycogen concentration in the mantle tissue of freshwater mussels (Bivalvia: Unionidae) during starvation and controlled feeding. MATTHEW A. PATTERSON, BRUCE C. PARKER, and RICHARD J. NEVES .00....occccccccceeesesseessessceseeseceseeseceeceseeseseaeennees 47 Variation in glycogen concentrations within mantle and foot tissue in Amblema plicata plicata: Implications for tissue biopsy sampling. TERESA J. NAIMO eNOS ECIVICYG IVES IV NOOIN RO sci Sees sew per ota yas arn Coeur ve beent oven neunei eo pseeoomnEC pease 51 Recruitment in a freshwater unionid (Mollusca: Bivalvia) community downstream of Cave Run Lake in the Licking River, Kentucky. STEPHEN E. McMURRAY, GUENTER A. SCHUSTER, and BARBARA A. RAMEY ..0.0......ccccccccccceccssccccccssssstsececeeesesentaees 57 Recovery Status of Freshwater Mussels (Bivalvia: Unionidae) in the North Fork Holston River, Virginia. WILLIAM F. HENLEY and RICHARD J. NEVES ..............ccccccccseceseseseeeeeees 65 Historical and ontogenetic changes in shell width and shape of land snails on the island of Kikai. EMIKO HAYAKAZE and SATOSHI CHIBA 0.00... ccccccesceseeceescseenscseeecseeseseeseesees 75 New acavid land snails from Madagascar. KENNETH C. EMBERTOON ...00o ccc ccecesceseeeeeseeeeees 83 Edentulina of Madagascar (Pulmonata: Streptaxidae). KENNETH C. EMBERTON...................00cc0c 97 PDITANI CLAIMING POU pac vers aser oun ne vice ees ances yop eens ra, val tA Ceca ca Sasa’ wed dees zuaene¥y Lane deus aataal@avenionl taneeaeniaiauca ridoeee: 109 PADMOUN GCM ID sero cse sani say int nu tuiiesepettnds etisans daa varsanieviuataiiin votesledesstaenraacsacdd wislegshay cones andeaeaiuiinveercehaniibaestonds 110 UITEIVTC MNO GLU cree cseeeretes estos s5 pete caete ciui ta ehenrre inccs ales oaindal Suiah MD wkeavneallabounercddiecta leseai iN satahesnencanl beskise tends 111 ¥v - + Sa at Maly oh) EDITORIAL COMMENT It is with sadness that I note the recent loss to the science of malacology and to the leadership and membership of the American Malacological Society, of a great friend, loyal devotee, and generous benefactor. Constance E. Boone passed away on September 14 in Australia. It is fitting that she left us while engaged in one of the activities that she loved best, traveling the world, oftentimes well off the beaten path, on collecting trips in search of mollusks. A past president of the American Malacological Society (then the American Malacological Union) and the Houston Conchology Society, Co-Editor of the Texas Conchologist and long-time Associate in the Malacology Department of the Houston Museum of Natural Science, Connie epitomized the knowledgeable and dedicated amateur and helped to strengthen the relationships between and among professional, paraprofessional, and amateur malacologists. An ardent supporter of undergraduate and graduate students and young professionals in malacology, Connie took great interest in seeing that the profession of malacology was in secure hands for the future. There are many of us who were privi- leged to witness and many others who were the direct beneficiaries of her many generous contributions to student paper awards and student research in malacology, which helped to ensure the future of her passion and our science and profession. On a personal level, my memories of two decades of AMS meetings are filled with recollections of Connie’s warmth, friendship, and caring. True to her great form and good nature, one of her first questions to me at the most recent AMS meet- ing this past summer in Pittsburgh was of my wife and three children and how they were adapting to our recent move from Georgia to Arkansas. Her questions then turned to how much time I would have to spend on my research interests and whether I would be able to continue training students in light of my new administrative duties. How typically Connie — thoughtful, supportive, concerned, direct. It is wonderfully appropriate that at the AMS meeting in Pittsburgh, her last AMS meeting, Connie helped to organize and was a contributing panelist for the session on “Women in Malacology”. I was fortunate to have been in the audience that day as she spoke eloquently and passionately about the history of malacology from the perspective of the contributions of those women who were ahead of their time as compared to contemporary standards. In recent years, Connie was particularly pleased to see that women were well represented among the ranks of young professional malacologists and her talk helped to inspire them further to continue in the field and to succeed. Love and appreciation of friends, family, colleagues, as well as adventure, the discovery of the natural world, and working toward the perpetuation of malacology as an attractive and active hobby and science were the passions that drove her and to which she dedicated much of her life and near boundless energies. Her legacy lives on through the many of us whose lives she touched, shared, and enriched. We will miss her. Ronald B. Toll October 1999 ill se 5 ’ cate ‘4 ‘ted CA ORLA : 72 within feta lB) alnene se is leoiepelenin ct? : Tea hd Le Oe - ~~ oi) vil UML) nyt] nate on on let bye ih I ey? enue Waele 4 “a ee Ley See Lymn Abusi* an ’ (Ty B ak? ot auth a fil fang east ub : ii, Salgrag i are Ne: FANE ot bs - epics TAL sal . a : { MM « wie eH <1, PU ae Fp ea otal ts age ante Fae Dini sard ‘all saat boasaly - - Ba 4 “ait Wyte: : - 2 Mle avErd ue fy) Gnitow S bei Gl bn 288 “|i deril seortin | The utility of fossil data in phylogenetic analyses: a likelihood example using Ordovician-Silurian species of the Lophospiridae (Gastropoda: Murchisoniina) Peter J. Wagner Department of Geology, Field Museum of Natural History, Chicago, Illinois 60605 U. S. A., pwagner@fmnh.org Abstract: Gastropods have a dense fossil record dating back to the Late Cambrian. Intuitively, this would appear to aid phylogenetic reconstructions. However, workers question both whether gastropod shell characters are phylogenetically informative and whether stratigraphic data can be used to test phy- logenetic hypotheses. Both questions are addressed with an analysis of 82 species of the Lophospiroidea (= Lophospiridae + Trochonematidae) from the Ordovician and Silurian. Compatibility analyses of 95 shell characters shows that characters are far more compatible than one would expect from homo- plasy-saturated data. However, compatibility among the 73 characters that vary among the earliest lophospirids decreases over time, which suggests that later species introduced homoplasy. Simulations using 95 characters and sampling similar to that observed for lophospirids show that parsimony performs poorly relative to methods that incorporate stratigraphic data such as stratocladistics. An alternative approach is used here. The first step is to estimate the likelihood of a hypothesized tree given observed character congruence (i. e. parsimony length) using simulations. The second step uses two different statisti- cal tests to estimate the likelihood of hypothesized trees given observed stratigraphic data. Likelihoods then are combined to evaluate trees. The resulting likelihood tree is nearly 30 steps longer (378.4 versus 350.5), but is considered more likely (given a 350.5-step matrix) than a tree of 350.5 steps. Both trees suggest that budding cladogenesis (where ancestors co-exist with descendants) was the most common pattern of speciation, although the likelihood tree is more emphatic on this point. Both suggest a trend towards increasing numbers of ornate species; however, whereas parsimony suggests the differential diversification of an ornate clade, likelihood suggests a strong tendency for inornate ancestors to have ornate descendants. A genus-level taxonomic revision is provided that is consistent with both trees and that attempts to reflect historical diversity patterns. Key words: phylogeny, likelihood, fossils, Paleozoic, gastropods Gastropods have a rich fossil record coupled with character congruence will reflect convergence rather than extensive ecologic and morphologic diversity among extant shared ancestry (Felsenstein, 1984; Archie, 1996). The species. These factors make the clade an excellent model dense fossil record of gastropods offers a potential antidote for testing macroevolutionary hypotheses (Bieler, 1992). to the homoplasy problem because phylogenetic hypotheses Most macroevolutionary hypotheses make predictions make necessary predictions about durations, which in turn about phylogenetic patterns. Therefore, robust phylogenet- make probabilistic predictions about stratigraphic ranges. ic estimates offer means of testing alternative macroevolu- Accordingly, several workers have used stratigraphic data tionary hypotheses. Examples of such hypotheses that have to test whether characters linking taxa are better explained been tested in a phylogenetic context with fossil gastropod as convergences than as homologies (e. g. Fisher, 1991, data include hypotheses about speciation patterns (Wagner 1994; Huelsenbeck, 1994; Cheetham and Jackson, 1995; and Erwin, 1995), predator-prey escalation (Carlson and Wagner, 1995b). Vermeij, 1996), phylogenetic constraints (Wagner, 1995a), This paper will attempt to address whether shell and long-term trends (Wagner, 1996). character data retain phylogenetic signal and whether incor- Unfortunately, fossilized character data for gas- porating stratigraphic data improves estimates of phyloge- tropods are limited almost entirely to shell characters, nies. Shell character and stratigraphic data for Ordovician which likely are highly homoplastic (e. g. Harasewych, and Silurian members of the Lophospiroidea (= 1984; Kool, 1993). Phylogenetic methods such as parsimo- Lophospiridae + Trochonematidae) are analyzed using ny likely will perform poorly with shell data because much compatibility, parsimony, and maximum likelihood tech- niques to contrast alternative hypotheses about lophospiroid This paper is a contribution to the 1997 AMU Symposium on Traditional relationships and character evolution. The implications of Versus Phylogenetic Systematics of Mollusks. See American the different results for macroevolutionary hypotheses and Malacological Bulletin 14(2):189. lophospiroid classification also are discussed. American Malacological Bulletin, Vol. 15(1) (1999):1-31 1 Bi AMER. MALAC. BULL. 15(1) (1999) DATA AND ANALYSES SHELL CHARACTER DATA, ANALYZED SPECIES, AND PARSIMONY A previous phylogenetic analysis of lophospiroids used 57 characters and 150 character states to estimate rela- tionships among 55 Ordovician species (Wagner, 1995b). The analysis presented here uses 95 characters encompass- ing 257 character states for 82 Ordovician and Silurian species. The increased number of character states reflects in part the addition of Silurian species with characters and character states unobserved among Ordovician species. Also, the characters were re-coded following Wagner [in press (a)]. Many characters from the 1995 analysis are divided into two or more characters that better reflect varia- tions observed among Early Paleozoic gastropods. Nevertheless, 95 characters might appear to be a large num- ber for a group of gastropods that previously were divided into only two families. However, a survey of numerous phylogenetic studies of gastropods shows that the ratio of shell characters to analyzed taxa increases as the taxonomic level of the analysis becomes finer [Wagner, in press (a)]. Also, lophospiroids are a morphologically diverse clade. Many more characters are needed to described this diversity than are needed to describe any one species. The characters are listed and described in Appendix 1. The character matrix is given in Appendix 2. Continuous characters are divided into ordered series using segment coding and then de-weighted so that the maximum difference equaled one step (Chappill, 1989). As a result, alternative trees often have fractional lengths. Other multi- state characters are treated as ordered if they represented a logical geometric series; otherwise, they are considered unordered. Some characters, such as the shapes and dimen- sions of the right and left ramps, vary both in symmetry (i. e. left and right being nearly identical) and dimensions. For such characters, symmetry versus asymmetry is coded as one character. Left and right characters are coded sepa- rately if they varied independently among asymmetric species. However, left and right characters necessarily covary on symmetric species, which leads to a conundrum: devising a coding scheme where the difference in left and right ramp shape accounts for one difference between two symmetrical species, two differences between two asym- metrical species, and two or three differences between a symmetrical and an asymmetrical species (one difference reflecting symmetry:asymmetry, with the third difference apparent only if both the left and right ramps differ). Step matrices (Swofford and Olsen, 1990) offer a potential solu- tion, but these result in exorbitant run times. Instead, asym- metrical characters are weighted one-half of presence/absence characters, making changes among sym- metrical species equal to one step. Trees are rescaled using a separate program that insures no “chimera” reconstruc- tions (e. g. hypothesized symmetrical ancestors with differ- ent left and right sides), with the tree lengths now scaled as they would be by a step matrix. Broader phylogenetic analyses [Wagner, in press (a)] suggest that lophospiroids are nested within the Murchisoniina, with species assigned to the genus Ectomaria representing the immediate outgroup. Therefore, Ectomaria adelina, which is both the earliest known Ectomaria and contemporaneous with the oldest lophospiroids, is included as an outgroup. Another contem- poraneous murchisoniinae species, Hormotoma simulatrix, also is included as an outgroup. Heuristic parsimony analyses using PAUP 4.0* (Swofford, 1998) found 36 trees of 350.8 steps. The illus- trated tree (Fig. 1) is the parsimony tree with the fewest stratigraphic gaps, which satisfies the secondary role for stratigraphy advocated by Smith (1994). [This is equiva- lent to Clyde and Fisher’s (1997) “Analysis 2.”] The tree is discussed in greater detail below, but it is presented here to introduce the potential errors for the following sections. The 50 random addition replicates found an additional six islands of trees (Maddison, 1991) with minimum length trees of 380.5 or fewer steps. The phylogenetic reconstruc- tion (Fig. 1) accounts for both stratigraphic ranges and character distributions. Both are important because species identical to reconstructed ancestral morphologies obviate the need for hypothesized range extensions (Fisher, 1991; Smith, 1994). DO SHELL CHARACTERS CONTAIN PHYLOGENETIC SIGNAL? Bandel and Geldmacher (1996) criticized a previous phylogenetic analysis of Ordovician lophospiroids by Wagner (1995b) because the resultant phylogenetic esti- mate linked Trochonema and allies to lophospirids. Bandel and Geldmacher (1996) considered these to be distant rela- tives based on Triassic specimens. Those authors attributed Wagner’s (1995b) results to pervasive homoplasy among shell characters. There is some support for Bandel and Geldmacher’s (1996; hereafter B&G) hypothesis. Parsimony reconstructions based on soft anatomy suggest high frequencies of homoplasy among shell characters (e. g. Carlson and Vermeij, 1996; Haas], 1997). The utility of shell characters (on which this analysis necessarily relies) must be questioned and the B&G hypothesis must be tested before proceeding with further analyses. Hypothesized phylogenies predict hierarchical structure among homologies. Conversely, a hypothesis of pervasive homoplasy (such as the B&G hypothesis) pre- dicts little hierarchical signal among the characters. However, the B&G hypothesis does not predict an absence of congruence. Although the terms often are confounded WAGNER: LIKELIHOOD, STRATIGRAPHY, AND LOPHOSPIROID PHYLOGENY snoipunj108 vidsoydoy SUDLLaGD BUOZOYI DISNQO4qns DiuuDWMapany DIDILINU DUJUNT wee = - == 2-2-2 -- ~~ =e Daplouo) Djjaipyouoqg DSandap DIplapjpuoq ~ 1a400u vaidsoydoT psoydopidoay plied! 1uamo piidsoydoT Sisuasauuuns vitdsoydoT oe DIDI4ISINUU} D411dSOYCO] —-=+ 2-220 sisuauising pardsoydoT. 1U2PMOQ 0]]21P[0U0 | s+ - 22 —fo- 2-2 e eee DIVAUI]LAY D1] A1P[OUC CT mmm - = ~~ = 2 2 eee eee pene | ete e ee eee Dayjo10op Daidsoposvg DSOUY, D]JaIpjouUogqum- +--+ 2-2-2 eee eee eee impimiap vaidsopospg Dsoj141uaa v4sidsoydoy Sisuauosupny piuupUDlIy me = - 222 ee eee eee SUDJIadxaUI DIUUDWOLY = -- == ~~~ Ipupjpoom vruupwolsy. DjD]jaIuD) DuuouDLiy: SISUBATDAY] DIUUDUDILY um = = pipuipoiyjaq vaidsoydoT “ds ay[taspana vyjaippouoq pjonpoid vijaipouog = sngnjos iil gus, Sommer iene eset 1y21madpas viidsoydoT sisurysuizapdy{, vitdsoydoy ppunut vaidsoydo7] vioouqui v11dS0/«0|]se—_—_——,""---""~~ "7-776 o 110314902 DAIGSOYdOT] mmm = 2 = = = = += tuqoy viidsoydo mm ---------+--2------ DSO[NGILIL OUff0SIUO = - - - = - = = fe = puanby viffojssuoy: DUgUdjIK) DIffDISSUCT] mm --------- SISUI]IIAUIIUII IPOS SU ————————— = = = = = 2ISNgGOL DIUUDWUAapANY Sisuauqoyysog;, vsidsoydoT DUNASSIAID] DIUUDUapaN Yum - - -- ee ee ee ee eee eee DIDI] DIuUDWApaNny Simumy Duupwapany mulau0sids 041dSOYdO] mmmm= = = = = 2-2 ence cece nes pyoQnsuponb vsidsoydo] mm ----------+---- pynuiguoo vitdsoydo7] mat as DIDIS1U41s Dwauny Sasajonay buidsoydoT sypaquad pacouaey DIDULADINUAS DIuaUoIA) DyJay2]nd viuaUos{t) om - - - 1. DIDULMDNG DIJAT4ANIOL Majin vsidsoydoT] isakvy Duauny DjLLIN] DUZUOYIOL| ——<1—< snjouuDpsiq miulauogo]y pnyo{ Duauoyr04 J, uayjps DUIaUoYIO4 I « DIDIIIqun \] “jyB DUaUOYIOAT we e-em e ee en asuaUosipou DUaUOYIOL J, DIN]]9q DuaUOYIOL iL a 1uopund "J “Jy DUIJUOYIOL] mmm == = eee ee ee eee ee ee Saplowauoyro.s D]]aUusaUoYyIos J wtunqjayoind “p,, Djjauauoyso4] Dypnisas ~] “yye vsidsoydoT 1duisnas D]JaWaUOYIOL 1uIyANYD DjjwWauoYyIo4 T S1]qQDjOUu DjJauMaUuoYyIOL V ‘ds'u ppjauauoys0s 7 SISUI]DAATUOUL D]]IUIUOYION [mead © ~~ SISUI][IAXOUY DIJaWuMaUoYyIOs | pyojnssas vardsoydoT snqo1sisijga4 vaidsoydoT Dynz4044108 Daidsoydoy vjojn8uvsad vsidsoydoT Tan ana OPeae TI qa Al d at aq I CIE IW a a a T Topyd = Moppny yoopaam, = Ayoaopuey T= fiSysy oopere) UAUeL] Stuy ueLIN|IS UISIAOPIO) Ww Thin Fig. 1. Phylogenetic interpretation of one of 36 parsimony cladograms. This tree implies less stratigraphic debt than does any of the other 35 trees. solid lines show estimated phylogenetic connections. Dashed lines show hypothesized unsampled lineages and taxa that are implicit to the phylogenetic 1993). The tree posits 350.5 steps [consistency index (CI) ’ hypothesis (i. e. range extensions sensu Smith, 1988, 1994; = ghost lineages and taxa sensu Norell 0.316; retention index (RI) = 0.807) and 102 units of stratigraphic debt, with debt units based on the substages marked on the time scale. Species identical to hypothesized ancestors are linked directly to nodes and are considered ancestral to their apomorphic sister taxa here. identical to Trochonemella BMNH 36364 in Appendix 2.] “G. pulchellum” is [Trochonemella important assumption of phylogenetic methods, i. e. the independence of characters.) (e. g. Bryant, 1992; Carpenter, 1992), hierarchy and congru- ence are not synonyms: congruence exists in non-hierarchi- Several workers (e. g. Meacham, 1984, 1994; Sharkey, 1989, 1994; Alroy, 1994) have suggested using cal matrices (KAllersjo et al., 1992; Alroy, 1994) and even in randomly generated matrices (Archie, 1989; Faith and character compatibility (Le Quesne, 1969; Estabrook et al., Cranston, 1991). However, whereas hypotheses of phyloge- ny and homology predict hierarchical congruence among characters, hypotheses of homoplasy predict non-hierarchi- Two 1975) to evaluate hierarchical content of matrices. binary characters with pairings of (00), (10), (01), and (11) must have homoplasy in one or both characters. Such cal congruence. (The one exception to this expectation is when multicharacter complexes are homoplastic and pat- incompatible characters necessarily lack global hierarchy. Two binary characters with only three of the possible pairs terned so that character combinations appear in the same [e. g. (00), (10), and (01) or (11)] are considered compatible order multiple times; however, this begins to violate another 4 AMER. MALAC. BULL. 15(1) (1999) because they can be reconstructed on some trees without homoplasy. Characters with little homoplasy should form many compatible pairs whereas those with numerous homoplasies should form many incompatible pairs. Alroy’s (1994) Permutation Compatibility (PC) test is Operationally similar to the Permutation Tail Probability test (Faith, 1991). PC tests whether the hierarchical struc- ture of a character matrix is significantly greater than that of a randomly produced matrix. Alroy (1994) divided com- patibility into two types: hierarchical [= “direct” of Sharkey (1994)] and general (“indirect”). The former describes pairs where combinations show a hierarchical arrangement [e. g. (00), (01), and (11)]. The latter describes pairs where there is neither necessary homoplasy nor implied hierarchy [e. g. (00), (01), and (10)]. Because phylogenetic hypothe- ses explicitly predict hierarchy, I used the PC test with 1,000 randomized matrices to determine if randomly per- muted matrices ever retained the same level of hierarchical compatibility as did the real matrix. Lophospiroid character data yielded 1,389 hierarchically compatible pairs out of 8,930 possible pairs. (The test has been amended here to compare characters, not character states, as was done in Alroy’s implementation). Permuted matrices typically showed approximately half as many pairs (Fig. 2). Thus, the matrix has more hierarchical structure than expected if congruence largely reflected random homoplasy, allowing us to reject the B&G hypothesis. 0.12 0.10 0.08 0.06 0.04 0.02 Proportion of Permuted Matrices 0.00 700 800 900 1000 1100 1200 1300 1400 Hierarchically Compatible Character Pairs Fig. 2. Results of the Permutation Compatibility test (Alroy, 1994). The arrow denotes the observed value whereas the histogram gives the values generated by 1,000 random permutations of the character data. Hypothesized phylogeny predicts hierarchical compatibility whereas other hypotheses of character congruence (e. g. general convergence) do not. Thus, these results support (but do not demonstrate) the idea that there is phylogenetic signal among these characters. DO WE NEED STRATIGRAPHIC DATA TO IMPROVE OUR ESTIMATES OF LOPHOSPIROID PHYLOGENY? The presence of some hierarchical structure is no guarantee that there is not sufficient homoplasy to mislead parsimony. Simulation studies show that even low rates of randomly accrued homoplasy will result in inaccurate par- simony reconstructions of phylogeny (e. g. Mooers et al., 1995). Finite numbers of character states (see Wagner, 1998c) insure that the probability of chance homoplasies confounding parsimony is not infinitesimal (Felsenstein, 1978). Patterned homoplasies (e. g. functional complexes) violate assumptions of character independence and also lead to inaccurate parsimony reconstructions (Lamboy, 1994; Archie, 1996). Finally, rates of character change that are not adequately reflected by character weighting (see Felsenstein, 1981) also yield inaccurate parsimony trees (Kuhner and Felsenstein, 1994). Neontologists can test phylogenetic hypotheses derived from one data set by examining different character sets (e. g. different types of molecules). The hypothesis that congruence from the original data set reflects phyloge- ny predicts very similar patterns of congruence in other data sets (assuming that changes in one character set do not affect changes in the other; Templeton, 1983). Unfortunately, only a single character set is available for extinct taxa such as lophospiroids. However, phylogenetic hypotheses also make necessary predictions about temporal durations of taxa (see, e. g. Smith, 1988). Predictions about durations make probabilistic predictions about stratigraphic ranges (Paul, 1982; Strauss and Sadler, 1989). If a hypoth- esis makes predictions (necessary or probabilistic) about a data set, then those data offer a test of the hypothesis. Additional factors suggest that stratigraphic data might improve estimates of lophospiroid phylogeny. Simulation studies [Wagner, in press (b)] show that phylo- genetic error exaggerates the phylogenetically implied range extensions (Smith, 1988, 1994; “ghost” lineages and taxa sensu Norell, 1993) far more often than error underes- timates range extensions. The exaggeration becomes worse as frequencies of change per character per branch increase, but it is pronounced at the frequencies of change posited by the parsimony tree for lophospiroids (i. e. f = 0.042 per character per branch). Thus, the long gaps in sampling posited in Fig. 1 are evidence that synapomorphies (sensu Sober, 1988) are homoplastic rather than homologous. For lophospirids, most gaps are highly improbable given observed sampling distributions (Wagner, 1995b). Another concern is that the hierarchical signal of the character matrix might decrease as geologically younger taxa are added. Suppose that geologically older species have morphotypes (00), (01), and (01), but that later species introduce morphotype (10). The characters no longer are WAGNER: LIKELIHOOD, STRATIGRAPHY, AND LOPHOSPIROID PHYLOGENY 5 Proportion of Compatible Character Pairs 0 10 20 30 40 50 60 70 80 90 Number of Species Fig. 3. Effects of geologically younger taxa on overall compatibility. As taxa are added over time, compatibility decreases. The white line gives the compatibility among all characters after X species are added to the matrix. Crosses give the compatibility among only the 73 characters informative for “early” (i. e. pre-Caradoc) relations. Thin black line gives the compati- bilities for all characters when stratigraphic ranges are assigned at random. compatible, and the younger species necessarily imply reversal or convergence. Conversely, suppose that the older species had states (00) and (11) and some geologically- younger species had (10). This suggests that the “younger” species is descended from a phylogenetic intermediate that was not sampled initially. The former scenario predicts decreased character compatibility whereas the latter does not. I examined the effect of geologically younger taxa on character compatibility among lophospirids using all character and also only the 73 characters that vary among early lophospiroid species (Fig. 3). (Late-appearing char- acters are not relevant to relationships among the oldest taxa.) Compatibility decreases markedly over time. Sampling so poor that species were effectively sampled at random through time relative to their phylogenetic position (as implied by abundant ghost lineages) predicts a very dif- ferent pattern. One thousand randomizations of stratigraph- ic ranges show that compatibility should begin lower than observed and that the decrease over time should be much less precipitous (Fig. 3). The observed patterns are consis- tent with the idea that Late Ordovician and Silurian lophos- piroids exhibited homoplasy among the characters that dis- tinguish Early-Middle Ordovician lophospiroids. A final concern is whether methods incorporating stratigraphic data perform better than does parsimony. (“Parsimony” here and throughout this paper denotes mini- mum steps evolution [Edwards and Cavalli-Sforza, 1964; Kluge and Farris, 1969]; other methods using parsimony criteria [e. g. stratocladistics] are labeled differently.) To test this, I simulated morphologic evolution using the num- ber of characters and states per character apparently avail- able to lophospiroids. Character evolution was ordered or unordered reflecting the assumptions of the phylogenetic analysis (Appendix 1). In one round of analyses, log-prob- abilities of character change were proportional to the rela- tive weighting of characters (1. e. “equiprobable change”; see Felsenstein, 1981). In a second round, rates of change varied among characters at random (i. e. “variable change’’). Two different speciation models were used, one in which ancestral morphotypes survived cladogenesis and could produce any number of descendants (“budding clado- genesis”) and one in which ancestral morphotypes became pseudo-extinct while giving rise to two distinct descendants (“bifurcating cladogenesis”). Sampling parameters derived from lophospiroid data (see Appendix 3) were used to sam- ple six species. (A six-species limit was imposed so that exhaustive searches could be used.) Parsimony analyses then were run using the simulated character matrices. Matrices were maintained only if parsimony estimated the same steps per sampled taxon as the real lophospiroid tree (1. e. 350.5 steps per 82 taxa ? 25.5 steps per six taxa; here- after: a 25.5-step matrix). The simulations then had character matrices and fossil records that were similar to those observed for lophospiroids. The stratigraphic data then were used for three additional phylogenetic analyses: reweighting with the Stratigraphic Consistency Index (SCI; Huelsenbeck, 1994), stratocladistics (Fisher, 1994), and sieving with 95% confidence intervals (CIS; Wagner, 1995b). The results of the phylogenetic analyses were contrasted using Robinson and Foulds’ (1981) metric, which measures the proportion of nodes that agree on two trees. (If multiple optimal trees were found, then the mean error was used.) All methods incorporating stratigraphic data outper- formed parsimony when simulated character matrices and simulated sampling matched observed parameters for lophospiroids (Fig. 4). In the best case, stratocladistics cor- rectly reconstructed over 50% of the cladograms. Parsimony, meanwhile, correctly reconstructed under 15% of the cladograms. Stratigraphic methods are less success- ful when rates of character evolution are not proportional to character weighting, something that is true of parsimony and other phylogenetic methods (Kuhner and Felsenstein, 1994). However, stratigraphic methods are slightly less affected by such variation than is parsimony. In summary, lophospiroid character data show too much compatibility to be dismissed as noise. However, reasons to question the parsimony estimate of phylogeny include: (1) many implied but statistically improbable sam- pling gaps, (2) decreasing hierarchical signal over time, and (J Parsimony SCI Reweighted B83 Stratocladistic CI Sieved ‘“Equiprobable” Change Relative Frequency gy WWMM, VLE E_ Variable Change Relative Frequency . N N N N \ \ \ \ \ \ N \ 0 1 2 3 4 Number of Incorrect Nodes Fig. 4. Success of four phylogenetic methods for six-taxon simulations. The numbers reflect the number of incorrect nodes (“0” indicates a correct tree). The simulations use 95 characters and the same number of possible character states and character state orderings per character as lophospiroid data. Sampling intensities mimic observed sampling. “Equiprobable” evo- lution used probabilities of character change proportional to the character weight, whereas “variable” evolution used randomly assigned probabilities of change. Results are from budding cladogenesis, in which ancestral species can survive to yield any number of descendants, but the simula- tions produce nearly identical results when using bifurcating cladogenesis. Note that all three methods using stratigraphic data outperform parsimony. See text for additional discussion. (3) superior performance of methods incorporating strati- graphic data in simulations. WHY MAXIMUM LIKELIHOOD APPROACHES ARE USEFUL How best to incorporate stratigraphic data into phy- logenetic analyses is a non-trivial issue. I reject the posi- tion that stratigraphic data cannot improve phylogenetic estimates because such data are not hierarchically distrib- uted (e. g. Eldredge and Cracraft, 1980; Rieppel and Grande, 1994; Smith, 1994). This only means that strati- graphic data do not offer inductive statements about phy- AMER. MALAC. BULL. 15(1) (1999) logeny. If the goal is to test the predictions of phylogenetic statements, then one criterion applies: do estimated phylo- genies make predictions about stratigraphic data? The syl- logism is straightforward: if an inferred phylogeny is accu- rate, then taxa originated by particular times; if taxa origi- nated by a particular time, then there is a some probability that the known fossil record would be observed. Of course, the “deduction” is necessarily fuzzy, as stratigraphic data cannot demonstrate that a lineage had originated by any point in time, only that it is highly unlikely to have done so given observed data. Unfortunately, the methods incorporating strati- graphic data used above all are unsatisfactory. Reweighting trees with SCI suffers because the SCI is greatest on pecti- nate trees (Siddall, 1996). Sieving trees with confidence intervals is not obviously predisposed toward favoring par- ticular topologies. However, CIS suffers from focusing on particular nodes rather than the whole tree (Wagner, 1998b). Multiple independent tests and a traditional signif- icance value of 0.05 should yield Type I errors (i. e. incor- rectly rejecting the null) one time in every 20 tests. More damning, multiple tests should yield Type II errors (i. e. incorrectly accepting the null) even more frequently. For example, if there are ten gaps that are significant at a p- value of 0.10 each, then the CIS method will accept all of them. However, we expect only one of those gaps to be real. An alternative approach with sounder logical justifi- cations is to use likelihood to evaluate hypothesized trees. Stratocladistics actually represents a simple likelihood test. Summing steps is equivalent to summing negative logs of character change probabilities (Felsenstein, 1981) if: (1) probabilities for each character are the same among all clade members, (2) characters never change twice on the same branch, and (3) characters always evolve indepen- dently. The most likely tree then is the one invoking the fewest steps. Summing the negative log-probabilities of implied gaps (i. e. stratigraphic debt sensu Fisher, 1991, 1994) estimates the log-probability of a tree. Again, the most likely tree is the one implying the fewest stratigraphic gaps. The log-likelihood of the entire tree now equals the sum of stratigraphic debt and morphologic steps if the stratigraphic debt is weighted as: -In (1 - R) = W * -In (P [c]) where | - R is the probability of not sampling a taxon over a particular interval (Foote and Raup, 1996), W is the weight of stratigraphic debt relative to characters, and P [c] is the probability of character change along a branch. Stratocladistics will retain nodes implying improba- ble gaps if they are supported by numerous synapomor- phies. This offers a control for Type I errors which is lack- ing in the confidence interval method. Clyde and Fisher (1997) further justified stratocladistics on the grounds that WAGNER: LIKELIHOOD, STRATIGRAPHY, AND LOPHOSPIROID PHYLOGENY 7 it uses the same optimality criterion to evaluate all of the data, which is the justification of “total evidence” analyses (e. g. Kluge and Wolf, 1993). However, stratocladistics requires a priori assumptions about sampling intensity (R) and frequencies of character change. As unknowns, both should be tested rather than assumed. Fortunately, the stratigraphic debt for any one tree implies an R, the likeli- hood of which can be assessed analytically (Foote, 1997). R’s predicting some debt are more likely than R’s predict- ing no debt (Wagner, 1998a), which means that statistically optimal trees can be “suboptimal” by stratocladistic criteria. Nevertheless, the basic format of stratocladistics serves as a useful logical template. Because stratigraphy and morphology are independent data sets, the overall likelihood of any hypothesized tree is the likelihood of the tree given the character data times the likelihood of the tree given the stratigraphic data (Edwards, 1992). This is the approach that 02 HO NOS — i) S nN .04 02 | Known Length [ 23 25 00+rrF 19 21 27 AD7 21 will be taken in this paper (see also Wagner, 1998a). CALCULATING LIKELIHOOD OF A HYPOTHE- SIZED TREE GIVEN THE CHARACTER DATA Congruence is an observable datum of a character matrix that is summarized by the parsimony length. The trees of interest are treated as hypotheses to be tested here. The likelihood of a hypothesized tree length (HL) yielding a parsimony length (PL) is proportional to the probability of observing the datum (PL) given the hypothesis (true tree length) (Edwards, 1992). Thus: L [HL | PL]

LAL =25.5 \PL = 25.5) This is best appreciated by examining the relative heights of the solid gray bars in Fig. 5, which approximate the like- lihood of HL’s given the observed congruence. The bar at 27.25 is the highest, indicating that this is the length most likely to have yielded the observed congruence. One point not discussed above is that the simula- tions estimate P [PL | KL] only if a particular hypothesis of character evolution is assumed. Fig. 5 illustrates the results when simulated character evolution is derived from the weighting scheme (see Felsenstein, 1981). Deviations from this models resulted in longer trees lengths becoming more likely and the shortest tree length becoming less likely (Wagner, 1998a). Thus, these results represent the highest possible likelihoods for parsimony hypotheses. We need to evaluate 82-taxon phylogenies to evalu- ate lophospiroid phylogenies. This is accomplished by ran- domly selecting 13 six-taxon clades and one four-taxon clade (derived from a separate set of simulations not illus- trated here). Parsimony lengths among random six-taxon clusters within lophospiroids varied from 15.8 to 29.6 steps. Therefore, true tree lengths were randomly selected from the range of true lengths that yielded parsimony lengths of 15.8 to 29.6 steps (i. e. 15.8 to 54.1 steps). Note that the use of a range of true lengths means that estimated proba- bility does not assume clock-like rates of morphologic change (contra Norell and Novacek, 1997). Parametric bootstrapping (see Huelsenbeck et al., 1996) then assigned parsimony lengths to each true length. Assignments used distributions such as those illustrated in Fig. 5. The true lengths and parsimony lengths of the 14 subclades are then summed, which yields the likelihood distribution for hypothesized tree lengths given 82 taxa and lophospiroid character data (Fig. 6). Log-likelihoods are illustrated because the likelihoods of shorter trees are too low to be visible on a histogram. Parametric bootstrapping never reconstructed any trees with both true and parsimony lengths of 350.5 steps. (This is not surprising, as the probability of doing so is approximately 10-14, which requires several orders of mag- nitudes more replications than performed.) Because the likelihoods of tree lengths under 356 steps could not be estimated from parametric bootstrapping, I assigned log- likelihoods of that length (-10.415) to all shorter lengths. Alroy (pers. comm., 1998) suggested that the likeli- hood estimates for the parsimony length might be too low because heuristic searches were used to estimate the initial parsimony length of 350.5 steps. However, for a hypothe- vy 0 S LV ay ioe) -2 s lov 0) & -4 3 al S -6 = a 3 4710 = -12 350 360 370 380 390 86400 Hypothesized Length (HL) Fig. 6. Log-likelihoods of hypothesized tree lengths given a 350.5 step matrix for 82 taxa with lophospiroid characters. Note that trees 20-30 steps longer than the parsimony length are orders of magnitude more likely than are trees of 350.5 steps. sized length of 350.5 to be as likely as a tree length of 370, the actual parsimony length must be around 330 steps. It seems improbable that heuristic searches could have missed by that much. Other aspects of the test overestimate the likelihoods of parsimony lengths (Wagner, 1998a). The test does not allow parsimony to confound homoplasy among subclades. Also, it uses the best-case model of character evolution (see above). All other models result in a lower probability of parsimony lengths equaling hypothesized lengths. These factors, coupled with the deliberate overes- timate of parsimony length likelihood described above, mean that the tests used herein are quite conservative if one wishes to treat the parsimony tree as a null hypothesis. CALCULATING LIKELIHOOD OF A HYPOTHE- SIZED TREE GIVEN STRATIGRAPHIC DATA Continuous Stratigraphic Data Wagner (1995b) used continuous stratigraphic data to estimate confidence intervals for lophospiroids. Huelsenbeck and Rannala’s (1997) maximum likelihood test was applied using these data (see Appendix 3). As in the 1995 analysis, the “time” scale used here is actually the number of sampled units (“horizons”) per stratigraphic interval. Thus, gaps through poorly sampled intervals are more likely than are gaps through well-sampled intervals, even if both intervals are of the same duration. Huelsenbeck and Rannala’s (1997) test is modified slightly here. The most important modification is that the WAGNER: LIKELIHOOD, STRATIGRAPHY, AND LOPHOSPIROID PHYLOGENY 9 phylogeny is no longer considered a predictor of species’ extinctions. This would be true if species evolved anage- netically. However, budding cladogenesis patterns predict only the latest possible origins of species. The parsimony tree (Fig. 1) illustrates possible examples of budding clado- genesis, as parsimony finds several plesiomorphic species known after their apomorphic sister taxa appear. (‘“Plesiomorphic species” are those that match inferred ancestral morphotypes.) Previous phylogenetic analyses suggest that budding cladogenesis was the predominant speciation pattern among lophospiroids (Wagner and Erwin, 1995). The exact probability of data given a hypothesized origination now is: P [FKA, LKA, H | HFA, f] = f ( (LKA - FKA)H-2 fl e-f (LKA - HFA), (H - 2)! ) when H _ 2 = f ef (LKA- HFA) when H = 1 where FKA is the first-known appearance, LKA is the last- known appearance, HFA is the hypothesized first appear- ance, H is the number of horizons from which the species is known, and f is the proportion of horizons within a species range from which it is sampled (amended from Huelsenbeck and Rannala, 1997: equations 2 and 4). The likelihood of a hypothesized origination time implicit to the inferred phylogeny now is: L [HFA, | FKA, LKA, H ] =c * IPR(I= 1,1 = nodes + OTUs - 1, P (FKA, LKA, H | HFA, A) ) (amended from Huelsenbeck and Rannala, 1997: equation 5) where c is an arbitrary constant that rescales each likelihood (L), so that the maximum likelihood is 1.0 (Edwards, 1992). Discrete Stratigraphic Data The stratigraphic debt implicit to an inferred phy- 6 5 = 3 > 2 8 more, 4 x © 5 am ‘S = Al i] 0 RON pS BC Duration (in Substages) logeny implies a particular R (Wagner, 1998a), the likeli- hood of which can be evaluated (Foote, 1997). (“Sampling intensity” here denotes the average proportion of species per interval that are sampled; Foote and Raup, 1996). Ifa species invokes four units of stratigraphic debt, then the phylogeny implies that the species was sampled in its fifth unit after diverging from its closest sampled relative. The phylogeny therefore implies a sampling intensity of 0.2 (i. e. one find in five tries). For an entire phylogeny, 90 units of stratigraphic debt distributed among ten species suggests an average sampling intensity of 0.1. Because phylogenies infer latest necessary divergences (Smith, 1988), stratigraphic debt gives the minimum implied gaps. A simple analytic estimate therefore is biased toward over- estimating R, especially on large phylogenies (Wagner, 1998a), so R is best estimated using simulations. Stratigraphic debt was calculated using the sub- stages marked in Fig. 1. For the parsimony tree, there are 102 units of stratigraphic debt. Stratigraphic debts of 102 for 82 species (two of which are present in the earliest sub- stage) imply an R = 0.33. (Note that an analytic R = f (80, 182) = 0.44 is substantially greater.) Taxon ranges reflect both R and durations (which reflects extinction intensity, 4) (Sepkoski, 1975; Foote and Raup, 1996). As sampling becomes worse, ranges will decrease regardless of actual durations. In the extreme case of nearly infinitesimal R, the few observed species will each be known from single localities and all species will have ranges of one unit. Such clearly is not the case for lophos- piroids (Fig. 7A). The most likely hypothesis for observed lophospiroid ranges is R = 0.68 per substage (Fig. 7B, with yu = 0.52 per substage). Given the R associated with given 1.0 0.9 oo 0.8 0.7 0.6 0.5 0.4 0.3 nAnnn Wu NNN 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0.2 B 0.2 0.3 Sampling Intensity (R) Fig. 7. Estimates of the likelihood of sampling intensities (R) and extinction intensities (u). A. Distributions of observed ranges, with ranges measured by the major divisions shown in Fig. 1. B. Log-likelihoods for hypotheses that particular combinations of y and R yielded the distribution shown in A. The maxi- mum likelihood estimate is R = 0.68 per substage and uy = 0.52 per substage. 10 AMER. MALAC. BULL. 15(1) (1999) amounts of stratigraphic debt, it now is possible to assess the likelihood of any given amount of stratigraphic debt (Fig. 8). RESULTS BASIC ASPECTS OF THE ALTERNATIVE TREES Likelihoods were calculated using both methods for 100,000 trees from several different islands. Both methods found strongest support for the same tree (Fig. 9; hereafter, the ML tree). Traditional tree statistics are presented in Table 1 and the figure captions. The ML tree is generally similar to the parsimony tree (Fig. 1) and also to the estimates provided in Wagner (1995a). Trochonema is again placed among Troch- onemella spp., whereas other taxa previously assigned to the Trochonematidae (e. g. Eunema, Gyronema, Pro- turritella) are derived separately from traditional lophos- pirids. This corroborates previous several previous hypotheses (e. g. Ulrich and Scofield, 1897; Wenz, 1938; Knight et al., 1960; Erwin, 1990). The ML tree (as the CIS tree before it) derives the problematic Lophospira serrulata directly from the con- temporaneous L. perangulata. That species was previously linked by parsimony to the clade of L. rectistriata; here, parsimony links it to the Trochonemella-Trochonema clade. L. serrulata shares several homoplasies with other species regardless of where it is placed, including open coiling of the gerontic whorls, pronounced sutural and umbilical cari- nae, and sharp right and left ramp carinae. Notably, these features are shared with species known from the same sedi- ments, such as Trochonema umbilicata, Trochonemella — Analytical Estimate — 95% UB from Simulations —= Simulation Estimate 0 10 20 30 40 50 60 70 80 90100110 A Stratigraphic Debt Table 1. Properties and statistics of alternative phylogenetic hypotheses. (HG, hypothesized gaps in horizon-level sampling; HL, hypothesized tree length; L [H | r], likelihood of a hypothesis given the measurable data; ML, maximum likelihood; Pars, parsimony; PL, parsimony reconstruction of the character matrix; SD, hypothesized stratigraphic debt, with debt measured in the same stratigraphic units as “Ranges’’). Tree HL L[HL!IPL] SD L{[SD1Ranges] L [HG| Horizons] Pars 350.50 <3xl0° 102 5.95x10° 1.04x 107)! ML 378.37 0598 oF. 0252 4.96 x 10° notablis, L. helicteres, and Eunema strigillata. These fea- tures are also homoplastic among these same taxa, which makes the assessment of these features as homoplastic on L. serrulata all the more plausible. Multicharacter homoplasies suggest that characters did not evolve independently. However, the homoplastic characters appear in different orders among the different lineages (1. e. morphotype 11 might be derived from 00 > 01 in some cases, 00 — 10 in others, and directly from 00 in still other cases). Also, the characters in question all vary freely in other parts of the clade. Finally, these characters sometimes present the only discernible differences among morphospecies. The initial assumptions of character inde- pendence almost certainly are violated, albeit “fuzzily.” Sophisticated likelihood tests examining varying degrees of interdependence among characters offer one solution for this problem. However, it bears stressing that the likeli- hood method used here recovers a pattern that violates the initial assumptions whereas parsimony does not. Regardless of the implications for character coding, j=) ' i) -4 In L[Stratigraphic Debt | Ranges] -14 0 10 20 30 40 50 60 70 80 90100110 Stratigraphic Debt Fig. 8. A. Estimated sampling intensity (R) given stratigraphic debt distributed among 82 species (with two present in the first interval). Shown are the ana- lytic solution, the average from 1,000 simulations and the 95%. B. Log-likelihoods of stratigraphic debts (see Fig. 7). WAGNER: LIKELIHOOD, STRATIGRAPHY, AND LOPHOSPIROID PHYLOGENY = 3 tc a ES 2 3 i] — s g Bx S te aS A: : os : 2 8 32 23 = S 3 2 8 ~ DES 3 sR Sis 33 see = fra 3 SSE F = eg Sse 3 a vu is G0 KR 23 BERS 23 8 84S sy 3 e &8SSe BE § Ssue 2se iS 2 eps SE 8 sg os3 z a cy ar 4 = 3 os £2] Sse Sas ERS & 5&5 3 g 3 a4 ees Sfp a EeEase ES § iS 4 = Sst Sess ars 2 S8ss 84 8 = fH, Sss Ges Sot SG os8 ‘= 8 3 Ra; Sk shea ssa Seas eee - = 3 Srl OSs < § = = BoNT 1 GERSSI ERS 2 ypRONR BEE Bs 8 ‘es S 83 se ~s[Si88 § 3 3 § ah a 3 EF SA poeVeSsn = cS 2 Bes a 5 eg: SSTaiSss ss. 8 258 < qa fs “]s]ssss .§.8 3 gs § §$ ae Sisses #8 3 ; & | = 8 8 Sas 8 P sg = Se ~ a iS 8 SS % 4g S28 = 3 Be 3 8 aes ED ae Ss of 4s = Anak a 3 sf 8S 8 3 $ ES s 2 3 6 ff << 8 ey iS seach 2 6 sy = iS] 3 es 8 8 ES Ss QM €£ se & ao 3 & gs Q S} s S So: 2 2£< 3 2 ls8 E iS 2 2S 3 of 5 38 S$ s .ES aS @ 938 as Ele ry Ss = fois As 8 3 Sy Sy S28 go§ ws (hs LE ae © ra 238 28 < ae) . 8 3 3 t< Ss Es . ES) . 3 S Ses safe ac) ESS FS:8 3 Sv$ #15 4 ges Soe ELSESTEVE se8gezs's& SeoRxrsie eSsSes'3 SPS ses SERSSE:2 Sse cess SSS EREIS SOS 8338 Q Lophospira rectistriatus Proturritella bicarinata Fig. 9. Maximum likelihood estimate of lophospiroid phylogeny. The tree posits 378.4 steps (CI = 0.293; RI = 0.784) and 25 units of stratigraphic debt, with debt units based on the substages marked on the time scale. The tree is considered more likely not simply based on stratigraphic debt (Fig. 10) and continu- ous stratigraphic data (Table 1), but also on morphologic data (Fig. 8). As in Fig. 1, solid lines show estimated phylogenetic links and dashed lines denote ghost lineages or taxa. the overall pattern is consistent with a hypothesis of strong functional or ecological convergence among these taxa. (Lophospiroids lacking these features are known from the same strata, so the characters cannot be dismissed simply as ecophenotypic variants.) Returning to the particular ques- tion of Lophospira serrulata, that species retains many ple- siomorphic features. Small specimens are very similar to L. perangulata, save that the medial lira of the sinus keel is serrated and a prominent right ramp carina is present. (Note that “sinus keel” is used here instead of selenizone, as the latter term is used to denote a morphogenetic artifact of a slit. Sinus keels appear long before slits and the fea- ture represents a separate homology that happens to be in the same location as a slit.) If the additional synapomor- phies realized on larger specimens are associated with a functional complex, then L. serrulata might be separated from L. perangulata by fewer evolutionary innovations than implied by the character coding. Parsimony links Donaldiella bowdeni and other high-spired lophospiroids to trochiform species such as Lophospira burginensis which differs from the trees pre- sented by Wagner (1995b). The ML tree differs by (1) placing only D. bowdeni and the Silurian D. trilineata in that clade, and (2) considering the characters associated with the high-spired morphology to be secondarily derived rather than primitive. All trees suggest that the earliest 12 AMER. MALAC. BULL. 15(1) (1999) lophospiroids were high-spired, corroborating Grabau’s (1922) hypothesis about lophospiroid origins. However, the ML tree corroborates Ulrich and Scofie!d’s (1897) hypothesis that D. bowdeni was derived from L. oweni rather than from older high-spired species. LIKELIHOOD RATIO TESTS Despite being nearly 10% longer than the most-par- simonious tree, the ML tree is orders of magnitude more likely than the parsimony tree given morphologic data alone (Table 1). (Again, the likelihood was calculated assuming that character evolution matched the weighting scheme of parsimony.) If variable rates of character change are used or if patterned homoplasies are included (both of which are biologically realistic), then shorter trees become even less likely whereas longer ones become more likely. The only model of evolution yields most likely lengths greater than 370 steps (i. e. the approximate most likely length gives a parsimony length of 350.5 steps among inde- pendently evolving lophospiroid characters): one in which there is no homoplasy. However, this hypothesis is falsified by the demonstration of incompatibility among lophos- piroid characters. (In essence, this also falsifies the abduc- tive predication of the parsimony proposition “if phyloge- netic topology, then maximum congruence topology”). Likelihood ratio tests can be used to evaluate the relative support for competing hypotheses. The test statis- tic is: 6 =2 * (InL; - InLg) where Lg is the likelihood of the null hypothesis (here, the parsimony tree) and L; is the likelihood of the test hypothe- sis (here, the likelihood tree) (Sokal and Rohlf, 1981: 695). The results of this test are given in Table 2. Ordinarily 6 is evaluated using a chi-square distribution. Goldman (1993) cautioned against doing so when evaluating phylogenies and suggested using simulations to determine the distribu- tion of 6 instead. However, simulated 6-distributions differ from a chi-square distribution at low 6-values, but not at high ones (see Wagner, 1998a). The extremely high 6-val- ues indicate that both morphologic and stratigraphic data provide significantly more support for the ML tree than they do for the parsimony tree. Table 2. Log-likelihood ratio tests contrasting the parsimony estimate with the maximum likelihood estimate. “8” gives the ratio test statistic and is evaluated using a Chi-square distribution. In L[tree | PL & Strat. Data] Stratigraphic Likelihood Parsimony ML rs) i) Stratigraphic Debt 2245 -189 41.11 144x10!° Horizon Data 496.22 -9792 796.59 2.97x 1077 fa ML (Continuous) ML (Discrete) 2 [_] Parsimony 3 F Ba Stratocladistics o" CI Sieved sa re g 3 = C 2 wl 0 0 1 2 3 4 Number of Incorrect Nodes Fig. 10. Performance of maximum likelihood methods. “Continuous” uses the method of Huelsenbeck and Rannala (1997) to calculate stratigraphic likelihood; “Discrete” uses the method of Foote (1997) to calculate strati- graphic likelihood. The simulations used the “variable” model of character evolution, but the maximum likelihood estimates for morphologic change assumed an equiprobable model. Nevertheless, both methods outper- formed parsimony, stratocladistics, and sieving with confidence intervals. THE EFFICACY OF MAXIMUM LIKELIHOOD METHODS An important consideration when contrasting differ- ent methods is relative performance under simple cases (Felsenstein, 1981, 1984). To this end, I conducted an additional set of simulations that contrasted the relative efficacy of the two maximum likelihood methods with that of parsimony, stratocladistics, and confidence interval siev- ing. Simulations were conducted using the variable rate model of character evolution described above (see Fig. 4). Note that the model of character change did not match the assumptions of the likelihood estimates of morphologic change. In addition, simulated sampling was made more variable, with sampling densities and sampling opportuni- ties varying from interval to interval. Both methods perform far better than does parsimo- ny (Fig. 10), and both also outperform the SCI, CIS, and stratocladistics. Huelsenbeck and Rannala’s (1997) method does somewhat better than does the debt-likelihood test. This is unsurprising because Huelsenbeck and Rannala’s (1997) method can distinguish the relative importance of particular gaps whereas the former cannot. Interestingly, stratocladistics frequently derived the same trees for simu- lated clades as did maximum likelihood. The stratocladis- tic and CIS trees were both very similar to the maximum likelihood trees with lophospiroid data. The general con- gruence among these methods and the superiority of these methods over parsimony in simplified simulations both suggest that the ML estimate to be a better summary of lophospiroid evolution than is the parsimony tree. WAGNER: LIKELIHOOD, STRATIGRAPHY, AND LOPHOSPIROID PHYLOGENY 3 DISCUSSION EVOLUTIONARY IMPLICATIONS OF THE PHYLOGENETIC ESTIMATES Accurate phylogenies are premises of many tests of macroevolutionary hypotheses (Harvey and Pagel, 1991). When presented with rival estimates of phylogeny, it is important to explore whether the topologic differences affect general macroevolutionary interpretations (Donoghue and Ackerly, 1996; Wagner, 1997). SPECIATION PATTERNS Initial phylogenetic estimates suggested that bud- ding cladogenesis was the most common speciation pattern within lophospiroids (Wagner and Erwin, 1995). This study also found a positive association between species’ durations and numbers of apparent descendant species. Few patterns were found consistent with either bifurcating cladogenesis or anagenesis, both of explicitly predict that ancestral and derived species did not co-occur. (Note that a “species” here is a unique combination of character states.) Budding cladogenesis does not demonstrate any particular speciation process. However, it is consistent with processes such as peripheral isolation (Mayr, 1963), shifting-balances (Wright, 1931; but see Provine, 1986), and punctuated equilibrium (Gould, 1982) and inconsistent with processes such as selective divergence (Darwin, 1859) and vicariance (Brooks and McLennan, 1991). Both the parsimony and ML trees corroborate the original conclusions about speciation patterns. There are 18 taxa with plesiomorphic sister species on the parsimony tree. In 15 of those cases, the plesiomorphic taxon co- occurs with the derived sister taxon. There are 49 species with plesiomorphic sister species on the ML tree, 36 of which co-occur with their putative ancestor. On both trees, significantly more cases are consistent with budding clado- genesis than with anagenesis or bifurcating cladogenesis (parsimony: p= 1.32 x 10°; ML: p=1.42x 10°, based on a standard binomial test assuming a 50:50 distribution). However, one major difference exists. The parsimony tree posits no species with multiple descendants whereas the ML tree posits several species with two or more descen- dants. Moreover, there is a strong association between duration and the apparent number of descendants on the ML tree (Kendall’s t= 0.391; p = 2.44 x 107; Sokal and Rohlf, 1981: 429). Models predicting budding cladogenic patterns are generally consistent with the pattern shown by the ML tree. However, it is not clear that any speciation models predict the pattern shown by the parsimony tree, where species such as Lophospira perangulata produce a single daughter taxon early and then persist many millions of years with no additional daughter taxa. TRENDS Wagner (1996) documented active trends for several shell features among Ordovician-Silurian gastropods, including lophospiroids. One of these trends, the reduction of the sinus, is clearly observable among lophospiroids. Early species such as Lophospira perangulata have wide, deep sinuses that curve continuously back toward the sinus keel. Similar sinuses are retained among species classified as Donaldiella and in some Lophospira subclades (e. g. the L. burginensis clade). However, reduction of the sinus hap- pens in parallel among derived Trochonemella (i. e. the clade including T. montrealensis and derived Trochonema (i. e. the clade including 7. umbilicata). Note that the com- mon ancestor of those two clades (i. e. species similar to Trochonemella knoxvillensis and T. trochonemoides), retained a L. perangulata-like sinus. A similar pattern of sinus reduction is observed among species in the clade including L. milleri). It should be noted that the trend as described at this level can be inferred from the parsimony and ML trees. Additional parallelisms concerning the sinus exist on both estimated trees. In the case of derived Trochonema and the clade including Lophospira centralis, sinuses lose most of their curvature and retreat nearly straight to the sinus keel. A parallel reduction of the sinus keel is observed within both clades. Lophospiroids primitively possess a trilineate sinus keel, with a strong, sharp, medial lira bordered by two sharp and somewhat weaker peripheral lira. Broader analyses [Wagner, in press (a)] imply that the peripheral lirae of lophospiroids are homologous with the peripheral lira of other early Murchisoniinae (e. g. Hormotoma and Eotomaria) whereas the medial lira is a synapomorphy of early lophospiroids. The peripheral lirae are lost on Trochonema and twice lost within the L. cen- tralis clade (1. e. Eunema and Gyronema spp.). The periph- eral lirae also are lost in Proturritella and the Kiviasukkaan-Loxoplocus clade. Thus, the single peripher- al keel of species previously assigned to the Trochonematoidea (e. g. Knight et al., 1960) appears to be the medial lira of primitive lophospiroids. The reduction of the sinus keel precedes the reduction of the sinus among Trochonema spp., because species such as T: bellula and T. pandori possess only the medial lira but retain curved and fairly wide sinuses. Conversely, the reduction of the sinus precedes the reduction of the sinus keel among the L. cen- tralis clade, because species such as L. centralis and L. helicteres possess shallow, straight sinuses but retain a tri- lineate sinus keel. Similar mosaic patterns of sinus and sinus-keel reduction are observed within other gastropod clades [Wagner, in press (a)]. Thus, the characters clearly are independent homologues in that a change in one does not necessitate an immediate change in the others. 14 AMER. MALAC. BULL. 15(1) (1999) However, the overall pattern suggests some sort of biologi- cal (e. g. functional) association. Bandel and Geldmacher (1996) have stated that species such as Trochonema and Eunema could not be derived from ancestors with sinuses and/or slits and thus could not be lophospiroids. However, early Trochonema spp. possess a lophospiroid sinus that is reduced or entirely lost in derived species. Eunema spp. retain a derived sinus observed in species previously assigned to Lophospira (e. g. “L.” centralis). Sinus reduction obviously was com- mon among Ordovician and Silurian lophospiroids. A slit is an uncommon feature among lophospiroids, but likely evolved at least three times within the clade. Derived Trochonemella spp. such as T: notablis are one such group of taxa, but primitive species such as T: knoxvillensis and T. trochonemoides clearly did not possess slits. Thus, whether 8 Ss ) i i = ee Po s 8 ES g x 3 E sa 3 oo & 5 4 2 ts 3 2 48S So g i S 8 is a 7:8 «8 a es ‘A f 5 5 ‘ = 87 ae 3 f ee 33 3» = |. w 4 a 3 zu! g weES § & 2/8: Rj SEsa& ~ AS + 3 i bo Se w 88 wept . s pal § SeirPig 5 “S73: 3 : So 4 8 gaug 2 ES AH ee Sets 8 ‘ 3 8 g EESe & g& eeye : § us) & Sess 8& SB ssye: a 5 a eEReae § E64 g gs 1 8 = a 3 ses sig = sss ; S a 2 S25 8 S35 8. SSIS: & a ssSs & sey SE sole: 8g 3 gs eee © S63 SS Gh:0 | s = gis e 8 82 Til: Su EB s £ sible aa gis? S Fo. 3 3. ‘Bee Ssss8 < os 8 g . E:3 » &ss& ' Zsss td b< ges »esee8 ' SEST SZESS3 ‘ 2. 'SSseos: ssce 4 Beeay ' SRP SSsse: RSgg ok aR : 3s: i 2ees sr PPE gSE3e8 si] ¢:ekee §! eS SRReSs§ § Srspipepy si gsiegs 23 ESsbSs : Se sp Psey ss SSSES oz = 8 3 = Ras 3 ByPSssE : B:spiletts: BSSLS Bsls| $/s2 Er SP yOiy sg: goo = eis © = Cel sere Boots 33 3| spss ge S Soda 3s. 3. sa & a 84 3. 3 ais i (=I mysPc< ope Pond ae ary Se 3 a £2 ray . ee 1 £ 5 iE SS: Ae ee : = — a“ ‘ Ss ~ rs = & Sy: Aas rE si! : Oz S|: ; - — ‘ . a ‘ : Arenig ++ <= Lophospira quadrisulcata ronema Eunema strigillata Lophospira concinula = === Lophospira spi Trochonema or Eunema could have slit-bearing ancestors is irrelevant, as neither ML nor parsimony trees posit slit- bearing ancestors for either taxon. Parsimony and ML trees suggest very different pat- terns of ornament evolution. The increasing number of ornate gastropod taxa over the Phanerozoic is a well-docu- mented trend (e. g. Vermeij, 1977). Ornament is uncom- mon among lophospiroids until the Late Ordovician. Conversely, most Silurian lophospiroids had at least some ornament. The parsimony tree suggests that ornament arose four times and was subsequently lost four times (Fig. 11). (The reconstructions on the parsimony tree are unam- biguous and thus not affected by optimization assump- tions.) Parsimony also suggests that ornate taxa were sub- stantially more diverse in the Middle Ordovician than sam- pling alone would imply. Hypothesized ornate range exten- m as a 53 = ge 5 £ey gs S G2 g $38 Sy SES S$ sr 3 geo 2 S83 & 3 git 2 S& = 8 wy ss =] <= BS & =: 5 2 23 o 8b 3 & 4 s so a 8 = wo £5 H $s 2 Siiies £ SS H 8 Bees § 33 os 8 47 -38s is & e] rece tls [E Ei ey) el [Pf] og ~j ; ae : g RAE. VZ08 a bays: |S S]8: S858 :S88 Fee |e gps: Ss s 3 "B7 Sis. | Sy 3: 58g 3 g ‘Sele. Ss Sys: sees g we SIs SiS spe: 3g = }2 ' =: 2 BE | = q = 3 ‘So fe Sie se Sak o: = 3 . . eX 3: . 2 Ss Eos ods yD SP Seu seasr: S. § 3. g's: |: : $e sics ss &. 32 B25 ¢ tale > S§i ss : Sz? 23 $8 eae :Sy] SSi ye: 2,50. 35 i . i] . « = a4 ° geek: : §] & : os P2es FESS g HS:78 : 8 a 5 B Jesh §shs ° g§is : 34 3 : § TSSees8s o.2 = aes = = . gS | fas SSRASLYOS ss: : : = ee ig QS i SPR SSSS sss Ras , S| S< -3 38 1Os soe ses S s * l q cay 3 . 3 ss ese Sf ibe £2 :/2)8pses Ss ' : H Bs i 78! Ss] $888 ms : j £8 tfsl Pe] sss : eT i/Elyy| S888 : ee erie (eS) RS a. ‘a : ssss S:otiy: aaaa Ro : Sucre b] ) Sig ¢ ‘ See : oe ne ea | 38 ge | Bie S's o's ' ie aut F ; ' Fig. 11. Patterns of ornament evolution as posited by parsimony. Black branches denote primitively inornate species, gray branches denote secondarily inor- nate species, and white branches denote ornate species. Note that the tree posits a much earlier divergence of ornate species than observed in the fossil record, resulting in hypotheses of many ornate ghost taxa (i. e. white dashed lines). WAGNER: LIKELIHOOD, STRATIGRAPHY, AND LOPHOSPIROID PHYLOGENY sions (Fig. 11; dashed white lines) represent approximately one-quarter of the inferred diversity during the Llanvirn. However, only one ornate species is known before the Middle Caradoc. Parsimony also implies that several inor- nate Ordovician species are secondarily inornate. Suspiciously, ornate relatives always appear after the “sec- ondarily” inornate taxa. The ML tree suggests that ornament arose nine times (Fig. 12), which implies a frequency of change of 0.098 per branch. This is over twice the typical frequency (0.040 per branch). However, ornament apparently is local- ly conservative, as there are no apparent reversals in sub- clades of up to seven ornate species. This pattern suggests a driven trend (i. e. the biased production of a particular morphologic type; McShea, 1994). Despite the low sample = io} Ke} - AW S| Ea & 3 3 a a 8 = ae 8 : er Ss Fs Bs : sg 7 st iS} a : 3 3 33 3 Q $8 3 8 MES s £3. 32 Seo Sra 8 on Se = BOG * SS 3 STss & $356 23 8 pa ESSS 34 5 Ee 3 Se 3 5 ssQs 22 § Ssis 885 s 2 S&_ps SER gua 3E2 S ie a NS ee Eeee ss = g 3 RA SEEEzE a SEGRE FS§ R oe) geo. S8t sce bes 3 S832 S85 : ' 3s a = « S , $$! PS8Se8 988 Sei: S23 22 4 = : BoST 1 SESS EN SSB 8 yeeSs ah ges bj ec = be) SS FE gs SS, ([S885 2 subg $33 7 2 BE S8 a eee te) 3 8 3 SEs < & ss s[S8S8 .§ 2 { #8 55 aa 2s] S225 ate as S23 gg SPs os ots . eS als af ?§& PUS]SSSS Sess : 3] ESS! f Bs Eree = : . g #2eR8 ' $$ SSSS page | Se] Ses As : SS Sacegsess : gs ss] a elles mets SSEES : Ys] sys 3 tg Sy age . £5 ER fs gs: £3 gn sf Ss{Ssrt Suse ge & oO SEs Seis 8 : $ 25's goes = outs eS = 8 8 = myse.s gs & = a) Ra' 8 2 3) eei8 RR, = VIis Sy. Es = 2a: ss Ss ' OD peecia : o) q an 2d i=] a = Trochonema aff. T. umbilicata Trochonema salteri § — 3 2 g 8 g 3 2 3 8 g3 =F 835 = 3 33 & = 3 e Vs So 62-3 3 aS} = s &§ age BAG 2 s§ i 3 aes 2 8 3 s— § 8 gs ~ 2 3 3 ff = 3 ay a3 & gi 3 3 ey Sa : 3 SS 3 ES gS = 2 Ba 4 $98 a o 8 Q0 B85 iS) xu = we 8 © ge As R i$ 8 1 eR soa 8 33 S Ex § “Sy Gig S g 2 sok 208 = 3 5 ame s 4.8 238 2s 5 2 gu Bs Ss ss ii Jz oe SSeS i ss 5. 0°; (2) ~10°; (3) ~20°; (4) ~30°; (5) ~40°. Ordered; weight 0.25. 3. Sinus width [as described by angle passing from center of aperture through top and bottom of sinus; continuous character, with states broken into just above/below sinus apex (1), halfway between top/bottom of ramp and sinus apex (2), coinciding with top/bottom of ramp (3), and beyond top/bottom of ramp (4); coding the position of the onset of the sinus relative to the top/bottom of the ramps rather than the absolute width accommodates the fact that changes in ramp lengths and orientations also will change sinus width as an artifact; this coding scheme focuses only on differences clearly attributabie to the sinus itself]: (1) just above SK; (2) between SK and RR; (3) at RR. Ordered; weight 0.5. 4. Sinus shape [describing general trend of sinus; nearly straight up to sinus apex (i. e. V-shaped) or curving back continuously; this describes the shape of the curve whereas sinus depth (character 2) describes the general angle of retreat]: (2) straight; (3) continuous. Weight 0.5. 5. Crenulated aperture [in which apertural margin generates “zig-zag” growth lines]: (1) absent; (2) present. Weight 1. 6. Sigma-shaped aperture: (1) absent; (2) intersection between base and alveozone curved (yielding reverse sigma) instead of planar or angu- lar. Weight 1. 7. Prominence of growth lines (GL) [evaluated only by comparison with both different specimens known from same preservational conditions and similar specimens from different preservational conditions]: (2) weak; (3) fine sharp; (4) strong. Ordered; weight 0.5. 8. Imbricated GLs [alternating growth lines appreciably stronger than oth- ers; considered present only if imbrications obviously are patterned, as random imbrications can occur due to shell repair or on gerontic whorls of some species; growth line prominence is coded based on weaker growth lines]: (1) absent; (2) weak; (3) moderate; (4) strong. Unordered; weight 1. 9. Sinus keel (SK) width [originally thought to be morphologic artifact of a slit or a “notch,” however, appearing on slitless specimens and the seemingly mythical “notch” has never been documented; continuous character, where 360( equals circumference of aperture]: (1) ~5°; (2) ~10°; (3) ~15°; (4) ~20°; (5) ~25°; (6) ~30°. Ordered; weight 0.2. 10. Peripheral lirae (PL) [two lirae on bilineate peripheral bands (or outer two on trilineate bands of lophospiroids); in some cases, might be produced by placing a slit within a medial lira (see below); however, not occuring on all species with slits and many species with periph- eral lira clearly do not have slits]: (1) absent; (2) present. Weight 1. 11. PL type: (1) threads with round profile (e. g. Ruedemannia); (2) threads with sharp profile (e. g. Ectomaria or Lophospira). Weight 1. 12. PL strength: (2) weak (clearly visible but casting little relief); (3) moderate (casting noticeable relief); (4) strong (clearly projecting from shell). Ordered; weight 0.5. 13. Medial lira (ML) [= “notch keel” or “sinus keel’; present with periph- eral lirae (character 10) only if there is ontogenetic change (character 25) or if peripheral band is trilineate; carina at the apex of the sinus]: (1) absent; (2) present. Weight 1. 14. ML type [as seen in profile]: (1) round, of equal “height” and width (e. g. Ruedemannia); (2) sharp, of equal “height” and width (e. g. Lophospira). Weight 1. 15. ML strength: (1) extremely weak (barely visible, casting almost no relief); (2) weak (clearly visible but casting little relief); (3) moder- ate (casting noticeable relief); (4) strong (clearly projecting from shell). Ordered; weight 0.33. 16. Imbricated ML: (1) consistent; (2) flaring periodically (e. g. Lophospira serrulata). 17. SK prominence [not equal to strength, as prominence refers to whole structure projecting from rest of shell; sometimes coincides with channel underneath peripheral band; however, species with channels and non-prominent peripheral bands and species with prominent peripheral bands and no obvious channel are both known]: (1) peripheral band not altering profile of whorl; (2) whole peripheral band projecting slightly, creating slight ridges between shell and peripheral band; (3) whole peripheral band projecting strongly, creat- ing strong ridges between shell and peripheral band. Ordered; weight 0.5. 18. Slit: (1) absent; (2) present. Weight 1. 19. Slit depth: (1) shallow (extending <10° behind aperture, measured looking down coiling axis); (2) deep (extending ~20° behind aper- ture). Weight 1. 20. Maintenance of slit: (1) periodically erased or reduced, with shell material deposited within slit less frequently than on rest of shell, resulting in slit depth (and sometimes presence/absence) varying over time (e. g. Trochonemella; recognized in part by greater dis- tance between lunulae than observed between growth lines); (2) con- tinuously maintained (e. g. “Schizolopha” moorei). Weight 1. 21. Lunulae shape [lunulae here are growth lines within sinus keel]: (1) concentric (shallow U-shape); (4) V-shaped (e. g. Lophospira cen- tralis and relatives). Weight 1. 22. Lunulae strength: (1) weaker than GLs; (2) same as GLs; (3) stronger than GLs (i. e. imbricated; e. g. Trochonemella spp.). Unordered; weight 1. 23. Imbricated lunulae type: (1) deep and obtuse; (2) serrated. Weight 1. 24. Ontogenetic change in imbricated lunulae: (1) juvenile whorls only; (2) throughout. Weight 1. 25. Reduction to monolineate SK over ontogeny: (1) absent; (2) monolin- eate on final whorls. Weight 1. 26. Peripheral band attitude: (1) projecting straight from shell, as seen on most species; (2) curving adapically, e. g. Arjamannia and relatives. Weight 1. 27. Midwhorl (MW) channel [groove underneath medial lira; often but not always associated with prominent band, and so coded separate- ly]: (1) absent; (2) present. Weight 1. 28. MW channel strength: (1) weak; (2) strong. Weight 1. 29. SK position [described relative to aperture centroid, with 90( indicat- ing that a plane is passing through centroid and that SK is perpendic- ular to coiling axis; O( indicating that plane is parallel to coiling axis]: (2) ~100°; (3) ~90°; (4) ~80°; (5) ~70°; (6) ~60°; (7) ~50°. Ordered; weight 0.2. 30. Ramp shape symmetry [on primitive bilaterally symmetrical gas- tropods, nght and left ramp shapes are symmetrical; derived species are asymmetrical]: (1) right rounder; (2) symmetrical; (3) left rounder. Unordered; weight 1. 24 31. 32. 33: 34. 35. 36. 58. 59. AMER. MALAC. BULL. 15(1) (1999) Right ramp (RR) shape: (1) globular; (2) convex; (3) flat; (4) slightly concave; (5) concave. Ordered; weight 0.125 (reflecting both con- tinuous nature and de-weighting for asymmetry). RR and LR lengths [on primitive bilaterally symmetrical gastropods, right and left ramp lengths are equal ; lophospirds with longer right and longer left ramps are known]: (1) longer RR; (2) equal lengths; (3) longer LR. Unordered; weight 1. RR length [describing angle from “top” of ramp to sinus apex, based on triangle passing through aperture centroid]: (3) ~50°; (4) ~60°; (5) ~70°; (6) ~80°; (7) ~90°. Ordered; weight 0.25. LR length [see character 33]: (3) ~50°; (4) ~60°; (5) (7) ~90°. Ordered; weight 0.25. RR:LR projection [ramp projection describing angle of “nse” of ramp from sinus keel, based on plane passing through SK and aperture centroid]: (1) RR projection higher; (2) LR and RR projections equal. Weight I. RR projection [see character 35]: Ordered; weight 0.25. 70°; (6) ~80°; (5) ~60°; (6) ~70°; (7) ~80°(. . LR projection [see character 35]: (3) ~40°; (4) ~50°; (5) ~60°; (6) ~70°; (7) ~80°. Ordered; weight 0.125. . Sutural carina (SC): (1) absent; (2) present. Weight 1. . SC strength: (1) weak (creating only a weak profile); (2) moderate (partly filling up the suture). . RR carina (RRC) [strong carina usually located at top of sinus}: (1) absent; (2) present. Weight 1. . RRC strength: (1) weak (creating no profile); (2) moderate (roughly equal to weak-to-moderate peripheral band); (3) strong (roughly equal to strong peripheral band). Ordered; weight 0.5. . RRC type: (2) thin local thickening; (3) round profile; (4) sharp pro- file. Unordered; weight 1. . Ontogenetic change in RRC strength: (1) absent; (2) becoming weak- er on adult whorls. Weight 1. . Channel beneath RC: (1) absent; (2) present. Weight 1. RRC location: (1) ~75° toward suture from SK; (2) ~45° toward suture from SK. Weight 1. . RRC attitude: (1) carina projecting perpendicularly to RR; (2) carina curving abapically. Weight 1. . RRC: (1) plain thread; (2) serrated. Weight 1. . Shape of shell at top of right ramp [usually at suture]: (1) oblique; (2) acute, with channel underneath suture. Weight 1. . LR shape [see character 31]: (1) globular; (2) very convex; (3) slight- ly convex; (4) flat. Ordered; weight 0.167. . Swelling at intersection of left ramp and base [might be primitively homologous with left ramp carina (character 51), and present on ear- liest gastropods]: (1) absent; (2) present. Weight 1. . LR carina (LRC): (1) absent; (2) present. Weight 1. . LRC type: (1) thick contusion, but creating more distinct profile than simple swelling; (2) sharp profile. Weight 1. . LRC strength [see character 41]: (1) weak; (2) moderate (~ sinus); (3) strong. Ordered; weight 1. . LRC [see character 47; species with serrated right carinae but plain left carina exist, indicating that the two evolve independently]: (1) plain thread; (2) serrated. Weight 1. Second LRC: (1) one carina; (2) two carinae (e. g. Lophospira quadrisulcata). Weight 1. Channel beneath LRC [see character 44]: (1) absent; (2) present. Weight 1. . Columella thickness: (1) no thicker than rest of shell; (2) slightly thicker than rest of shell; (3) much thicker (partly filling umbilicus); (4) extremely thick (filling umbilicus). Ordered; weight 0.33. Umbilical carina (UC) [carina at base of shell, circling umbilicus]: (1) absent; (2) present. Weight 1. UC type: (1) thick, dull protrusion; (2) sharp extension accomodating 60. 63. 64. 65. 66. 67. 68. 69. 70. WW. 72. 73. 74. 1D: 76. 77. 78. 79. 80. 81. 82. 83. a channel; (3) lirum with sharp profile. Unordered; weight 1. UC strength [see characters 41 and 53]: (1) very weak; (2) weak; (3) strong. Ordered; weight 0.5. . Ontogenetic change in UC strength: (1) constant; (2) becoming weak- er on adult whorls. Weight 1. . UC location [with larger angle indicating UC closer to coiling axis]: (1) ~120° below SK; (2) ~90° below SK. Weight 1. Angle at base of columella [narrower angle indicates sharper, more siphonate base of shell]: (3) ~60°; (4) ~75°; (5) ~90°; (6) ~105°; (7) ~120°. Ordered; weight 0.25. Shape of columella on inner margin [can differ from shape on outer margin, so the two coded separately]: (1) arching like half-circle; (2) arching in obtuse curve; (3) trending toward straight; (4) curving slightly into the aperture. Ordered; weight 0.5. Outer margin shape: (1) more obtuse than inner margin; (2) same as inner margin; (3) more acute than inner margin. Weight 1. Ontogenetic change in margin shape: (1) none; (2) becoming rounder over ontogeny. Weight 1. Columella attitude [describing main trend of columella relative to coiling axis]: (1) 0° (i. e. perpendicular to coiling axis); (2) 15°; (3) 30°; (4) 45°; (5) 60°. Ordered; weight 0.25. Columella lira [carina in middle of columella, visible in umbilicus]: (1) absent; (2) present. Weight 1. Parietal inductura thickness (silicification can occur differently among different shell layers, which can affect characters such as this; there- fore, “relative” states were coded based on comparisons among taxa found from the same beds, with comparisons among conspecifics from different beds used to establish the final character code]: (1) absent; (2) thinner than rest of shell; (3) thickness same as shell. Unordered; weight 1. Columella reflected around coiling axis: (1) absent; (2) present. Weight 1. Whole aperture inclined (tangential): (1) absent; (2) present. Weight 1. Degree of whole aperture inclination (InAn): (1) ~10°; (2) ~20°; (3) ~30°; (4) ~40°; (5) ~50°; (6) ~60°. Ordered; weight 0.2. Left side of aperture only inclined [or inclined at different angle than night side]: (1) absent; (2) present. Weight 1. Left InAn: (1) ~10°; (2) ~20°; (3) ~30°. Ordered; weight 0.5. Right side of aperture only inclined [or inclined at different angle than left side]: (1) absent; (2) present. Weight 1. Right InAn: (1) ~10°; (2) ~20°; (3) ~30°. Ordered; weight 0.5. Anterior projection of aperture [best observed from base, because growth lines will project forward instead of radially; it is important that they do this at the onset, as tangential aperture will cause growth lines on the periphery to slope forward as well]: (1) absent; (2) pre- sent. Weight 1. Degree of anterior projection [relative to plane passing radially through coiling axis]: (1) 10°; (2) 20°; (3) 30°; (4) 40°; (5) 50°. Ordered; weight 0.25. Aperture expansion [expansion of shell in radians as a “tube”]: (2) 0.05-0.10; (3) 0.10-0.15; (4) 0.15-0.20. Ordered; weight 0.5. Curvature about coiling axis [in radians]: (3) 0.65 < K < 0.75; (4) 0.75 < K < 0.85; (5) 0.85 < K < 0.95. Ordered; weight 0.5. Translation (T) [vector of downwards growth (in radians)]: (5) 0.16 < T < 0.47 (low dextral); (6) 0.47 < T < 0.79 (moderate dextral); (7) 0.79 < T < 1.10 (high dextral); (8) 1.10 < T < 1.41 (very high dex- tral). Ordered; weight 0.33. Continuous ontogenetic change in T: (2) isometric; (3) continuously increasing, resulting in ever-increasing apical angle. Weight 1. Early ontogenetic change in T: (1) early decrease of T (e. g. Trochonema, Trochonemella), (2) isometric; (3) early increase in T. Unordered; weight 1. 84. 85. 86. 87. 88. 89. WAGNER: LIKELIHOOD, STRATIGRAPHY, AND LOPHOSPIROID PHYLOGENY Late ontogenetic change in T [often resulting in open-coiled “cork- screw”’-like shells, e. g. Lophospira helicteres, Eunema strigillata]: (2) isometric; (3) late increase in T. Weight 1. Punctuated late ontogenetic change in T: (1) occuring over 1-2 whorls (e. g. Trochonema, Trochonemella),; (2) punctuated, occuring over less than one revolution. Weight |. Magnitude of late ontogenetic change in T: (1) slight, changing suture point; (2) major, resulting in open coiling. Weight 1. Ornament on left side of aperture: (1) absent; (2) present on left ramp and base; (3) present on left ramp only. Unordered; weight 0.5. LR ornament density [based on average angular distance between threads]: (1) 1 per 20°; (2) | per 10°; (3) 1 per 5°; (4) I per 1°. Ordered; weight 0.5. LR ornament type: (1) thin local thickenings, little stronger than growth lines and with no profile (e. g. Ruedemannia), (2) lirae with weak profile (e. g. Arjamannia); (3) thick lirae with strong profile (e. 90. APPENDIX 2 CHARACTER MATRIX Character data for lophospiroid species. See Appendix | for characters and character descriptions. [?, characters that could not be observed; -, characters that do not pertain to a species (e. g. ornament type for an inornate species); A-D denote polymorphic species [A, states 1 + 2; B, states 1 + 3; C, states 2 + 3; D, states 6 + 7]; BMNH, The Natural History Museum, London]. Oo No Oo Zz g. Longstaffia). Weight 1. RR ornament [because species with left ramp ornament sometimes lack right ramp ornament, these characters coded separately; com- pletely inornate to completely ornate represents one step]: (1) absent; (2) present. Weight 0.5. . RR ornament density [see character 88]: (1) 1 per 20°; (2) 1 per 10°; (3) 1 per 5°; (4) 1 per 1°. Ordered; weight 0.5. RR ornament strength: (1) thin threads; (2) weak lirae; (3) strong lirae. Ordered; weight 0.5. . Ornament as changes in aperture shape: (1) simple local thickenings of shell; (2) representing local flarings of aperture. Weight 1. Ontogenetic change in RR ornament: (1) constant; (2) weaker on adult whorls. Weight 1. . Size: (1) <10 mm? (micro-mollusk); (2) small (>10 to <10° mm’): (3) moderate (>10° to <10° mm’); (4) large (>10? to <104 mm’). Ordered; weight 0.33. — — — WOIDRNAWN+QOQ 10 Trochonemella trochonemoides (Ulrich in Ulrich and Scofield, 1897) 11 12 Trochonema bellula Ulrich and Scofield, 1897 13 14 Trochonemella montrealensis Okulitch, 1935 15 16 Trochonemella n. sp. 17 Lophospira serrulata (Salter, 1859) 18 Eunema strigillata Salter, 1859 19 Trochonema umbilicata (Hall, 1847) 20 T. canadensis Wilson, 1951 21 22 Lophospira ventricosa (Salter, 1859) 23 Trochonemella notablis (Ulrich in Ulrich and Scofield, 1897) 24 Gyronema pulchella Ulrich and Scofield, 1897 25 G. semicarinata (Ulrich and Scofield, 1897) 26 G. liljevalli Rohr, 1980 27 Trochonema madisonense Ulrich and Scofield, 1897 28 Lophospira burginensis Ulrich and Scofield, 1897 29 L. oweni Ulrich and Scofield, 1897 30 L. concinula Ulrich and Scofield, 1897 31 32 Donaldiella decursa (Ulrich and Scofield, 1897) 33 34 D. producta (Ulrich and Scofield, 1897) 35 D. Curdsville sp. 36 Lophospira sumnerensis (Safford, 1869) 37 Trochonema salteri Ulrich and Scofield, 1897 38 Ruedemannia humilis (Ulrich and Scofield, 1897) 39 R. lirata (Ulrich and Scofield, 1897) 40 Lophospira tenuistriata Ulrich and Scofield, 1897 Hormotoma simulatrix (Billings, 1865) Ectomaria adelina (Billings, 1865) Pagodospira cicelia (Billings, 1865) Lophospira perangulata (Hall, 1847) L. sorrorcula (Billings, 1865) L. rectistriata (Billings, 1865) Pagodospira derwiduii Grabau, 1922 Lophospira milleri (Hall in Miller, 1889) Trochonemella knoxvillensis Ulrich in Ulrich and Scofield, 1897 Proturritella bicarinata (Koken, 1889) Pagodospira dorothea Grabau, 1922 Lophospira centralis Ulrich and Scofield, 1897 T. eccentrica Ulrich and Scofield, 1897 Lophospira helicteres (Salter, 1859) T. wilsoni Steele and Sinclair, 1971 L. spironema Ulrich and Scofield, 1897 D. conoidea (Ulrich and Scofield, 1897) RBNYNNHNNYNHNNNNNNNNNNNNNNYNYNNNNNNNNNNNNNNNNNN LD ARRWUNAUAAHNEFHYUUWEWWWARUNWWWHWUNAWRWAWUUNAUWNDRUWBUUUAUY WNHNN WW WWWN KF WWNN KK BK WNK NNW KE NK NNN NWN WH W ND W WH WH WH W WWWN WW WW WWNYWWWNNN WWNNN DN WWN WHY WD WW WWW WW WW Ww ww /lRp oO mee ee ee NNR MK NM RK NN DN RK RK NHK NN DN & DNR Re i i ~ANME NMP NYVNNNNNNYNN YE — wre Wr WW WWW WW Ww Ww '‘NNNN! Whew NWN YM WNNN NH YCWNYYNYWWNHKY AN WW WW BPW PW WW WN WH BW NN NH KN KW PLO bO ee Ne ce ce ce ce oe ee oe ce oe rr BNP REN NNNNHNNNDN— K— KK NN RK RK KB KE NNNNK KEP NNNNNNNNNNH A A RWNNENNYNNHNHNNNNE EEE NNE SBE ENNNNE SE NENENNNNNHNNNDLHD ti ‘oo NNNNMNNNNNKYMNNNNNNN EK KNNMYNNNN KEK NEF NNNNNNNNNNNNN KH —& NRK KBPNNNNNNMNNNNNNN! NNNNNN! NF NNNNNNNNNNNNNW : NVNNWENNENNWENWWNN! NWWWWW! WI WWNYWNWNNNNNNN YY RNNN!E NK eee NNNNYM? WNN! WNNWWWHYWWWw (continued) 26 APPENDIX 2. (continued) 41 L. tropidophora Meek, 1873 42 Donaldiella filosa (Donald, 1902) 43 D. bowdeni (Safford, 1869) 44 Lophospira bellicarinata Donald, 1906 45 L. quadrisulcata Ulich and Scofield, 1897 46 “Schizolopha” moorei Ulrich and Scofield, 1897 47 Lophospira aff. serrulata [in Rohr, 1988] 48 Trochonemella BMNH 36364 49 T. churkini Rohr, 1988 50 Trochonemella reusingi Rohr and Blodgett, 1985 51 Trochonema aff. umbilicata [in Rohr, 1988] 52 Arjamannia thraivensis (Longstaff, 1924) 53 Globonema bicarinata (Wenz, 1938) 54 Lophospira ?borkholmensis (Koken, 1925) 55 Trochonema pandori Koken, 1925 56 T. aff. pandori Koken, 1925 57 Lophospira sedgewicki Donald, 1905 58 L. ?hyaecinthinsis Foerste, 1924 59 Longstaffia centervillensis (Foerste, 1923) 60 Arjamannia cancellata (M’Coy in Sedgwick and M’Coy, 1852) 61 A. woodlandi (Longstaff, 1924) 62 Donaldiella trilineata (Foerste, 1923) 63 Lophospira gotlandica Ulrich in Ulrich and Scofield, 1897 64 Longstaffia laquetta (Lindstrém, 1884) 65 Arjamannia inexpectans (Hall and Whitfield, 1872) 66 Lophospira holmi (Lindstrém, 1884) 67 Kiviasukkaan nelsonae Peel, 1975b 68 Lophospira munda (Lindstrém, 1884) 69 L. imbricata (Lindstr6m, 1884) 70 Trochonema turrita (Lindstrém, 1884) 71 Eunema kayesi Rohr, 1981 72 Ruedemannia laevissima (Lindstrém, 1884) 73 Arjamannia aulangonensis (Peel, 1975a) 74 Ruedemannia robusta (Lindstrém, 1884) 75 Longstaffia tubulosa (Lindstrém, 1884) 76 Longstaffia cyclonema (Salter, 1873) 77 Trochonema fatua (Whiteaves, 1895) 78 Loxoplocus soluta (Whiteaves, 1884) 79 Lophospira cochleata (Lindstrém, 1884) 80 Eunema muricata (Lindstrém, 1884) 81 Ruedemannia subrobusta (Perer, 1907) 82 Ptychozone aberrans Peer, 1907 My a IN > ico Lt ee cl cl el ee eo meee eo ee H. simulatrix E. adelina P. cicelia L. perangulata L. sorrorcula L. rectistriata P. derwiduii L. milleri T. knoxvillensis P. bicarinata P. dorothea 10 T. trochonemoides 11 L. centralis 12 T. bellula 13 T. eccentrica 14 T. montrealensis 15 L. helicteres 16 T. n. sp. 17 L. serrulata 18 E. strigillata 19 T. umbilicata 20 T. canadensis 21 T. wilsoni 22 L. ventricosa 23 T. notablis 24 G. pulchella @ @ 1 2 3 4 5 6 7 8 9 NO ee er) NVNNNNNNNNNNNNNNWNHNNWNNNNNN ON nO ' ' Bee eee hee Pee eae Pee eee eee eee NVWBNNNNNYVYNNNNNNNNNNNNNNNXYN ' ' ' | el ell cel ee ee ee ce ee ce ce cee oe oe oe ee ee NNNNNNNNNNNNNNNNNNNNNNN ND — NNKENNNNNNNNKFP KN NNNNNNNNNNNNNNNNNNNN NN NNN DN DN LV RFP NENNNNEK NE NNNKE NK KE NK ee eee ‘rphPHAHAUNUADPAHRPAARPAHDK ANHAAAAUNWAUUNN as hh HAHAH WrNAIWIY YY RP WYW DY YW DH WW DH WW WD WH WH WwW WD Ww NRK WNNNNWKRKNNNNNN W WW WW WW WWD ee EWE BB eB EB NNNEP SE NNNKE EWE ewww NM! KK NOK KH NOK ND! WNNWWWWUNWWWWWWWARNHBRUHR RUNS rWWWWW WW WW WWW WWW WN WD WW WW WN WW WH WwW WWt WWNWWWWWws BWBNNNNNWNNWNNN WN HD WHY WN NNN NH LD LY mM RON SE SB ENE eB EB NNNN HK KEE NK Kee Se ee ee Be ee eee eS Se Se Se Se Se eee WNIWNIYYWW YW DIY YW UW WW WWW WWW W Pa Og nr gg a a eT YL nT ee AMER. MALAC. BULL. PAIWAINANA WIPE ANP AWHRAAWWWWWWA 15(1) (1999) j'NHeE NIE NNNNNNNNN NV AN! Wr WWWWH WW WWD tNUNNNNNNNNNNNN ! ARK NNRK NHN AW PKK NPNINNNNNNWNNNNN | RK PK NK NNNNNN NHN WN LO PWWWNNNWWNHNN WW! NNKHK NN KNYNNNNK KN NNNNNNNNNNNN KK NK NK NNNNNN NNN LV YBHYNPRWWRANWNH PRA WWWWARWNHWWRWWWWNWWW HSK HhA HR HW WH NLD OS ee 0 ee ee OO Ne ee NNMOtENN! NNNNN NN! WWt NYUONNNNE eee eee NRENYNNYNNNNNNNNNNNNBP SHEP eee PRDWANAWWAWAUNUWAWWWWWWWWHNwa PNEFPNNNNNNNNNNKR KR Be eee eee ee MAMAAYYWWAIWAIIYIWIWNWHUNHDUNHDUNUHUUAD foe IE NYVNNK Leeann NNNNNNNNNNNNNNKHNNNNNNNNNNNNNNNNNNK KEK NNNNNNN LY NYUENNYWVBWUnnnnNnnNHennn 'NNNNNNNN =H NNNNK KR KB NK NNN NNNN KE NNNNNNNNNK NNN! WWE WWWWWWWWWWWWe WWW it tWWeWRNND — =“H—KSWWNWRARKP RYE WWWWWPRHhA HK HAPHPRARWWNHPAW! oe ee ee ee oo Ke TF NNN DK KH NK ND HR BR BR See eS BNVNNNEF ENE EP NNR RE ee eB eB VV meee eee i YD ee ee ee ee ee OY Cc ee ee eee ee eo any WWW WWW WW WW WWW WW WW WD WW WY WY WY Ww Ww te RHE N BE ee ee ee eee EE NE NENNNNNNND NNRENNNNNNNNNNNN KH NNNKYNNNNN — 12222222222333333333344444444445 555 5 5 NN! NNNNNNNNNNNY?D ! NPN' Fe DRDe! NwWIE NYVNNNNWNWNNNK I NYNN! Bee ees (continued) 21 WAGNER: LIKELIHOOD, STRATIGRAPHY, AND LOPHOSPIROID PHYLOGENY 25 G. semicarinata APPENDIX 2. (continued) 26 G. liljevalli 2 1 1 1 1 2.2: 7 2 3.27 7 7 1 27 T. madisonense 2 2:2 1 3:3 13 332331 2 28 L. burginensis 29 L. oweni 30 L. concinula 31 L. spironema 32 D. decursa 33 D. conoidea 34 D. producta 35 D. Curdsville sp. 36 L. sumnerensis 37 T. salteri 38 R. humilis 39 R. lirata Zod. 2 40 L. tennuistriata 41 L. tropidophora 42 D. filosa 43 D. bowdeni 1 4353371642222 3 1 pe) am | 44 L. bellicarinata 45 L. quadrisulcata 46 L. moorei yao 2 ee ae 47 L. aff. L. serrulata 48 T. BMNH 36364 49 T. churkini 50 T. reusingi 1 1 i 12 2 51 T. aff. umbilicata 52 A. thraivensis 53 G. bicarinata 1 1 1 2 4343371642222 31 224 23 5 426622 1 1 1 2 2 54 L. ?borkholmensis 55 T. pandori 1 1 1 1 1 1 2,222, 3: «3. 1 2 14353574164 22 1 15 Dei (Dad mee he Qe 2 2 fo al 1 1 2 2 2: 32 253. 3 1 1 1 7S 56 T. aff. pandori 57 L. sedgewicki 64 22 - 4353571 3 933346164 143333741 58 L. 2hyaecinthinsis 59 L. centervillensis 60 A. cancellata 61 A. woodlandi 62 D. trilineata - 2 2 3 1 1 1 64222 2 3 - 2 2 1 3.3.3 2. 3 3 63 L. gotlandica 64 L. laquetta Zz 2, 2 1 2 3 2-2 12 4 22 3 1 164 22 2 220-1 4 2.3 3° 3-77 65 A. inexpectans 66 L. holmi 67 K. nelsonae 1 222 1 1 68 L. munda 69 L. imbricata 70 T. turrita 71 E. kayesi 5 3 2.2. 7 7 2 1 7 2,-2.2, 2 3 14 23 1 2. 2 1 1 31:22. <= 1 2 2-2 DMD 2 1 1 1 1 2 2 eS 2 273 1 1 2 PZ Ss 2h G27 7 22724242. <3, 72 R. laevissima 22 1-42.33 3:7 1 64 73 A. aulangonensis 74 R. robusta ? 2 2 75 L. tubulosa 32346164 1 5 4266 2 2 3 1 22 4 23 76 L. cyclonema 77 T. fatua 78 L. soluta 1 1 1 79 L. cochleata 80 E. muricata 1 P2200 2 Pe 2, 2 2 2. 3). 2 2 3 1 81 R. subrobusta 82 P. aberrans 5555666666666677777777778 8888288 «8 8899999 9 @ H. simulatrix @ E. adelina P. cicelia 2 L. perangulata 3 L. sorrorcula 4 L. rectistriata 5 P. derwiduii 6 L. milleri 1 1 7 ‘T. knoxvillensis 8 P. bicarinata (continued) AMER. MALAC. BULL. 15(1) (1999) APPENDIX 2. (continued) 9 P. dorothea 1 23 44 5 3 1 23-2 2.2 4 2 2 1 2 10 7. trochonemoides 11 L. centralis 12 T. bellula 1 3 1 23 445 3 1252215 2441264121 23 2 13 T. eccentrica 1 1 2; 2. 2 14 T. montrealensis 15 L. helicteres 16 7. n. sp. 22 2 17 L. serrulata 18 E. strigillata 142 1 2243222 1 1 2 2 2 19 T. umbilicata 20 T. canadensis 21 T. wilsoni 2444 5 3 2. 3° 2 2 22 L. ventricosa 23 T. notablis 24 G. pulchella 2 2.92 1 1 1 25 G. semicarinata 26 G. liljevalli 2 5°4 4-53 5 2 4 27 T. madisonense 28 L. burginensis 29 L. oweni 2 2 2 2.2.4 3: 2 1 2) 22.3 30 L. concinula 3 663 2 2 1 1 31 L. spironema 32 D. decursa 33 D. conoidea 34 D. producta 1 2. 3:28) 3; 2.435272 35 D. Curdsville sp. 36 L. sumnerensis 37 T. salteri 41 L. tropidophora 40 L. tennuistriata 42 D. filosa 38 R. humilis 39 R. lirata 1 2 5: 8 3 22 43 D. bowdeni 44 L. bellicarinata 1 2 by 2 2y43e 221 22 2 3 45 L. quadrisulcata 46 L. mooret 47 L. aff. serrulata 1 1 1 2.2.3) <3); '6".3-.24 3 1 1 1 2223 1 2 2 2 22 2 1 48 T. BMNH 36364 49 T. churkini 50 T. reusingi 224 463 2 3 2 2 224463 2 3 1 25 1 3 1 51 T. aff. umbilicata 52 A. thraivensis 53 G. bicarinata 1 54 L. 2borkholmensis 55 T. pandori 56 T. aff. pandori 57 L. sedgewicki 1 3 673 2 3 1 1 $8 L. ?hyaecinthinsis 59 L. centervillensis 60 A. cancellata 61 A. woodlandi 62 D. trilineata 3232 1 1 1 12. 4. 23) 2 63 L. gotlandica 64 L. laquetta 2 65 A. inexpectans 66 L. holmi 3:6: 73420 3-252 1 2 67 K. nelsonae 68 L. munda 69 L. imbricata 70 T. turrita 71 E. kayesi 2 2 3 2 1 72 R. laevissima 1 ) aie dae ame ee ae a ee ee ae’ 4 73 A. aulangonensis 74 R. robusta 76 L. cyclonema 75 L. tubulosa 77 T. fatua 78 L. soluta 79 L. cochleata 80 E. muricata 81 R. subrobusta 82 P. aberrans WAGNER: LIKELIHOOD, STRATIGRAPHY, AND LOPHOSPIROID PHYLOGENY Gondwana (Gond), Laurentia (Laur), Toquima-Tablehead (ToqTab)]. No. Species Pagodospira cicelia Lophospira perangulata L. sorrorcula L. rectistriata Pagodospira derwiduii Lophospira milleri Trochonemella knoxvillensis Prowrritella bicarinata 9 Pagodospira dorothea 10 Trochonemella trochonemoides 11 Lophospira centralis 12. Trochonema bellula 13. T. eccentrica 14. Trochonemella montrealensis 15. Lophospira helicteres 16 Trochonemella n. sp. 17 Lophospira serrulata 18 Eunema strigillata 19 Trochonema umbilicata 20 T. canadensis 21 ‘iT. wilsoni 22 ‘L. ventricosa 23. Trochonemella notablis 24 Gyronema semicarinata 25. G. pulchella 26 G._liljevalli 27. Trochonema madisonense 28 Lophospira burginensis 29 L. oweni 30 =-L. concinula 31 -L. spironema 32 Donaldiella decursa 33D. conoidea 34 Dz producta 35 __~+D« Curdsville sp. 36 Lophospira sumnerensis 37. = Trochonema salteri 38 Ruedemannia humilis 39 R. lirata 40 Lophospira tenuistriata 41 L. tropidophora 42D. filosa 43D. bowdeni 44 Lophospira bellicarinata 45 L. quadrisulcata 46 L. moorei 47 L. aff. serrulata 48 Trochonemella BMNH 36364 49 T. churkini ONAN MNHPWN — H Ww — W nN \o Were WAIDKDN NK NY WD WwW OR RR re KS NY Pe KEINE HRW H eK — FKA 19 132 132 155 155 155 187 138 138 155 187 187 187 187 187 210 210 221 221 221 221 LKA APPENDIX 3 STRATIGRAPHIC DATA Range and sampling data for lophospiroid species. Separate horizon scales are used for each realm when calculating the likelihood of gaps (Wagner, 1995a). Stratigraphic debt and discrete ranges were calculated using the scales on Fig. 1, where Cassinian = Middle Arenig, Fennian = Late Arenig, Llanvirn = Early Llanvirn, Llandeilo = Late Llanvirn, Ashbian-Black Riveran = Early Caradoc, Rocklandian-Kirkfieldian = Middle! Caradoc, Shermanian-Edenian = Middle2 Caradoc, Maysvillian-Richmondian = Ashgill, Rhuddanian-Aeronian = Early Llandovery, Telychian = Late Llandovery, Sheinwoodian = Early Wenlock, Homerian = Late Wenlock, Gorstian = Early Ludlow, and Ludfordian = Late Ludlow. [FKA, oldest sampled horizon, counted as number of older lophospiroid horizons known up to that point (see Wagner, 1995a); FSA, first stage from which a species is known; H, number of horizons from which a species is known; LKA, latest sampled horizons; LSA, last stage from which a species is known; Realm, biogeographic province: Baltica (Balt), Realm Laur Laur ToqTab Laur ToqTab Laur Laur Balt ToqTab Laur Laur Laur Laur Laur Laur Laur Laur Laur Laur Laur Laur Laur Laur Laur Laur ToqTab Laur Laur Laur Laur Laur Laur Laur Laur Laur Laur Laur Laur Laur Laur Laur Laur Laur Laur Laur Laur Laur Laur Laur FSA Cassinian Cassinian Fennian Fennian Fennian Llanvirn Llanvirn Llandeilan Llandeilo Llandeilo Ashby Ashby BlackRiveran BlackRiveran BlackRiveran BlackRiveran BlackRiveran BlackRiveran BlackRiveran BlackRiveran BlackRiveran BlackRiveran BlackRiveran Rocklandian BlackRiveran Rocklandian Kirkfieldian Rocklandian Rocklandian Rocklandian Rocklandian Kirkfieldian Kirkfieldian Kirkfieldian Kirkfieldian Shermanian Shermanian Edenian Edenian Edenian Edenian Edenian Edenian Maysvillian Maysvillian Richmondian Richmondian Richmondian Richmondian LSA Llanvirn Richmondian Llandeilo Llandeilo Llanvirn Richmondian Llandeilan Llandeilan Llandeilan Ashby Rocklandian Black Riveran Black Riveran Black Riveran Rocklandian Black Riveran Kirkfieldian Black Riveran Shermanian Shermanian Shermanian Kirkfieldian Black Riveran Rocklandian Rocklandian Rocklandian Richmondian Edenian Edenian Rocklandian Rocklandian Shermanian Shermanian Shermanian Shermanian Richmondian Richmondian Richmondian Richmondian Richmondian Richmondian Edenian Richmondian Richmondian Maysvillian Richmondian Richmondian Richmondian Richmondian 29 (continued) 30 APPENDIX 3. (continued) AMER. MALAC. BULL. 15(1) (1999) 50 T. reusingi 3 221 256 Laur Richmondian Richmondian 51 Trochonema aff. umbilicata 3 221 256 Laur Richmondian Richmondian 52. Arjamannia sybellina 5 221 256 Laur Richmondian Richmondian 53 Globonema bicarinata 2 223 239 Balt Richmondian Richmondian 54 Lophospira ?borkholmensis 2 223 239 Balt Richmondian Richmondian 55 Trochonema pandori 2 223 239 Balt Richmondian Richmondian 56 T. aff. pandori 1 223 239 Balt Richmondian Richmondian 57 Lophospira sedgewicki 5 223 273 Laur Richmondian Telychian 58 L.?hyacinthensis 2 223 256 Laur Richmondian Richmondian 59 Longstaffia centervillensis 3 257 273 Laur Rhuddanian Telychian 60 Arjamannia cancellata 1 257 261 Laur Rhuddanian Telychian 61 A. woodlandi 5 257 273 Laur Rhuddanian Telychian 62 ‘OD. trilineata 3 257 265 Laur Aeronian Aeronian 63 Lophospira gotlandica 30 257 337 Laur Aeronian Ludfordian 64 Longstaffia laquetta 6 257 301 Laur Rhuddanian Homerian 65 A. inexpectans 3 262 273 Laur Telychian Telychian 66 Lophospira holmi 1 275 277 Laur Sheinwoodian Sheinwoodian 67 = Kiviasukkaan nelsonae 1 275 277 Laur Sheinwoodian Sheinwoodian 68 Lophospira munda 1 282 294 Laur Sheinwoodian Homerian 69 _L. imbricata 7 278 330 Laur Sheinwoodian Gorstian 70 =Eunema turrita 5 278 308 Laur Sheinwoodian Homerian 71 E. kayesi 1 282 294 Laur Sheinwoodian Gorstian 72 Ruedemannia laevissima 1 295 301 Laur Homerian Homerian 73 = Arjamannia aulongensis 1 295 301 Laur Homerian Whitwellian 74 Ruedemannia robusta 7 282 330 Laur Homerian Gorstian 75 Longstaffia tubulosa 3 295 330 Laur Homerian Gorstian 76 ~L. cyclonema 5 302 308 Laur Homerian Homerian 77 ~~ Trochonema fatua 3 309 330 Laur Gorstian Gorstian 78 Loxoplocus soluta 4 309 330 Laur Gorstian Gorstian 79 Lophospira cochleata 3 309 330 Laur Gorstian Gorstian 80 Trochonema muricata 1 331 332 Laur Ludfordian Ludfordian 81 Ruedemannia subrobusta 3 331 337 Gond Ludfordian Ludfordian 82. Ptychozone aberrans 2 331 337 Gond Ludfordian Pridoli APPENDIX 4 SYSTEMATIC PALEONTOLOGY In the interest of conserving space, taxonomic revisions are pre- Additional Species: ‘“Lophospira” centralis, “L.” helicteres, “L.” concinula Ulrich and Scofield, 1897, “L.” quadrisulcata. sented only for new genera and for those requiring rediagnoses. Genera merely redefined but retaining traditional diagnoses are reclassified as per Fig. 14. SUPERFAMILY LOPHOSPIROIDEA NOM. TRANS. WENZ, 1938 FAMILY LOPHOSPIRIDAE WENZ, 1938 Diagnosis: Lophospiroids primitively featuring a trilineate sinus keel, a strong left ramp carina, and a deep sinus that curves back to the sinus keel. All of these features change within the family in at least one subclade. Included Genera: Lophospira, Donaldiella, Loxoplocus, Eunema, Gyronema, Ruedemannia, Arjamannia, Paupospira gen. nov., Frodospira gen. nov. Genus Eunema Salter, 1859 Diagnosis: Prominent, sharp right and left ramp carina. Strong medial keel usually bordered by peripheral lira (although not in types species). Shallow sinus trending nearly straight into sinus keel. Increasing translation during final whorls, sometimes resulting in disjunct coiling. Type Species: Eunema strigillata Salter, 1859. Discussion. Knight et al. (1960) considered Eunema to be a poly- phyletic subgenus of Trochonema, meant to describe high-spired variants of that genus. However, phylogenetic analyses indicate that the synapo- morphies of Eunema and Trochonema are parallelisms. Knight et al. (1960) acknowledged the polyphyletic nature of the Trochonematidae by noting that the diagnostic characters (e. g. shallow sinus and a medial keel) appeared to be convergent among many forms. This analysis takes their conclusions a step further by positing that the form is polyphyletic among lophospiroids. The much more restrictive definition presented here excludes several species previously assigned to Eunema, which have been reassigned to genera such as Globonema. First-known appearance (FKA): “Lophospira” centralis: Murfreesboro Limestone (Early Caradoc [Ashbyan]). Last-known appearance (LKA): “Lophospira” quadrisulcata: Maquoketa Formation (Early Ashgill [Maysvillian]). Genus Loxoplocus Fischer, 1885 Syn. Kiviasukkaan Peel, 1975b Diagnosis: Sharp and strong medial keel that hooks slightly adapically, typically bordered by sharp but much weaker peripheral lira (although seemingly not on the type species). Narrow, shallow sinus, WAGNER: LIKELIHOOD, STRATIGRAPHY, AND LOPHOSPIROID PHYLOGENY 31 curving back towards sinus keel. Dull but strong left ramp carina, with right ramp concave and featuring a sharp carina at the suture, which cre- ates an acute sutural margin. Increasing translation during final whorls, sometimes resulting in disjunct coiling (with the type species showing dis- junct coiling throughout life). Type Species: “Murchisonia” soluta Salter, 1859. Additional Species: “Lophospira” sedgewicki Donald, 1906, “L.” 2bellicarinata Donald, 1906, “L.” gotlandica Ulrich and Scofield, 1897, “Kiviasukkaan” nelsonae Peel, 1975b. Discussion: Knight et al. (1960) originally defined Loxoplocus as a broad genus that included Lophospira and other subgenera, an inter- pretation rejected by more recent authors (e. g. Tofel and Bretsky, 1987). It now is restricted to a paraclade of lophospiroids whose most prominent features include increasing translation over ontogeny, a very acute suture, and a very concave right ramp. FKA: “Lophospira” sedgewicki: “Starfish Beds,” Girvan District (Late Ashgill [Rawtheyan]). LKA: “Lophospira” gotlandica: Kopanina Formation (Late Ludlow [Ludfordian]); Genus Paupospira gen. nov. Fig. 14 Diagnosis: Thick columella, filling the entire umbilicus and pro- ducing a shovel-like siphon in extreme cases. Deep and wide sinus curv- ing back strongly to sinus keel. Sinus keel trilineate on early whorls, but with trilineations often vague and weak on adult whorls. Type Species: Lophospira oweni Ulrich and Scofield, 1897. Additional Species: “Lophospira” burginensis, “L.” tenuistriata Ulrich and Scofield, 1897, “L.” sumnerensis Ulrich and Scofield, 1897, “L.” tropidophora Meek, 1873, “Schizolopha” moorei, “Donaldiella” bowdeni (Safford, 1869), “Hormotoma” trilineata Foerste, 1923. Etymology: For David Swofford’s computer program PAUP (“Phylogenetic Analysis Using Parsimony”; Swofford, 1998), which helped to diagnose and define this clade. Discussion: Paupospira matches a monophyletic group original- ly hypothesized by Ulrich and Scofield (1897). A new genus is erected to recognize that group. The most distinguishing feature is the extremely thick columella, which often preserves as a core without the rest of the shell. Cladistic depictions of relationship highlighting the key features are given in Fig. 14. FKA: “Lophospira” burginensis: Leray beds, Rockland Formation (Middle Caradoc [Rocklandian]). LKA: “Hormotoma” trilineata; Saugh Hill Group (Middle Llandovery [Aeronian]). Genus Frodospira gen. nov. Diagnosis: Very small shells, with strong ornament that leaves imbrications on the growth lines, a strongly adapically curved medial keel, and highly tangential apertures. Type Species: “Murchisonia” imbricata Lindstrém, 1884. Additional Species: “Lophospira” munda (Lindstrom, 1884), “L.” cochleata (Lindstrom, 1884). Etymology: After J. R. R. Tolkein’s character from Lord of the Rings, reflecting the unusually small size of the known species. Hypotheses about extraneous epidpodial tentacles are purely speculative. Discussion: This closely matches Ulrich and Scofield (1897: 963) imbricata subsection of Lophospira. The genus is remarkable in that all known species are known only from very small shells. As these shells are found in the same assemblages as large gastropods, it appears to repre- sent a character of the genus rather than a taphonomic artifact. The illus- trations of these species, provided by Lindstrém (1884) are very accurate and convey the general characters very well. FKA: “Lophospira” munda: C beds, Visby Formation (Early Wenlock [Sheinwoodian]). LKA: “Lophospira” imbricata: Upper Hemse Beds (Early Ludlow [Gorstian]). FAMILY TROCHONEMATIDAE ULRICH AND SCOFIELD, 1897 Diagnosis: Trochiform, widely umbilicate lophospiroids. Early forms retain wide, and deep sinus curving back to a trilineate selenizone, but later forms possess either the medial lira only or the peripheral lira only. Sharp left and nght ramp carina with strong umbilical and sutural carina found on most species. Included Genera: Trochonemella, Trochonema, Globonema. Discussion: Knight et al.’s (1960) reduction of the Trochonematidae is advanced here. Preliminary analyses (Wagner, unpub. data) indicate that Devonian genera assigned to the Trochonematidae such as Trochonemopsis belong to the Euomphalinae. At this time, it is not clear if any post-Silurian genera assigned to this fam- ily belong here. 1 eee yrhiy TS. wird nena that e los Tee BE, ote “Fy arti Te sii AV yalficr: jan cies he A uy j ; hte Dy A Ve , Lia A, 9 CATON, oo ici 'y tiene haaeeiet ery hh ‘Sn Lyatiie: Pip veh jel! (An fly iy | v , § OTats | - vey haraxkt ion’ boatenaed | Sore Lait Bad eth Y aS. Sh PAS aA! 7 - + bapa oh if “ai lhe Me Pr cat ahh, ‘ s 1 er t ‘ f Vian, AW ARS ~ ites ny 1 bd ay od ‘ ; my, pays / The Tonicella lineata (Wood, 1815) species complex (Polyplacophora: Tonicellidae), with descriptions of two new species Roger N. Clark* Field Associate in Malacology, Natural History Museum of Los Angeles County, 900 Exposition Boulevard, Los Angeles, California 90007, U.S. A. Abstract: Four species of lined chitons from the Pacific coast of North America (two of them new) formerly regarded as Tonicella lineata (Wood, 1815) are described and discussed. Tonicella lineata; T. undocaerulea Sirenko, 1973; T. lokii Clark, spec. nov.; and T. venusta Clark, spec. nov. are differentiated by characteristics of their shells, girdle elements and radulae. Key Words: Tonicella lineata complex, chitons, Polyplacophora For many years chiton workers on the Pacific coast of North America have been aware of two or three varieties of lined chitons of the genus Jonicella Carpenter, 1873 that differed markedly from typical T. lineata (Wood, 1815). In 1973, Sirenko described 7: undocaerulea, a species from the northern Sea of Japan that he believed to be the western Pacific Ocean counterpart of the eastern Pacific 7: lineata. However, soon after its description, several workers recog- nized 7: undocaerulea as one of the forms they had been encountering in the waters of Washington and British Columbia. In his revision of the family Lepidochitonidae of the eastern Pacific Ocean, Ferreira (1982) concluded that T. undocaerulea and the other lined forms were varieties of a single, widely distributed species properly referred to as T. lineata, a view echoed by Kaas and Van Belle (1985). My own preliminary investigations revealed many consis- tent differences between the various forms. These differ- ences suggested the probability of a complex of several closely related species. Further investigation has produced evidence for the presence of four sibling species that over- lap broadly in their ecological habitats as well as in their geographic and bathymetric ranges. MATERIALS and METHODS Several hundred specimens of Tonicella were exam- *Mailing Address: 1839 Arthur Street, Klamath Falls, Oregon 97603-4617 U.S.A. email: insignis@cdsnet.net ined and their characters and distributional patterns were compared. Valves, girdle elements and radulae were stud- ied with light and scanning electron microscopes. Several specimens of each of four nominal species were immersed in a heated 10% solution of KOH until all that remained were the shell plates, radulae and epidermal layers (hyper- notum & hyponotum) of the girdle; these were then rinsed thoroughly in distilled water. The shell plates were placed in a 50% solution of household bleach for 30 minutes, rinsed and air dried. Radulae and epidermal layers of the girdle were separately rinsed, dehydrated in an acetone series, and air dried. The specimens were then mounted on SEM stubs with colloidal silver paint, sputter coated with gold for two minutes, and examined at 5 or 1O KV with an Hitachi S-2100 scanning electron microscope at the Depart- ment of Biology at Southern Oregon University, Ashland, Oregon. Abbreviations used in the text are: LACM, Los Angeles County Museum of Natural History; LACMIP, Los Angeles County Museum of Natural History-Invertebrate Paleontology; USNM, United States National Museum of Natural History, Washington, D.C.; CAS, California Acade- my of Sciences, San Francisco; SBMNH, Santa Barbara Museum of Natural History; ZISP, Zoological Institute, Saint Petersburg, Russia; RBCM, Royal British Columbia Museum, Victoria British Columbia, Canada; RMNH, Rijksmuseum van Natuurlijke Historie, Leiden; UMMZ, University of Michigan Museum of Zoology, Ann Arbor; UAF, University of Alaska, Fairbanks; RNC, Private col- lection of the author. American Malacological Bulletin, Vol. 15(1) (1999):33-46 33 34 AMER. MALAC. BULL. 15(1) (1999) SYSTEMATICS Higher chiton systematics are presently in a state of flux. The widely used and variously modified scheme of Thiele (1909-1910) is based primarily on the characters of the shell plates. However, Sirenko (1993) has proposed a new scheme based on gill placement and the shape of egg hull processes. A very similar system of relationships had been noted earlier by Eernisse (1984). The new scheme appears to be an advancement and is herein adopted. Order: CHITONIDA Thiele, 1929 Suborder: ACANTHOCHITONINA Bergenhayn, 1930 Superfamily: Tonicelloidea Simroth, 1894 Family: Tonicellidae Simroth, 1894 Subfamily: Tonicellinae Simroth, 1894 Genus: Tonicella Carpenter, 1873 Type species: Chiton marmoreus Fabricius, 1780, subse- quent designation by Dall (1878). Tonicella lineata (Wood, 1815) (Figs. 1-8, 33) Chiton lineatus Wood, 1815:15, pl. 2, figs. 4-5 Chiton (Stenosemus) lineatus Wood var. fusca Von Midden- dorff, 1847:110 Lepidochitona lineata: Oldroyd, 1927: 255, 256 (in part). Tonicella lineata: Berry, 1917: 233 (in part); Sirenko, 1974: 995; Ferreira, 1982: 124, figs. 78-81 (in part); Baxter, 1983: 66; Kaas & Van Belle, 1985: 142-144, fig. 65, map 24 (in part); Baxter, 1987: 105. Diagnosis: Chitons of medium size (to 5.0 cm), oval to elongate-oval, valves convex to subcarinate, beaked; tegmentum smooth (shiny) except for faint growth lines; color orange-salmon or maroon, with alternating white and dark maroon-brown lines; lines on head valve forming forwardly directed gothic arch or arrowhead shape. Description: Medium-sized chitons, often reaching 40-45 mm in length; largest specimen examined 52.5 mm (RNC 959, Ogden Point Breakwater, Victoria, British Columbia, Canada, Jeg. George P. Holm, 17 May, 1991). Shell (neotype, herein designated) (fig. 1) oval to elongate- oval, valves subcarinate to convex, beaked; tegmentum smooth (shiny), with faint growth lines. Head valve (fig. 2) semicircular, posterior margin widely V-shaped. Intermedi- ate valves (fig. 3) broadly rectangular, lateral areas very slightly raised, side margins slightly rounded. Tail valve (figs. 4, 5) less than semicircular, about twice as wide as long, with mucro in anterior 1/5-1/3, post-mucronal slope straight to convex. Articulamentum white, with pink stain under central areas; sutural laminae short, wide, about 1/2 length of valve five tegmentum, moderately to strongly rounded; insertion teeth short, thick, finely rugose on ante- rior surface. Slit formula 8/1/9. Color: orange-salmon or maroon, with alternating white and dark, maroon-brown (often nearly black) lines on terminal valves and latero- pleural areas of intermediate valves; lines on head valve always forming forwardly directed arrowhead shape; jugal triangles delineated, usually of darker hues than ground color. Girdle of medium width, about 1/4 width of valve five tegmentum; usually light brown (in alcohol), often with paler bars but occasionally unicolored; dorsal surface covered with very minute, fairly broad, pointed scales (fig. 6), scales strongly ribbed on distal 1/2-3/4, about 28 um long and 13 um wide at base; ventral surface covered with similar but slightly larger scales (fig. 7) about 32 um long and 15 um wide. Radula (fig. 8) rachidian tooth narrow, elongate, sides nearly parallel along distal 3/4, distal end rounded; denticle cap of major lateral teeth (fig. 8a) broad, rounded, about 200 um x 175 um with deep notch on inner edge defining a small, thumb-like denticle, often with a smaller secondary notch above it. Ctenidia merobranchial, abanal, extending 4/5 length of foot. Range of morphological vatiation: Some speci- mens have one or more dark maroon valves. Some speci- mens are nearly totally (except for subjugal markings) orange-salmon or maroon [var. fusca von Middendorff, 1847 (fig. 33), see discussion], but specimens nearly always have at least one white, dark maroon-brown, or black line on the head valve and often on the tail valve as well. Completely white “albino” specimens are infrequent. Type locality: “Their country is unknown” (Wood, 1815: 16). Subsequently designated as Sitka, Alaska (Sirenko, 1974: 994). Here restricted to Old Sitka (Starri- gaven Bay), 10 km N of Sitka, Baranof Id., Alexander Archipelago, SE Alaska (57°08’N, 135°55’W). Ferreira (1982: 125) designated Monterey Bay, California as the type locality, apparently unaware that Sirenko had already designated Sitka as the type locality. Type material: Lost (Ferreira, 1982). Neotype LACM 2734 (fig. 1) here designated from type locality (leg. RNC, intertidal beneath cobble encrusted with coralline algae; 19 May, 1983); 35 mm x 22.5 mm x 6.5 mm (ex-RNC 496). Additional material: Alaska: 1, CAS 41449, Montague Id.; 1, RNC 659, Navy breakwater, Kuluk Bay, Adak Id., Aleutian Islands, +0.5 m; 1, RNC 666, Nazan Bay, Atka Id., Aleutians, 0 m; 3, RNC 398, Chernofski CLARK: TONICELLA LINEATA COMPLEX 55 Figs. 1-8. Tonicella lineata (Wood, 1815). 1. Whole animal, Neotype, LACM 2734; Old Sitka, Baranof Island, Alaska. 35 mm x 22.5 mm x 6.5 mm. 2-8. RNC 430; Cape Arago, Oregon. 2. Head Valve; width 9.0 mm. 3. Intermediate valve five; width 12.2 mm. 4, 5S. Tail valve; width 6.3 mm. 6. Dorsal girdle scales. 7. Ventral girdle scales. 8. Radula. 8a. Denticle Cap of Major Lateral Tooth. Bar = 200 um. 36 AMER. MALAC. BULL. 15(1) (1999) Harbor, Unalaska Id., Aleutians, 0 m; 5, Eider Point, Unalaska Id., Aleutians, 0 m; 2, RNC 901, Herendeen Bay, SE Bering Sea, 0 m; 3, RNC 146, Chiniak Bay, Kodiak Id., Om; 4, RNC 442, Kachemak Bay, Kenai Peninsula, 0 m; 2, RNC 496, Old Sitka, Baranof Id., 0 m; 9, RNC 975, Petersburg, Mitkof Id., 0-1 m; 5, RNC 1295, Petersburg, Mitkof Id., | m; 6, CAS 083452, Edena Bay, Kosciusko Id.; 4, RNC 1236, N. Vallinar Rocks, Gravina Id., 3-5 m; 77, CAS 013454, N of Ketchikan, Revillagigedo Id.;_ 3, RNC 970, Saxman, Revillagigedo Id., 0-1 m; 3, RNC 969, Rotary Beach, Revillagigedo Id., 0-1 m; 3, RNC 1191, Mountain Pt., Revillagigedo Id., 0-1 m; 12, RNC 5450, Metlakatla, Annette Id., 0-2 m; 9, RNC 544, Brownson Bay, Prince of Wales Id., 0-1 m; 24, CAS 013121, Rose Inlet, Dall Id., British Columbia: 2, CAS 013590, Portage Inlet, | m; 5, RNC 934, Bowen Id., 3-5 m; 3, RNC 320, Victoria, Vancouver Id., 0-2 m; 1, RNC 959, Victoria, Van- couver Id., 0-1 m. Washington: 2, CAS 013346, Marrow- stone Id., Jefferson Co.; 2, RNC 344, Indian Id., Jefferson Co., 0-1 m; 2, RNC 642, Neah Bay, Clallum Co., 0-1 m; 5 RNC 13. Tacoma Narrows, Pierce Co., 0-1 m. Oregon: 2, RNC 430, N Cove, Cape Arago, Coos Co., +1 m; 5, RNC 4, Cape Blanco, Curry CO., 0m; 1, RNC 979, Port Orford, Curry Co., 2m; 1, RNC 999, Island Rock, Curry Co., 17 m; 4, RNC 963, Brookings, Curry Co., 0 m. California: 4, CAS 060413, Trinidad, Humboldt Co,; 28, CAS 009294, S of Cape Mendocino, Humboldt Co.; 3, CAS 013396, Shelter Cove, Mendocino, Co., 0-1 m; 1, CAS 013374, Fort Ross, Sonoma Co.; 2, CAS 077112, Bodega Head, Sonoma Co,; 3, CAS 069442, Bodega Bay breakwater, Sonoma Co.; 4, CAS 034724, Duxbury Reef, Marin Co.; 2, CAS 078781, Marin Co.; 1, CAS 077028, Angel Id., San Francisco Bay; 3, CAS 001933, Fort Mason , San Francisco Co.; 2, CAS 007149, Franklin Point, San Mateo Co.; 1, CAS 007148, Ano Nuevo Point, San Mateo Co.; 1, CAS 000219, Davenport Landing, Santa Cruz Co.; 3, CAS 013361, China Point, Monterey Co.; 2, RNC 1233, Coast Guard breakwater, Monterey Bay, 3-5 m. Distribution: Tonicella lineata is a North American boreal species endemic to the Aleutian and Oregonian provinces (Fig. 37). The species occurs continuously from the central Aleutian Islands to central Californica in depths ranging from +2 to 17 m. The westernmost record is Navy breakwater, Kuluk Bay, Adak Island, Aleutian Islands, Alaska (51°45°N, 176°45’°W) (LACM 86-00.0, RNC 659, leg. Rae Baxter, 12 August, 1286, intertidal); the northern- most record is Montague Island, Prince William Sound, Alaska (60°10°N, 147°15°W) (CAS 41449), and the south- ernmost record is the United States Coast Guard breakwa- ter, Monterey Bay, Monterey County, California (36°45’°N, 121°S55°W)(RNC 1060; leg. RNC, with SCUBA, 23 March, 1992, 5 m). These animals are abundant at some localities, with population densities often exceeding 50 per square m. This species fairly common at the Navy breakwater at Adak Island (Rae Baxter, pers. comm., Nov. 1988) but is abruptly absent west of that locality. In the southern portion of its range, it becomes uncommon from Mendocino County, California southward, where it begins to overlap with the similar Tonicella lokii; it is quite rare in Monterey County. Habitat and ecology: Lives on cobbles, boulders and bedrock encrusted with purple-pink coralline red algae (Lithothamnion spp.) or unidentified rust-brown encrusting bryozoans Fossil record: Valves of Tonicella lineata have been identified from Pleistocene deposits in southern Ore- gon and northern California (LACMIP loc. 2636, Coquille Point, Coos Co., Oregon; LACMIP loc. 2641, Cape Blan- co, Curry Co., Oregon; LACMIP loc. 3955, Battery Forma- tion, Crescent City, Del Norte Co., California; LACMIP loc. 4816 & 10770 & USGS loc. M7824, Point Arena, Mendocino Co., California). These specimens have been radiometricaly dated at 80,000-85,000 years before present (BP). Specimens from Moonstone Beach, Humboldt Co., California (LACMIP loc. 3942) have been dated at 700,000-1,000,000 years BP (George L. Kennedy, pers. comm. Oct. 1992). Remarks: Tonicella lineata is remarkably similar to T. undocaerulea_ Sirenko, 1973 and T. lokii Clark, spec. nov. (see comments for those species). Tonicella undocaerulea Sirenko, 1973 (Figs. 9-16, 35, 36) Lepidochitona lineata, non Wood: Oldroyd, 1927: 255, 256 (in part). Tonicella lineata, non Wood: Taki, 1938: 331; Yakovleva, 1952: 56, 61-62 (in part); Itoigawa et al., 1978 (Fossil): 150-153, pl. 16, figs. 2-7; Ferreira, 1982: 124-126 (in part); Kaas & Van Belle, 1985: 142- 144, fig. 65-5, 6, map 24 (in part). Tonicella undocaerulea Sirenko, 1973: 663, figs. 3/1-7, 9- 13; Baxter, 1983: 66; Baxter, 1987: 105. Diagnosis: Chitons of relatively small size [(to 3.8 cm), Asian specimens smaller]; shell oval to elongate-oval; valves subcarinate to round, beaked; tegmentum smooth, shiny; color (in alcohol) light orange with white zigzag lines, often with short, dark maroon streaks on edges of pleural areas. Description: Relatively small chitons, rarely exceeding 25 mm in North American waters and 16 mm in CLARK: TONICELLA LINEATA COMPLEX 37 te ‘ Fis TS x 4.5 mm. 10-16 RNC 255; Neah Bay, Clallum County, Washington. 10. Head valve; width 4.5 mm. 11. Intermediate valve five; width 6.7 mm. 12, 13. Tail valve; width 4.0 mm. 14. Dorsal girdle scales. 15. Ventral girdle scales. 16. Radula. 16a. Denticle Cap of Major Lateral Tooth (Asian specimen). 16b. Denti- cle Cap of Major Lateral Tooth (North American specimen). Bar = 100 um. 38 AMER. MALAC. BULL. 15(1) (1999) Asian waters. Largest specimen examined 38.0 mm (RNC 287b, Ogden Point breakwater, Victoria, British Columbia, leg. RNC, 12 July, 1987). Shell oval to elongate-oval (often egg-shaped)(fig. 9), valves subcarinate to round, beaked; tegmentum smooth, shiny, usually lacking notice- able growth lines. Head valve (fig. 10) semicircular, poste- rior margin widely V-shaped to nearly straight. Intermedi- ate valves (fig. 11) broadly rectangular, lateral areas indis- tinct to slightly raised, side margins slightly rounded. Tail valve (figs. 12, 13) oval, mucro in anterior 1/3, post- mucronal slope convex to (rarely) slightly concave. Articu- lamentum white, stained with intense rose-maroon under central areas; sutural laminae short, wide, about 1/3 length of valve five tegmentum, rounded; insertion teeth short, solid; slit formula 8/1/8-9. Color light orange, terminal valves with concentric, white zigzag lines radiating from apices; lateropleural areas with white zigzag lines, those on pleural areas often bordered beneath with short, dark maroon streak; jugal areas with delineated triangles of maroon, yellow, orange, or white. Girdle narrow, less than 1/4 width of valve five tegmentum, appearing nude; dorsal surface covered with minute, rather narrow, pointed, smooth scales (fig. 14), about 18-20 um long and 5 um wide, with single groove (often obscured) extending from apex to about mid length of scale; ventral surface covered with minute, tear-drop shaped scales (fig. 15), about 25 um long and 12 um wide, ribbed along proximal 1/2-1/3. Radu- la (fig. 16) rachidian tooth very broadly dilated anteriorly, forming distinctive spoon shape; denticle cap of major later- al teeth (16a & b) tridentate, about 120 um x 130 um, cen- tral cusp the largest, inner cusp slightly smaller, and outer cusp about 1/2 length of the central one. Ctenidia mero- branchial, abanal, extending about 3/4 to 4/5 length of foot. Range of morphological variation: Tonicella undocaerulea varies somewhat in shape and size from one locality to another in American waters, some specimens being quite egg-shaped in outline, whereas others are very elongate. The slope of the shell also varies from straight and nearly carinate to very convex. The valves of some specimens may be completely maroon, or suffused with maroon on the lateropleural areas; the ground color of some specimens may be yellow. Type Locality: Russia, Bay of Minonosok, Posjet Strait, Sea of Japan (42°39’N, 130°54’E), by original desig- nation (Sirenko, 1973). Type material: Holotype ZISP 1905; paratypes ZISP 1906 & 1907. Additional material: Russia: 3, RNC 103, Vostok Bay, Sea of Japan, 1-2 m; 3, RNC 118, Vostok Bay, Sea of Japan, 1 m; 2, CAS 077081, Vostok Bay, Sea of Japan, 1- 22 m; 7, CAS 013448, Vostok Bay, Sea of Japan, 2-3 m; 12, LACM 91-91, Vostok Bay, Sea of Japan, 0-3 m; 4, LACM 91-94, Vostok Bay, Sea of Japan, 3-10 m; 1, LACM 91-89, Popova Id., Amuskly Bay, Sea of Japan, 0-2 m. Japan: 8, LACM 82-12, Mutsu Bay, Honshu Id., 3-10 m; 1, RNC 725, Oshoro, Hokkaido, 1-3 m; 1 valve, RNC F-1 (fossil), Soli, Kisarazu City, Chiba Prefecture, Kioroshi Formation, Shimosa group, Middle Pleistocene. Alaska: 2, RNC 501, Naked Id., Prince William Sound, 3-6 m; 1, CAS 013449, Kodiak Id.; 1, RNC 1196, Chiniak Bay, Kodiak Id., | m; 1, RNC 679, Hesketh Id., Kachemak Bay, Kenai Peninsula, 9 m; 6, RNC 348, Kachemak Bay, Kenai Peninsula, 1-2 m; 6, RNC 348, Kachemak Bay, Kenai Peninsula, 1-2 m; 6, RNC 144, Sitka, Baranof Id., 1-2 m; 1, LACM 73-13, Sitka, Baranof Id., 1-5 m; 1, LACM 73- 15, Pirate Cove, Baranof Id., 3-12 m; 2, RNC 1192, Peters- burg, Mitkof Id., 5-6 m; 25 CAS 013331, Edena Bay, 940, Saxman, Revillagigedo Id., 0-2 m; 5, RNC 1031, Rotary Beach, Revillagigedo Id., 0-1 m; 4, RNC 241, Mountain Pt., Revillagigedo Id., 1-3 m; 10, RNC 1189, Mountain Pt., Revillagigedo Id., 0-5 m; 9, RNC 425, Met- lakatla, Annette ID., 0-2 m; 1, RNC 1198, Washington Monument, Revillagigedo Channel, 5-8 m. British Colum- bia: 1, RNC 222, Port Hardy, Vancouver Id., 1-2 m; 1, CAS 002414, Quiet Bay, Vancouver Id.;_ 11, RNC 287, Victoria, Vancouver Id., 0-2 m. Washington: 5, RNC 136, San Juan Id., 12-18 m; 2, RNC 337, Port Gable, Hood Canal, Kitsap Co., | m; 2, RNC 255, Neah Bay, Clallum Co., 1-2 m. Oregon: 5, RNC 973, Island Rock, Curry Co., 30 m; 1, RNC 141, Brookings, Curry Co. California: 9, LACM 71-106, Tolo Bank, Mendocino Co., 21-30 m; 1, LACM 64-20, Salt Point Ranch, Sonoma Co., 3-5 m; 1, CAS 008948, SE Farallon Id.; 4, CAS 013350, Monterey Bay, Monterey Co., 18 m; 9, LACM 60-22, Del Monte, Monterey Bay, 18 m; 6, RNC 1027, Coast Guard breakwa- ter, Monterey Bay, 3-15 m; 3, LACM 63-3, Coast Guard breakwater, Monterey Bay, 4-12 m; 2, CAS 083441, Chace Reef, Monterey Co., 12-13 m; 1, LACM 60-24, Carmel Submarine Canyon, Monterey Bay, 3-15 m; 3, LACM 63- 3, Coast Guard breakwater, Monterey Bay, 4-12 m; 2, CAS 083441, Chace Reef, Monterey Co., 12-13 m; 1, LACM 60-24, Carmel Submarine Canyon, Monterey Bay, 12-38 m; 2, LACM 38-153, San Luis Obispo Bay, San Luis Obis- po Co., 15-26 m; 1, LACM 38-162, San Miguel Id., Chan- nel Islands, 9-27 m. Distribution: Tonicella undocaerulea has a discon- tinuous distribution. In Asian waters it has been reported from the Sea of Japan (Sirenko, 1973; Saito, 1994) from Mutsu Bay, Honshu Id. Japan (42°20’N, 140°55’E) to near Uglegorsk, SE Sakhalin Id., Russia (49°00°N, 142°31°E) CLARK: TONICELLA LINEATA COMPLEX 39 and the Okhotsk Sea (Sirenko, 1973) from Iturup Id., Kurile Is., Russia (approx. 45°00’N, 147°30’E) at depths of 1-27 m (Sirenko, 1973). In American waters it is found from Kodiak Id., Alaska (57°50’°N, 152°30’W)(RNC 1196) and Naked Id., Prince William Sound, Alaska (60°37.8’N, 146°23’W)(RNC 501) to San Miguel Id., California (34°O1'N, 120°24°N)(LACM 38-162) (Fig. 37) at depths of 0-38 m. Habitat: Tonicella undocaerulea lives on encrust- ing coralline algae (Lithothammon spp.) on pebbles, cob- bles, boulders and bedrock. Fossil record: Valves of 7. undocaerulea have been found in the middle Pleistocene deposits on the Boso Peninsula, Honshu Id., Japan (Itoigawa ef al., as T. lineata). Remarks: Tonicella undocaerulea has long been confused with the similar 7: lineata and T. lokii because of its lined color pattern and habitat of encrusting Lithotham- nion. However, it may be distinguished from both of theses species by its lack of maroon-brown “black” lines on the valves. North American specimens of Tonicella undo- caerulea are considerably larger than Asian specimens, attaining a length of up to nearly 40 mm, whereas the largest specimen recorded from Asia is 16 mm (Sirenko, 1973). Also, the denticle cap of the major lateral teeth of North American specimens have very rounded cusps, in Asian specimens they are quite triangular. However, all other characters of North American and Asian specimens are identical. Because of the differences in size and distrib- ution, and the shape of the denticle caps, North American specimens should perhaps be considered a subspecies. Tonicella lokii sp. nov. (Figs. 17-24) Tonicella lineata, non Wood: Berry, 1922: 433, pl. 2 figs. 1-5 (fossil); Burghardt and Burghardt, 1969: 36 (in part, pl. 4, no. 76); of authors (from Monterey Bay and central California). Lepidochitona lineata, non Wood: Oldroyd, 1927: 255, 256 (in part). Diagnosis: Chitons of medium size (to 5.0 cm), shell low to moderately elevated; valves salmon to light orange, with white and dark maroon-brown (black), zigzag lines on terminal valves and lateropleural areas of interme- diate valves; jugal areas with dark orange, pink, or maroon triangles; tail valve with concave post-mucronal slope; gir- dle with alternating light and dark bars. Description: Holotype (fig. 17) 32.3 mm x 21.0 mm x 5.5 mm; body oval in outline, slightly elevated; valves subcarinate, smooth. Head valve (fig. 18) semicir- cular, posterior margin widely V-shaped. Intermediate valves (fig. 19) rectangular, length about 1/3 width, slightly beaked. Tail valve (figs. 20-21) oval, more than twice as wide as long, mucro in anterior 1/3, post-mucronal slopes concave. Articulamentum white, with light brown rays on slit rays, often with triangular pink stain under central areas; slit formula 9/1/11; sutural laminae short, about 1/2 length of valve five tegmentum; jugal sinus rather narrow; insertion teeth short, thick, those on head valve crenulated on exterior surface. Girdle moderately wide, about 1/4 width of valve five tegmentum; dorsal surface appearing smooth, actually clothed with very minute non-imbricating, broad, smooth, bluntly pointed scales (fig. 22) 22-25 um long and 10-12 um wide, rarely with a barely perceptible groove extending from apex to mid-length of scale; ventral surface covered with minute, densely packed, broad scales (fig. 23) 2/3 length. Radula (fig. 24) 9.7 mm long, bearing 53 mature rows of teeth; rachidian tooth about 100 um long, narrow at base dilated anteriorly to broad, spatula shape about 50 um wide and slightly recurved at distal end; denticle cap of major lateral teeth (24a) broad, rounded, slightly longer than wide, about 120 um x 135 um, with deep notch on inner edge defining a small thumb-like den- ticle. Ctenidia merobranchial, abanal, extending 4/5 length of foot, about 25 plumes per side. Range of morphological variation: As is the case with several members of this genus, individual valves may be uniformly white or dark maroon. However, unlike other Tonicella spp., the central portion of the articulamentum is often unstained, and when stain is present it may be pink, rose, or light brown. Type locality: Coast Guard breakwater, Monterey Bay, Monterey County, California (36°45’N, 121°55’W), 0- 13 m. Type material: Holotype (LACM 2623) and 36 paratypes (2, LACM 2624 & 2625); (2, USNM 880067); (2, CAS 103560); (2, SBMNH 141109); (1, UMMZ 252868); (2, RMNH 9360); (3, ZISP 1933); (22, RNC 1056 & 1191). Type material preserved flat and fully extended, all but seven in 70% ethanol or 70% isopropyl alcohol; remaining seven specimens glycerin-dried. Type material collected 16 & 23 March, 1992 by RNC (9 specimens) and 26 March, 1993 by RNC and Bob Abela (28 specimens). Additional material: California: 1, CAS 013583, Shelter Cove, Humboldt Co.; 1, CAS 013798, N of West- 40 AMER. MALAC. BULL. 15(1) (1999) Figs. 17-24. Tonicella lokii Clark, spec. nov. 17. Whole animal, Holotype, LACM 2623; 32.3 mm x 21.0 mm x 5.5 mm. 18-24. Paratype, RNC 1056. 18. Head valve: width 9.4 mm. 19. Intermediate valve five; width 13.2 mm. 20, 21. Tail valve; width 8.8 mm. 22. Dorsal girdle scales. 23. Ventral girdle scales. 24. Radula. 24a. Denticle Cap of Major Lateral Tooth. Bar = 100 um. CLARK: TONICELLA LINEATA COMPLEX 4] port Landing, Mendocino Co.; 3, CAS 013459, Buckhorn Cove, Mendocino Co., 0.5 m; 5, CAS 013453, Albion Head, Mendocino Co.; 10, LACM 49-3, Salmon Point, Mendocino Co., 0.5 m; 20, LACM 50-11, Fort Ross, Sono- ma Co., 0.5 m; 8, CAS 043848, Stewart’s Point, Sonoma Co.; 3, CAS 069441, Bodega Bay breakwater, Sonoma Co.; 1, CAS 043485, Tomales Point, Marin Co.; 18, CAS 008948, SE Farallon Id.; 1, CAS 069445, S of Pescadero Point, San Mateo Co.; 1, CAS 007065, Pigeon Point, San Mateo Co.; 1, CAS 000217, Davenport Landing, Santa Cruz Co.; 5, CAS 013350, Monterey Bay, Monterey Co., 18 m; 6, CAS 013361, China Point, Monterey Co.; 2, CAS 013364, Cannery Row, Monterey Bay, 10 m; 7, LACM 63-58, Mission Point, Monterey Co., 0.5-2 m; 2, RNC 342, Point Pinos, Monterey Co., 0-1 m; 4, RNC 656, Carmel, Monterey Co., 0-1 m; 2, CAS 056335, 6.4 km N of San Simeon, San Luis Obispo Co.; 2, RNC 402, Cayu- cos, San Luis Obispo Co., 0.5 m; 1 CAS 013586, Purisima Point, Santa Barbara Co.; 1, CAS 020482, Surf, Santa Bar- bara Co.; 1, LACM 64-28, S side of Anacapa Id., Channel Is., 15-24 m; 7, LACM 67-38, SE of Bay Point, San Miguel Id., Channel Is., 0-8 m. Distribution: Tonicella lokii occurs between lati- tudes 40°N and 34°N , on the northern and central coast of California (Fig. 37) at depths of 0-24 m. The northernmost record is Shelter Cove, Humboldt County (40°01.5’N, 124°45’°W)(CAS 013583), and the southernmost record is SE of Bay Point, San Miguel Id., California (34°02’N, 120°18°W)(LACM 67-38). Only one specimen of T. lokii has been recorded from Humboldt County, and the species is not common in adjacent Mendocino County to the south, but it becomes more abundant closer to the Monterey County distribution center. Habitat and ecology: Tonicella lokii lives on cob- bles, boulder, and bedrock encrusted with coralline algae (Lithothamnion spp.). Fossil record: Valves of T. lokii have been identi- fied from Pleistocene deposits in southern California (LACMIP Loc. 11004, First Terrace, Army Camp Beach, San Nicholas Island), at a San Pedro area terrace (Berry, 1922, as T. lineata), and from northern Baja California, Mexico (LACMIP loc. 10131, Lighthouse Terrace, Bahia el Playon, Punta Banda, and LACMIP loc. 10619, Lighthouse Terrace, near tip of Punta Banda). The specimens have been dated at 80,000-85,000 years BP (G. L. Kennedy, pers. comm., 1992). Etmology: Named for Loki, the Norse God of mis- chief and deception, appropriate for a species that long has deceived biologists as to its true identity. Remarks: Tonicella lokii is often found with the morphologically similar and apparently closely related T. undocaerulea, but it can be easily distinguished from the latter by the presence of dark maroon-brown lines on its valves. Tonicella lokii may also be found with T. lineata, from which it can be distinguished by the concentric zigzag lines on the head valve (those on T: lineata form a very dis- tinctive gothic arch) and by the concave post-mucronal slope of its tail valve. Tonicella venusta sp. nov. (Figs. 25-32, 34) Tonicella rubra, non Linnaeus: Berry, 1917: 233 (in part); Smith and Gordon, 1948: 205; Burghardt and Burghardt, 1969: pl. 4, fig. 79. Lepidochitona lineata: Oldroyd, 1927: 255, 256 (in part). Lepidochitona ruber, non Linnaeus: Oldroyd, 1927: 256- 257 (in part). Tonicella marmorea, non Fabricius: Baxter, 1983: 66. Diagnosis: Chitons of small size (to 1.7 cm), oval, shell low to moderately elevated, subcarinate; valves light orange or pink, terminal valves and lateral areas of interme- diate valves with white, zigzag lines; pleural areas with 2-5 large white flammules; jugal areas with orange, pink, white or maroon triangles; post-mucronal slope of tail valve con- cave. Description: Holotype (fig. 25) preserved dry, flat and fully extended, 10.5 mm x 6.0 mm x 1.8 mm, body oval in outline, slightly elevated; valves subcarinate, smooth. Head valve (fig. 26) semicircular, posterior mar- gin widely V-shaped. Intermediate valves (fig. 27) rectan- gular, length about 1/3 width, beaked; lateral areas poorly defined. Tail valve (figs. 28-29) oval, length about 1/2 with; mucro anterior 1/3, post-mucronal slope concave. Articulamentum white or pale pink, with triangular maroon stain under central area; sutural laminae short, about 1/3 length of valve five tegmentum; jugal sinus about 1/5 valve width; insertion teeth short, fairly thick; slit formula 8/1/11. Girdle of moderate width, about 1/4 width of valve five tegmentum, of sandy appearance, clothed dorsally with rel- atively large, closely packed, erect, smooth, rotund, mam- milate scales (fig. 30) about 45 um long and 28 um wide; mammilae with 8-10 heavy ribs; ventral surface of girdle covered with closely packed, rather broad, bluntly tapering scales (fig. 31) about 30 um long and 16 um wide, strongly ribbed on proximal 1/3-1/2. Radula (fig. 32) 3.2 mm long, with 53 mature rows of teeth; rachidian tooth spatulate, about 75 um long and 12 tm wide at base, dilating proxi- mally to about 30 tm wide at working edge, thickened on 42 AMER. MALAC. BULL. 15(1) (1999) 32a 28 Figs. 25-32. Tonicella venusta Clark, spec. nov. 25. Whole animal, Holotype, LACM 2626: 10.5 mm x 6.0 mm x 1.8 mm. 26- Head valve: width 4.0 mm. 27. Intermediate valve five; width 5.0 mm. 32. Radula. 32a. Denticle cap of major lateral tooth. Bar = 50 pm. 32. Paratype, RNC 1033 26. 28, 29. Tail valve; width 3.2 mm. 30. Dorsal girdle scales. 31. Ventral girdle scales. CLARK: TONICELLA LINEATA COMPLEX 43 36 Fig. 33. Tonicella lineata var. fusca (von Middendorff, 1847) ? RNC 934, Tunstal Bay, Bowen Id., near Vancouver, British Columbia, 6 m. 38.0 mm x 23.0 mm x 6.0 mm. Fig. 34. Tonicella venusta Clark, spec. nov., RNC 194, Neah Bay, Clallum Bay, Washington, 0-1 m. 8.0 mm x 4.0 mm x 2.0 mm (Albino). Figs. 35 & 36. Tonicella undocaerulea Sirenko, 1973, fossil intermediate and tail valves, middle Pleistocene, Boso Peninsula, Honshu Id., Japan (Photos: courtesy of A. Naruse). Scale bar 5.0 mm. lateral edges and notched at center at proximal end; denti- cle cap of major lateral teeth (32a) broad, tridentate, about 65 um x 75 um, rounded on outer edge, denticles about equal in length (inner one only slightly smaller). Ctenidia merobranchial, abanal, extending about 3/4 length of foot, 16 plumes per side. Range of morphological variation: Some speci- mens have a dark maroon-brown, central area bordered by a diagonal white band at the edge of the lateral areas on valve two. Rare albinistic specimens (fig. 34) are gray- white with a few darker gray blotches at the posterior edges of the plates. Some specimens are nearly uniformly, light pink, and one southern specimen has the color pattern of the lateropleural areas extending across the jugal areas. Type locality: Mountain Point, 8 km S of Ketchikan, Revillagigedo Island, Alexander Archipelago, SE Alaska (55°17°35”N, 131°32’20’W), 1-10 m. Type Material: Holotype (LACM 2626) and 41 paratypes (2, LACM 2627, 2628); (2, CAS 103559); (1, SBMNH 141110); (2, USNM 880068); (1, RMNH 9361); (1, ZISP 1934); (2, UAF MO-5572); (1, UMMZ 252870); (29, RNC 1026, 1033, 1140). Type material preserved flat and fully extended. Holotype and 17 paratypes preserved dry (with glycerin), collected 24-25 June, 1991 by RNC & Alan Murray; Twenty-four additional paratypes preserved in 70% ethanol, 8 collected 3 September, 1992 by RNC, Alan Mur- ray, Kurt Morin & David Zwick, and 16 collected 26-30 September, 1993 by RNC, David Zwick & Kurt Morin. Additional Material: Alaska: 1, CAS 025595, Knight Id., Prince William Sound; 5, LACM 83-106, Hes- keth Id., Kachemak Bay, Kenai Peninsula, 12-16 m; 3, RNC 435, Kachemak Bay, 0-2 m; 2, RNC 1193, Chiniak Bay, Kodiak Id., 1 m; 7, RNC 200, Sitka, Baranof Id., 0-1 m; 1, RNC 1193, Chiniak Bay, Kodiak Id., | m; 7, RNC 44 AMER. MALAC. BULL. 15(1) (1999) 200, Sitka, Baranof Id., 0-1 m; 1, RNC 1197, Petersburg, Mitkof Id., 8-10 m; 2, SBMNH 36067, Forrester Id.; 6, RNC 1010, Rotary Beach, Revillagigedo Id., 0-1 m; 9, RNC 1033, Mountain Pt., Revillagigedo Id., 1 m; 9 RNC 579, Metlakatla, Annette Id., 42 m; 1, RNC 912, Saxman, Revillagigedo Id., 1 m; 9, RNC 1199, Washington Monu- ment (submerged pinnacle), Revillagigedo Channel, 5-30 m. British Columbia: 1, SBMNH 36069, Departure Bay, Vancouver Id.; 3, RNC 251, Tofino Harbor, Vancouver Id., 1-2 m; 8, RBCM 7516, Tofino Harbor, Vancouver Id., 1-2 m; 2, RNC 397 Saanach Inlet, Vancouver Id., 18 m; 2, RNC 349, Victoria, Vancouver Id., 1-2 m. Washington: 6, RNC 194, Neah Bay, Clallum Co., 0-1 m. Oregon: 1, RNC 972, Port Orford breakwater, Curry Co., 5 m; 1 RNC 971, Island Rock, Curry Co., 30 m. California: 1, CAS 013476, Shelter Cove, Mendocino Co.; 1, RNC 1226, Coast guard breakwater, Monterey Bay, Monterey Co., 10- 12m; 3 SBMNH 36073, N of San Simeon, San Luis Obis- po Co.; 1, Hanselman Coll., Gaviota, Santa Barbara Co., 18 m; 1, LACM 41-195, E of Carwell Point, San Miguel Id., Channel Is., 38 m: 3, SBMNH 36066, Dago Bank, off San Pedro, Los Angeles Co.; 1, CAS 025200, Catalina Id., 79-140 m; 1, CAS 083447, Point Fermin, Los Angeles Co., 30 m; 3, LACM 65-3, S of Los Angeles Harbor, Los Angeles Co., 27 m; 2, LACM 65-2, SW of Point Fermin, Los Angeles Co., 29 m; 1, LACM 57-53, NW of Point Vincent Light, Los Angeles Co., 18 m; 24, LACM 72-91, off San Pedro, Los Angeles Co., 22 m; 1, CAS 083442, Cortez Bank, Channel Is., 12-27 m; 1, RNC, 264, Point Loma, San Diego Co., 15 m; 2, Hanselman Coll., Point Loma, San Diego Co., 15 m. Baja California, Mexico: 1, a at ceo CAS 083446, Bahia de Todos Santos; 1, LACM 67-52, Arbolito, S side of Punta Banda, 15 m; 1, LACM 67-46, Puerto Santo Tomas, 13 m; 1, LACM 71-151, NE end of Isla Cedros, 5-12 m. Distribution: Tonicella venusta occurs continuous- ly from south-central Alaska to Isla Cedros, Baja Califor- nia, Mexico (Fig. 37) at depths of 0-140 m. The northern- most record is Knight Island, Prince William Sound, Alas- ka (60°15’N, 147°44’W)(CAS 025595), the westernmost record is Chiniak Bay, Kodiak Id., Alaska (57°41°15”N, 152°24°00"W), and the southernmost record is Isla Cedros, Baja California (28°17’N, 115°8°W)(LACM 71-151). Habitat and Ecology: Tonicella venusta lives on pebbles, cobbles, boulders, shells, and bedrock encrusted with coralline algae (Lithothamnion spp.) and bryozoans of unknown identity. Fossil record: No fossil specimens of Tonicella venusta have been found. This is probably due to the small size and fragile nature of the valves. Etymology: The name is Latin, and means lovely. The species is so named for the brilliant coloration of the valves in live animals, which are pink or lavender with blue or purple markings. Remarks: Because of its small size and sandy gir- dle, Tonicella venusta has previously been confused with Tonicella beringensis (Yakovleva, 1952) [as Tonicella ALASKA e an \stae Aleuti oo North American distribution of the 0o Tonicella lineata complex. {e] Tonicella lineata [0] Tonicella undocaerulea [©] Tonicella loki *00 a xO*NPt. Conception x00 ¥¢ [x] Tonicella venusta Fig. 37. Distribution of the Tonicella lineata complex. CLARK: TONICELLA LINEATA COMPLEX 45 rubra (Linnaeus, 1767), see discussion] from which it may be distinguished by its anterior mucro, concave post- mucronal slope and color pattern. Tonicella venusta has also been confused with 7: marmorea (Fabricius, 1780), from which it is distinguished by the sandy girdle and the color pattern; with juveniles of 7: lineata, from which it may be distinguished by the post-mucronal slope of the tail valve, sandy girdle and the color pattern; and with Lepido- chitona beanii (Carpenter, 1855), from which it is distin- guished by lacking large spicules in the girdle at the valve sutures. DISCUSSION The similarities in color pattern and habitat between the members of the Tonicella lineata complex are remark- able, and explain why the four taxa have been confused. Yet they differ in line pattern, post-mucronal slope of the tail valve and the morphology of the girdle elements and radula (Table 1). In addition to morphological characters of the valves, girdle elements and radula, these species also have distinct differences in the number of gills. In speci- mens of the same size, Tonicella lokii has slightly more numerous gills than 7: lineata, and these two species both have more numerous gills than T. undocaerulea and T. venusta. Tonicella venusta has the fewest gills of the four (figure 38). Berry (1917) was apparently misled by these simi- larities, as an examination of Berry’s material from For- rester Island, Alaska revealed that he identified specimens of Tonicella undocaerulea as juveniles of T. lineata, but remarked that the “juveniles” were found off shore in 15-20 fathoms (27-36 m) and “adults” were found only on the shore. Additionally Berry identified both “Tonicella rubra” (see below) and Tonicella venusta as ‘“Tonicella Animal length (mm) Fig. 38. Number of gills as a function of length (N = 15 specimens). Toni- cella lineata [¢]; T. undocaerulae [0]; T. lokii [©]; T. venusta [*]. Table 1. Characters of taxa in the Tonicella lineata species complex. T. undocaerulea _T. loki T. venusta T Taxa/Character r T. lineata Dorsal girdle ribbed smooth w/ smooth rotund w/ apical scales groove pleats Dimensions of 40x 18 20 x 10 22x12 30 x 18 scales (um) + Central tooth narrow/ broad/ narrow/ narrow/ of radula elongated spoon-shaped spoon-shaped | cupped head Major lateral rounded/w tridentate rounded/w tridentate teeth 1-2 notches 1 notch Brown lines yes no yes no on head valve + a gothic arch | zigzag z19zag z19Zag Head valve line pattern Flammules on yes pleural areas Post-mucronal +/- variable concave concave slope of tail valve straight rubra.” Berry’s specimens are preserved at the Santa Bar- bara Museum of Natural History. The identity of Tonicella lineata var. fusca (von Middendorff, 1847) is questionable, as the type is lost (Sirenko, pers. comm., 1997), however the name, fusca (from the Latin fuscus) means dark or dusky, and most like- ly refers here to the dark, delineated form of this species, which is not uncommon throughout most of it’s range, and which is particularly common from southeastern Alaska to Puget Sound, Washington. The type came from Sitka (Old Sitka). An example of this variation is illustrated in figure 33. This report brings the total number of species in the genus TJonicella in the northeastern region of the Pacific Ocean to six. These include: Tonicella lineata (Wood, 1815), Tonicella undocaerulea Sirenko, 1973, Tonicella lokii Clark, spec. nov., Tonicella venusta Clark, spec. nov., Tonicella insignis (Reeve, 1847) and Tonicella submar- morea (Middendorff, 1846) [here considered distinct from Tonicella marmorea (Fabricius, 1780) on the basis of valve and radula morphology]. I follow Yakovleva (1952) and Sirenko (1974) in considering Tonicella rubra (of authors) from the Pacific as 7. beringensis Yakovleva, 1952. Both of these species will soon be reclassified in another genus (Sirenko, pers. comm.). ACKNOWLEDGMENTS For their help and encouragement in various stages of this work, I am grateful to the following people: James H. McLean and Lindsey Groves (LACM); Paul H. Scott (SBMNH); Elizabeth Kools (CAS); Boris I. Sirenko (ZISP); Douglas J. Eernisse, California State University, Fullerton, California; Ian McTaggart Cowan, Victoria, British Columbia; George A. Hanselman, San Diego, California; George L. Kennedy (for- 46 AMER. MALAC. BULL. 15(1) (1999) merly LACM); Darlene Southworth, Southern Oregon University, Ash- land, Oregon; Hiroshi Saito, National Science Museum, Tokyo, Japan; Atsushi Naruse, Mizunami Fossil Museum, Mizunami City, Japan; Alan J. Murray, David Zwick and Kurt Morin, Ketchikan, Alaska; Robin C. Harrison National Marine Fisheries Service, Seattle, Washington. A speci- men and numerous photographs (figs. 35 & 36) were received through the kindness of Dr. Atsushi Naruse of the Mizunami Fossil Museum, Mizuna- mi City, Japan. I also Thank James H. McLean and Douglas J. Eernisse for critically reviewing the manuscript. The comments of three anony- mous reviewers were also very helpful. LITERATURE CITED Baxter, R. 1983. Mollusks of Alaska. China Poot Society Publ., Homer, Alaska. 96 pp. Baxter, R. 1987. Mollusks of Alaska. (revised ed.) Shells & Sea Life Publ., 163 pp. Berry, S. S. 1917. Notes on West American chitons. I. Proceedings of the California Academy of Sciences (4)7(10):229-248. Berry, S. S. 1922. Fossil chitons of Western North America. Proceedings of the California Academy of Sciences (4)11(18):399-526, pts. 1- 16. Burghardt, G. E. and L. E. Burghardt. 1969. A collectors Guide to West Coast Chitons. Special Publication No. 4, San Francisco Aquari- um Society. 45 pp., 4 color pls. Eernisse, D. J. 1984. Lepidochitona Gray, 1821 (Mollusca: Polyplacopho- ra), from the Pacific Coast of the United States: systematics and reproduction. Ph.D. Dissertation. University of California, Santa Cruz. 358 pp. Ferreira, A. J. 1982. The family Lepidochitonidae Iredale, (Mollusca: Polyplacophora) in the Northeastern Pacific. The Veliger 25(2): 93-138, figs. 8 pls. Itoigawa, J., M. Kuroda, A. Naruse, H. Nishimoto, T. Asada, T. Iwai and K. Hayashi. 1978. Polyplacophora assemblages from the Pleis- tocene formations of Ksarazu, Ichihara and their environs, Boso Peninsula, Japan. Mizunami Fossil Museum Bulletin, 5:143-155, pls. 14-16 [in Japanese, abstract in English]. Kaas, P. and R. A. Van Belle. 1985. Monograph of Living Chitons (Mol- lusca: Polyplacophora). Vol. 2, Suborder Ischnochitonina, Ischnochitonidae: Schizoplacinae, Callochitoninae and Lepido- chitoninae. E. J. Brill/Dr. W. Backhuys, Leiden. 198 pp. Middendorff, A. T. von. 1847. Beitrage zu einer Malacozoologica Rossi- ca. I. Chitonen. Mémemoires de Science Naturelle, Academie Imperiale des Sciences, St. Pétersburg 6:3-151; 14 pls. Oldroyd, I. S. 1927. The Marine Shells of the West Coast of North Ameri- ca. Stanford University Publications, University Series, Geologi- cal Sciences. Vol. 2, part 3:1-339 [602-941], pls. 73-108. Saito, H. 1994. The Shallow-water Chiton Fauna of Eastern Hokkaido, Japan. Memoirs of the National Science Museum, Tokyo, (27):93- 104. Sirenko, B. I. 1973. Amphipacific distribution of Chitons (Loricata) and their new species in the north-west section of the Pacific Ocean. Zoological Journal (Moscow, USSR) 52(5):659-667 [trans. & ed. by I. McT. Cowan and A. G. Smith, Of Sea and Shore 5 (2):59-63 (summer, 1974)]. Sirenko, B. I. 1974. Taxonomy of Chitons of the genus Tonicella Carpen- ter (Ischnochitonidae). Zoological Journal (Moscow, USSR) 53(7):988-997. Sirenko, B. I. 1993. Revision of the system of the order Chitonida (Mol- lusca: Polyplacophora) on the basis of correlation between the type of gills arrangement and the shape of the chorion processes. Ruthenica 3(2): 93-117. Smith, A. G. and M. Gordon, Jr. 1948. The marine Mollusks of Monterey Bay, California and vicinity. Proceedings of the California Acad- emy of Sciences (4)26(8): 147-245, pls. 3-4. Taki, I. 1938. Report of the biological survey of Mutsu Bay 31. Studies on chitons of Mutsu Bay with general discussion of chitons of Japan. Tohoku Imperial University, series 4, 12:323-423, pls. 14- 34, Thiele, J. 1909-1910. Revision des Systems der Chitonen. Zoologica Stuttgart 22:1-70, pls. 1-6 (1909):71-132, pls. 7-10 (1910). Wood, W. 1815. General Conchology; or, a description of shells arranged according to the Linnean system. London. i-xi, 1-7, i-iv, pls. 1-60. Yakovleva, A. M. 1952. Shell bearing Mollusks (Loricata) of the seas of the USSR. Keys to the fauna of the USSR, Academy of Sciences, USSR, no. 45, 127 pp. [English trans., Jerusalem, 1965]. Date of manuscript acceptance: 31 August 1998 Glycogen concentration in the mantle tissue of freshwater mussels (Bivalvia: Unionidae) during starvation and controlled feeding Matthew A. Patterson!, Bruce C. Parker!, and Richard J. Neves” 1 Department of Biology, Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24061, U.S. A. 2 Virginia Cooperative Fish and Wildlife Research Unit, Department of Fisheries and Wildlife Sciences, Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24061, U.S. A. Abstract: The effects of controlled feeding versus starvation during quarantine on mantle tissue glycogen concentration (represented as milligrams glycogen per gram of mantle tissue + SD) of two freshwater mussel species were compared. Starved individuals were not provided with supplemental food during quarantine, while fed specimens were provided with 10° algal cells/ml, twice per day. Initial mean glycogen levels for Amblema plicata (Say, 1817) (9.4 + 2.4 mg/g) and Quadrula pustulosa (1. Lea, 1831) (7.9 + 1.8 mg/g) collected from Ohio River Mile (ORM) 175.5 in July 1997 were not significantly different (p > 0.3) from mean glycogen levels of A. plicata (8.1 + 4.2 mg/g) and Q. pustulosa (6.2 + 2.9 mg/g) collected from the same site in July 1996. Initial glycogen concentrations of quarantined mussels, therefore, were similar in both the starved and fed groups. After seven days of feeding in quarantine, mean glycogen levels of A. plicata (12.3 + 2.3 mg/g) and Q. pustulosa (7.1 + 3.7 mg/g) did not change significantly (p > 0.1) relative to wild-caught speci- mens, and were significantly larger (p < 0.05) than mean glycogen levels of starved individuals (3.6 + 1.8 mg/g and 3.5 + 2.3 mg/g, respectively). Similarly, mean glycogen levels of A. plicata and Q. pustulosa after 14 days (8.1 + 3.3 mg/g and 7.7 + 3.3 mg/g, respectively) and 30 days (9.9 + 4.8 mg/g and 8.4 + 2.7 mg/g, respectively) of feeding were significantly larger (p < 0.01) than mean glycogen levels of starved specimens after 14 days (3.27 + 1.74 mg/g and 5.37 + 3.06 mg/g, respectively) and 30 days (1.2 + 0.5 mg/g and 1.9 + 1.4 mg/g, respectively). Adequate feeding of unionids in quarantine is essential to maintain animals in a condition that will increase the likelihood of survival following relocation. Key Words: glycogen, zebra mussels, Ohio River, quarantine, Dreissenidae Relocation has been used widely as a management relocation. tool to re-establish or assist in the recovery of declining Condition, as defined by Mann (1978: 490), is the populations of terrestrial and aquatic organisms (Griffith et “ability of an animal to withstand an adverse environmental al., 1989; Cope and Waller, 1995). In response to recent stress, be this physical, chemical or biological.” Glycogen declines in the unionid fauna of North America (Williams is the primary energy store in bivalves (Bayne, 1976; et al., 1993), relocation is being tested as a management Gabbott, 1983), and the relative amount of glycogen stored technique to (1) re-introduce extirpated populations, (2) in bivalve tissues is considered a good indicator of body remove unionids from project construction zones, (3) aug- condition (Galtsoff, 1964; Walne, 1970). Significant reduc- ment populations to increase genetic diversity, and (4) sal- tions in unionid glycogen stores prior to and during reloca- vage unionids from zebra mussel-infested areas (Cope and tion, therefore, could greatly reduce their ability to cope Waller, 1995). The ultimate goal of any relocation project is with natural stressors present in the new environment. to establish a self-sustaining population of the target organ- Collection, handling, aerial exposure, abrupt tem- ism, and several authors have developed a list of factors perature changes, holding, and transport of unionids during that should extend population persistence after relocation, relocation can result in varying degrees of stress, depending including (1) the presence of suitable habitat and refugia, on the length of exposure, that can in turn adversely affect (2) the release of large founder groups with high genetic body condition (Cope and Waller, 1995). Relocation of diversity, (3) low levels of competition, and (4) high rates unionids to avoid the adverse effects of bridge construction, of population growth (Griffith et al., 1989; Cope and for example, could have little or no effect on condition if Waller, 1995). Little attention, however, has been given to individual specimens are removed from their natural habitat the biochemical and physiological condition of an organism for short time periods (several hours or days). The presence prior to relocation and its influence on survival and the abil- of zebra mussels, however, has been shown to cause ity to develop reproductively viable populations after decreased glycogen reserves in unionids (Haag et al., American Malacological Bulletin, Vol. 15(1) (1999):47-50 47 48 AMER. MALAC. BULL. 15(1) (1999) 1993). In addition, zebra mussel-infested unionids presently require a 30-day quarantine period to ensure that zebra mussel adults and juveniles are removed prior to relocation (J. Clayton, pers. comm.). In addition to the stress of relo- cation, unionids can experience nutritive stress in quaran- tine because little or no information exists on their nutri- tional requirements (Gatenby et al., 1997). Under laborato- ry or hatchery conditions, bivalve energy stores have been shown to decline without proper feeding (Calvin, 1931; Pora et al., 1969; Bayne and Thompson, 1970; Gabbott and Walker, 1971), and recent studies reveal that unionid glyco- gen stores can decline as much as 80% after only 30 days of starvation in quarantine (Patterson et al., 1997). Consequently, the development of a feeding regime for mussels held in quarantine could be a critical link in the success of future relocation projects. The objectives of this experiment were to (1) monitor the glycogen levels of unionids during controlled feeding in quarantine, and (2) compare these results to reported changes in unionid glyco- gen levels during starvation in quarantine, as reported by Patterson et al. (1997). METHODS On 20 July 1997, ten specimens each of Amblema plicata (Say, 1817) and Quadrula pustulosa (I. Lea, 1831) were collected from Ohio River Mile (ORM) 175.5 near Parkersburg, West Virginia. Mussels were removed from the shell and placed in 95% ethanol for the determina- tion of initial glycogen levels. Additional specimens of A. plicata and Q. pustulosa (N = 200 and 80, respectively) were collected from ORM 175.5 for placement in individ- ual quarantine tanks. Unionids salvaged from zebra mussel- infested waters were thoroughly scrubbed to remove zebra mussels. Cleaned unionids were then hand-inspected before being placed in aerated quarantine tanks without substra- tum for a minimum of 30 days. During this 30-day period, water temperatures were maintained around 20°C to allow juvenile zebra mussels missed during the scrubbing proce- dure to become visible. At the end of 30 days, individual unionids were inspected under 10X magnification to ensure the absence of zebra mussels prior to relocation. Ten speci- mens of A. plicata and Q. pustulosa were sacrificed after seven, 14, and 30 days of quarantine, and preserved in 95% ethanol for subsequent glycogen analysis. After the experi- ment, all remaining specimens were returned to the Ohio River. In 1997, unionids were fed cultures of the chloro- phyte, Neochloris oleoabundans Chantanachat and Bold, 1962. This species was chosen because previous experi- ments indicated that juvenile unionids readily ingest and assimilate this alga (Gatenby et al., 1997). Initial stock cul- tures of N. oleoabundans were grown in Bold’s Basal Medium (Nichols, 1973) under continuous cool white fluo- rescent light (photon flux: 60-100 uE/m?/s) at 20°C. Stock cultures were then transferred to the quarantine facility and used to inoculate a 20 | carboy containing Fritz F2 algal medium (Fritz Aquaculture, Mesquite, Texas). When cell densities in the carboy reached 10° cells/ml, 7 1 aliquots were used to inoculate three 250 | algal culture tanks (Aquatic Ecosystems, Inc., Apopka, Florida), that also were fertilized with Fritz F2 algal medium, aerated, and placed outside the quarantine facility in direct sunlight. Culture tanks were placed outside because algal cultures continual- ly failed inside the quarantine facility, possibly due to insufficient light or high water temperatures. Algal cell densities in the 250 | culture tanks reached 10° cells/ml in ca. 4 days. Water samples from the culture tanks were col- lected daily and fixed with acid Lugol’s solution (Saraceni and Ruggiu, 1969) for enumeration and identification of the algae. Water samples from the Ohio River in 1996 and 1997 showed that algal cell densities ranged from 104 - 10°/cells/ml during the summer (Parker et al., 1998), thus unionids in quarantine were fed ca. 10° cells/ml twice per day, at 8:00 AM and 5:00 PM. Each day, 75% of the water in the quarantine tanks was drained; fresh water and food were added, and vigorous aeration was applied to maintain algal cells in suspension. The glycogen content of all preserved specimens was determined from mantle tissue using the technique of Keppler and Decker (1974) as described in Patterson et al. (1997). Mean glycogen levels were expressed in milligrams of glycogen per gram of preserved mantle tissue + standard deviation (SD). It should be noted that preserved tissue weights overestimate dry weights and underestimate wet tissue weights because 95% ethanol dehydrates tissue. Dehydration by 95% ethanol also reduces error that can result from changes in tissue water levels during stress. Mean glycogen levels were compared using ANOVA and, if significant differences were detected, the Scheffe F-test was used to determine the statistical significance of individ- ual treatments. RESULTS AND DISCUSSION During the first 14 d of quarantine, Neochloris oleoabundans comprised > 95% of the algae in the 250-1 culture containers. Because culture tanks were maintained outside the quarantine facility, cultures were contaminated with low densities of two green algae, Scenedesmus sp. and Ankistrodesmus sp. Once cultures were contaminated, den- sities of these contaminants continued to increase. After 30 d, the algal community in the culture containers had changed significantly, with Scenedesmus and PATTERSON ET AL.: GLYCOGEN LEVELS IN UNIONIDS 49 Ankistrodesmus comprising ca. 40% of the available algae, and N. oleoabundans comprising the remaining 60%. Despite changes in the algal community, cell densities remained at 10° cells/ml throughout the experiment. Both the fed and starved treatments had some intial mortality likely as a result of collection and handling, how- ever, no mortality occurred during the remainder of the quarantine period. Mean glycogen levels of Amblema plica- ta and Quadrula pustulosa during controlled feeding in quarantine remained the same, whereas mean glycogen lev- els declined in a previous starvation experiment by Patterson et al. (1997) (Fig. 1). Initial mean glycogen levels for A. plicata (9.4 + 2.4 mg/g) and Q. pustulosa (7.9 + 1.8 mg/g) collected from ORM 175.5 in July 1997 were not significantly different (p > 0.3) from the mean glycogen levels of A. plicata (8.1 + 4.2 mg/g) and Q. pustulosa (6.2 + 2.9 mg/g) collected from the same site in July 1996 (Patterson et al., 1997). Glycogen stores of unionids enter- ing quarantine, therefore, were similar in both the starvation and controlled feeding experiments. After seven days of feeding in quarantine, mean glycogen levels of A. plicata (12.3 + 2.3 mg/g) and Q. pustulosa (7.1 + 3.7 mg/g) did not change significantly (p > 0.1) relative to wild-caught specimens, and were significantly larger (p < 0.05) than mean glycogen levels of starved individuals (3.6 + 1.8 mg/g and 3.5 + 2.3 mg/g, respectively). Similarly, mean glycogen levels of A. plicata and Q. pustulosa after 14 days (8.1 + 3.3 mg/g and 7.7 + 3.3 mg/g, respectively) and 30 days (9.9 + 4.8 mg/g and 8.4 + 2.7 mg/g, respectively) of feeding were significantly larger (p < 0.01) than mean glycogen levels of starved specimens after 14 days (3.27 + 1.74 mg/g and 5.37 GE 1. plicata (starved) GEE) QO. pustulosa (starved) ME 4. plicata (fed) Q. pustulosa (fed) Glycogen (mg/g) o 1 REE eae 7 o Time (days) Fig. 1. Glycogen levels of Amblema plicata and Quadrula pustulosa at 0, 7, 14, and 30 d of starvation and controlled feeding in quarantine (N = 10/sampling period). + 3.06 mg/g, respectively) and 30 days (1.2 + 0.5 mg/g and 1.9 + 1.4 mg/g, respectively). Glycogen, the primary energy reserve in bivalves, drives many important physiological processes and can be used to endure short-term exposure to anoxia, emersion, or reduced food supplies (Bayne, 1976; Gabbott, 1983; Bayne et al., 1985; Hummel et al., 1988). Although exposure to anoxia and emersion can be limited if unionids are relocat- ed to suitable habitat, unionids will likely experience short- term, localized shifts in food abundance and long-term food shortages during the winter months. Normally, by accumu- lating glycogen when food is abundant, bivalves are able to withstand these food shortages (Gabbott, 1983; Hummel et al., 1988). However, unionid survival after relocation could be greatly reduced if glycogen stores are depleted in quar- antine and do not recover prior to the onset of winter. Regardless of effects on survival, decreased glycogen levels in adult bivalves also can have sublethal effects including reduced fecundity and reduced growth rates of developing offspring (Bayne, 1972; Helm et al., 1973; Bayne et al., 1975). Thus, the provision of adequate food resources for unionids in quarantine is pertinent to maintaining glycogen levels and enabling mussels to develop reproductively viable populations after relocation. Results from this study indicate that relatively high cell densities of the green alga, Neochloris oleoabundans, in combination with Scenedesmus and Ankistrodesmus, is an adequate food resource for the maintenance of unionid glycogen stores in the short term (30 d). Currently, no infor- mation exists on the ability of unionids to digest and assim- ilate Scenedesmus and Ankistrodesmus, but recent studies show that unionids assimilate Neochloris oleoabundans with relatively high efficiency (> 50%, M. Patterson, unpub. data). Additional studies to determine which algal species are readily digested and assimilated by unionids are critical, because some algal species might not be readily digested and assimilated by certain bivalve species (Peirson, 1983). Regardless of the digestibility of a particular algal food resource, large amounts of algae will be required if manage- ment agencies hope to relocate large numbers of native mussels away from zebra mussel-infested areas. In this study, the quarantine of 300 mussels required constant cul- ture of 750 1 of living algae. Discovering ways to shorten the quarantine period would make the production of algae more feasible and decrease the time that unionids must be maintained in captivity, outside their natural habitat. Studies dealing with more efficient methods of removing zebra mussels from the shells of unionids could prove to be the best avenue for reducing the quarantine period. Ultimately, a short quarantine period along with the provi- sion of food will improve the body condition of unionids and improve the success of future relocations. 50 AMER. MALAC. BULL. 15(1) (1999) ACKNOWLEDGMENTS We would like to thank Patty Morrison, Mitch Ellis, and every- one else at the Ohio River Islands National Wildlife Refuge Office, as well as Dr. Andrew Miller and the United States Army Corps of Engineers for their help in collecting mussels from the Ohio River. We would also like to thank the many volunteers that assisted with collection and process- ing of mussels in the field. Thanks also go to Li-Yen Chen for assistance in developing the glycogen assay and Catherine Gatenby for building the quarantine facility and assisting with field work, algae culture, and manu- script revisions. Finally, very special thanks go to Jim Dotson for ensur- ing that mussels received plenty of food and for his incredible ability to fix anything in quarantine. This study was funded by Quick Response Funds of the Biological Resources Division of USGS. LITERATURE CITED Bayne, B. L. 1972. Some effects of stress in the adult on the larval devel- opment of Mytilus edulis. Nature 237:459. Bayne, B. L. 1976. Aspects of reproduction in bivalve molluscs. Jn: Estuarine Processes, M. L. Wiley, ed. pp. 432-448. Academic Press, New York. Bayne, B. L., D. A. Brown, K. Burns, D. R. Dixon, A. Ivanovici, D. R. Livingstone, D. M. Lowe, and M. N. Moore. 1985. The Effects of Stress and Pollution in Marine Animals. Praeger, New York. 384 Pp. Bayne, B. L., P. A. Gabbott, and J. Widdows. 1975. Some effects of stress in the adult on the eggs and larvae of Mytilus edulis. Journal of the Marine Biological Association of the United Kingdom 55:675-689. Bayne, B. L. and R. J. Thompson. 1970. Some physiological conse- quences of keeping Mytilus edulis in the laboratory. Helgolander wissenschaftliche Meeresuntersuchungen 20:528-552. Calvin, D. B. 1931. Glycogen content of fresh-water mussels. Proceedings of the Society of Experimental Biology and Medicine 29:96-97. Cope, G. W. and D. L. Waller. 1995. Evaluation of freshwater mussel relocation as a conservation and management strategy. Regulated Rivers: Research and Management 11(2):147-155. Gabbott, P. A. 1983. Developmental and seasonal metabolic activities in marine molluscs. In: The Mollusca, Vol. 2, Environmental Biochemistry and Physiology, P. W. Hochachka, ed. pp. 165-217. Academic Press, New York. Gabbott, P. A. and A. J. M. Walker. 1971. Changes in the condition index and biochemical content of adult oysters (Ostrea edulis L.) main- tained under hatchery conditions. Journal du Conseil 34:99-106. Galtsoff, P. S. 1964. The American oyster (Crassostrea virginica). United States Fish and Wildlife Service Fisheries Bulletin 64, 480 PP. Gatenby, C. M., B. C. Parker, and R. J. Neves. 1997. Growth and survival of juvenile rainbow mussel, Villosa iris (Lea, 1829) (Bivalvia: Unionidae), reared on algal diets and sediment. American Malacological Bulletin 14(1):57-66. Griffith, B., J. M. Scott, J. W. Carpenter, and C. Reed. 1989. Translocation as a species conservation tool: status and strategy. Science 245:477-480. Haag, W. R., D. L. Berg, D. W. Garton, and J. L. Farris. 1993. Reduced survival and fitness in native bivalves in response to fouling by the introduced zebra mussel (Dreissena polymorpha) in western Lake Erie. Canadian Journal of Fisheries and Aquatic Sciences 50 (1):13-19. Helm, M. M., D. L. Holland, and R. R. Stephenson. 1973. The effect of supplementary algal feeding of a hatchery breeding stock of Ostrea edulis L. on larval vigor. Journal of the Marine Biological Association of the United Kingdom 53: 673-684. Hummel, H., L. de Wolf, and A. W. Fortuin. 1988. The annual cycle cf glycogen in estuarine benthic animals. Hydrobiological Bulletin 22:199-202. Keppler, D. and K. Decker. 1974. Glycogen determination with amyloglu- cosidase. In: Methods of Enzymatic Analysis, H. U. Bergmeyer, ed. pp. 11-17. Academic Press, New York. Mann, R. 1978. A comparison of morphometric, biochemical and physio- logical indexes of condition in marine bivalve molluscs. In: Energy and Environmental Stress in Aquatic Systems, J. H. Thorpe and J. W. Gibbons, eds. pp. 484-497. Technical Information Center, United States Department of Energy, Washington DC. Nichols, H. W. 1973. Growth media - freshwater. In: Handbook of Phycological Methods, J. R. Stein, ed. pp. 7-24. Cambridge University Press, New York. Parker, B. C., M. A. Patterson, and R. J. Neves. 1998. Feeding interactions between native freshwater mussels (Bivalvia: Unionidae) and zebra mussels (Dreissena polymorpha) in the Ohio River. American Malacological Bulletin 14(2):173-179. Patterson, M. A., B. C. Parker, and R. J. Neves.1997. Effects of quarantine times on glycogen levels of native freshwater mussels (Bivalvia: Unionidae) previously infested with zebra mussels. American Malacological Bulletin 14(1):75-79. Peirson, W. M. 1983. Utilization of eight algal species by the bay scallop, Argopecten irradians concentricus (Say). Journal of Experi- mental Marine Biology and Ecology 68:1-11. Pora, E. A., C. Wittenberger, G. Suarez, and N. Portilla. 1969. The resis- tance of Crassostrea rhizophorae to starvation and asphyxia. Marine Biology 3:18-23. Saraceni, C. and D. Ruggiu. 1969. Techniques for sampling water and phytoplankton. In: A Manual on Methods for Measuring Primary Production in Aquatic Environments, R. A. Vollenweider, ed. pp. 5-7. Blackwell Scientific Publications, Oxford. Walne, P. R. 1970. The seasonal variation of meat and glycogen content of seven populations of oysters Ostrea edulis L. and a review of the literature. Fisheries Investigation Series II 26, 35 pp. Williams, J. D., M. L. Warten, Jr., K. S. Cummings, J. L. Harris, and R. J. Neves. 1993. Conservation status of freshwater mussels of the Unites States and Canada. Fisheries 18 (9):6-22. Date of manuscript acceptance: 2 October 1998 Variation in glycogen concentrations within mantle and foot tissue in Amblema plicata plicata: Implications for tissue biopsy sampling Teresa J. Naimo! and Emy M. Monroe U.S. Geological Survey, Biological Resources Division, Upper Midwest Environmental Sciences Center, 2630 Fanta Reed Road, La Crosse, Wisconsin 54603 U. S. A. Abstract: With the development of techniques to non-lethally biopsy tissue from unionids, a new method is available to measure changes in biochemi- cal, contaminant, and genetic constituents in this imperiled faunal group. However, before its widespread application, information on the variability of bio- chemical components within and among tissues needs to be evaluated. We measured glycogen concentrations in foot and mantle tissue in Amblema plicata plicata (Say, 1817) to determine if glycogen was evenly distributed within and between tissues and to determine which tissue might be more responsive to the stress associated with relocating mussels. Glycogen was measured in two groups of mussels: those sampled from their native environment (undisturbed mussels) and quickly frozen for analysis and those relocated into an artificial pond (relocated mussels) for 24 months before analysis. In both undisturbed and relocated mussels, glycogen concentrations were evenly distributed within foot, but not within mantle tissue. In mantle tissue, concentrations of glyco- gen varied about 2-fold among sections. In addition, glycogen varied significantly between tissues in undisturbed mussels, but not in relocated mussels. Twenty-four months after relocation, glycogen concentrations had declined by 80% in mantle tissue and by 56% in foot tissue relative to the undisturbed mussels. These data indicate that representative biopsy samples can be obtained from foot tissue, but not mantle tissue. We hypothesize that mantle tissue could be more responsive to the stress of relocation due to its high metabolic activity associated with shell formation. Key Words: glycogen, tissue variation, Amblema plicata, biopsy, relocation With the development of techniques to obtain biop- sy samples from mantle (Berg et al., 1995; Byrne and Vesk, 1997) and foot tissue (Naimo et al., 1998) in unionids, a new tool has emerged to measure the biochemical, contami- nant, and genetic constituents in this imperiled faunal group. These techniques permit the measurement of indices of an organism’s physiological condition without adversely affecting survival. Recently, biopsy samples have been removed from four unionid species (Berg et al., 1995; Byrne and Vesk, 1997; Naimo et al., 1998) and survival of biopsied and non-biopsied mussels has been similar for up to 19 months following biopsy. Furthermore, because of their nondestructive nature, these techniques have potential for use on endangered, threatened, or otherwise sensitive populations. However, possible variation in concentration of a given constituent within a tissue type is unknown. Because only a small mass of biopsied tissue is taken (about 10 to 40 mg wet weight), the representativeness of this sample must be determined. 1Author to whom all correspondence should be addressed. E-mail: Teresa_Naimo @usgs.gov In the past few years, considerable effort has been invested into the relocation of unionids as a conservation strategy, principally in response to the threat of the exotic zebra mussel (Cummings et al., 1997). As a consequence, numerous researchers are developing methods for reloca- tion and for evaluating the success of relocations (Dunn and Layzer, 1997; Dunn and Sietman, 1997; Patterson et al., 1997). However, because of the long life span of mussels, traditional measures of condition, such as changes in shell length or tissue weight, are often inappropriate for short- term observations. Thus, the identification of sublethal indi- cators of stress in unionids could provide valuable data on physiological condition that precedes measurable changes in survival. We measured glycogen, because it is the principal storage form of carbohydrates in many aquatic invertebrates (Stetten and Stetten, 1960; De Zwaan and Zandee, 1972; Hummel et al., 1989) and it has been used as an indicator of the energetic status of individual mussels (Holopainen, 1987; Hemelraad et al., 1990). Alterations in glycogen con- centrations in unionids can be observed long before changes in either growth or survival are observed. For American Malacological Bulletin, Vol. 15(1) (1999):51-56 Dh 52 AMER. MALAC. BULL. 15(1) (1999) example, 3-month exposure to zebra mussels significantly reduced glycogen concentrations in Amblema plicata but did not adversely affect survival (Haag et al., 1993). Furthermore, measurement of glycogen concentration is a good indicator of nutritional stress in unionids during quar- antine periods before relocation (Patterson et al., 1997). Our goal was to determine the representativeness of biopsied tissue for the analysis of glycogen concentrations in foot and mantle tissue in Amblema plicata plicata (Say, 1817). Our objectives were (1) to determine if glycogen concentrations were evenly distributed within mantle and foot tissue; (2) to determine if glycogen concentrations var- ied between tissues; and (3) to qualitatively determine which tissue might be more responsive to the stress associ- ated with relocating mussels from their native environment into an artificial environment. METHODS TISSUE SAMPLING To estimate the distribution of glycogen in foot and mantle tissue in Amblema plicata plicata, we relocated individuals from Pool 9 of the Upper Mississippi River in May 1995, and placed them into mesh bags in a 0.04 hectare earthen pond. The pond was filled with well water (retention time of about 1 week) and was continuously aerated. No supplemental food was added to the pond dur- ing the study. After 24 months (June 1997), 5 individuals were removed, and mantle and foot tissue were dissected and subsequently frozen at -84°C until analysis for glyco- gen. To determine if the relocation altered the distribution of glycogen within and between tissues, we also deter- mined the distribution of glycogen in undisturbed Amblema plicata plicata that were removed directly from the Upper Mississippi River in May 1997, but not subjected to relocation. In both groups (relocated and undisturbed), we sac- rificed 5 individuals (shell length 80-90 mm) for determina- tion of glycogen. All unionids were free of zebra mussels. In each individual, the mantle and foot tissue were dissect- ed into three sections. In mantle, the three 5-mm sections extended ventral (section A) to dorsal (section C) and in foot tissue, the three 15-mm sections extended anterior (section A) to posterior (section C; Fig. 1). All foot sections were removed from the ventral margin to exclude reproduc- tive tissue from the sample. Within each section, we ran- domly removed five 5-mm2 samples, mean 10.1 + 0.1 mg wet weight in both foot and mantle tissue. Because glycogen is frequently reported on a wet- weight basis, spatial differences in water content within a given tissue may bias estimates of glycogen concentrations. To determine if water content varied within a tissue, we measured the percent water in each section of foot and Dorsal Ventral Posterior Fig. 1. Sample locations for foot and mantle tissue in Amblema plicata plicata. In the mantle, the three 5-mm sections extended ventral (section A) to dorsal (section C) and in the foot, the three 15-mm sections extended from anterior (section A) to posterior (section C). NAIMO AND MONROE: GLYCOGEN VARIATION IN AMBLEMA 55 (Hemminga et al., 1985). Similarly, in the marine mussel Mytilus edulis (Linnaeus, 1758), the hepatopancreas and the mantle have been identified as the primary storage organs for glycogen (Bayne, 1973a, b; Zaba et al., 1981). However, comparisons of glycogen concentrations between mantle tissue in freshwater and marine mussels should be made cautiously because mantle tissue in Mytilus is also the site for gonad development (Zaba et al., 1981). Although we observed elevated concentrations of glycogen in mantle tissue, relative to foot tissue, in undisturbed Amblema pli- cata plicata this does not necessarily suggest that mantle tissue is a storage site for glycogen. Because the mass of foot tissue likely exceeds that of mantle tissue, the mass of glycogen in the foot could easily surpass that in the mantle. The consistent reduction in glycogen concentrations in the relocated mussels, relative to the undisturbed mus- sels, suggests that the relocated mussels were not obtaining adequate nutritional requirements in this artificial environ- ment. For example, mean concentrations of chlorophyll a (+ 1 SEM) in the pond averaged 11 + 4 ug/L (n=11) during April through November, whereas concentrations in the river averaged 51 + 9 pg/L (n=15) during this same period. Furthermore, the magnitude of the reduction in glycogen was substantial--80% in mantle tissue and 56% in foot tis- sue. Similarly, Patterson et al. (1997) reported a 67% reduction in glycogen concentrations in mantle tissue in Amblema plicata from a site in the Ohio River heavily infested by zebra mussels (3,600 zebra mussels/m2) relative to glycogen concentrations in mussels from a lightly infest- ed site (0.3 zebra mussel/m2). These data suggest that the relocation of unionids into refugia could be unsuccessful if the new environment does not contain adequate nutritional resources. In addition, even if adequate food resources are available, mussels must be physiologically-capable of obtaining and processing the available food. Our criteria for selecting which tissue to biopsy for biochemical analysis were (1) within section variation in glycogen was minimal; (2) glycogen concentrations were similar across sections; and (3) it would be responsive to stress (in this case, relocation to an artificial pond). Using these criteria, the measurement of glycogen in foot tissue meets all three criteria. While the magnitude of the respon- siveness to stress was greater in mantle tissue than in foot tissue, glycogen concentrations were not evenly distributed and were highly variable in mantle tissue. However, this should not preclude the use of mantle tissue, it just suggests a need for consistency in the location of the biopsied sam- ple. In addition, the water content in a given section should be determined and used when reporting glycogen on a dry- weight basis in mantle tissue. If the mantle tissue is sam- pled, we recommend sampling the 5-mm section closest to the shell margin (section A) because it may be the most metabolically active and the easiest section to biopsy. Although this study was conducted on a few indi- viduals of Amblema plicata plicata, our preliminary data suggest that glycogen concentrations vary between and within certain tissues, and between relocated and undis- turbed mussels. However, the applicability of these data to other species at other times of the year remains unknown. Future studies need to examine the influence of confound- ing factors, such as reproductive status, seasonal trends, and differences between sexes, on the distribution and utiliza- tion of glycogen in several unionid species. ACKNOWLEDGMENTS Technical assistance in the field and laboratory was provided by Erika Damschen and Shari Greseth. Steve Gutreuter provided statistical guidance. We thank Thomas Dietz, Helen Kitchel, and James Layzer for helpful comments on an earlier version of the manuscript and two anony- mous reviewers for helpful comments on earlier drafts of the manuscript. LITERATURE CITED American Public Health Association, American Water Works Association, and Water Environment Federation (APHA). 1995. Standard Methods for the Examination of Water and Wastewater, 19th edition. American Public Health Association, Washington DC. 993 pp. Bayne, B. L. 1973a. Physiological changes in Mytilus edulis L. induced by temperature and nutritive stress. Journal of the Marine Biological Association of the United Kingdom 53:39-58. Bayne, B. L. 1973b. Aspects of the metabolism of Mytilus edulis during starvation. Netherlands Journal of Sea Research 7:399-410. Berg, D. J., W. R. Haag, S. I. Guttman, and J. B. Sickel. 1995. Mantle biopsy: A technique for nondestructive tissue-sampling of fresh- water mussels. Journal of the North American Benthological Society 14:577-581. Byme, M. and P. A. Vesk. 1997. Microanalysis of elements in granules in Hyridella depressa (Bivalvia): Multivariate analysis and biomoni- toring potential. Australasian Journal of Ecotoxicology 2:91-97. Cummings, K. S., A. C. Buchanan, C. A. Mayer, and T. J. Naimo. 1997. Conservation and Management of Freshwater Mussels IT: Initiatives for the Future. Proceedings of an Upper Mississippi River Conservation Committee symposium, October 1995, St. Louis, Missouri. Upper Mississippi River Conservation Committee, Rock Island, Illinois. 293 pp. De Zwaan, A. and D. I. Zandee. 1972. The utilization of glycogen and accumulation of some intermediates during anaerobiosis in Mytilus edulis L. Comparative Biochemistry and Physiology 43B:47-54. Dunn, C. S. and J. B. Layzer. 1997. Evaluation of various holding facili- ties for maintaining freshwater mussels in captivity. In: Conservation and Management of Freshwater Mussels II: Initiatives for the Future, K. S. Cummings, A. C. Buchanan, C. A. Mayer, and T. J. Naimo, eds. pp. 205-213. Proceedings of an Upper Mississippi River Conservation Committee symposium, October 1995, St. Louis, Missouri. Upper Mississippi River 56 AMER. MALAC. BULL. 15(1) (1999) Conservation Committee, Rock Island, Illinois. Dunn, H. L. and B. E. Sietman. 1997. Guidelines used in four geographi- cally diverse unionid relocations. In: Conservation and Management of Freshwater Mussels II: Initiatives for the Future, K. S. Cummings, A. C. Buchanan, C. A. Mayer, and T. J. Naimo, eds. pp. 176-183. Proceedings of an Upper Mississippi River Conservation Committee symposium, October 1995, St. Louis, Missouri. Upper Mississippi River Conservation Committee, Rock Island, Illinois. Haag, W. R., D. J. Berg, D. W. Garton, and J. L. Farris. 1993. Reduced survival and fitness in native bivalves in response to fouling by the introduced zebra mussel (Dreissena polymorpha) in Western Lake Erie. Canadian Journal of Fisheries and Aquatic Sciences 50:13-19. Hemelraad, J., D. A. Holwerda, H. J. Herwig, and D. I. Zandee. 1990. Effects of cadmium in freshwater clams. III. Interaction with ener- gy metabolism in Anodonta cygnea. Archives of Environmental Contamination and Toxicology 19:699-703. Hemminga, M. A., J. J. Maaskant, J. C. Jager, and J. Joosse. 1985. Glycogen metabolism in isolated glycogen cells of the freshwater snail Lymnaea stagnalis. Comparative Biochemistry and Physiology 82A:239-246. Holopainen, I. J. 1987. Seasonal variation of survival time in anoxic water and the glycogen content of Sphaerium corneum and Pisidium amnicum (Bivalvia, Pisidiidae). American Malacological Bulletin 5:41-48. Hummel, H., L. De Wolf, W. Zurburg, L. Apon, R. H. Bogaards, and M. Van Ruitenburg. 1989. The glycogen content in stressed marine bivalves: The initial absence of a decrease. Comparative Biochemistry and Physiology 94B:729-733. Littell, R. C., G. A. Milliken, W. W. Stroup, and R. D. Wolfinger. 1996. SAS System for Mixed Models. SAS Institute Inc., Cary, North Carolina. 633 pp. Naimo, T. J., E. D. Damschen, R. G. Rada, and E. M. Monroe. 1998. Nonlethal evaluation of the physiological health of unionid mus- sels: Methods for biopsy and glycogen analysis. Journal of the North American Benthological Society 17:121-128. Patterson, M. A., B. C. Parker, and R. J. Neves. 1997. Effects of quaran- tine times on glycogen levels of native freshwater mussels (Bivalvia: Unionidae) previously infested with zebra mussels. American Malacological Bulletin 14:75-79. Petit, H., W. L. Davis, and R. G. Jones. 1978. Morphological studies on the mantle of the fresh-water mussel Amblema (Unionidae): Scanning electron microscopy. Tissue and Cell 10:619-627. SAS Institute. 1989. SAS/STAT User’s Guide, Version 6, Fourth Edition, Volume 1. SAS Institute Inc., Cary, North Carolina. 943 pp. Stetten, D. and M. R. Stetten, 1960. Glycogen metabolism. Physiological Review 40:505-537. Wilbur, K. M. and A. S. M. Saleuddin. 1983. Shell formation. In: The Mollusca, Volume 4, Physiology Part I, A. S. M. Saleuddin and K. M. Wilbur, eds. pp. 235-287. Academic Press, New York. Zaba, B. N., P. A. Gabbott, and J. I. Davies. 1981. Seasonal changes in the utilisation of !4C- and 3H-labelled glucose in a mantle tissue slice preparation of Mytilus edulis L. Comparative Biochemistry and Physiology 70B:689-695. Date of manuscript acceptance: 15 April 1999 Recruitment in a freshwater unionid (Mollusca: Bivalvia) community downstream of Cave Run Lake in the Licking River, Kentucky Stephen E. McMurray!, Guenter A. Schuster, and Barbara A. Ramey Department of Biological Sciences, Eastern Kentucky University, Richmond, Kentucky 40475 U.S. A. Abstract: Unionids, fish, and glochidia were collected to determine why recruitment had ceased or had been dramatically decreased in a speciose union- id community in the Licking River at Moores Ferry, Kentucky, 35.4 km downstream of Cave Run Lake. Only six unionid glochidia were collected with drift nets, and only six fish collected had infestations of glochidia. A small percentage (10.1%) of the unionids observed had their gills modified as marsupia. An analysis of water temperature and discharge indicated no significant difference in average monthly discharge (p > 0.05) and a significant decrease in temper- ature (p < 0.05) between pre- and post-impoundment periods. Average monthly discharge and temperature may not be as biologically important as the spikes of discharge and corresponding sudden decreases in temperature that are caused by releases of hypolimnionic water from the reservoir. Key Words: Unionidae, recruitment, gametogenesis, impoundments, Kentucky North America has the richest unionoidean (mussel) (Bivalvia: Unionidae) fauna in the world. This fauna has disproportionately more endangered, threatened, and spe- cial concern taxa than all the groups of terrestrial organ- isms. Only 70 of the 297 taxa known from the United States are considered stable (Williams et al., 1993). Of the freshwater unionid taxa recognized from the United States and Canada, 35% (103 taxa) are known to occur, or have occurred in Kentucky, which ranks this state third in faunal richness behind Tennessee and Alabama (Cicerello et al., 1991). Human activities in the Commonwealth have severe- ly impacted unionid populations during the last 200 years, making this group of organisms the most endangered in the state (Cicerello et al., 1991). Freshwater mussels are impacted by anthropogenic factors such as the additions of toxic substances into aquat- ic systems, sedimentation, habitat destruction, loss of their host fish(es), introduced species, and commercial harvest- ing (van der Schalie and van der Schalie, 1950; Fuller, 1980; Bogan, 1993; Williams et al., 1993). The construc- tion of dams along the course of several rivers in North America has resulted not only in the loss of taxa, but also of whole beds (e. g., van der Schalie and van der Schalie, 1950; Bates, 1962; Miller et al., 1984; Miller et al., 1992; Williams et al., 1992; Layzer et al., 1993; Sickel and 1Present Address: Kentucky Division of Water, Water Quality Branch, 14 Reilly Road, Frankfort, Kentucky 40601 U.S. A. Chandler, 1996). One of the most severe and perplexing problems facing freshwater mussels is the documented loss of recent recruitment (reproduction) in unionid communi- ties that were previously thought to be healthy. Recent research hypothesized that recruitment had ceased or had been dramatically decreased in a diverse unionid community in the Licking River at Moores Ferry, Kentucky (Kane, 1990; Smathers, 1990). The present study was an attempt to determine if recruitment had ceased or had been decreased in this diverse unionid community, and if so, to determine where in the life cycle reproduction was breaking down. This study was also an attempt to deter- mine what effects, if any, the hypolimnionic discharges from Cave Run Lake were having on the mussels in the bed at Moores Ferry. It was hypothesized that the loss of recruitment in this bed was either directly or indirectly the result of the dam, located approximately 35.4 km upstream, which has altered both the natural temperature and flow regimes of the river. METHODOLOGY STUDY AREA The Licking River is a sixth order tributary to the Ohio River, originating on the Unglaciated Allegheny Plateau in the Appalachian Province of eastern Kentucky. The river flows through an extremely variable topography in a northwesterly direction through the Blue Grass region American Malacological Bulletin, Vol. 15(1) (1999):57-63 ay, 58 AMER. MALAC. BULL. 15(1) (1999) of the Commonwealth for 496 km until its confluence with the Ohio River near Covington, Kentucky, at Ohio River km 757.2 (Harker et al., 1979, Hannan, et al., 1982, Burr and Warren, 1986). This drainage encompasses approxi- mately 10% of the Commonwealth (9601 km2) and covers all or a portion of 21 counties (Harker et al., 1979). In 1974 the Licking River was impounded to form Cave Run Lake, a warm oligotrophic lake (Clinger, 1974). The lake has a surface area of 3347 ha that inundates 61 km of the mainstem and the lower reaches of several small trib- utaries (Burr and Warren, 1986). The Licking River drainage has a speciose unionid fauna with 53 taxa, over half of the state’s native mussel fauna, historically existing in the drainage (Cicerello et al., 1991). A recent survey below the reservoir indicated that 50 taxa still reside in that portion of the river (Laudermilk, 1993). Moores Ferry, Kentucky, is a ford approximately 35.4 km downstream from Cave Run Lake (Fig. 1). The watershed at this site is utilized mainly for agriculture. Discharge at this site is affected by releases of water from Cave Run Lake, which at times causes drastic water level fluctuations. Substrata consisted mainly of cobble, gravel, rubble, and some boulders with intermixed sandy areas (Smathers, 1990). The unionid community at this site had a rich assemblage of unionid species, with 35 known present or historical taxa (Smathers, 1990; Laudermilk, 1993). SAMPLING AND LABORATORY PROCEDURES Five collections of glochidia, fish, and unionids were made from July through October 1995. High water conditions prevented collections during Spring 1996. For each collection period a drift net was haphazardly placed in the bed to collect glochidia. After one hour, the contents of the drift net were removed and preserved in 70% ethanol and returned to the laboratory. Drift net collections were examined using cross-polarized light microscopy (Johnson, 1995), with a portion of the sample being delivered into a watch-glass with a gridded bottom. Each square in the grid was then systematically searched under cross-polarized light (10-20X) for glochidia, which were counted and removed along with any juvenile Corbicula fluminea (Miller, 1774). Fish were collected for one hour using a minnow seine. All fish retained were initially preserved in a 10% formalin solution and then transferred to 70% ethanol in the laboratory for final preservation. Following sorting and identification, the fins and scales of each individual were examined under a dissecting microscope (10-30X) for attached glochidia. The opercular flaps were removed, and each gill arch was carefully examined under a dissecting microscope for attached glochidia (Bruenderman and Neves, 1993). Unionids were collected by hand for one hour by snorkeling, or by wading with the use of water scopes. After identification, the shell of each unionid was carefully opened with a small screwdriver, and the gills examined for signs of gravidity. The species name was recorded and notes were made on the condition of the gills. Except for individuals of two target species retained for histological examination, all unionids were returned to the river. Three to five individuals of the two most common species in the bed at Moores Ferry (Smathers, 1990), Actinonaias liga- mentina (Lamarck, 1819) and Elliptio dilatata (Rafinesque, 1820), were chosen for histological examination from each collecting period. These species represented both of the breeding regimes of freshwater mussels, they were both commonly encountered throughout their respective ranges (Oesch, 1995), and neither had any federal or state protec- tion status in Kentucky (Kentucky State Nature Preserves Commission [KSNPC], 1996). Individuals were prepared for histological examina- tion by placing them into a 10% formalin solution, then transferring them to 70% ethanol in the laboratory. The valves were opened by cutting the adductor muscles and portions of the gonadal and gill tissues were removed and placed into either 70% ethanol or Bouins fixative. These were then dehydrated through a series of alcohols and embedded in paraffin (Humason, 1967). Sections were made at a thickness of 10 um using an American Optical 820 Microtome. Slides were stained with Ehrlich’s hema- toxylin and eosin was used as the counterstain (Humason, 1967). Sections were mounted with Permount. The sections were then examined under a compound microscope (400- 430X) to determine a sex ratio for both species; to deter- mine if gametogenesis was occurring, and, if so, to try to quantify it; and to determine the contents of the marsupia. All drift net, fish, and unionid collections were deposited in the Branley A. Branson Museum of Zoology, Eastern Kentucky University (EKU). Five cell types of spermatogenesis (Garner, 1993) were used to determine the stage of gametogenesis in males. Stage 1 males were those that had only spermatogo- nia present in their acini, and Stage 5 males had mature spermatozoa present. Stages 2, 3, and 4 corresponded respectively to sperm morulae, spermatocytes, and sper- matids being present in the acini. Three cell types of ooge- nesis (McMurray, 1997) were used to determine the stage of gametogenesis in female Elliptio dilatata. Stage 1 females were those with oogonia as the dominant cell type in their alveoli, Stage 2 were those with oocytes dominant, and Stage 3 were those with mature ova dominant. Marsupia were classified according to their contents as being empty (EM), or containing mature glochidia (MG), early embryos (EE), or advanced embryos (AE) (Garner, 1993). In the case of known females that did not have their gill tissues examined, the marsupia were consid- McMURRAY ET AL.: IMPOUNDMENTS AND FRESHWATER MUSSEL RECRUITMENT 59 Ohio River Licking River Moores Ferry Cave Run Lake Fig. 1. Location of the Moores Ferry study site. Inset shows the location of the Licking River drainage in Kentucky. ered to be empty since sections were made of any gill that showed signs of gravidity. Marsupia with mature glochidia were those that had glochidial shells and the single adductor muscle visible in the section. Temperature and discharge data were obtained from the United Stages Geological Survey (USGS) for the Licking River gauging station at Farmers, Kentucky, which is just upstream of Moores Ferry. Data were obtained for a pre-impoundment (1966-1970) and a post-impoundment (1991-1994) period. Daily temperature and discharge data from the months of May through September for each of these years, with the exception of 1994 (May-June), were analyzed to determine if any correlation existed between discharge and water temperature. RESULTS AND DISCUSSION A total of only six unionid glochidia were collected using drift nets, compared to 185 juvenile Corbicula flu- minea. The large number of juvenile C. fluminea was large- ly the result of a single collection of 127 individuals. Drift net collections were made between 1000 and 1700 hours (EST), which corresponds to the period when glochidial densities should have been at their highest (Kitchell, 1985). The high number of C. fluminea present may impact juve- nile unionids through resource competition (Neves and Widlak, 1987). Only 11.3% of the 399 fish collected were suitable hosts for unionids in the bed (Watters, 1994). Of the fish collected only six specimens had attached glochidia: one Micropterus punctulatus (Rafinesque, 1819); three Percina copelandi (Jordan, 1877); one P. evides (Jordan and Copeland, 1877); and one P. oxyrhyncha (Hubbs and Raney, 1939). Even though a small fish may successfully carry sev- eral hundred glochidia (Lefevre and Curtis, 1910), low infestation rates appear to be common (Kitchell, 1985; Neves and Widlak, 1988; Bruenderman and Neves, 1993; Weiss and Layzer, 1995). The attachment of glochidia to their hosts is dependent upon several factors such as infes- tations of hosts by copepod parasites (Wilson, 1916), age of the host, immunity caused by previous infestations (Parker et al., 1984), and water temperature (Matteson, 1948). A total of 288 unionids were collected from Moores Ferry (Table 1). Of these, only 10.1% (29 individuals) had their gills modified as marsupia, indicating an 8.93:1 male to female ratio if modified gills are taken to represent females. Marsupial condition is probably not a true repre- sentation of the actual male to female ratios, but most unionids are not sexually dimorphic (McMahon, 1991), and the only way to determine the sex of an individual without using histological techniques is to examine the gills for signs of gravidity. However, it would still be expected that at least 50% of the individuals collected would have had 60 Table 1. Unionids with and without gills modified as marsupia from the Licking River at Moores Ferry, Kentucky. With Without Modified Modified Taxa Gills Gills Totals Actinonaias ligamentina (Lamarck, 1819) 20 135 155 Alasmidonta marginata Say, 1818 0 0 0 Amblema plicata (Say, 1817) 0 45 45 Cyclonaias tuberculata (Rafinesque, 1820) 0 0 0 Elliptio dilatata (Rafinesque, 1820) 0 18 18 Fusconaia flava (Rafinesque, 1820) 0 2 2 Lampsilis cardium Rafinesque, 1820 2 6 8 Lasmigona complanata (Barnes, 1823) 0 0 0 L. costata (Rafinesque, 1820) 0 6 6 Leptodea fragilis (Rafinesque, 1820) 0 1 1 Megalonaias nervosa (Rafinesque, 1820) 0 9 9 Obliquaria reflexa Rafinesque, 1820 0 0 0 Potamilus alatus (Say, 1817) 1 2 3 Ptychobranchus fasciolaris (Rafinesque, 1820) 6 19 25 Quadrula metanevra (Rafinesque, 1820) 0 0 0 Q. nodulata (Rafinesque, 1820) 0 0 0 Q. pustulosa (Lea, 1831) 0 1 1 Q. quadrula (Rafinesque, 1820) 0 7 7 Tritogonia verrucosa (Rafinesque, 1820) 0 8 8 Totals: 29 259 288 their gills modified as marsupia if normal reproduction were occurring. Histological examination revealed that the male to female ratios for the two target species were not statistically different from 1:1. Both of the target species, as well as other species of unionids, usually maintain a 1:1 male to female ratio (Jirka and Neves, 1992). Most of the males of the two target species had more than one stage of spermatogenesis present in their gonads, but usually the most advanced stage present domi- nated the acini of the testes (Table 2). Spermatids were observed in most of the male Elliptio dilatata collected, but all the cell types except spermatozoa were also observed. All the cell types except spermatocytes were observed in male Actinonaias ligamentina, with sperm morulae and spermatids the most common. All of the female Elliptio dilatata had each of the three egg cell types present in their ovaries (Table 2). The most advanced stage present did not always dominate the AMER. MALAC. BULL. 15(1) (1999) alveoli of the ovaries, as was observed in the testes of the males. Most of the FE. dilatata were in the second stage of oogenesis with oocytes dominating the alveoli. One female was categorized as unknown because the stage of oogenesis could not be determined due to technical difficulties. Hermaphroditism, which has been reported for both of the target species (van der Schalie, 1970; Jirka and Neves, 1992), was observed in one Actinonaias ligamentina indi- vidual. Most of the acini in this individual contained male gametes, and those with female gametes seemed to be degenerate. Even though the male to female ratios of the target species were not significantly different from 1:1, a majority of the females had empty marsupia (Table 2). The post-impoundment water temperatures at Farmers, Kentucky, which is located between Moores Ferry and Cave Run Lake, were found to be significantly lower than pre-impoundment temperatures for the months of May through July. There was no significant change in water temperature for August, and a significant increase for September (Table 3). A change in the temperature regime of an aquatic system, especially a decrease, can have drastic effects on freshwater mussels. Water temperature influ- ences the length of the period of development, the time of fertilization (Matteson, 1948), the release of glochidia (Yokley, 1972; Jirka and Neves, 1992), the survival of glochidia before they attach to a host (Tedla and Fernando, 1969), and it is the major environmental factor that regu- lates the period of glochidial attachment (Matteson, 1948; Zale and Neves, 1982; Neves and Widlak, 1988). Decreases in water temperature can also result in a decreased production of the crystalline style (Allen, 1921), reduce the motility of adult unionids (Yokley, 1972) and influence the growth of adults (Hruska, 1992). The average monthly discharge at Farmers, Kentucky, increased from pre- to post-impoundment for the months of June through September and decreased during May, but none of these changes were statistically signifi- cant (Table 3). Even though these changes were themselves not statistically significant, they were directly related to sig- nificant changes in water temperature. The hydrological variability of a river influences the spatial distribution of individuals in a freshwater mussel community (Di Maio and Corkum, 1995), and stream velocity is a major factor Table 2. Stages of gametogenesis and marsupial contents of Actinonaias ligamentina and Elliptio dilatata from the Licking River at Moores Ferry, Kentucky. See text for stage and content descriptions. Stage of Spermatogenesis Taxa 1 2 3 4 Actinonaias ligamentina 1 4 0 4 Elliptio dilatata 1 2 3 7 Totals: 2 6 3 «ll 5 1 0 Marsupial Contents Stage of Oogenesis EE AE MG EM 1 2 3 UNK Ja eee 2 0°93. 12 24 1 ~«1 0 2 tecl 5 2.4: tN 2 2 2 AP AT McMURRAY ET AL.: IMPOUNDMENTS AND FRESHWATER MUSSEL RECRUITMENT 61 Table 3. Mean monthly water temperature and discharge for the Licking River at Farmers, Kentucky, during pre- (1966-1979) and post-impoundment (1991-1994) periods. Water Temperature (°C) Discharge (m3/s) Month 1966-79 1991-94 — Students-r 1966-79 1991-94 Students-r May 16.3 13.4 3.0291* 48.1 27.1 0.8716 June 22.1 17.0 3.4908* 5.6 14.7 -1.2913 July 24.5 21.3 2.0554* 4.0 8.3 -1.4505 August 23.0 22.9 0.1184 7.6 9.0 -0.3339 September 20.7 23.0 -3.1114* 3.4 5.0 -1.0613 *Statistically different at a = .0S determining the presence or absence of different species within a freshwater community (Horne and McIntosh, 1979). During periods of high discharge the dilution of sperm and glochidia may occur (Miller and Payne, 1988), and continuous high-turbulence flow can impact adult unionids by eroding the periostracum (Miller et al., 1984). Even though there were no changes in discharge, and significant decreases in temperature (p < 0.05) only during the months of May through July, the impacts of these changes were still biologically important. It is believed that the average monthly discharge and tempera- ture may not be as important as the spikes of discharge, and the corresponding sudden decreases in temperature, that are caused by the releases of hypolimnionic water from Cave (s/,Ul) aBseyostq Water Temperature (°C) Water Temperature (°C) (S/W) adseYyosIG —«— Temperature Fig. 2. Licking River at Farmers, Kentucky, mean water temperature (°C) and daily mean discharge (m3/s) during July 1966 and July 1991. Run Lake during important reproductive periods. For example, during July 1966, before the comple- tion of Cave Run Lake, water temperature averaged 26.4 + 0.85°C and the discharge averaged 3.2 + 3.2 m3/s (Fig. 2). During July 1991, after the construction of the lake, the average water temperature was 18.5 + 2.82°C and the aver- age discharge was 16.8 + 17.5 m3/s. There were large spikes of discharge during both of these months. During July 1966 discharge increased and then returned to normal seven days later, resulting in only a 3°C decrease in temper- ature that did not last long after the increased discharge had subsided. During July 1991 the spike in discharge returned to normal nine days later, and resulted in a 6.5°C decrease in temperature that did not return to its pre-release level until four days later. This single spike of discharge impact- ed the normal water temperature for a period of 13 days. Sudden, drastic, temperature changes such as these have been reported to cause the abortion of embryos and glochidia (Matteson, 1948; Bruenderman and Neves, 1993). ACKNOWLEDGMENTS The authors would like to acknowledge the assistance of C. Abbruzzese, J. S. Board, M. C. Compton, M. D. Moeykens, A. R. T. Nix, T. E. Oliver, M. A. Patterson, and D. Vey in the field and laboratory. R. R. Cicerello (KSNPC) and P. A. Ceas (EKU) assisted with fish identifica- tion. G. T. Watters (OSU) provided helpful tips on the use of cross-polar- ized light. We would also like to thank D. L. Batch (EKU) for serving on the primary author’s thesis committee. D. L. McClain, Water Resources Division, USGS, Louisville, provided the temperature and discharge data for the Licking River. Two anonymous reviewers are thanked for their comments on the manuscript. This research was funded by a grant from the Kentucky Department of Fish and Wildlife Resources (Project No. E-2-9). LITERATURE CITED Allen, W. R. 1921. Studies of the biology of freshwater mussels. Experimental studies of the food relations of certain Unionidae. Biological Bulletin 40:210-241. 62 AMER. MALAC. BULL. 15(1) (1999) Bates, J. M. 1962. The impact of impoundment on the mussel fauna of Kentucky Reservoir, Tennessee River. American Midland Naturalist 68(1):233-236. Bogan, A. E. 1993. Freshwater bivalve extinctions (Mollusca: Unionoida): a search for causes. American Zoologist 33:599-609. Bruenderman, S. A. and R. J. Neves. 1993. Life history of the endangered fine-rayed pigtoe Fusconaia cuneolus (Bivalvia: Unionidae) in the Clinch River, Virginia. American Malacological Bulletin 10(1):83-91. Burr, B. M. and M. L. Warren, Jr. 1986. A Distributional Atlas of Kentucky Fishes. Kentucky Nature Preserves Commission Scientific and Technical Series Number 4, Frankfort. 398 pp. Cicerello, R. R., M. L. Warren, Jr., and G. A. Schuster. 1991. A distribu- tional checklist of the freshwater unionids (Bivalvia: Unionoidea) of Kentucky. American Malacological Bulletin 8(2):113-129. Clinger, C. G. 1974. A pre-impoundment study of Cave Run Lake. Master’s Thesis, Eastern Kentucky University, Richmond, Kentucky. 43 pp. Di Maio, J. and L. D. Corkum. 1995. Relationship between the spatial dis- tribution of freshwater mussels (Bivalvia: Unionidae) and the hydrological variability of rivers. Canadian Journal of Zoology 73:663-671. Fuller, S. L. H. 1980. Historical and current distributions of fresh-water mussels (Mollusca: Bivalvia: Unionidae) in the Upper Mississippi River. In: Proceedings of the UMRCC Symposium on Upper Mississippi Bivalve Mollusks, J. Rasmussen, ed. pp. 72-119. Upper Mississippi River Conservation Committee, Rock Island, Mlinois. Garner, J. T. 1993. Reproductive cycle of Quadrula metanevra (Unionidae) in Kentucky Reservoir, Tennessee River, Hardin Co., Tennessee. Master’s Thesis, University of Alabama, Huntsville. 56 pp. Hannan, R. R., M. L. Warren, Jr., K. E. Camburn, and R. R. Cicerello. 1982. Recommendations for Kentucky’s Outstanding Water Classifications with Water Quality Criteria for Protection. Kentucky State Nature Preserves Commission Technical Report, Frankfort. 459 pp. Harker, D. F., S. M. Call, M. L. Warren, Jr., K. E. Camburn, and P. Wigley. 1979. Aquatic Biota and Water Quality Survey of the Appalachian Province, Eastern Kentucky. Kentucky Nature Preserves Commission Technical Report, Frankfort. 3 Volumes. 1152 pp. Horne, F. R. and S. McIntosh. 1979. Factors influencing the distribution of mussels on the Blanco River of central Texas. Nautilus 94(4):119- 133. Hruska, J. 1992. The freshwater pearl mussel in South Bohemia: Evaluation of the effect of temperature on reproduction, growth and age structure of the population. Archives Hydrobiologia 126(2):181-191. Humason, G. L. 1967. Animal Tissue Techniques. Second edition. W. H. Freeman and Company, San Francisco. 569 pp. Jirka, K. J. and R. J. Neves. 1992. Reproductive biology of four species of freshwater mussels (Mollusca: Unionidae) in the New River, Virginia and West Virginia. Journal of Freshwater Ecology 7(1):35-44. Johnson, L. E. 1995. Enhanced early detection and enumeration of zebra mussel (Dreissena spp.) veligers using cross-polarized light microscopy. Hydrobiologia 312:139-146. Kane, M. A. 1990. The predation of freshwater bivalves by muskrats (Ondatra zibethicus) in the Licking River, at Moores Ferry, Kentucky. Master’s Thesis, Eastern Kentucky University, Richmond, Kentucky. 71 pp. Kentucky State Nature Preserves Commission (KSNPC). 1996. Rare and extirpated plants and animals of Kentucky. Transactions of the Kentucky Academy of Science 57(2):69-91. Kitchell, H. E. 1985. Life history of the endangered shiny pigtoe pearly mussel Fusconaia edgariana in the North Fork Holston River, Virginia. Master’s Thesis, Virginia Polytechnic Institute and State University, Blacksburg. 118 pp. Laudermilk, E. L. 1993. A survey of the unionids (Bivalvia: Unionidae) of the mainstem Licking River and selected tributaries below Cave Run Reservoir, Kentucky. Master’s Thesis, Eastern Kentucky University, Richmond, Kentucky. 71 pp. Layzer, J. B., M. E. Gordon, and R. M. Anderson. 1993. Mussels: the for- gotten fauna of regulated rivers. A case study of the Caney Fork River. Regulated Rivers: Research and Management 8:63-71. Lefevre, G. and W. C. Curtis. 1910. Experiments in the artificial propaga- tion of fresh-water mussels. Bulletin of the United States Bureau of Fisheries [Doc. 671] 28:615-626. Matteson, M. R. 1948. Life history of Elliptio complanatus (Dillwyn, 1817). American Midland Naturalist 40(3):690-723. McMahon, R. F. 1991. Mollusca: Bivalvia. In: Ecology and Classification of North American Freshwater Invertebrates, J. H. Thorp and A. P. Covich, eds. pp. 315-399. Academic Press, San Diego. McMunmray, S. E. 1997. Reproduction in a freshwater mussel (Bivalvia: Unionidae) community downstream of Cave Run Reservoir in the Licking River at Moores Ferry, Kentucky. Master’s Thesis, Eastern Kentucky University, Richmond, Kentucky. 119 pp. Miller, A. C. and B. S. Payne. 1988. The need of quantitative sampling to characterize size demography and density of freshwater mussel communities. American Malacological Bulletin 6(1):49-54. Miller, A. C., B. S. Payne, and R. Tippit. 1992. Characterization of a freshwater mussel (Unionidae) community immediately downriv- er of Kentucky Lock and Dam in the Tennessee River. Transactions of the Kentucky Academy of Science 53(3-4): 154-161. Miller, A. C., L. Rhodes, and R. Tippit. 1984. Changes in the naiad fauna of the Cumberland River below Lake Cumberland in central Kentucky. Nautilus 98(3):107-110. Neves, R. J. and J. C. Widlak. 1987. Habitat ecology of juvenile freshwa- ter mussels (Bivalvia: Unionidae) in a headwater stream in Virginia. American Malacological Bulletin 5(1):1-7. Neves, R. J. and J. C. Widlak. 1988. Occurrence of glochidia in stream drift and on fishes of the Upper North Fork Holston River, Virginia. American Midland Naturalist 119:111-120. Oesch, R. D. 1995. Missouri Naiades: A Guide to the Mussels of Missouri. Missouri Department of Conservation, Jefferson City. 270 pp. Parker, R. S., C. T. Hackney, and M. F. Vidrine. 1984. Ecology and repro- ductive strategy of a south Louisiana freshwater mussel, Glebula rotundata (Lamarck) (Unionidae: Lampsilini). Freshwater Invertebrate Biology 3(2):53-58. Sickel, J. B. and C. C. Chandler. 1996. Unionid fauna of the Lower Cumberland River from Barkley Dam to the Ohio River, Kentucky (Mollusca: Bivalvia: Unionidae). Transactions of the Kentucky Academy of Science 57(1):33-46. Smathers, K. L. 1990. An analysis of a bivalve (Mollusca: Bivalvia) com- munity in the Licking River, at Moores Ferry, Kentucky. Master’s Thesis, Eastern Kentucky University, Richmond, Kentucky. 65 pp. Tedla, S. and C. H. Fernando. 1969. Observation on the glochidia of Lampsilis radiata (Gmelin) infesting yellow perch, Perca flavescens (Mitchell) in the Bay of Quinte, Lake Ontario. Canadian Journal of Zoology 47:705-712. van der Schalie, H. 1970. Hermaphroditism among North American fresh- McMURRAY ET AL.: IMPOUNDMENTS AND FRESHWATER MUSSEL RECRUITMENT 63 water mussels. Malacologia 10(1):93-112. van der Schalie, H. and A. van der Schalie. 1950. The mussels of the Mississippi River. American Midland Naturalist 44(2):448-466. Watters, G. T. 1994. An Annotated Bibliography of the Reproduction and Propagation of the Unionoidea (Primarily of North America). Ohio Biological Survey Miscellaneous Contributions No. 1. vi + 158 p. Weiss, J. L. and J. B. Layzer. 1995. Infestations of glochidia on fishes in the Barren River, Kentucky. American Malacological Bulletin 11(2):153-159. Williams, J. D., S. L. H. Fuller, and R. Grace. 1992. Effects of impound- ments on freshwater mussels (Mollusca: Bivalvia: Unionidae) in the main channel of the Black Warrior and Tombigbee Rivers in western Alabama. Bulletin of the Alabama Museum of Natural History 13:1-10. Williams, J. D., M. L. Warren, K. S. Cummings, J. L. Harris, and R. J. Neves. 1993. Conservation status of freshwater mussels of the United States and Canada. Fisheries 18(9):6-22. Wilson, C. B. 1916. Copepod parasites of fresh-water fishes and their eco- nomic relations to mussel glochidia. Bulletin of the United States Bureau of Fisheries 34:331-374. Yokley, P., Jr. 1972. Life history of Pleurobema cordatum (Rafinesque, 1820) (Bivalvia: Unionacea). Malacologia 11:351-364. Zale, A. V. and R. J. Neves. 1982. Fish hosts of four species of lampsiline mussels (Mollusca: Unionidae) in Big Moccasin Creek, Virginia. Canadian Journal of Zoology 60:2535-2542. Date of manuscript acceptance: 15 January 1999 ie ae Aid 5 > (LeRoi ya dare tt ahd aet PED a Roky an Fe ow oe ; yun i rans 4 X (I. Lea, 1838) 6.2 6.1 xX >, Gaim, Gane, Gam, © Xx Xx Xx X X Xx xX XxX X* xX 4 X X HENLEY AND NEVES: UNIONID RECOVERY IN VIRGINIA, U. S. A. 69 Table 3. Survey results from 19 sampling sites on the NFHR from June through August 1995. Transect CPUE, transect CPUE density, and quadrat density values are site means. (CPUE, no./h; density, no./m2). North Fork Holston River Survey Site (NFHRM) 79.9 78.0 77.0 73.9 68.6 60.7 59.9 56.4 53.3 53.2 48.6 45.8 37.7 306 135 11.0 93 62 61 Investigator CPUE 1 0 0 0 0 0 5 5 11 4 3 2 8 22 13 4 3 6 Random CPUE 50 0.0 00 50 2.0 3.0 110 80 86 343 165 29 49 136 308 17.0 2.3 26 5.3 Transect CPUE 2.96 3.75 8.28 5.90 186 11.2 * Transect CPUE Density Estimate 0.04 0.13 0.19 0.15 0.45 0.40 * Quadrat Density Estimate (+ 21%) 2.20 2.60 1.76 ** Juveniles Observed 2 7 =) 11 5 Species with Juveniles Observed g2 1 3 2 *Survey disrupted by the activity of another investigator. It should be noted that between 1991 and 1995, 852 adult mussels of 9 species were translocated from sites in the Clinch River and upstream of Saltville to NFHRM 13.5 (R. Neves, unpub. data). This was the site with the highest recruitment, as evidenced by the collection of 11 juvenile mussels (Table 3). Collection of juveniles of 3 species indi- cates that this location is being repopulated by reproduction from translocated animals, because of the prior absence of mussels at this site. The site will presumably become a source of recruitment of juvenile mussels to proximate river reaches in future years. In general, the number of age classes for all species was highest at NFHRM 53.2, and the number of age classes at downstream sites where random CPUE equaled or exceeded 5 mussels/h remained roughly constant. At most sites, Lampsilis fasciola, Villosa vanuxemensis, V. iris, and L. ovata were represented by multiple age classes. The number of age classes and relative abundance of L. fascio- la, V. iris, and L. ovata generally increased with distance from Saltville, while age classes and relative abundance of V. vanuxemensis decreased. Generally, the number of V. iris increased, while the number of V. vanuxemensis decreased proceeding downstream. The number of L. ovata and L. fasciola generally remained similar with downstream river mile, except that the number of L. fasciola peaked at NFHRM 53.2. A comparison of our survey results with those of Ortmann (1918) and Hill et al. (1974) revealed that the number of species in the river downstream of Saltville, Vir- ginia has decreased over time (Tables 3 and 4). Between the early 1900s and 1972, the number of reported mussel species decreased from 24 to 1 within the river reach of this study (Ortmann 1918; Hill et al., 1974). From the results of our survey we see that the number of species recorded in the reach has increased from 1 to 9 since 1972 (Table 3). Using maximum ages of live mussels per site, we conclude that recolonization or reproduction within aggregations began at least by the early to mid-1980s. In addition to an increase in the number of species observed downstream of Saltville since the Hill et al. (1974) survey, there also appears to have been a shift in the pattern of species rich- ness within the river this century (Ortmann, 1918). Where- as we found species richness to be highest at NFHRM 11.0, the highest species richness in the early 1900s occurred at NFHRM 59.3 (26 species) and 39.2 (24 species) (Table 4). Thus, not only has the number of mussel species reported from the river been drastically reduced, but the spatial dis- tribution of aggregations within the river has shifted. Descriptive characteristics of mussel assemblages at sites surveyed, such as the total number of mussels observed (r2 = 0.40, p = 0.18), the number of juveniles detected (r2 = 0.06, p = 0.35), number of age classes for all species (r2 = 0.24, p = 0.33), and number of age classes for L. fasciola and V. iris (r2 = 0.23, p = 0.33 and r2 = 0.34, p= 0.24, respectively) did not increase significantly with dis- tance downstream of Saltville. Also, random CPUE (no./h) (r2 = 0.10, p = 0.18) and mean site quadrat density (no./m2) (r2 = 0.06, p = 0.95) were not correlated to river mile loca- tion. Mean transect CPUE was marginally related to NFHRM (r2 = 0.64, p = 0.06). The Anderson-Darling test for normality (Sokal and Rohlf 1995) showed that all mus- sel assemblage characteristics were normally distributed (p > 0.05), except random CPUE (p < 0.001) and number of juvenile species observed (p < 0.01). Transformations indi- cated by the Box-Cox transformation procedure (Sokal and Rohlf, 1995) did not substantially increase r2 and associat- ed p values, when these variables were regressed on river mile. Thus, mussel assemblage characteristics did not seem to exhibit a curvilinear relationship to river mile location. Site quadrat data from surveyed mussel aggrega- tions followed a negative binomial distribution, using the Chi Square Goodness-of-Fit Test (Elliott 1977). Chi square values (x2), associated p value ranges, and k estimates were: NFHRM 53.3, x2 = 0.53 (0.90 < p < 0.95) with & = 10.000; NFHRM 13.5, x2 = 6.18 (0.25 < p < 0.50) with k = 0.495; and NFHRM 11.0, x2 = 2.95 (0.50 < p < 0.75) with k = 70 AMER. MALAC. BULL. 15(1) (1999) Table 4. Species presence at survey sites with comparisons to survey findings of Ortmann (1918) (O) and Hill et al. (1974) (H). Holding ponds at Saltville, Virginia are at NFHRM = 80.3. Ortmann site locations are approximate. Ortmann (1918) binomials were revised to conform to current taxonomic names according to Parmalee and Bogan (1998). North Fork Holston River Survey Site (NFHRM) 82.8 79.9 79.0 59.9 59.3 56.4 53.2 45.8 45.0 39.2 30.6 20.9 (O) (H) (O) Actinonaias ligamentina (Lamark, 1819) A. pectorosa (Conrad, 1834) Alasmidonta marginata Say, 1818 A. viridis (Rafinesque, 1820) Amblema plicata (Say, 1817) - - - - Cyclonaias tuberculata - - - - (Rafinesque, 1820) Elliptio dilatata (Rafinesque, 1820) - - - - Epioblasma brevidens (I. Lea, 1831) E. capsaeformis (I. Lea, 1834) - - - - E. haysiana (1. Lea, 1834)* - - - - E. torulosa gubernaculum (Reeve, 1865)* - - - - E. triquetra (Rafinesque, 1820) - - - - Fusconaia barnesiana (1. Lea, 1838) x - - - F. cor (Conrad, 1834) - - - - F. cuneolus (I. Lea, 1840) - - - - F. subrotunda (1. Lea, 1831) Lampsilis fasciola Rafinesque, 1820 L. ovata (Say, 1817) Lasmigona costata (Rafinesque, 1820) Lemiox rimosus (Rafinesque, 1831) - Lexingtonia dolabelloides (1. Lea, 1840) X - Medionidus conradicus (1. Lea, 1834) x - - - xX xX et x KM Pn) oe eo a oe Pegias fabula (I. Lea, 1838) Pleurobema oviforme (Conrad, 1834) Ptychobranchus fasciolaris - - - - (Rafinesque, 1820) P. subtentum (Say, 1835) Xx - - - Quadrula cylindrica (Say, 1817) - - - - Q. intermedia (Conrad, 1836) - Strophitus undulatus (Say, 1817) 4 - - - Toxolasma lividus Rafinesque, 1831 x Villosa fabalis (1. Lea, 1831) - - V. iris (I. Lea, 1829) X - - X V. perpurpurea (I. Lea, 1861) - - - - V. vanuxemensis (1. Lea, 1838) Xx - - Xx Total Species Observed 16 1 0 3 ~~! xx Pace] 2 >< > > *extinct 0.369. Because these data fit the negative binomial distribu- tion, it was appropriate to use k as an estimate of site aggre- gation. It is important to note that higher k values reflect lower levels of mussel aggregation or clumping (Elliott 1977). Values of k can range from 0 to +e, with values from 0 to about 8 indicating distributional aggregation (Poole 1974). Thus k values > 8 may indicate a distribution of mussels that approaches the Poisson or random distribu- tion. On this basis, mussel distributivns at NFHRM 13.5 and 11.0 may be viewed as highly aggregated. Since the k value associated with NFHRM 53.2 is > 8, mussels at this site were considered to be approaching a random distribu- x WK 13.5 110 88 63 62 61 (H) (O) (O) (O) (H) : zg : : z 2 E < x c = 2 eel me OR ow) OR ee als Sa peas: - z 2 xX - - 2 - - = : e3 : 7 7 xX z : - xX ‘ 2 : Bee OR a te. es 1 nts z z : - = - 7 xX - = E : - xX : 2 : : xX a - z - : - 2 - 2 3 xX i é : 2 : xX : E - Z xX is E g re ee ee ee ee ee ee ee te a oe oS XM 2 MS eh, em RE at eee Se = =X gi sg &. Ae “a, rer ty 3 7 4 7 : 7 X k 2 £ KN = K Xe Re! bs Cites KM ke RR oO 1.) hehe : = x Mth ke ORE Le pao - - zs 2 2 xX - : “ - 4 e xX : x z - 2 - ' : : X 4 x - - 2 : : 2 : - 4 - X 7 x 4 i s a ae” a a ee ds fe) Bee : : : xX - : : xX : uy 2 2 & we & Rie Ce. Vt Pas - 3 - xX 7 E B bs xX A - ‘ xX : : Xx Ff : x © 2 xX MOR Oe owe tees x : : : xX : 2 : a 2 Z = X 2. 3K. Ge OS xX, Jo \beeee a. are ee) ay a © i OS (ae CY | tion. Elliott (1977) noted that if the same number and size of quadrats are obtained at survey locations, then k may be used to compare aggregations. These requirements were fulfilled in our survey. Results of our study show no evidence of mussel assemblages immediately below Saltville, between NFHRM 79.9 and 59.9. Between Saltville and river mile 59.9, only isolated individuals of mussel species were observed. Possible explanations for this are the lingering effects of Hg, or insufficient numbers of mussels and suit- able fish hosts to support reproduction. A translocation of sufficient specimens of species such as Villosa iris and HENLEY AND NEVES: UNIONID RECOVERY IN VIRGINIA, U.S. A. 71 Lampsilis fasciola into this river reach would allow an assessment of its suitability for recolonization. Comparison of the body burdens of Hg in Corbicula fluminea and V. iris from the U. S. Geological Survey (G. Johnson, unpub. data), Woodward-Clyde (1993), Seivard et al. (1993), and our values shows the need for further Hg monitoring (Table 5 and Fig. 2). Although most studies show Hg levels > 1 mg/l dry weight, content values between studies are incon- sistent and likely vary between body burdens in C. fluminea versus V. iris. Further study is needed to show whether lev- els of Hg in C. fluminea and V. iris are comparable for specimens of similar age. Within this study, random CPUE and total number of species observed were the only mussel assemblage char- acteristics that were related to total Hg content in Corbicula fluminea, as measured by Seivard et al. (1993) at proximate survey sites (r2 = 0.69,p = 0.02 and r2 = 0.77, p = 0.13, respectively). There was a noticeable absence of younger age classes and a decrease in total mussels observed at sites surveyed at locations upstream of NFHRM 60.0. Total Hg content in C. fluminea was inversely correlated to river mile locations (r2 = 0.79, p = 0.003), whereas methylmercury content in C. fluminea was not related to river mile (r2 = 0.52, p = 0.04), as reported by Seivard et al. (1993). Two conclusions seem evident from this information. First, the probability of the effect of Hg on mussel recruitment (recent and historic) in the river increases with proximity to Saltville; and secondly, that effect is seemingly related to total Hg and not methylmercury. It is interesting that mea- sures of relative abundance of mussels estimated by our study were inversely correlated to total Hg content rather than methylmercury content, since methylmercury is known to produce toxic effects in aquatic organisms (Chang et al., 1974; Khan and Weis, 1987; Stinson and Mallatt, 1989; Chen and McNaught, 1992). Perhaps some form of Hg in the total Hg component that was not tested has had a contributing effect. When examining the relationships between site assemblage characteristics, such as the number of mussel species and random and transect CPUE versus methylmer- Table 5. Mean total mercury content (mg/l, dry weight) in Corbicula fluminea and Villosa iris from selected North Fork Holston River sites (G. Johnson, unpub. data; Woodward-Clyde, 1993; Seivard et al., 1993; this study). Settling ponds at Saltville are at NFHRM 80.3. Seivard et al. (1993) | Woodward-Clyde (1993) G. Johnson, unpub. data This Survey C. fluminea C. fluminea C. fluminea V. iris Site Mean s Mean s* Mean s** Mean s 91.5 NP 88.5 < 0.20 85.6 < 0.20 84.5 0.18 0.02 1.54*** 1.61 81.3 0.13* 79.9 3.34 0.42 76.7 NP 74.0 0.49* 71.0 0.52 0.12 69.9 3.46 0.21 1.50 66.3 NP 63.8 < 0.26 60.7 0.48* 60.2 2.65 0.04 58.2 < 0.20 56.4 < 0.28 53.2 0.51 0.08 2.67 0.53 51.3 1.79 0.15 39.7 2.38 0.23 30.6 5.27*** 3.94 30.1 1.94 0.20 22.8 1.89 0.14 7.6 1.47 0.13 4.8 1.40 *Most standard deviations for Woodward-Clyde data were not included because some data were presented as < 0.20 or of only one datum; **standard deviations of G. Johnson (unpub. data) not provided; ***one datum was comparatively high for the site (3.39 mg/l, dry weight); ****one datum was comparatively high for the site (9.73 mg/l, dry weight). a2 11 O = Thisdudy 5 ) 4 =G. Johnson, unpub. data s O = Seivard et al. (1993) ram 4 oO — 3 = eA 2 7 Ss l io) ia i) 0 5 10 15 20 25 W 35 40 45 50 55 60 65 70 75 8 85 9% NFHRM Fig. 2. Mean total mercury content (mg/l, dry weight) in Corbicula flu- minea and Villosa iris from selected sites on the North Fork Holston River (G. Johnson, unpub. data; Woodward-Clyde, 1993; Seivard er al., 1993; this study). Accuracy of data points of Woodward-Clyde (1993) are sus- pect (represented at < 0.20 or by one datum). Settling ponds at Saltville are at NFHRM 80.3. cury content in Corbicula fluminea (Seivard et al., 1993), an interesting phenomenon is seen. At NFHRM 45.8 through 30.6, decreases in site assemblage characteristics occur, and associated with these decreases are increases in total Hg in C. fluminea at NFHRM 39.7(Tables 3 and 5 and Fig. 2). Although total Hg is seemingly related to reduced mussel assemblages in this reach of the river, further survey would be required to confirm this apparent effect. Seivard et al. (1993) found a general decreasing trend in total Hg content in C. fluminea downstream of Saltville, Virginia. The exception to this decreasing trend was observed at NFHRM site 39.7, where total Hg content increased (Seivard et al., 1993). Mercury content in C. fluminea mea- sured by the U. S. Geological Survey (G. Johnson, unpub. data) also decreased with distance from Saltville, whereas similar data measured by Woodward-Clyde (1993) showed no trend (Table 5 and Fig. 2). Additional testing is needed to determine whether the decrease in total Hg with down- stream distance from Saltville is real or spurious. If the ele- vated Hg levels in C. fluminea near NFHRM 30.0 through 40.0 are representative of elevated levels of Hg in river sed- iment, then mussel aggregations downstream, such as at NFHRM 13.5 and 11.0, could be at future risk if such a contaminated sediment load is being transported down- stream. When considering the re-establishment of mussel species reported by Ortmann (1918) in the river down- stream of Saltville, Virginia, and the current absence of most of those species in the upper and lower river, the need for a mussel restoration program becomes apparent. Of the 33 species reported by Ortmann (1918) downstream of AMER. MALAC. BULL. 15(1) (1999) Saltville, Hill et al. (1974) and Barr et al. (1993) observed 11 and 10 of these species upstream, respectively. Conse- quently, translocation of mussels from another river within the same basin will be essential for the re-establishment of mussel species that were present historically. In addition to previous translocation, habitat requirements of translocated species, and possible contami- nant effects, the selection of sites should be guided by evi- dence of current reproduction, historical pattern of mussel aggregations, and the availability of suitable fish hosts. Our results indicate that NFHRM 11.0, 13.5, 30.6, and 53.2 are suitable sites for reintroduction. The river reach between NFHRM 40 through 60 should be evaluated in further assessments of site suitability, as this stretch of the river historically supported a high diversity of species (Ortmann, 1918). ACKNOWLEDGMENTS We thank Gregg Johnson of the U. S. Geological Survey, Knoxville, Tennessee for providing unpublished data of Hg content in Corbicula fluminea from the NFHR. We also thank Braven Beaty, Dan Dorosheff, Jess Jones, Debra Neves, Mike Pinder, and Michelle Steg for their help with the river surveys. LITERATURE CITED Adams, C. C. 1915. The variations and ecological distribution of the snails of the genus Io. Memoirs of the National Academy of Science XII (I1):1-92. Ahlstedt, S. A. 1979. Recent mollusk transplants into the North Fork Hol- ston River in southwestern Virginia. Bulletin of the American Malacological Union 1979:21-23. Barr, W. C., S. A. Ahlstedt, G. D. Hickman, and D. M. Hill. 1993. Cum- berlandian mollusk program. Activity 8: analysis of macrofauna factors. Walkerana 7(17/18):159-224. Carter, L. J. 1977. Chemical plants leave unexpected legacy for two Vir- ginia Rivers. Science 198:1015-1020. Chang, L. W., K. R. Reuhl, and A. W. Dudley. 1974. Effects of methylmercury chloride on Rana pipiens tadpoles. Environmental Research 8(1):82-91. Chen, T. Y. and D. C. McNaught. 1992. Toxicity of methylmercury to Daphnia pulex. Bulletin of Environmental Contamination and Toxicology 49(4):606-612. Downing, J. A. and W. L. Downing. 1992. Spatial aggregation, precision, and power in surveys of freshwater mussel populations. Canadian Journal of Fisheries and Aquatic Sciences 49(5):985-991. Elliott, J. M. 1977. Statistical Analysis of Samples of Benthic Inverte- brates. Freshwater Biological Association. Scientific Pub. No. 25, second edition. Ambleside, England. 144 pp. Hill, D. M., E. A. Taylor, and C. F. Saylor. 1974. Status of faunal recovery in the North Fork Holston River, Tennessee and Virginia. Pro- ceedings of the 28th Annual Conference of Southeastern Associa- tion of Game and Fish Commissioners 28:398-413. Khan, A. T. and J. S. Weis. 1987. Effects of methylmercury on sperm and HENLEY AND NEVES: UNIONID RECOVERY IN VIRGINIA, U.S. A. 12 egg viability of two populations of killifish (Fundulus heterocli- tus). Archives of Environmental Contamination and Toxicology 16(4):499-505. Krebs, C. J. 1989. Ecological Methodology. Harper and Row, Publishers, New York. 653 pp. Leftwich, K. N. 1994. Habitat models for predicting the occurrence of blotchside logperch (Percina burtoni) and tangerine darters (P. aurantiaca) in the North Fork Holston River and Littie River, Vir- ginia. Masters Thesis, Virginia Tech, Blacksburg, Virginia. 103 PP. Neves, R. J. and S. N. Moyer. 1988. Evaluation of techniques for age determination of freshwater mussels (Unionidae). American Mal- ocological Bulletin 6(2):179-188. Ortmann, A. E. 1918. The nayades (freshwater mussels) of the upper Ten- nessee drainage with notes on synonymy and distribution. Pro- ceedings of the American Philosophical Society 57(6):521-626. Parmalee, P. W. and A. E. Bogan. 1998. The Freshwater Mussels of Ten- nessee. The University of Tennessee Press, Knoxville, Tennessee. 328 pp. Poole, R. W. 1974. An Introduction to Quantitative Ecology. McGraw-Hill Inc., New York. 532 pp. Seivard, L. D., D. A. Stilwell, S. O. Rice, and K. R. Seeley. 1993. Geo- graphic distribution of mercury in asiatic clams, Corbicula flu- minea, from the North Fork Holston River, Virginia. U. S. Fish and Wildlife Service, Environmental Contaminants Division, Vir- ginia Field Office, White Marsh, Virginia. 23 pp. Sheehan, R. J., R. J. Neves, and H. E. Kitchel. 1989. Fate of freshwater mussels transplanted to formerly polluted reaches of the Clinch and North Fork Holston rivers, Virginia. Journal of Freshwater Ecology 5(2):139-149. Sokal, R. R. and F. J. Rohlf. 1995. Biometry. W. H. Freeman and Compa- ny, New York. 887 pp. Stansbery, D. H. 1972. The mollusk fauna of the North Fork Holston River at Saltville, Virginia. Bulletin of the American Malacologi- cal Union 1972: 45-46. Stinson, C. M. and J. Mallatt. 1989. Branchial ion fluxes and toxicant extraction efficiency in Lamprey (Petromyzon marinus) exposed to methylmercury. Aquatic Toxicology 15(3):237-252. Woodward-Clyde Engineering. 1993. Remedial Investigation Report: Operable Unit Three, Saltville Waste Disposal Site, Olin Chemi- cal Group. Woodward-Clyde Consultants, Franklin, Tennessee. 23 pp. Date of manuscript acceptance: 20 January 1999 —_—— ules BeBe oe. ys at Bn _ Wat ira siabta har wae ~ ee kee, ar rc : _ Re kg iy iy dat ; ed iy as ; Ripe ie Me vnearicinet ge 04 Ny gg a4 - aid, Ay ene yt Aner ang arte Way % catia ae hs eleciniaes 1) of WE eegaat 15) , HO AEE i ey EY Stay f Ne hail anf if BD, af hay ee weal 7 MM vein ee yy Sip bentnnne a ieee APES 7 et ns a re CP ee mie tie va in. ib die | PHyaik sf 7 Gi ee | ; ere a) a seeest//pdl Aw Rad “yp ane tibent Me, aihailtan sn 10) j - alt he ime ‘5 i Neo zi, | avant reed’ ppt Pale) viele Wa Vie u) ee ae Historical and ontogenetic changes in shell width and shape of land snails on the island of Kikai Emiko Hayakaze! and Satoshi Chiba2 IInstitute of Geosciences, Shizuoka University, 836 Oya, Shizuoka 422, Japan 2Biological Institute, Graduate School of Science, Tohoku University, Aramaki-Aza-Aoba, Aoba-ku, Sendai 980-8578, Japan Abstract: Patterns of change in shell width over a period of 35,000 years are documented in three fossil land snail species, Euhadra pachya (Pilsbry, Phaeohelix phaeogramma (Ancey), and Coniglobus mercatorius daemonorus (Pilsbry) on the island of Kikai in the Ryu-kyu Islands of the southwest part of Japan. The shell width of these species fluctuated through time, and the patterns were mostly synchronized among the three species. The temporal variations in shell width among fossil populations were far larger than geographical variations in shell width among modern populations. Since Kikai Island has been isolated from the other islands in the region, these morphological changes are regarded as genuine changes that have occurred within the island. The tempo- ral changes in width were accompanied by distinct changes in spire indexj (height/width) within a single species. The patterns of change in width are corre- lated with the pattern of change in the climate since 35 Ka (35,000 years ago). Key Words: body size, fossil, Kikai Island, land snail, shell morphology A number of studies on the patterns and mecha- nisms of phenotypic evolution have been undertaken on fossil land snails in island systems (e. g., Gould, 1969, 1984, 1989; Goodfriend and Gould, 1996; Chiba, 1996). Important contributions have been made to the study of pat- terns of Quaternary biogeographic change in island snails within an island (Goodfriend, 1987, 1993; Goodfriend and Mitterer, 1993; Goodfriend et al., 1994; Cook et al., 1993). In the present study, the morphologies of fossil and modern land snails from Kikai Island in the southwest part of Japan were studied, because it offers an excellent example to see morphological changes of lineages within an island. Kikai Island, in the northern part of Ryukyu Islands, is a small island, approximately 60 km2 in area and lower than 200 m above sea level. It has been isolated from other islands in the region since 100,000 years ago. The island has eolianite dunes that were produced during the late Pleistocene and Holocene no more than 40,000 years ago. Stratigraphic studies and 14C dating of the eolianite dunes on Kikai have been performed by Mitsui and Kigoshi (1966), Nakagawa (1967), Kakuta (1977), and Naruse and Inoue (1987), and these studies have shown that paleosoil layers have been produced four times: 32 - 31 Ka (thousand years ago) (period II), 29 - 27 Ka (period III), 22 Ka (peri- od IV) and 3 - 2.5 Ka (period V). Fossil land snails have been reported from these paleosoil layers. In addition, land snail shells are found in a calcareous sand layer just above the base deposits of the dune. This layer consists of coarse sand, blocks of beach rocks and marine shell fragments, and includes fossils of the land snails. The age of the snails from this calcareous sand layer is 35,600 yr (Naruse and Inoue, 1987), and the layer is designated period I. Fossil land snails are found in the dune deposits, and three species of pulmonates, Euhadra pachya (Pilsbry, 1902), Phaeohelix phaeogramma (Ancey, 1888), and Coniglobus mercatorius daemonorus (Pilsbry, 1901), are the most common species. E. pachya is endemic to Kikai, and C. mercatorius dae- monorus is an endemic subspecies of Kikai Island. Although it is not apparent how these species divide resources to avoid competition, Phaeohelix phaeogramma and Coniglobus mercatorius daemonorus seem to have sim- ilar niches: both live sympatrically under leaf litter or rocks on the ground and never climb trees. Euhadra pachya became extinct during the Holocene and so its habitat is not known. On the basis of observations of E. herklotsi, the most closely related species to E. pachya, it is assumed that the niche of this species is similar to that of P. phaeogram- ma and C. mercatorius daemonorus. Here, we show evidence of distinct changes in shell width that have occurred in the three species during the past 35,000 years. The morphological patterns are synchronized among the three species. Distinct changes in shell shape follow the increase or decrease in shell width. Variations in shell shape may have close relationships with ontogenetic changes in shell shape (Foote and Cowie, 1988; Gould, 1984). We show that patterns of historical changes in shell American Malacological Bulletin, Vol. 15(1) (1999):75-82 75 76 AMER. MALAC. BULL. 15(1) (1999) shape are associated with ontogenetic changes in shell shape. In addition, the correlates of climatic changes and other factors on the patterns of morphological changes are discussed. MATERIALS AND METHODS Twenty-two samples of fossil and living snails were collected from 18 localities (Fig. 1 and Table 1). Specimens used for the analysis were all adult snails that had a reflect- ed, thickened shell lip. Specimens in each modern sample were taken from an area of 25 m2. The vegetation at each modern site was categorized as shrub or forest (Table 1). Sites categorized as shrub were sites with plants lower than 1 m in height, and sites categorized as forest were sites with plants higher than 2 m in height. The leaf litter of the for- mer was more than 30 cm in depth, but that of the latter was very thin. Forest and shrub did not coexist in a sam- pling site, because the area of each sampling site was small. All of the samples used in this study were deposited in Uni- 13 16 14 15 SJ Fig. 1. Map of the island of Kikai showing the numbered sample loca- tions. versity Museum, Shizuoka University (SUM). Data of 14C ages previously reported for each layer are adopted for the age of the samples used in the present study (Table 1). These data were obtained by dating shells. In addition, the ages of seven fossil samples were examined by 14C dating. The !4C ages were determined by standard methods with benzene-liquid scintillation (Stipp et al.. 1974; Gupta, 1985). All fossil samples were collected from paleosoil layers except for one sample (P16L) that was from the cal- careous sand layer. Phaeohelix phaeogramma and Euhadra pachya were found in all fossil samples, but Coniglobus mercatorius daemonorus was not found in sample P16L. E. pachya was not found in the modern samples. In total, 951 specimens were measured. Characters measured were shell height (H), width (W) and number of whorls. An edge of the protoconch was defined zero whorls. Shell shape was represented by spire index (H/W) (Cain, 1977). All three species examined here have determinate growth. Shell width and height at each whorl number (Wn and Hn in Fig. 2) were measured on the shell which was cut along the coiling axis. Ten snails randomly selected from the set of modern and fossil shells were cut in order to mea- sure ontogenetic change for each species. Spire index at each whorl number was obtained by Hn/Wn, and ontoge- netical changes of these characters were examined. In addi- tion, relationships between shell width, number of whorls and spire index were examined in adult snails to estimate Fig. 2. Description of shell measurements on a cross section of a shell. W: width of an adult shell, H: height of an adult shell, WO: width of a proto- conch, Wn: width of the nth whorl (n=0.5, 1, 1.5, 2...), HO: height of pro- toconch, Hn: height of the nth whorl (n=0.5, 1, 1.5, 2...). HAYAKAZE AND CHIBA: HISTORICAL CHANGES IN LAND SNAIL SHELLS ad Table 1. Modern and fossil samples of Phaeohelix phaeogramma (P.p.), Coniglobus mercatorius daemonorus (C.m.), and Euhadra pachya (E.p.) from Kikai (H=Holocene, P=Pleistocene, M=modern, U=upper section, M=middle section, L=lower section). Accession number of University Museum, Shizuoka University (SUMEFC) for each sample is also presented. Asterisked ages are from Naruse & Inoue (1987). Vegetation and height above sea level for each sample site are also presented. Number of specimens E.p. Vegetation Loc. Sample Period Age P.p. C.m. 1 Ml Modern 22 13 2 M2 Modern 20 24 3 M3 Modern 24 20 4 M4 Modern 20 30 5 M5 Modern 20 20 6 M6 Modern 20 12 7 M7 Modern 20 17 8 M8 Modern 18 12 9 M9 Modern 20 14 9 P9 III 26800+500 46 16 10 M10 Modem 25 21 11 Mill Modern 21 10 12. H12 V 670+40 21 10 13. ~-H13 Vv 3030+90 24 5 14 H14 V 3200+60* 36 12 15 PISU IV =. 22200+300 24 10 15 PISL II 31300+700 26 8 16 PI16U Tl 27900+630* 10 10 16 PI6M = II 32600+870* 26 16 16 =PI6L I 35600+1500* 16 0 17. ~PI7 Ill 29100+400 45 18 18 P18 II 31700+850 13 6 effects of determinate growth of these species on their adult shell morphologies. Geographical variation in shell width of the modern populations was examined to estimate the relationship between geographical variation in shell width and variation of local environmental conditions. Differences of shell width among modern samples were analyzed statistically with ANOVA. Temporal changes of shell width and spire index were examined by placing the populations in their stratigraphic context. RESULTS The 1!4C age data presented in this study were con- sistent with the previous 14C age data for the dune deposits of Kikai Island (Kakuta, 1977; Naruse and Inoue, 1987). 14C age anomalies in snail shells were reported by Good- friend and Stipp (1983) and Goodfriend (1987). Age anom- alies in samples of living Phaeohelix phaeogramma and Coniglobus mercatorius daemonorus from Kikai have been examined and ranged from 300 to 800 years (Takahashi and Wada, 1998). This imples that the samples are 300 ~ 800 years younger than the 14C ages, and that the order of ages of two samples in which the interval of ages is less than io) =n oo ~ OPS ae te ee OO oS Height above Accession sea level (m) number forest 15 SUMFCC0051 shrub 25 SUMFCC0052 forest 8 SUMFCC0053 forest 103 SUMFCC0054 forest 150 SUMFCC0055 forest 198 SUMFCC0056 shrub 5 SUMFCC0057 shrub 43 SUMFCC0058 shrub 50 SUMFCC0059 unknown 50 SUMFCC0060 shrub 55 SUMFCC0061 forest a SUMFCC0062 unknown 4 SUMFCC0063 unknown 15 SUMFCC0064 unknown 14 SUMFCC0065 unknown 20 SUMFCC0066 unknown 14 SUMFCC0067 unknown 32 SUMFCC0068 unknown 30 SUMFCC0069 unknown 29 SUMFCC0070 unknown 46 SUMFCC0071 unknown 32 SUMFCC0072 1000 years is not clear. However, the affect of these age anomalies on the overall patterns of morphological changes are minor, because the intervals between most of the fossil samples examined in this study are greater than 1000 years except for those between the three Pleistocene samples (P15L, P16M, P18) and between the Holocene samples. Shell width of the three species showed similar changes over the past 35,000 years (Fig. 3). Shells reached a maximum width at approximately 28 Ka, and stayed wide until 22 Ka. Specimens in samples H12 - H14 (2.6 - 3.2 Ka) were far smaller than the specimens of the sample P15U (22 Ka), indicating that a reduction in shell width had occurred between 22 and 3.2 Ka. The shell width of the modern samples was similar to that of samples from the Holocene deposits (2.6 - 3.2 Ka) and that of the samples during 29 - 35 Ka. Because of the historical changes in shell width, the mean widths in some samples of Phaeohelix phaeogramma (28 ~ 22 Ka) were greater than the mean width of modern Coniglobus mercatorius daemonorus, which was approxi- mately 1.2 times greater than the mean width of modern P phaeogramma. In spite of the changes in width, however, the ratios of widths among C. mercatorius daemonorus, P. phaeogramma and Euhadra pachya within the same sample remained mostly constant. This constancy in ratios of width 78 AMER. MALAC. BULL. 15(1) (1999) C. mercatorius 40 wo (on) Width (W) P. phaeogramma wo oO nm a 20 36 32 28 24 Age (Ka) Fig. 3. Patterns of temporal change of sample means of shell width for three species through 35,000 years. Closed circles: Phaeohelix phaeogramma, open circles: Coniglobus mercatorius daemonorus, closed triangles: Euhadra pachya. Each bar indicates one standard deviation. The closed bar at top indicates periods of wetter climates than other periods. resulted from simultaneous changes in width in the same direction in all three species. The spire indices of Coniglobus mercatorius dae- monorus and Phaeohelix phaeogramma also showed a pat- tern of change that was similar to that of shell width (Fig. 4). Spire index of the shells of C. mercatorius daemonorus increased at 28 Ka, and then it became low at 3.5 Ka. Spire index of P. phaeogramma also became low at 3.5 Ka. Espe- cially, the spire index of modern C. mercatorius dae- monorus 1s distinctly different from that of 28 - 22 Ka (Fig. 4). Within this species, the spire index of the modern sam- 0 95 —___ 4 C. mercatorius 0.8 é 0.75 a od 0.7 Bios s P. phaeogramma = 0.65 cD) £ io) 0.6 E. pachya = - 0.55 - a8 . 36 32 28 24 20 4 0 Age (Ka) Fig. 4. Patterns of temporal change of sample means of spire index for three species through 35,000 years. Closed circles: Phaeohelix phaeogramma, open circles: Coniglobus mercatorius daemonorus, closed triangles: Euhadra pachya. Each bar indicates one standard deviation. The closed bar at top indicates periods of wetter climates than other periods. ples and those of the fossil samples before 28 Ka were clos- er to each other than the samples of other ages (Fig. 4). Euhadra pachya also showed temporal changes in spire index, but the specimens in the fossil samples from 28 Ka to 22 Ka had a lower spire index than the samples of other ages (Fig. 4). Ontogenetic changes of spire index showed that the spire index (HO/WO, H0.5/W0.5, H1/W1 ... H/W) of Coniglobus mercatorius daemonorus and Phaeohelix phaeogramma decreased from juvenile to middle stage (whorl numbers fewer than 3), but increased after the mid- dle stage (Fig. 5). However, the spire index of Euhadra pachya monotonically decreased during ontogeny (Fig. 5). All of the specimens of a species showed consistent ontoge- netic patterns. A similar pattern is seen in spire index ver- sus width. There were nearly perfect correlations among width, number of whorls and spire index of adult shells (Fig. 6). There was a positive correlation between shell 1 . 0.9 é C. mercatorius — x oO ao) & = a n 0 1 2 3 4 5 Whorl number C. mercatorius Spire index (H/W) 0 10 20 30 40 50 Width (W) mae Fig. 5. Ontogenetic changes of spire index versus whorl number and shell width for representative specimens of the three species. Closed circles: Phaeohelix phaeogramma, open circles: Coniglobus mercatorius dae- monorus, closed triangles: Euhadra pachya. HAYAKAZE AND CHIBA: HISTORICAL CHANGES IN LAND SNAIL SHELLS Phaeohelix phaeogramma Spire index (H/W) 25 Width (W) Coniglobus mercatorius Spire index (H/W) Width (W) Euhadra pachya Spire index (H/W) 0.5 35 40 Width (W) 45 50 mm 19 3.6 37 38 3.9 4 Number of Whorls Width (W) 3.8 4 4.2 4.4 Number of whorls 4.6 4.8 46 B.S [o) Width (W) 3.8 Number of whorls 3.9 4 4] Fig. 6. Relationships among sample means of adult shell width (W), spire index (H/W), and number of whorls. A positive correlation between the number of whorls and the width is found in all species. A positive correlation between width and spire index is found in Phaeohelix phaeogramma and Coniglobus mercatorius daemonorus, but that of Euhadra pachya shows a negative correlation. Correlation coefficients (R) and P values are shown. width and the number of whorls in the three species. Coniglobus mercatorius daemonorus and Phaeohelix phaeogramma showed a positive correlation between width and spire index. Euhadra pachya had a negative correlation between width and spire index. Correlation coefficients for these relationships (see Fig. 6) were all statistically signifi- cant (P<0.01 in all cases, see Fig. 6). Geographical variations in shell width observed in the modern samples of Phaeohelix phaeogramma are shown in Fig. 7. Shell widths of samples M8, M9, and M10 collected from shrub were significantly smaller than all of the samples collected from forest (P<0.05). Shell width of sample M2 from shrub did not differ significantly (P>0.05) from any forest samples except sample M6 (P<0.05), and sample M7 from shrub did not differ significantly from any of the forest samples (P>0.05). As a whole, however, shell widths of snails from forest tend to be larger than those from shrub. Two modern samples of Coniglobus mercato- rius daemonorus (M9 and M10) from shrub had statistical- ly smaller shell widths than samples from forest (P<0.05), 80 AMER. MALAC. BULL. 15(1) (1999) P. phaeogramma Width (W) C. mercatorius 28 30 32 Width (W) Fig. 7. Geographical variation in sample means of shell width among modern samples of Phaeohelix phaeogramma and Coniglobus mercatorius dae- monorus. Open circles: samples collected from forest, closed circles: samples collected from shrub. Each bar indicates one standard deviation. but shell width of other samples from shrub did not differ significantly from forest samples (P>0.05). A significant association of shell width and the height above sea level of the modern sampling site (correla- tion coefficient R=0.70, P<0.05) was found in Coniglobus mercatorius daemonorus (Fig. 8). However, the scatter is wide and the significance is probably dependent on only a few of the outlying points. In addition, association of these was not clear in Phaeohelix phaeogramma (R=0.28, P>0.05). DISCUSSION The synchronized patterns of change in shell size and shape of Euhadra pachya and Coniglobus mercatorius daemonorus are regarded as genuine changes that have occurred within the island and not a reflection of migration within the island or from elsewhere, because these species are endemic species and subspecies respectively, and because the morphological variation among the modern populations was far smaller than that among samples with different ages. Because Kikai Island has been isolated from other islands since the last ice age (Nakagawa, 1969), the pattern of change documented in Phaeohelix phaeogramm also does not reflect immigration from other areas. The morphological patterns showed that shell width of Phaeohelix phaeogramma seemed to decrease during 36 - 32 Ka. Shell widths of the three species remained low during 32 - 29 Ka, and increased after 29 Ka, and reached maximum during 28 - 22 Ka. This latter period corresponds to the time when sea level was lowest. Although there is no fossil record for the period 22 - 4 Ka, reduction of shell width occurred during this period when sea level rose. Yasuda (1987) has suggested that the Japanese islands were wetter 50 - 33 Ka and 28 - 25 Ka than other periods. Although climate of the main parts of the Japanse islands became slightly dry after 25 Ka, the relative increase of the land height of Kikai (approximately 100 m) due to the fall of sea level during 28 - 20 Ka may be responsible for the wettest climate in the island during this period. The periods with wet environments (50 - 33 Ka and 28 - 20 Ka) roughly correspond to the periods when shell width was larger (periods older than 32 Ka and 28 - 22 Ka). Thus, increased shell width appears to be correlated with wet conditions. There are many environmental factors that can affect shell size, e. g., moisture, temperature, density, pre- dation, and random effects, but temperature and moisture may be the most important factors (Goodfriend, 1986; Emberton, 1994). Moisture level is the best documented environmental correlate of shell size, and a positive correla- tion between moisture and size or aperture area has been HAYAKAZE AND CHIBA: HISTORICAL CHANGES IN LAND SNAIL SHELLS 81 De) oS oO C. mercatorius Height above sea level (m) 50 (@) 27 27.5 28 28.5 29 Width (W) cas 200 e P. phaeogramma o 150 > 2 i] 2 ” 100 > 8 fr i} = 50 ay o jen 0 20.5 21 21.5 22 22.5 23 23.5 mm Width (W) Fig. 8. Relationship between mean shell width of a sample and the height above sea level for samples of modern Phaeohelix phaeogramma and Coniglobus mercatorius daemonorus. reported by many authors (e. g., Rensch, 1932; Heller, 1975, 1979; Goodfriend, 1986; Emberton, 1994). In this study, larger shell widths of living snails from forest than from shrub suggest a similar relationship with moisture. The correlation of the morphological pattern with the sup- posed pattern of climatic change suggests that the historical change in shell width may have been induced by the climat- ic change. There are several possible processes by which envi- ronmental changes could induce changes in shell width. These changes may be ecophenotypic responses to the changing environment. Distinct non-genetic changes in shell morphology can be created by environmental changes without genetic changes (Gould, 1984). However, these changes may be genetic and evolutionary changes. Breed- ing experiments on a number of land snail species have shown that size differences among populations have a high heritability (Cook, 1967; Murray and Clarke, 1968; Cook and Cain, 1980; Johnson et al., 1993). Cain (1977) suggest- ed that natural selection for differences in feeding behavior have resulted in divergence of shell height (and see Cook and Jaffar, 1984; Heller, 1987; Emberton, 1994). It is diffi- cult to determine whether the observed changes are genetic or non-genetic. Regardless of the basis of the changes, however, the relationship of climate and shell morphology in the history of these land snail species is affirmed. There are at least two possible causes for the pat- terns of historical changes in spire index. First, these changes may reflect adaptation for environmental condi- tions. Shells with high spires are adaptive in environments with high humidity, because high moisture conditions per- mit more diverse foraging activities and should mechanical- ly favor higher spires (Cain, 1977; Emberton, 1994). How- ever, in the present examination, shells with a high spire index were not necessarily found in the fossil samples of the period with wet environments. For examples, shells of Euhadra pachya became flatter during the period with wet- ter environments. Second, these changes may reflect ontogenetical relationships among characters. There was a nearly perfect correlation between shell width and number of whorls, and this relationship implies that a larger shell is a shell with more whorls. Climate affects the number of whorls or shell width at which growth ceases, thereby affecting spire index. A species that increases its spire index with increasing num- ber of whorls and shell width after the middle stage pro- duces an adult shell with low spire index by becoming adult at a reduced number of whorls, thereby decreasing shell width. Therefore, patterns of change in spire index can be interpreted as by-products of the change in numbers of whorls and shell width because of the correlation between width and spire index that is determined ontogenetically. A positive or negative correlation between spire index and shell width observed in the samples of adult snails reflects this ontogenetical relationship between characters. A positive correlation between spire index and shell width in Phaeohelix phaeogramma and Coniglobus merca- torius daemonorus and a negative correlation between these parameters in Euhadra pachya imply that the high spire index of P. phaeogramma and C. mercatorius daemonorus and low spire index of FE. pachya of 28 ~ 22 Ka were creat- ed by increasing the shell width by increasing the number of whorls (Fig. 6). The temporally fluctuating patterns of shell shape presented in this study may be induced by fluc- tuations of shell width that are induced by climatic change. A correlation between spire index and shell width is found in most land snails (e. g., Goodfriend, 1986). It has been claimed that a non-adaptive change of morphology can occur when the characters are closely linked to size (Gould, 1966, 1969, 1971, 1984). For example, heterochrony, an important cause of evolutionary novelty (Gould, 1977; McKinney, 1986), could produce shape variations as a result of a change in size or a change in the time to maturation. The present study suggests that temporal change in shell 82 AMER. MALAC. BULL. 15(1) (1999) form can occur in association with a change in shell width. This idea implies that changes in shell form of the three species could simply be a direct consequence of a change in number of whorls, and may not be adaptive. Although it is difficult to demonstrate that they are non-adaptive at present, the presence of opposite trends in species with similar life styles suggests that the occurrence of non-adaptive change in shell shape can not be ruled out. ACKNOWLEDGMENTS We thank R. Tanaka, H. Wada, T. Ohji, and S. Takahashi for their advice and help on the !4C dating analyses, which were performed at Shizuoka University and University of Tokyo. LITERATURE CITED Cain, A. J. 1977. Variation in the spire index of some coiled gastropod shells, and its evolutionary significance. Philosophical Transac- tions of the Royal Society of London (B) 277:377-428. Chiba, S. 1996. A 40000-year record of discontinuous evolution of island snails. Paleobiology 22:177-188. Cook, L. M. 1967. The genetics of Cepaea nemoralis. Heredity 22:397- 410. Cook, L. M. and A. J. Cain. 1980. Population dynamics, shell size and morph frequency in experimental populations of the snail Cepaea nemoralis. Biological Journal of the Linnean Society 14:259-292. Cook, L. M. and N. Jaffar. 1984. Spire index and preferred surface orien- tation in some land snails. Biological Journal of the Linnean Society 21:307-313. Cook, L. M., G. A. Goodfriend, and R. A. D. Cameron. 1993. Changes in the land snail fauna of eastern Madeira during the Quaternary. Philosophical Transactions of Royal Society B 339:83-103. Emberton, K. C. 1994. Partitioning a morphology among its controlling factors. Biological Journal of the Linnean Society 53:353-369. Foote, M. and R. H. Cowie. 1988. Developmental buffering as a mecha- nism for stasis: evidence from the pulmonate Theba pisana. Evo- lution 42:396-399. Goodfriend, G. A. 1986. Variation in land-snail shell form and size and its causes: a review. Systematic Zoology 35:204-223. Goodfriend, G. A. 1987. Radiocarbon age anomalies in shell carbonate of land snails from semi-arid areas. Radiocarbon 29:159-167. Goodfriend, G. A. 1993. The fossil record of terrestrial mollusks in Jamaica. Geological Society of America Memoir 182:353-361. Goodfriend, G. A. and S. J. Gould. 1996. Paleontology and chronology of two evolutionary transitions by hybridization in the Bahamian land snail Cerion. Science 274:1894-1897. Goodfriend, G. A. and R. M. Mitterer. 1993. A 45,000-yr record of tropi- cal lowland biota: The land snail fauna from cave sediments at Coco Ree, Jamaica. Geological Society of America Bulletin 105:18-29. Goodfriend, G. A. and J. J. Stipp. 1983. Limestone and the problem of radiocarbon dating of land-snail shell carbonate. Geology 11:575- 577. Goodfriend, G. A., R. A. D. Cameron, and L. M. Cook. 1994. Fossil evi- dence of recent human impact on the land snail fauna of Madeira. Journal of Biogeography 21:309-320. Gould, S. J. 1966. Allometry in Pleistocene land snails from Bermuda: the influence of size upon shape. Journal of Paleontology 40:1131- 1141. Gould, S. J. 1969. An evolutionary microcosm: Pleistocene and Recent history of the land snail P. (Poecilozonites) in Bermuda. Bulletin of the Museum of Comparative Zoology 138:407-532. Gould, S. J. 1971. Precise but fortuitous convergence in Pleistocene land snails from Bermuda. Journal of Paleontology 45:09-418. Gould, S. J. 1977. Ontogeny and Phylogeny. Belknap Press, Cambridge. 501 pp. Gould, S. J. 1984. Morphological channeling by structural constraint: Convergence in style of dwarfing and gigantism in Cerion, with a description of two new fossil species and a report on the discov- ery of the largest Cerion. Paleobiology 10:172-194. Gould, S. J. 1989. A developmental constraint in Cerion, with comments on the definition and interpretation of constraint in evolution. Evolution 43:516-539. Gupta, S. K. 1985. Radiocarbon dating practices at ANU. Radiocarbon dating research. Garran, Australia. Heller, J. 1975. The taxonomy, distribution and faunal succession on Buliminus (Pulmonata: Enidae) in Israel. Zoological Journal of the Linnean Society 57: 1-57. Heller, J. 1979. Distribution, hybridization and variation in the Israeli landsnail Levantina (Pulmonata: Helicidae). Zoological Journal of the Linnean Society 67: 115-148. Heller, J. 1987. Shell shape and land snail habitat in a Mediterranean and desert fauna. Biological Journal of the Linnean Society 31:257- 272. Johnson, M. S., J. Murray, and P. Clarke. 1993. The ecological genetics and adaptive radiation of Partula on Moorea. In: Oxford Surveys in Evolutionary Biology 9, D. Futuyma and J. Antonovics, eds. pp. 167-238. Oxford University Press, Oxford. Kakuta, K. 1977. 14C age of palaeosoil in the dune of Suitengu in Kikai. Chikyu-Kagaku 31:134-135. McKinney, M. L. 1986. Ecological causation of heterochrony: test and implications for evolutionary theory. Paleobiology 12:282-289. Mitsui, H. and K. Kigoshi. 1966. 14C age of coastal deposits and dune in Kikai, Amami Groups. Chikyu-Kagaku 82:43-45. Murray, J. and B. Clarke. 1968. The genus Partula on Moorea: Speciation in progress. Proceedings of the Royal Society of London B 211:83-117. Nakagawa, H. 1967. Geology of Tokunoshima, Okierabujima, Yoronto, and Kikaijima, Amami Gunto (1). Tohoku University Institute of Geology and Palaeontology Contribution 63:1-39. Nakagawa, H. 1969. Geology of Tokunoshima, Okierabujima, Yoronto, and Kikaijima, Amami Gunto (2). Tohoku University Institute of Geology and Palaeontology Contribution 68:1-17. Naruse, T. and K. Inoue. 1987. !4C age of eolianites in Kikai, Ryukyu Islands. Chikyu-Kagaku 41:198-201. Rensch, B. 1932. Uber die Abhangigkeit der Grosse, des relativen Gewichtes und der Oberflachenstruktur der Landschnecken- schalen von Umweltsfaktoren (Okologische Molluskenstudien I.). Zeitschrift fuer Morphologie und Oekologie der Tiere 25:757-807. Stipp, J. J., K. L. Eldridge, S. J. Cohen, and K. Webber. 1974. University of Miami radiocarbon dates I. Radiocarbon 16:402-408. Takahashi, S. and H. Wada. 1999. Radiocarbon age determination at Shizuoka University (2). Geoscience Reports of Shizuoka Univer- sity 25 (in press). Yasuda, Y. 1987. The cold climate of the last glacial age in Japan. Daiyonki-Kenkyu 25:277-294. Date of manuscript acceptance: 15 January 1999 New acavid land snails from Madagascar Kenneth C. Emberton Molluscan Biodiversity Institute, 110 Old Airport Road, Concord, NC 28025, U. S. A. emberton@concordne.com Abstract: Descriptions are given of Ampelita akoratsara sp. nov., A. ambanianae sp. nov., A. analamerae sp. nov., A. anjanaharibei sp. nov., A. ivohibei sp. nov., A. josephinae sp. nov., A. masoalae sp. nov., A. ranomafanae sp. nov., A. raxworthyi sp. nov., A. (Eurystyla) griffithsi sp. nov., Clavator griffiths- jonesi sp. nov., C. masoalae sp. nov., and Helicophanta gargantua sp. nov. Discovery of A. (Eurystyla) griffithsi sp. nov. greatly expands the geographical, ecological, and morphological ranges of its subgenus. Madagascar’s 115 species of acavids are in drastic need of revision. Key words: Gastropoda, Stylommatophora, Acavacea, Ampelita, Clavator, Helicophanta This paper is the fifth in a series on the Acavidae of Madagascar (Emberton, 1990, 1994, 1995a, 1995b). A recent survey and inventory of Madagascar has yielded extensive collections of acavids; identifications are incom- plete but so far have yielded the thirteen new species described herein. With these additions, Madagascar’s rec- ognized species of acavids (Fischer-Piette et al., 1994) now total 115. METHODS AND MATERIALS Materials were collected 1992-1995. Identifications and comparisons were made using Fischer-Piette et al. (1994) and available collections. Measurements were made using vernier calipers. Due to ongoing habitat destruction and the urgency of making this fauna known to conserva- tionists and systematists, only conchological descriptions have been given, except for one case in which dissection of the genitalia was required for generic assignment. SYSTEMATICS Higher classification follows Nordsieck (1986). Latitudes and longitudes are given in degrees and minutes. Types are placed in the United States National Museum, Washington, D.C. (USNM); the Australian Museum, Sydney (AMS); the Muséum national d’ Histoire naturelle, Paris (MNHN, which does not assign catalog numbers to types); and the Academy of Natural Sciences of Philadelphia (ANSP). Prior, working catalog numbers of the Molluscan Biodiversity Institute (MBI), are also given, as they provide access to an ecological database of stations, available on request. MBI catalog numbers consist of sta- tion number, reference number of the species within the station, D (dry) or A (alcohol-preserved), and H (holotype) or P (paratype) or PR (paratype “representative” that is illustrated and/or described). In lot descriptions, ad means adult(s), juv juvenile(s). Class GASTROPODA Subclass PULMONATA Order STYLOMMATOPHORA Suborder SIGMURETHRA Infraorder ACHATINIDA Superfamily ACAVOIDEA Family ACAVIDAE Pilsbry, 1895 The Faune de Madagascar’s monograph on terres- trial pulmonate gastropods (Fischer-Piette et al., 1994) was a delayed, posthumous publication that overlooked some prior systematic changes within the Madagascan acavids. The most important of these were Mead’s (1985) transfer of Leucotaenius Martens, 1860, from Achatinidae to Acavidae, and Emberton’s (1990) reduction of Eurystyla Ancey, 1887, to a subgenus of Ampelita Beck, 1837; estab- lishment of the new subgenera Ampelita (Vesconis) Emberton, 1990, and A. (Xystera) Emberton, 1990; and transfer of Ampelita covani (Smith, 1879) to Rhytididae (provisionally to Rhytida Albers, 1860). Genus Ampelita Beck, 1837 Emberton’s (1990) four subgenera of Ampelita were based on single anatomical differences, most of his species assignments to those subgenera were provisional and based on shells, and one such assignment was later contradicted American Malacological Bulletin, Vol. 15(1) (1999):83-96 83 84 AMER. MALAC. BULL. 15(1) (1999) by allozyme evidence (Emberton, 1995a). The genus Ampelita is seriously in need of revision. Such a project would be aided by the anatomical and frozen-tissue materi- als now available for many species (Emberton, unpub- lished, this paper). In the meantime, however, there is no point in trying to assign these new species to subgenera, except for the one A. (Eurystyla). Fischer-Piette et al. (1994:87-89, 147), in contrast, put Ampelita species into six phenetic groups—counting A. (Eurystyla) as the sixth—based on gross shell morphology. Despite some inconsistencies, these are useful for identifi- cation purposes, so have been followed here. Species are arranged alphabetically within Fischer-Piette et al.’s (1994) groups. Written definitions of the groups were sketchy and inconsistent, so are clarified and corrected here, based on included species. Except for Group 6, these groups have no taxonomic validity. Group 1. Non-carinate, rounded periphery; low spired; umbilicus broad to narrow; peristome slightly to moderate- ly reflected; no dorsal, spiral gutter. Ampelita anjanaharibei sp. nov. Fig. | HOLOTYPE. USNM 880350 (ex MBI 693.50DH and AH, ad shell and pulled body in alc): 14°44’S, 49°26’E: Madagascar: Anjanaharibe Sud Reserve, 1750 m. 20 Oct. 1995. PARATYPE. AMS C.203526 (ex MBI 697.50DP, | ad): 14°44’S, 49°26’E: Madagascar: Anjanaharibe Sud Reserve, 1650 m. 20 Oct. 1995. DESCRIPTION OF HOLOTYPE. Diameter 41.5 mm, height 24.0 mm, whorls 4.6. Body-whorl periphery round- ed; suture moderately impressed, faintly crenulate; shoulder narrow, nearly a flat shelf. Umbilicus narrowly funneled, faintly enlarged by a rounded rim; width 5.4 mm (0.13 shell diameter). Aperture compressed-elliptical, height 11.7 mm, width 19.4 mm. Aperture downward deflection moderate, 0.1 whorl; face angle (relative to axis of coiling) 45°. Apertural lip reflection narrow to moderately wide, widest above; sharp edge rolled back and nearly under. Embryonic whorls 2.2; embryonic sculpture nearly smooth, then with minute pustules. Shell with a satin-like sheen, covered with small pustules, and with low, transverse growth lines and faint spiral lines. Color chocolate brown with rare, light- yellow, transverse-linear flecks, with a brownish yellow subsutural band that has a distinct upper edge and an indis- tinct lower edge, and with a narrow yellow-brown band on the umbilical rim. VARIATION. Single paratype virtually identical to holo- type, but with broken lip. COMPARISONS. Very similar in size and shape to some Ampelita subatropos (Dautzenberg, 1894), but with a much narrower umbilicus and more rapid whorl expansion, and with a glossier, more rugose, less spirally engraved sculp- ture. Rounder whorls, tighter coiling, and smaller umbilicus than A. gaudens (Mabille, 1884). Completely lacks the spi- ral, dorsal gutter of other Ampelitas of the same size. ETYMOLOGY. For Anjanaharibe Reserve, the type local- ity. Ampelita ivohibei sp. nov. Fig. 2 HOLOTYPE. USNM 880351 (ex MBI 1499.50DH and AH, ad shell and pulled body in alc): 24°34’S, 47°12’E: Madagascar: Ivohibe Forest, 570 m. 30 Oct. 1992. PARATYPE. ANSP 401985 (ex MBI 1504.50DP, subadult shell): 24°34’S, 47°12’E: Madagascar: Ivohibe Forest, 400 m. 30 Oct. 1992. DESCRIPTION OF HOLOTYPE (preserved in ethanol before the body was removed, therefore with some dissolu- tion pits and breakage). Diameter 20.3 mm, height 10.8 mm, whorls 3.8. Body-whorl periphery rounded; suture very strongly and deeply impressed, simple; shoulder broad, flatly rounded. Umbilicus funneled, faintly enlarged by a slight, rounded rim; width 2.3 mm (0.11 shell diame- ter). Aperture compressed-elliptical, height 5.1 mm, width 8.9 mm. Aperture downward deflection moderate, <0.1 whorl; face angle (relative to axis of coiling) 55°. Apertural lip reflection narrowly and evenly rolled, thin edged. Embryonic whorls 2.2; embryonic sculpture apparently smooth. Shell somewhat glossy, with faint, irregular growth lines—otherwise smooth. Coloration consisting of trans- verse, irregular stripes of dark brown and ivory, overlain with four interrupted, peripheral bands of ivory. COMPARISONS. Unique for its tiny size and flat spire. Most similar to Ampelita anosiana Fischer-Piette, Blanc, Blanc, and Salvat, 1994, which is larger, with looser coil- ing, broader umbilicus, and sunken apex. ETYMOLOGY. For Mount Ivohibe, also known as Mount Varabe, near the type locality. Ampelita masoalae sp. nov. Fig. 3 HOLOTYPE. USNM 880352 (ex MBI 309.50DH and AH, ad shell and pulled body in alc): 15°33’S, 50°0’E: Madagascar: Masoala National Park, 1000 m: hardwood rainforest with palms, pandanus, and tree moss. 26 Sep. 1995: PARATYPES. AMS C.203484 (ex MBI 309.50DP, 1 ad): type locality. MNHN (ex MBI 309.50DP, 1 ad): type locality. ANSP 401986 (ex MBI 309.50DP, 1 ad): type locality. USNM 880353 (ex MBI 309.50DPR [1 juv], MBI 309.5O0DP [18 ad, 18 juv], and MBI 309.50AP [1 juv]): type locality. USNM 880400 (ex MBI 644.50DP, 4 ad, 2 juv & frag): 14°28’S, 49°34’E: Madagascar: Marojejy Reserve, incidental collecting along trail, 1050 m aver- age elevation, 27 Sep. 1995. EMBERTON: ACAVID LAND SNAILS FROM MADAGASCAR Figs. 1-2. Fig. 1. Ampelita anjanaharibei sp. nov. holotype. Fig. 2. A. ivohibei sp. nov. holotype. Scale bars 5 mm. 85 86 AMER. MALAC. BULL. 15(1) (1999) Figs. 3-4. Fig. 3. Ampelita masoalae sp. nov. holotype. Fig. 4. A. raxworthyi sp. nov. holotype. Scale bars 5 mm. EMBERTON: ACAVID LAND SNAILS FROM MADAGASCAR 87 DESCRIPTION OF HOLOTYPE (apex eroded). Diameter 30.0 mm, height 17.3 mm, whorls 4.6. Body- whorl periphery sharply rounded, with a faint trace of angu- lation; suture moderately impressed, simple; shoulder broadly rounded. Umbilicus broadly funneled, with only a slight trace of a rim; width 4.9 mm (0.16 shell diameter). Aperture compressed-elliptical, height 7.0 mm, width 11.7 mm. Aperture downward deflection moderate, 0.1 whorl; face angle (relative to axis of coiling) 40°. Apertural lip reflection narrowly to moderately wide, widest above; sharp edge rolled back and nearly under. Embryonic whorls 1.9. Shell with a faint sheen, bearing moderate-sized pustules, faint growth lines, and traces of dense spiral lines. Color dark yellow with three narrow bands of reddish brown; umbilicus with diffuse red-brown splotches; peristome and apertural interior white. EMBRYONIC SCULPTURE (Paratype USNM 880353). Embryonic sculpture nearly smooth, but with faint trans- verse ribs and even fainter traces of spiral lines. VARIATION. Shells from MBI 644 considerably flatter (H/D 0.50) and with a proportionally wider aperture (AW/D 0.42) and slightly more flared upper apertural lip; some paratypes lack color bands; adult diameters range 27.6-33.1 mm. COMPARISONS. More tightly coiled than Ampelita futu- ra Fischer-Piette and Garreau, 1965, the holotype of which has about three whorls (Fischer-Piette et al., 1994:plate V, fig. 12), not four as described (Fischer-Piette et al.. 1994:90). Much more compressed whorls than A. subsepul- chralis (Crosse, 1868). Differs from A. consanguinea (Férussac, 1851) and others of similar size and shape by its relatively wide, white peristome, in addition to other details. ETYMOLOGY. For Masoala National Park, the type locality. COMMENTS. Apparently restricted to high elevations. The Masoala and Marojejy populations thus appear to be extremely isolated and could be separate species or sub- species. Ampelita raxworthyi sp. nov. Fig. 4 HOLOTYPE. USNM 880354 (ex MBI 6.50DH and AH, ad shell and pulled body in alc): 24°46’S, 47°9’E: Madagascar: Forét Sainte Luce, 10 m: coastal rainforest. 29 Jan. 1995. PARATYPES. AMS C.203485 (ex MBI 6.50DP, | ad): type locality. MNHN (ex MBI 6.50DP, | ad): type locality. ANSP 401987 (ex MBI 6.50DP, 1 ad): type locality. USNM 880401 (ex MBI 6.50DP [4 ad, 6 juv] and MBI 7.50AP [1 juv, beat from veg- etation, in alc]): type locality. DESCRIPTION OF HOLOTYPE. Diameter 22.0 mm, height 14.5 mm, whorls 4.2. Body-whorl periphery round- ed; suture moderately impressed, simple; shoulder narrow, nearly flat. Umbilicus a narrow pit, with only a slight trace of a rim; width 2.9 mm (0.13 shell diameter). Aperture compressed-elliptical, height 6.2 mm, width 9.1 mm. Aperture downward deflection moderate, 0.2 whorl; face angle (relative to axis of coiling) 50°. Apertural lip reflec- tion narrowly and evenly rolled, thin edged. Embryonic whorls 2.1; embryonic sculpture nearly smooth in the first whorl, then with faint riblets. Shell slightly glossy, almost smooth, but with faint, irregular growth lines and obscure traces of spiral lines. Color yellow, with three dark reddish brown bands: one narrow subsutural, one broad suprape- ripheral, and one medium-broad subperipheral. VARIATION. Smallest shell diameter 18.7 mm. COMPARISONS. The small, compact, globose, umbili- cate shape is approached only by Ampelita parva Fischer- Piette and Garreau, 1965, which is more domed and tightly coiled, and A. petiti Fischer-Piette, 1952, which is also more tightly coiled and has an unreflected upper peristome. A. globulus Fischer-Piette, Blanc, and Vukadinovic, 1974, bears some resemblance, but has a proportionally much larger aperture and looser coiling. ETYMOLOGY. For Dr. Chris Raxworthy, herpetologist and biogeographer, who suggested collecting at the type locality. Group 2 (no new species). Sub-carinate to non-carinate; low spired; umbilicus moderate to broad; peristome moder- ately reflected; no dorsal, spiral gutter. Group 3. Carinate; low spired; umbilicus narrow or imper- forate; peristome moderately reflected; no dorsal, spiral gutter. Ampelita ambanianae sp. nov. Figs. 5, 6 HOLOTYPE. USNM 880355 (Fig. 5, ex MBI 315.50DH, ad shell): 15°40’°S, 49°58’°E: Madagascar: Masoala Peninsula: near Mount Ambaniana, trail to Andranobe, 0 to 300 m. 29 Sep. 1995. PARATYPES. AMS C.203486 (ex MBI 315.50DP, | ad): type locality. MNHN (ex MBI 315.50DP, | ad): type locality. ANSP 401988 (ex MBI 315.50DP, 1 ad): type locality. USNM 880356 (Fig. 6, ex MBI 315.50DPR [ad shell]; and ex MBI 315.50DP [2 frag]): type locality. DESCRIPTION OF HOLOTYPE (a weathered shell retaining about half of the body-whorl periostracum). Diameter 39.4 mm, height 23.2 mm, whorls 4.7. Body- whorl periphery carinate, carina wide, blunt-edged, and shallowly guttered above and below; suture shallowly impressed, bordered on either side by broad, shallow gut- ters; shoulder narrow, rounded, then dropping off steeply into a shallow gutter. Umbilicus narrow, steep-sided, faintly enlarged by a slightly angular rim; width 2.9 mm (0.07 88 AMER. MALAC. BULL. 15(1) (1999) Figs. 5-7. Figs. 5-6. Ampelita ambanianae sp. nov.: Fig. 5 holotype in three views; Fig. 6 paratype small adult in one view. Fig. 7. A. analamerae sp. nov. holotype. Scale bars 5 mm. EMBERTON: ACAVID LAND SNAILS FROM MADAGASCAR 89 shell diameter). Aperture compressed-elliptical, height 10.8 mm, width 18.9 mm. Aperture downward deflection mod- erate, <0.1 whorl; face angle (relative to axis of coiling) 40°. Apertural lip reflection narrow, thin, incompletely rolled, narrow above, wider below. Shell sculpture consist- ing of parallel, slightly wavy, spiral lines; faint, broad growth ridges; and obscure, short, oblique cut marks that sometimes give a herringbone appearance. Color yellowish brown, sometimes with a faintly greenish cast. VARIATION. Largest diameter 41.3 mm, most elevated shell H/D 0.60. COMPARISONS. Differs from all others of Fischer-Piette et al.’s (1994) group 3 species by its two broad, dorsal, spi- ral channels, and by its round-edged, cord-like carina. In general shape it most resembles Ampelita stumpfii (Kobelt, 1880), but it has much looser coiling. Its size, coiling, color, and sculpture are similar to those of A. lancula (Férussac, 1821), from which it differs—in addition to its unique characters—by its more rounded aperture and greater pre-apertural downward deflection. ETYMOLOGY. For Mount Ambaniana, near the type locality. Group 4. Carinate; low spired; umbilicus broad; peristome narrowly reflected; no dorsal, spiral gutter. Ampelita akoratsara sp. nov. Fig. 8 HOLOTYPE. USNM 880357 (ex MBI 657.50DH and AH, ad shell and pulled body in alc): 140°32’S, 49°42’E: Madagascar: near Marojejy Reserve, Ambatosorotra Mountain, 800 m: rainforest. 4 Oct. 1995. PARATYPES. AMS C.203487 (ex MBI 648.50DP, 1 ad): 14°29°S, 49°33’E: Madagascar: Marojejy Reserve W, 805 m: rainforest; 28 Sep. 1995. MNHN (ex MBI 626.50DP, | ad): 14°20’S, 49°35’E: Madagascar: Marojejy Reserve W, incidental along trail, 950 m average elevation: rainforest; 24 Sep. 1995. ANSP 401989 (ex MBI 626.50DP, 1 ad): 14°20°S, 49°35’E: Madagascar: Marojejy Reserve W, incidental along trail, 950 m average elevation: rainforest; 24 Sep. 1995. USNM 880402, 880404 (ex MBI 626-631DP, 2 lots; total 2 ad, 2 juv): 140S, 490E: Madagascar: Marojejy Reserve W, incidental along trail, 950-1125 m: rainforest; Sep. 1995. USNM 880403, 880405, 880406 (ex MBI 657-674DP, 3 lots; total 2 ad, 1 juv): 14°32’S, 49°42’°E: Madagascar: near Marojejy Reserve, Ambatosorotra Mountain, 800-940 m: rainforest; Oct. 1995. USNM 880407, 880408 (ex MBI 704-705DP, 2 lots; total 1 ad, 1 juv): 14°45’S, 49°28°E: Madagascar: Anjanaharibe Sud Reserve, 1100-1185 m: rainforest; 23 Oct. 1995. DESCRIPTION OF HOLOTYPE. Diameter 27.6 mm, height 13.1 mm, whorls 4.0. Body-whorl periphery sharply carinate; suture shallowly impressed, simple; shoulder nar- row, flattish. Umbilicus funneled, enlarged by a sharply angular rim; width 3.8 mm (0.14 shell diameter). Aperture compressed-elliptical, height 5.7 mm, width 11.7 mm. Aperture downward deflection great, 0.1 whorl; face angle (relative to axis of coiling) 60°. Apertural lip reflection nar- row, thin, rolled, even. Embryonic whorls 2.0; embryonic sculpture nearly smooth, with slight, irregular pitting. Shell slightly glossy, with parallel spiral grooves transected by slightly fainter, more irregular growth lines. Ground color light brown; periphery and suture with a single thick band of white edged with dark brown; umbilicus slightly darker brown than ground color and edged with a band of light beige; peristome light beige. VARIATION. Adult paratype diameters range 26.3-29.5 mm; less extreme preapertural downward deflections occur in several paratypes. COMPARISONS. Among carinate, widely umbilicate Ampelita (Fischer-Piette et al.’s, 1994, group 4), most simi- lar in size, shape, and general sculpture to A. ranomafanae sp. nov., but differs in its smooth (vs. pustulose) embryonic sculpture, its less shelved carina, and its greater pre-aper- tural deflection. Within group 4, A. akoratsara sp. nov. and A. ranomafanae sp. nov. share their spiral-groove sculpture only with A. namerokoensis Fischer-Piette, 1952, which is much more tightly coiled. ETYMOLOGY. For the beautiful (Malagasy “tsara”’) shell (Malagasy “akora”’). Ampelita analamerae sp. nov. Fig. 7 HOLOTYPE. USNM 880358 (ex MBI 201.50DH, ad shell): 12°44’S, 49°30°E: Madagascar: Analamera Reserve, 315 m: dry deciduous forest. 15 Jul. 1995. PARATYPES. AMS C.203488 (ex MBI 201.50DP, 1 ad): type locality. MNHN (ex MBI 201.50DP, | ad): type locality. ANSP 401990 (ex MBI 201.50DP, | ad): type locality. USNM 880409 (ex MBI 201.50DP, 14 ad, 45 juv & frag): type locality. DESCRIPTION OF HOLOTYPE. Diameter 27.2 mm, height 13.6 mm, whorls 4.8. Body-whorl periphery cari- nate; suture deeply impressed, very slightly crenulate; shoulder moderate, flat. Umbilicus funneled, faintly enlarged by a slightly angular rim; width 4.5 mm (0.17 shell diameter). Aperture compressed-elliptical, height 5.3 mm, width 11.3 mm. Aperture downward deflection strong, <0.1 whorl; face angle (relative to axis of coiling) 55°. Apertural lip reflection grading from no reflection at the suture, to narrowly reflected at the columella; edge sharp. Embryonic whorls 2.2; embryonic sculpture nearly smooth, with very faint riblets. Shell sculpture consists of closely spaced, even, sharp, transverse ridges that bear, when fresh, periostracal lamellar-like extensions. Color an even yellow- ish brown, lighter in the umbilicus. VARIATION. Adult paratype diameters range 25.2-28.5 mm. COMPARISONS. Among Fischer-Piette ef al.'s (1994) 90 AMER. MALAC. BULL. 15(1) (1999) Figs. 8-9. Fig. 8. Ampelita akoratsara sp. nov. holotype. Fig. 9. A. ranomafanae sp. nov. holotype. Scale bars 5 mm. EMBERTON: ACAVID LAND SNAILS FROM MADAGASCAR 91 group 4 (carinate, widely umbilicate) species, unique for its distinctive sculpture. Most closely resembles Ampelita namerokoensis Fischer-Piette, 1952, in size and coiling tightness, but is higher spired and more narrowly umbili- cate. In its strong pre-apertural deflection it resembles A. bathiei Fischer-Piette, 1952, and A. akoratsara sp. nov., but is much more tightly coiled, and its upper apertural lip is unreflected. ETYMOLOGY. For Analamera Reserve, the type locality. Ampelita ranomafanae sp. nov. Fig. 9 HOLOTYPE. USNM 880359 (ex MBI 459.50DH, ad shell): 21°13’S, 47°25’E: Madagascar: Ambatolahy, adja- cent to Ranomafana National Park, 850 m: rainforest. 9 Oct. 1995. PARATYPE. USNM 880410 (ex MBI 459.50DP, 1 juv): type locality. DESCRIPTION OF HOLOTYPE (originally with about a third of its periostracum, most of which has flaked off; shell somewhat eroded). Diameter 28.2 mm, height 12.9 mm, whorls 4.1. Body-whorl periphery very sharply cari- nate, upper surface of carina broad and nearly flat; suture shallowly impressed, simple; shoulder narrow, flatly round- ed. Umbilicus steep-sided, enlarged by an angular, raised rim; width 4.4 mm (0.16 shell diameter). Aperture com- pressed-elliptical, height 7.0 mm, width 11.1 mm. Aperture downward deflection slight, <0.1 whorl; face angle (relative to axis of coiling) 45°. Apertural lip reflection narrow, thin, incompletely rolled, even in width. Embryonic whorls 2.0; embryonic sculpture apparently initially smooth, then pus- tulose. Shell slightly glossy, with parallel spiral grooves transected by slightly stronger growth lines; appearance almost beaded in places. Color yellowish brown, with cari- na, suture, and peristome a dark, purplish brown. COMPARISONS. See comparisons under Ampelita ako- ratsara sp. Nov. ETYMOLOGY. For Ranomafana National Park, adjacent to the type locality. Group 5. Non-carinate, rounded periphery; low spired; umbilicus broad; peristome broadly flared; with or without dorsal, spiral gutter. Ampelita josephinae sp. nov. Fig. 10 HOLOTYPE. USNM 880360 (ex MBI 357.50DH and AH, ad shell and pulled body in alc): 16°19°S, 49°46’E: Madagascar: W of Sahasoa, 330 m: hardwood rainforest. 21 Oct. 1995. PARATYPES. AMS C.203489 (ex MBI 355.50DP, | ad): 16°19°S, 49°44’°E: Madagascar: W of Mount Andaitra, 510 m: hardwood and pandanus; 19 Oct. 1995. MNHN (ex MBI 351.50DP, 1 ad): 16°19’S, 49°44’E: Madagascar: summit Mount Andaitra, 515 m: hardwood and pandanus; 18 Oct. 1995. ANSP 401991 (ex MBI 351.50DP, | ad): 16°19°S, 49°44’E: Madagascar: summit Mount Andaitra, 515 m: hardwood and pandanus; 18 Oct. 1995. USNM 880411-880420 (ex MBI 347-357DP, 10 lots; total 24 ad, 5 juv): 16°S, 49°E: Madagascar: Mount Andaitra area, 300- 515 m; Oct. 1995. DESCRIPTION OF HOLOTYPE. Diameter 37.2 mm, height 19.7 mm, whorls 4.1. Body-whorl periphery broadly rounded; suture moderately impressed, simple; shoulder broad, gently rounded. Umbilicus a pinhole, then rapidly expanding in the last half whorl; artificially enlarged by a rounded rim offset internally by two parallel grooves; width 2 mm (0.07 shell diameter). Aperture broadly auriculate, height 9.3 mm, width 15.5 mm. Aperture downward deflec- tion strong, 0.1 whorl; face angle (relative to axis of coil- ing) 60°. Apertural lip very broadly and flatly reflected, wider above; edge sharp, rolled back. Embryonic whorls 2.2; embryonic sculpture nearly smooth, then with faint, interrupted, tranverse riblets. Shell with a satin-like sheen, covered with small, low pustules, and with faint, irregular growth lines. Color a dark, vivid yellow, with light-yellow flecks; apex and inner and outer peristome dark brownish purple; outer edge of peristome white; apertural interior white; the shell is white where peristome has flaked off. VARIATION. All adult specimens very similar in size and shape. COMPARISONS. Other than Ampelita perampla Dautzenberg, 1907 (as redefined by Fischer-Piette et al., 1994), this is the only flared-lip Ampelita without a con- spicuous dorsal, spiral gutter. It differs from A. perampla in its more rapid post-embryonic whorl expansion (diameter of first three whorls 12.8 mm vs. 9.9 mm), its lower spire, its fainter and sparser pustulate sculpture, its proportionally smaller, less flared aperture, and its subtly different umbili- cus. ETYMOLOGY. For Josephine Djaohasara Emberton, the author’s wife, who helped him collect this lovely species. Group 6 = Ampelita (Eurystyla) Ancey, 1887. Non-cari- nate, rounded periphery; high spired; imperforate to creviced; peristome moderately reflected; no dorsal, spiral gutter. Ampelita (Eurystyla) griffithsi sp. nov. Figs. 11, 12 HOLOTYPE. USNM 880361 (ex MBI 250.50DH and AH, ad shell and pulled body in alc): 19°8°S, 44°49°E: Madagascar: S Bemaraha Reserve, 80 m: river gallery for- est. 16 June 1995. PARATYPES. AMS C.203490 (ex MBI 247.50DP, 3 ad, 6 juv and frag): 20°3’S, 44°39’E: Madagascar: Kirindy, 40 m: river gallery forest; 15 June 1995. AMS C.203491 (ex MBI 249.50DP, 92 AMER. MALAC. BULL. 15(1) (1999) Figs. 10-12. Fig. 10. Ampelita josephinae sp. nov. holotype. Figs. 11-12. A. (Eurystyla) griffithsi sp. nov. holotype: Fig. 11 shell in two views; Fig. 12 geni- talia (abbreviations: e epiphallus, g genital pore, m penial retractor muscle, p penis, r prostate, s spermatheca, u uterus, v vas deferens). Scale bars 5 mm. EMBERTON: ACAVID LAND SNAILS FROM MADAGASCAR 93 4 ad, 5 juv and frag): 19°8’S, 44°50’E: Madagascar: Bemaraha Reserve: river gallery forest; 16 June 1995. AMS C.203492 (ex MBI 494.50DP, 2 ad, | juv): 18°45’S, 440°45’E: Madagascar: N Bemaraha Reserve, 280 m: semideciduous forest; 29 June 1996. AMS C.203493 (ex MBI 495.50DP, 1 ad, 2 juv): 18°47°S, 44°47°E: Madagascar: N Bemaraha Reserve, 300 m: deciduous scrub; 29 June 1996. MNHN (ex MBI 249.50DP, 1 ad): 19°8°S, 44°50’E: Madagascar: Bemaraha Reserve: river gallery forest; 16 June 1995. ANSP 401992 (ex MBI 247.50DP, 1 ad): 20°3’S, 44°39’E: Madagascar: Kirindy, 40 m: river gallery forest; 15 June 1995. Nationnal Natuurhistorisch Museum, Leiden, the Netherlands 59149 (1 ad in alcohol): 19°9’S, 44°49°E: Madagascar: Bemaraha Reserve. USNM 880421-880426 (ex MBI 247-251DP and 488-495DP, 6 lots; total 10 ad, 21 juv and frag): 20°3’S, 44°39°E: Madagascar: Kirindy and Bemaraha Reserve; 1995, 1996. DESCRIPTION OF HOLOTYPE. SHELL (a very thin, brownish yellow periostracum remains on some paratypes but is lost from the holotype). Diameter 22.2 mm, height 23.6 mm, whorls 4.2. Body-whorl periphery broadly round- ed; suture moderately impressed, simple; shoulder falling off steeply, broadly rounded. Umbilicus a crevice; width 0.8 mm (0.04 shell diameter). Aperture oval, height 10.6 mm, width 10.8 mm. Aperture downward deflection slight, 0.2 whorl; face angle (relative to axis of coiling) 30°. Apertural lip reflection narrowly and evenly rolled, expand- ed at the columellar insertion. Embryonic sculpture smooth, then with faint, wavy riblets. Shell with a faint sheen, cov- ered with small pustules, and with low growth lines and traces of spiral lines. Color whitish yellow, with four thin, dark-brown bands. GENITALIA. Genital pore far forward, just behind the right tentacle. Right tentacular retractor muscle passes between the penis and the vagina. Atrium small, without appendages. Penis 10 mm in length, about six times as long as wide. Penial retractor muscle insertion on the penial apex, origin on the interior body wall near the junction of the left mantle collar, origin apparently enveloped by a small, thin sheath. Vas deferens long, convoluted, bound to the penis by connective-tissue strands. Epiphallus bulbous, thin-walled, adherent to the penis. Vagina about two-thirds the length of the penis. Lower spermathecal (bursal, game- tolytic) duct slightly swollen; upper duct slender, opening into a small, globular spermatheca (bursa copulatrix, game- tolytic gland); spermatheca plus duct about the same length as the penis. VARIATION. Adult shells more elevated at MBI 247 and 249 (H/D 1.1), smaller at MBI 494 and 495 (smallest diam- eter 19.2 mm); all paratypes lack columellar apertural node. COMPARISONS. In size and shape most similar to Ampelita (Eurystyla) viridis (Deshayes, 1838), but with a pronouncedly sharper spire, tighter coiling, exposed umbili- cus, and much different sculpture. In penial morphology, very similar to A. (E.) cerina (Morelet, 1877) (Fischer- Piette and Garreau de Loubresse, 1965:fig. 20). ETYMOLOGY. For Owen Griffiths, collector of this species. COMMENTS. Discovery of Ampelita (Eurystyla) griffithsi sp. nov. greatly expands the geographical, ecological, and morphological ranges of its subgenus. Genus Clavator von Martens in Albers, 1860 Clavator griffithsjonesi sp. Nov. Figs. 13, 14 HOLOTYPE. USNM 880362 (Fig. 14, ex MBI 483.50DH, ad shell; and ex MBI 483.50AH (pulled body in alc): 18°1°S, 44°31°E: Madagascar: N Bemaraha Reserve. 22 June 1996. PARATYPES. AMS C.203494 (ex MBI 483.50DP, 2 ad, 2 juv): type locality. MNHN (ex MBI 483.50DP, | ad): type locality. ANSP 401993 (ex MBI 483.50DP, | ad): type locality. USNM 880363 (Fig. 13, ex MBI 483.50DPR, juv shell; and ex MBI 483.50DP [2 juv & frag] and AP [1 ad in alc]): type locality. USNM 880427 (ex MBI 484.50DP, 1 ad): 18°3°S, 44°317E: Madagascar: N Bemaraha Reserve; 23 June 1996. DESCRIPTION OF HOLOTYPE (apex broken). Diameter 32.7 mm, height 91.4 mm, whorls 9.6 (estimat- ed). Body-whorl periphery flatly rounded; suture strongly impressed, very slightly guttered, slightly crenulate; shoul- der narrow, steeply sloped. Umbilicus a narrow crevice; width 0.3 mm (0.01 shell diameter). Aperture slightly auriculate, height 28.3 mm, width 15.5 mm. Aperture downward deflection slight, <0.01 whorl; face angle (rela- tive to axis of coiling) 10°. Apertural lip reflection slight, thick, and even but broadened at the columella. Shell some- what glossy, with fine, closely spaced riblets and irregularly spaced growth-interruption lines; a faint subsutural line sometimes detectable. Color light yellow-brown to almost white, with occasional streaks and splotches of darker, somewhat reddish brown. EMBRYONIC-SHELL CHARACTERS (Paratype USNM 880363). Embryonic whorls 4.2; embryonic sculp- ture nearly smooth in the first whorl; subsequent whorls with closely spaced riblets; subsuture can be slightly cord- ed. VARIATION. Only slight variation in height and apertural shape. COMPARISONS. Unique in its shape and strongly ribbed sculpture; superficially somewhat like a giant C. moreleti (Férussac, 1851); much more tightly coiled and straight- sided than C. clavator (Petit de la Saussaye, 1844), C. grandidieri (Crosse and Fischer, 1868), and C. anteclavator Fischer-Piette, 1963, which it can resemble somewhat in color. ETYMOLOGY. For Owen Griffiths (Bioculture Mauritius) and Dr. Carl Jones (Jersey/Mauritius Wildlife Appeal Fund), collectors of this species. 94 AMER. MALAC. BULL. 15(1) (1999) Figs. 13-19. Figs. 13-14. Clavator griffithsjonesi sp. nov.: Fig. 13 juvenile paratype with complete apex; Fig. 14 holotype adult with broken apex. Fig. 15. Helicophanta gargantua sp. nov. holotype in two views. Figs. 16-19. Clavator masoalae sp. nov.: Fig. 16 holotype adult with naturally truncate apex; Fig. 17 paratype adult with truncate apex; Fig. 18 paratype small juvenile with complete apex; Fig. 19 paratype large juvenile with broken apex. Scale bars 5 mm. EMBERTON: ACAVID LAND SNAILS FROM MADAGASCAR 2. Clavator masoalae sp. nov. Figs. 16-19 HOLOTYPE. USNM 880364 (Fig. 16, ex MBI 309.51DH, ad shell; and ex MBI 309.51AH, ad body, partially decayed, in alc): 15°33’S, 50°0’E: Madagascar: Masoala National Park, 1000 m: hardwood rainforest with palms, pandanus, and tree moss. 26 Sep. 1995. PARATYPES. AMS C.203495 (ex MBI 309.51DP, 1 ad, 1 juv): type locality. MNHN (ex MBI 309.51DP, | ad, | juv): type locali- ty. ANSP 401994 (ex MBI 309.51DP, 1 ad, 1 juv): type locality. USNM 880365 (Fig. 17, ex MBI 744.50DPR, 1 ad; and ex MBI 744.50AP, pulled body in alc): 16°46’S, 49°8’E: Madagascar: Ambatovaky Reserve, 1025 m: rainforest with pandanus; 21 Nov. 1995. USNM 880366 (Fig. 19, ex MBI 296.50DPR, | juv; and ex MBI 296.50AP, 1 juv and pulled juv body in alc): 15°47’S, 50°3’E: Madagascar: Masoala National Park, 310-450 m; 19 Sep. 1995. USNM 880367 (Fig. 18, ex MBI 605.50DPR, | juv): 14°26’S, 49°45’E: Madagascar: Marojejy Reserve, 1200 m: rain- forest. ; 16 Sep. 1995. USNM 880428 (ex MBI 308.50DP [1 ad, 2 juv] and AP (1 ad, partially decayed, in alc): 15°33’S, 50°0’E: Madagascar: Masoaia National Park, 680-1000 m; 26 Sep. 1995. USNM 880429 (ex MBI 309.51DP [5 ad, 7 juv] and AP (1 juv in alc): type locality. USNM 880430-880434 (ex MBI 593-644DP, 5 lots; total 5 ad, 18 juv): 14°S, 49°E: Madagascar: Marojejy Reserve, 900-1350 m: rainforest; Sep. 1995. USNM 880435- 880440 (ex MBI 741-749DP, 6 lots; total 9 ad, 5 juv): 16°S, 49°E: Madagascar: Ambatovaky Reserve, 870-1055 m; Nov. 1995. DESCRIPTION OF HOLOTYPE. Diameter 36.0 mm, height 105.9 mm, apex naturally truncated, whorls estimat- ed at approximately 9.4. Body-whorl periphery flatly rounded; suture strongly impressed, slightly crenulate; shoulder narrow, steeply sloped. Umbilicus imperforate, with a translucent callus. Aperture slightly auriculate, height 35.2 mm, width 19.3 mm. Aperture downward deflection slight, <0.01 whorl; face angle (relative to axis of coiling) 10°. Apertural lip reflection slight, thick, even but broadened at the columella. Shell somewhat glossy, with frequent, low, fairly regular growth lines crossed by fairly regularly spaced spiral lines. Color of embryonic whorls light yellow-brown; later whorls becoming progressively more reddish-brown with darker transverse streaks; body whorl reddish brown; apertural lip white; apertural interior bluish white. EMBRYONIC-SHELL CHARACTERS (Paratypes USNM 880366 and 880367). Embryonic whorls 4.9; embryonic sculpture of faint riblets in the first whorl; sub- sequent whorls with closely spaced riblets cut by spiral lines to produce a strongly beaded appearance. VARIATION. Aperture wide in some specimens, mini- mum H/W 1.6; shells from Ambatovaky Reserve are slen- derer (Fig. 17). COMPARISONS. In form, color, and sculpture, superfi- cially resembles C. eximius (Shuttleworth, 1852); C. din- geoni Fischer-Piette, Blanc, and Salvat, 1975; C. pauliani Fischer-Piette, 1963; and C. bathiei Fischer-Piette, 1963; but conspicuously tighter coiling than the first three and slightly tighter coiling, a more acute apex, a proportionally larger aperture, and stronger spiral sculpture than C. bathiei. ETYMOLOGY. For Masoala National Park, the type locality. Genus Helicophanta Férussac, 1821 Helicophanta gargantua sp. nov. Fig. 15 HOLOTYPE. USNM 880368 (ex MBI 1402.50DH and AH, ad shell and pulled body in alc): 22°4’S, 46°54’E: Madagascar: near Andringitra Reserve. 3 Oct. 1992. PARATYPES. AMS C.203758 (ex MBI 1402.50DP, | ad): type locality. MNHN (ex MBI 1402.50DP, 1 ad): type locality. ANSP 401995 (ex MBI 1401.50DP, 1 ad): type locality. pre-1992. ANSP 401996 (ex MBI 1402.50DP, 17 ad, 11 juv): type locality. ANSP A19000 (ex MBI 1402.50AP, 2 ad, 2 juv): type locality. USNM 880369 (ex MBI 1402.50DPR, juvenile broken shell): type locali- ty. DESCRIPTION OF HOLOTYPE (apex eroded, so juve- nile paratype USNM 880369 consulted to confirm embry- onic sculpture). Diameter 90.9 mm, height 79.3 mm, whorls 4.5. Body-whorl periphery broadly rounded; suture strongly impressed, simple; shoulder steeply sloped, flat- tened. Umbilicus imperforate, with a thick white callus. Aperture broadly elliptical, height 60.6 mm, width 55.8 mm. Aperture downward deflection extreme, 0.5 whorl; face angle (relative to axis of coiling) 40°. Apertural lip reflection slight, thick, even. Embryonic whorls 3.5; embry- onic sculpture consisting of elongate, transverse pustules arrayed as growth lines. Shell glossy, with strong, nearly regular growth lines; faint, supraperipheral spiral cords; and small, obscure, subperipheral, oblique cut marks. Color very dark brown, with reddish undertint; spiral cords darker brown; light beige where periostracum eroded; apertural lip and interior white. COMPARISONS. Most similar in form and sculpture to H. gloriosa (Pfeiffer, 1856), from which it differs in its tighter initial coiling (10.0 vs 11.2 mm diameter of first 2.5 whorls) but much larger embryonic shell (3.6 vs. 3.2 whorls) and adult shell. ETYMOLOGY. For the very large (Rabelais’s fictional giant, Gargantua) shell size. ACKNOWLEDGMENTS Funded by the U. S. National Science Foundation (DEB 9201060), with some additional funding provided by Owen Griffiths. Permits were issued by the Madagascar government agencies DEF and 96 AMER. MALAC. BULL. 15(1) (1999) ANGAP. Ranomafana National Park Project gave logistic support. Marojejy and Anjanaharibe expeditions were led by Dr. Tim Pearce; Bemaraha by Owen Griffiths; Ivohibe by Ruffin Arijaona; Ambatolahy- Ranomafana National Park by Roger Randalana; and the remainder by the author. Major assistance on one or more of the Masoala, Analamera, Sahasoa-Andaitra, Sainte-Luce, and Andringitra expeditions was provided by Max Felix Rakotomalala (requiescat in pace), Tim Pearce, Jean Rakotoarison, and Ruffin Arijaona. Many local Malagasy guides and col- lectors also rendered service. Dr. A. J. de Winter lent the Netherlands’ Nationnal Natuurhistorisch Museum’s specimen of Ampelita (Eurystyla) griffithsi sp. nov. LITERATURE CITED Emberton, K. C. 1990. Acavid land snails of Madagascar: subgeneric revi- sion based on published data (Gastropoda: Pulmonata: Stylommatophora). Proceedings of the Academy of Natural Sciences of Philadelphia 142:15-31. Emberton, K. C. 1994. Morphology and aestivation behaviour in some Madagascan acavid land snails. Biological Journal of the Linnean Society 53:175-187. Emberton, K. C. 1995a. Phylogenetic analysis of 18 species of Madagascan acavid land snails using allozyme characters. The Veliger 38:1-7. Emberton, K. C. 1995b. Distributional differences among acavid land snails around Antalaha, Madagascar: inferred causes and dangers of extinction. Malacologia 36:67-77. Fischer-Piette, E., C. P. Blanc, F. Blanc, and F. Salvat. 1994. Gastéropodes terrestres pulmonés. Faune de Madagascar 83:1- 551. Fischer-Piette, E. and N. Garreau de Loubresse. 1965. Mollusques ter- restres de Madagascar Famille Acavidae. Journal de Conchyliologie, Paris 104:129-160. Mead, A. R. 1985. Anatomical studies transfer Leucotaenius from Achatinidae to Acavidae (Pulmonata: Sigmurethra). Archiv fiir Molluskenkunde 116:137-155. Nordsieck, H. 1986. The system of the Stylommatophora (Gastropoda), with special regard to the systematic position of the Clausiliidae, Il. Importance of the shell and distribution. Archiv fiir Molluskenkunde 117:93-116. Date of manuscript acceptance: 15 January 1999 Edentulina of Madagascar (Pulmonata: Streptaxidae) Kenneth C. Emberton Molluscan Biodiversity Institute, 110 Old Airport Road, Concord, NC 28025, U.S.A. emberton @concordne.com Abstract: Edentulina Pfeiffer, 1856, contains some of the largest species of the diverse, carnivorous, land-snail family Streptaxidae and seems to be restricted to tropical Africa, Madagascar, and some other Indian-Ocean islands. Based on extensive collections made in 1992-1995, and on the 1994 Faune de Madagascar pulmonate monograph, 11 native Madagascan species of Edentulina Pfeiffer, 1856, can be recognized: E. ambongoaboae sp. nov.; E. ambra sp. nov.; E. analamerae sp. nov.; E. ankaranae sp. nov.; E. antankarana sp. nov.; E. arenicola (Morelet, 1860); E. battistinii Fischer-Piette, P. Blanc, and Salvat, 1975; E. bemarahae sp. nov.; E. bobaombiae sp. nov.; E. florensi sp. nov.; E. minor (Morelet, 1851); E. nitens (Dautzenberg, 1895); and E. rugosa sp. nov. Five species are synonymized under Edentulina minor: E. alluaudi (Dautzenberg, 1895); E. gaillardi Fischer-Piette and Bedoucha, 1964; E. inter- media (Morelet, 1851); E. montis Fischer-Piette, F. Blanc, and Salvat, 1975; and E. stumpfii Kobelt, 1904. Three species are transferred to a new genus described in a separate paper: Edentulina (?) glessi Fischer-Piette, Blanc, Blanc, and Salvat, 1994; E. (?) metula (Crosse, 1881); and E. (?) simeni Fischer- Piette, Blanc, Blanc, and Salvat, 1994. A Seychellean species, E. dussumieri (Dufo, 1840), is deleted from Madagascar’s faunal list. A dichotomous key is given to the native species plus the reportedly introduced Edentulina ovoidea (Bruguiere, 1792). Conchological descriptions are given of all native species. Many promising regions of Madagascar remain uncollected. Further exploration should yield additional new species of Edentulina. Key words: Gastropoda, Stylommatophora, land snails, taxonomy, shell variation This paper is the first in a series on the conchologi- Piette et al. (1994) and available collections. Authors and cal identification of Madagascar’s lesser-known land-snail dates of species are given as in Richardson (1988). Whorl groups, based on extensive collections made in 1992-1995, counts were made in apical view by the widely used and supplemental to the Faune de Madagascar monographs method of detecting the earliest sutural notch (using inci- of Fischer-Piette et al. (1993, 1994). dental lighting at 40x), extrapolating that notch as a line Edentulina Pfeiffer, 1856, contains some of the tangent to and projecting past the suture’s initial right-hand largest species of the diverse, carnivorous family curve, counting off whorls as that line crosses successive Streptaxidae and seems to be restricted to Africa, sutures, and rounding the final fraction of a whorl—ending Madagascar, and some other Indian-Ocean islands (Zilch, at the suture’s end, regardless of any distortion due to aper- 1959-1960:563). The pulmonate gastropod volume of the tural-lip reflection—to the nearest tenth (Emberton, Faune de Madagascar (Fischer-Piette et al., 1994) summa- 1985:fig. 1; 1989). Measurements were made using an ocu- rized knowledge of that island’s Edentulina, listing 14 lar micrometer and, rarely, vernier calipers. As an index of species. Among those species, three were listed as tenuous coiling tightness, the number of whorls was divided by the (Fischer-Piette et al., 1994) and have subsequently been natural logarithm (In) of the shell’s height. Because of transferred to a new genus (Emberton and Pearce, in press): Madagascar’s continuing environmental crisis and the E. (?) glessi Fischer-Piette, Blanc, Blanc, and Salvat, 1994; urgency of providing data to conservationists and systema- E. (?) metula (Crosse, 1881); and E. (?) simeni Fischer- tists, only conchological descriptions are given here. Piette, Blanc, Blanc, and Salvat, 1994. KEY TO SPECIES la. Subsutural spiral cord present ................... 2 METHODS AND MATERIALS 1b. Subsutural spiral cord absent ................... 3 2a; SCUMpUITG SIROOMN 52s oe oe ee OS Sse 4 ovoidea Materials were collected in 1992-1995. (Fischer-Piette et al., 1994:plate IV, figs. 4, 5) Identifications and comparisons were made using Fischer- 2b. Sculpture ribbed s. 005. sana ee minor (Figs. 1, 2, 3, 4) American Malacological Bulletin, Vol. 15(1) (1999):97-108 97 98 AMER. MALAC. BULL. 15(1) (1999) Figs. 1-5. Figs. 1-4. Edentulina minor (Morelet, 1851): Figs. 1-2 Analamera Reserve, Fig. 3 Montagne d’ Ambre National Park, Fig. 4 Namoroka Reserve. Fig. 5. E. nitens (Dautzenberg, 1895), Analamera Reserve. Scale bars 3 mm. EMBERTON: EDENTULINA LAND SNAILS OF MADAGASCAR 99 3a. Sculpture smooth or with only subsutural traces of ribs 3b. Sculpture weakly to strongly ribbed .............. 6 4a. No preapertural deflection of body whorl Reh rane Wisin abt R ene Ga as nitens (Fig. 5) 4b. Upward preapertural deflection .................5 5a. Diameter of first 1.5 whorls about 2.3 mm, no trace of subsutural ribs ........... ambongoaboae (Fig. 6) 5b. Diameter of first 1.5 whorls 1.7-1.9 mm, trace subsutur- al ribs present ............... ankaranae (Fig. 7) 6a. Ribs faint to weak ....... 0.0.0... eee eee 7 6b. Ribs strong to moderate .................0.0045. 8 7a. No preapertural deflection of body whorl, coiling loose (whorls/In height about 2.3) ........ ambra (Fig. 8) 7b. Upward preapertural deflection, coiling tight (whorls/In height-about 2.8)......2...<+: analamerae (Fig. 12) 8a. Ribs and sutural crenulation very strong, embryonic whorls smooth ............ 0.000 ce eee eee 9 8b. Ribs moderate and sutural crenulation weak to moder- ate, embryonic whorls sculpted with riblets...... 10 9a. Coiling tight (whorls/In height about 2.7-3.0), shell bar- rel-shaped to ovate, sutures fairly deeply impressed Beis Sete May Aisa ewe y 2 ays arenicola (Figs. 9, 10) 9b. Coiling loose (whorls/In height about 2.5), shell bullet- shaped, sutures shallowly impressed ............. 10a. Diameter of first 1.5 whorls 1.6-1.9mm......... 11 10b. Diameter of first 1.5 whorls 2.2-2.3............. 12 lla. Aperture narrow (height/width 1.3) and small (0.4 Shell width) ............... antankarana (Fig. 13) 11b. Aperture broader (height/width 1.0-1.2) and larger (05-00; Sne Width). arise en ev a ea aa» 12a. Shape fusiform-oval, never any pre-apertural duplicate perstome(S).. 2.760. ea weet battistinii (Fig. 14) 12b. Shape pyramidal, pre-apertural duplicate peristome(s) GHLEM DESEM(, 2 .y5ates «en ews florensi (Figs. 19, 20) SYSTEMATICS To aid users seeking only to verify a previous iden- tification, species descriptions are ordered alphabetically as in the Abstract. Higher classification follows Vaught (1989). Type materials are placed in the United States National Museum, Washington, D.C. (USNM); the Australian Museum, Sydney (AMS); the Muséum national d’Histoire naturelle, Paris (MNHN, which does not assign catalog numbers to types); and the Academy of Natural Sciences of Philadelphia (ANSP). Prior, working catalog numbers of the Molluscan Biodiversity Institute (MBI) are also given, because they provide reference to an ecological database available on request. MBI catalog numbers consist of station number, species reference number within that sta- tion, D (dry) or A (alcohol-preserved), and when appropri- ate H (holotype), P (paratype), or R (representative). As an aid to future workers, paratypes and vouchers that are illus- trated and/or described herein are listed separately as “‘rep- resentatives,” and alcohol-preserved materials for anatomi- cal/biochemical study are listed separately (even though USNM does not assign them separate catalog numbers from dry materials of the same lots). In lot descriptions, “ad” refers to adult(s), “juv” to juveniles(s). Class GASTROPODA Subclass PULMONATA Order STYLOMMATOPHORA Superfamily STREPTAXOIDEA Family STREPTAXIDAE Gray, 1860 Genus Edentulina Pfeiffer, 1856 Edentulina ambongoaboae sp. nov. Fig. 6 HOLOTYPE. USNM 880370 (ex MBI 407.02DH, | ad): 12°15’S, 49°15’E: Madagascar: Cap d’Ambre, Ambongoabo, 290 m: dry deciduous forest; 26 Aug. 1995. DRY PARATYPES. AMS C.203515 (ex MBI 407.02DP, | ad): type locality. MNHN (ex MBI 405.01DP, 1 ad): 12°15°S, 49°15°E: Madagascar: Cap d’Ambre, Ambongoabo, 320 m: baobab deciduous forest; 25 Aug. 1995. ANSP 401997 (ex MBI 405.01DP, 1 ad): 12°15’S, 49°15°E: Madagascar: Cap d’ Ambre, Ambongoabo, 320 m: baobab deciduous forest; 25 Aug. 1995. USNM 880441-880443 (ex MBI 404-407DP, 3 lots; total 5 ad): 12°15°S, 49°15°E: Madagascar: Cap d’Ambre, Ambongoabo; Aug. 1995. DESCRIPTION OF HOLOTYPE (a weathered shell). Shell elongate-ovoid, the aperture protruding slightly out- side the ovoid profile. Height 23.8 mm, diameter 11.0 mm, whorl count 6.5, coiling tightness (whorls/In height) 2.05. Body-whorl periphery gently rounded; suture moderately impressed, simple. Umbilicus a very narrow crevice, over half masked by reflected columellar peristome; umbilicus maximum diameter 0.5 mm. Aperture shape broad upright oval. Apertural lip reflected throughout, narrow at upper suture, then widening to a moderate reflection at the col- umellar insertion, with a narrow triangular, sloping shelf inside the columella. Aperture height 7.2 mm, width 6.7 mm. Preapertural deflection moderately upward, 0.1 whorl. Aperture side shape a reversed, very shallow comma. 100 AMER. MALAC. BULL. 15(1) (1999) Figs. 6-12. Fig. 6. Edentulina ambongoaboae sp. nov. holotype. Fig. 7. E. ankaranae sp. nov. holotype. Fig. 8. Edentulina ambra sp. nov. holotype. Figs. 9- 10. E. arenicola (Morelet, 1860): Fig. 9 southeast of Diego Suarez, Fig. 10 Cap d’ Ambre. Fig. 11. E. rugosa sp. nov. holotype. Fig. 12. E. analamerae sp. nov. holotype. Scale bars 3 mm. EMBERTON: EDENTULINA LAND SNAILS OF MADAGASCAR 101 Embryonic whorl count 2.6; diameter of first 1.5 whorls 2.3 mm. Embryonic sculpture smooth. Post-embryonic sculp- ture smooth, with faint traces of growth lines. VARIATION. All adult specimens are remarkably uniform in size and shape. ETYMOLOGY. For Mount Ambongoabo, the sole known locality for this species. Edentulina ambra sp. nov. Fig. 8 HOLOTYPE. USNM 880371 (ex MBI 191.01DH, 1 ad): 12°35’°S, 49°9°E: Madagascar: Montagne d’ Ambre National Park, 1260 m: rainforest; 11 July 1995. DRY PARATYPES. AMS C.203516 (ex MBI 193.03DP, | juv): 12°34’S, 49°9’E: Madagascar: Montagne d’Ambre National Park, 1305 m: rainforest; 12 July 1995. MNHN (ex MBI 193.03DP, 1 juv): 12°34’S, 49°9’E: Madagascar: Montagne d’Ambre National Park, 1305 m: rainforest; 12 July 1995. USNM 880444-880456 (ex MBI 169-193DP, 8 lots; total 2 ad, 20 juv): 12°S, 49°E: Madagascar: Montagne d’ Ambre National Park; July 1995. ALCOHOL PARATYPES. USNM 880446-880457 (ex MBI 172-194AP, 10 lots; total 8 ad, 30 juv): 12°S, 49°E: Madagascar: Montagne d’ Ambre National Park; July 1995. DESCRIPTION OF HOLOTYPE. Shell a tapered col- umn with bluntly fusiform apex, aperture scarcely protrud- ing outside the shell’s tapered-columnar profile. Height 16.6 mm, diameter 8.5 mm, whorl count 6.5, coiling tight- ness (whorls/In height) 2.31. Body-whorl periphery gently rounded; suture moderately impressed, weakly crenulate. Umbilicus a circular well, fairly evenly expanding; umbili- cus maximum diameter 1.1 mm. Aperture shape broad oval. Apertural lip scarcely reflected at the upper suture, broad- ening slightly below, then widening at the columella, where it grades into a wide, triangular, inward sloping shelf. Aperture height 4.6 mm, width 4.4 mm. No preapertural deflection. Aperture side shape nearly straight, but gently arching backward. Embryonic whorl count 2.5; diameter of first 1.5 whorls 2.1 mm. Embryonic sculpture low, narrow- ish, moderately spaced riblets. Post-embryonic sculpture weak ribs, nearly effaced below the periphery, but stronger at the umbilicus. Coloration pale straw-yellow. VARIATION. Some specimens have a narrower aperture, with a height/width of up to 1.5. ETYMOLOGY. For Montagne d’ Ambre National Park. Edentulina analamerae sp. nov. Fig. 12 HOLOTYPE. USNM 880372 (ex MBI 210.02DH, | ad): 12°44’S, 49°29’°E: Madagascar: Analamera Reserve, 35 m: dry deciduous floodplain; 16 July 1995. DRY PARATYPES. AMS C.203517 (ex MBI 201.02DP, | ad): 12°44’S, 49°30°E: Madagascar: Analamera Reserve, 315 m: dry deciduous forest; 15 July 1995. USNM 880458 (ex MBI 201.02DP, 1 juv): 12°44’S, 49°30°E: Madagascar: Analamera Reserve, 315 m: dry deciduous forest; 15 July 1995. DESCRIPTION OF HOLOTYPE (apex somewhat erod- ed). Shell fusiform-oval, aperture protruding only slightly outside the fusiform-oval profile. Height 18.8 mm, diameter 10.1 mm, whorl count 8.1, coiling tightness (whorls/In height) 2.76. Body-whorl periphery gently rounded; body- whorl shoulder extremely narrow, sloped about 45 degrees; suture moderately impressed, weakly crenulate. Umbilicus a nearly circular, narrow well, widened by and half covered by rapid expansion in the final 0.2 whorl; umbilicus maxi- mum diameter 1.8 mm. Aperture obliquely egg-shaped. Apertural lip reflected throughout, narrowly at the upper suture, moderately thereafter, widening at the columella, with a narrow, steeply sloping triangular shelf below the columellar peristome. Aperture height 5.0 mm, width 5.3 mm. Preapertural deflection gradually upward, 0.3 whorl. Aperture side shape nearly straight, but slightly arching bacward. Embryonic whorl count 2.8; diameter of first 1.5 whorls 2.0 mm. Embryonic sculpture smooth. Post-embry- onic sculpture smoothish, with low, faint ribs. Coloration beigish white. ETYMOLOGY. For Analamera Reserve. Edentulina ankaranae sp. nov. Fig. 7 HOLOTYPE. USNM 880373 (ex MBI 580.01DH, 1 ad): 12°58°S, 49°5°E: Madagascar: Ankarana Reserve, 95 m: dry deciduous forest; 26 Aug. 1995. DRY PARATYPES. AMS C.203518-C.203522 (ex MBI 802- 815DP; 5 lots; total 9 ad, 14 juv): Madagascar: Ankarana Reserve; Oct. 1994. MNHN (ex MBI 564.02DP, | ad): 12°55’S, 49°5’E: Madagascar: Ankarana Reserve, 95 m: dry deciduous forest; 22 Aug. 1995. ANSP 401998 (ex MBI 564.02DP, 1 ad): 12°55’S, 49°S°E: Madagascar: Ankarana Reserve, 95 m: dry deciduous for- est; 22 Aug. 1995. USNM 880459-880472 (ex MBI 557-580DP, 802-815DP, 14 lots; total 23 ad, 54 juv): Madagascar: Ankarana Reserve, 90 m: dry deciduous forest; 1994, 1995. DESCRIPTION OF HOLOTYPE. Shell fusiform, aper- ture protruding slightly outside the fusiform profile. Height 19.5 mm, diameter 8.0 mm, whorl count 6.4, coiling tight- ness (whorls/In height) 2.15. Body-whorl periphery very gently rounded; suture mildly impressed, simple. Umbilicus a narrow well, partially masked by reflected columellar peristome; umbilicus maximum diameter 0.5 mm. Aperture shape auriculate-oval. Apertural lip thinly reflected along outer margin, funneled basally, broadly reflected along col- umella and broadening greatly to the columellar insertion. Aperture height 6.1 mm, width 4.3 mm. Preapertural deflection gradually, then--abruptly--moderately upward, 0.3 whorl. Aperture side shape a reversed very shallow comma, but straight and just slightly recurved below. Embryonic whorl count 2.3; diameter of first 1.5 whorls 1.9 mm. Embryonic sculpture smooth. Post-embryonic sculp- 102 AMER. MALAC. BULL. 15(1) (1999) ture smooth, with faint growth lines and subsutural traces of ribs. Coloration ivory. VARIATION. Adult whorls range 5.9-6.9, height 13.6- 22.5 mm, coiling tightness (whorls/In height) 2.1-2.3; shell height/diameter 2.1-2.4; embryonic whorls 2.3-2.5; diame- ter of first 1.5 whorls 1.7-1.9 mm. ETYMOLOGY. For Ankarana Reserve. Edentulina antankarana sp. nov. Fig. 13 HOLOTYPE. USNM 880374 (ex MBI 810.02DH, 1 bro- ken ad): 12°54’S, 49°6’E: Madagascar: Ankarana Reserve, 90 m; 10 Oct. 1994. DRY PARATYPES. AMS C.203523 (ex MBI 810.02DP, 2 juv): type locality: 12°54°S, 49°6’E: Madagascar: Ankarana Reserve, 90 m; 10 Oct. 1994. USNM 880473 (ex MBI 810.02DP, 2 juv): type locality. DESCRIPTION OF HOLOTYPE (adult shell from which the final 0.5 whorls are broken off, leaving some traces of adult morphology). Shell fusiform-oval. Height 22.0 mm, diameter 10.8 mm, whorl count 7.8, coiling tight- ness (whorls/In height) 2.54. Body-whorl periphery gently rounded; suture well impressed, faintly crenulate. Umbilicus a narrow well with tear-drop-shaped periphery, widening abruptly (judging from scars on shell). Aperture shape uncertain, but elongate. Aperture height 6.0 mm, width 4.6 mm. Preapertural deflection weakly upward, 0.2 whorl (judging from scars on shell). Embryonic whorl count 2.8; diameter of first 1.5 whorls 1.9 mm. Embryonic sculpture very faint riblets. Post-embryonic sculpture strong, low, moderately wide, irregularly spaced ribs. ETYMOLOGY. For the tribe of people (Antankarana) liv- ing near Ankarana Reserve. Edentulina arenicola (Morelet, 1860) Figs. 9, 10 REPRESENTATIVES. USNM 894278 (Fig. 9, ex MBI 217.03DR, 1 ad): 12°23’S, 49°19’E: Madagascar: Montagne des Orchides, 360 m: dry deciduous forest; 20 July 1995. USNM 894279 (Fig. 10, ex MBI 400.07DR, 1 ad): 12°10’S, 49°13’E: Madagascar: Cap d’ Ambre, la Butte Bobaomby, 70 m: dry deciduous forest; 24 Aug. 1995. OTHER DRY VOUCHERS. AMS C.203556 (ex MBI 217.03D, 1 ad): 12°23°S, 49°19°E: Madagascar: Montagne des Orchides, 360 m: dry deciduous forest; 20 July 1995. AMS C.203548 (ex MBI 401.01D, 3 ad): 12°11’S, 49°13’E: Madagascar: Cap d’ Ambre, la Butte Bobaomby, 205 m: dry deciduous-baobab for- est; 24 Aug. 1995. MNHN (ex MBI 217.03D, 1 ad): 12°23’S, 49°19’E: Madagascar: Montagne des Orchides, 360 m: dry decid- uous forest; 20 July 1995. MNHN (ex MBI 401.01D, 1 ad): 12°11°S, 49°13’E: Madagascar: Cap d’Ambre, la Butte Bobaomby, 205 m: dry deciduous-baobab forest; 24 Aug. 1995. ANSP 401999 (ex MBI 217.03D, 1 ad): 12°23’°S, 49°19’E: Madagascar: Montagne des Orchides, 360 m: dry deciduous for- est; 20 July 1995. ANSP 402000 (ex MBI 401.01D, 1 ad): 12°11°S, 49°13’E: Madagascar: Cap d’Ambre, la Butte Bobaomby, 205 m: dry deciduous-baobab forest; 24 Aug. 1995. USNM 880474-880478 (ex MBI 215-219D, 4 lots; total 10 ad, 31 juv): Montagne des Orchides; 20 July 1995. USNM 880479 (ex MBI 222.02D, 10 ad, 22 juv): 12°19°S, 49°20’E: Madagascar: Montagne des Francais, 230 m: dry deciduous forest; 21 July 1995. USNM 880480 (ex MBI 401.01D, 52 ad, 103 juv): 12°11’S, 49°13’E: Madagascar: Cap d’Ambre, la Butte Bobaomby, 205 m: dry deciduous-baobab forest; 24 Aug. 1995. ALCOHOL VOUCHER. USNM 880477 (ex MBI 218.02A, 1 juv): 12°23’S, 49°19°E: Madagascar: Montagne des Orchides, 385 m: dry deciduous forest; 20 July 1995. DESCRIPTION. Shell barrel-shaped, with aperture not protruding to conspicuously protruding outside the shell profile. Height 9.6-13.8 mm, height/diameter 1.7-2.0, whorl count 6.3-7.5, coiling tightness (whorls/In height) 2.7-3.0. Body-whorl periphery flattened; suture moderately impressed, with a crenulate appearance due to rib sculpture. Umbilicus initial shape ranging from a tiny, round pinhole to a narrow well with tear-drop-shaped periphery; with abrupt widening in the final 0.2 whorls, in which the body whorl masks the preceding umbilicus; umbilicus maximum diameter approximately 0.9 mm. Aperture shape squarish- ovate to very broadly auriculate. Apertural lip reflected throughout, narrow at the upper suture, broad at the col- umella. Aperture size ranges from fairly small and narrow (Fig. 9; one extreme, aberrant specimen--not illustrated-- with an arched columella has an aperture height/width of 2.0) to large and broad (Fig. 10; aperture height/width approaching 1.0). Preapertural deflection weakly to moder- ately upward, <0.05 to 0.1 whorl. Aperture side shape a reversed, very shallow comma. Embryonic whorl count 1.9- 2.2; diameter of first 1.5 whorls 1.2-1.5 mm. Embryonic sculpture smooth, sometimes with a faint trace of close-set spiral striae detectable at 40x magnification. Post-embryon- ic sculpture consisting of strong, broad ribs. Color ivory. COMMENTS. Disposition of the type is unknown. The type locality is Port Leven, east of the two widely separated populations represented in Figs. 9 and 10, this paper. Previously published illustrations of E. arenicola are poor and misleading. Edentulina battistinii Fischer-Piette, F. Blanc, and Salvat, 1975 Fig. 14 REPRESENTATIVE. USNM 894280 (ex MBI 245.03DR, 1 ad): 19°8’S, 44°48’E: Madagascar: S Bemaraha Reserve: dry forest; 14 June 1995. OTHER DRY VOUCHERS. AMS C.203524, C.203525, C.203527-C.203534 (ex MBI 245-254D, 484-494D; 10 lots; total 38 ad, 18 juv): Madagascar: Bemaraha Reserve; 1995, 1996. MNHN (ex MBI 245.03D, 1 ad): 19°8’S, 44°48’E: Madagascar: S Bemaraha Reserve: dry forest; 14 June 1995. ANSP 402001 (ex EMBERTON: EDENTULINA LAND SNAILS OF MADAGASCAR 103 Figs. 13-18. Fig. 13. Edentulina antankarana sp. nov. holotype. Fig. 14. E. battistinii Fischer-Piette, F. Blanc, and Salvat, 1975, Bemaraha Reserve. Figs. 15-18. E. bemarahae sp. nov.: Fig. 16 holotype; Figs. 15, 17, 18 paratypes. Scale bars 3 mm. 104 AMER. MALAC. BULL. 15(1) (1999) MBI 254.02D, | ad): 19°8’S, 44°52’E: Madagascar: S Bemaraha Reserve, 100 m: forest; 18 June 1995. USNM 88048 1-880492 (ex MBI 245-254D, 483-494D, 11 lots; total 40 ad, 17 juv): Madagascar: Bemaraha Reserve; 1995, 1996. ALCOHOL VOUCHER. USNM 880492 (ex MBI 494.08A, 1 juv): 18°45’S, 44°45°E: Madagascar: N Bemaraha Reserve, 280 m: semideciduous forest; 29 June 1996. DESCRIPTION. Shell bluntly fusiform, aperture protrud- ing greatly outside the fusiform profile. Height 17.9-33.5 mm, height/diameter 1.8-2.5 mm, whorl count 6.8-8.8, coil- ing tightness (whorls/In height) 2.4-2.5. Body-whorl periphery gently rounded, slightly flattened; body-whorl shoulder extremely narrow, sloped about 60 degrees; suture moderately impressed, very faintly crenulate. Umbilicus a moderately narrow well with tear-drop-shaped periphery, with abrupt widening in final 0.3 whorl. Aperture shape very broad auriculate-oval. Apertural lip thickly reflected, from narrow at the upper suture, through moderate and somewhat funneled at the base, to broad at the columella, with a triangular, gently sloping shelf inside the columella. Preapertural deflection gradually, then--abruptly--moder- ately upward, approximately 0.3 whorl. Aperture side shape a broad, distorted S, the upper curve long, the bottom curve short, both curves very shallow. Embryonic whorl count Embryonic sculpture faint riblets. Post-embryonic sculpture consists of strong, somewhat thin, moderately spaced ribs. Edentulina bemarahae sp. nov. Figs. 15, 16, 17, 18 HOLOTYPE. USNM 880375 (Fig. 16, ex MBI 247.01DH, 1 ad): 19°8°S, 44°52°E: Madagascar: S Bemaraha Reserve: riverine gallery forest; 15 June 1995. DRY PARATYPE REPRESENTATIVES. USNM 880376 (Fig. 17, ex MBI 247.01DPR, | ad): type locality. USNM 880377 (Fig. 15, ex MBI 245.02DPR, | ad): 19°8’S, 44°048’E: Madagascar: S Bemaraha Reserve: dry forest; 14 June 1995. USNM 880378 (Fig. 18, ex MBI 254.01DPR, | ad): 19°2’S, 44°48°E: Madagascar: S Bemaraha Reserve: forest. OTHER DRY PARATYPES. AMS C.203535-C.203547 (ex MBI 245-255DP, 483-495DP; 13 lots; total 24 ad, 28 juv): Madagascar: Bemaraha Reserve; 1995, 1996. MNHN (ex MBI 247.01DP, 1 ad): type locality. ANSP 402002 (ex MBI 247.01DP, 1 ad): type locality. USNM 880493-880507 (ex MBI 245-255DP, 483-495DP, 14 lots; total 25 ad, 31 juv): Bemaraha Reserve; 1995, 1996. ALCOHOL PARATYPES. USNM 880497 (ex MBI 250.01 AP, | juv): 19°8°S, 44°49°E: Madagascar: S Bemaraha Reserve; 16 June 1995. USNM 880503 (ex MBI 489.02AP, 2 juv): 18°41’S, 44°43°E: Madagascar: N Bemaraha Reserve, 150 m: semidecidu- ous forest; 27 June 1996. DESCRIPTION OF HOLOTYPE. Shell bluntly fusiform, aperture protruding greatly outside the fusiform profile. Height 18.2 mm, diameter 8.0 mm, whorl count 7.9, coiling tightness (whorls/In height) 2.72. Body-whorl periphery gently rounded; body-whorl shoulder very nar- row, sloped about 30 degrees; suture moderately impressed, simple to weakly crenulate. Umbilicus a narrow well with tear-drop-shaped periphery, with abrupt widening in final 0.1 whorl; umbilicus maximum diameter 0.8 mm. Aperture shape broadly auriculate. Apertural lip thickly reflected, narrow at upper suture, broad at columella with sloping tri- angular shelf below. Aperture height 4.5 mm, width 4.3 mm. Preapertural deflection gradually, then—briefly— strongly upward, 0.5 whorl. Aperture side shape reversed very shallow comma. Embryonic whorl count 2.7; diameter of first 1.5 whorls 1.9 mm. Embryonic sculpture broad, close-set riblets. Post-embryonic sculpture strong, moder- ately wide, moderately spaced ribs. Coloration ivory. VARIATION (Figs. 15-18). Adult whorls range 6.5-7.9, height 13.5-18.2, coiling tightness (whorls/In height) 2.5- 2.8; shell height/diameter 1.8-2.3; umbilicus/shell diameter 0.1-0.2; aperture height/width 1.0-1.2; aperture width/shell diameter 0.5-0.6; pre-apertural deflection 0.2-0.5 whorl; embryonic whorls 2.7-2.9; diameter of first 1.5 whorls 1.6- 1.9; post-embryonic sculptural ribs vary considerably in width and density. ETYMOLOGY. For Bemaraha Reserve. Edentulina florensi sp. nov. Figs. 19, 20 HOLOTYPE. USNM 880379 (Fig. 19, ex MBI 490.41DH, 1 ad): 18°45’°S, 44°45’E: Madagascar: N Bemaraha Reserve, 280 m: semideciduous forest. 27-Jun-96. DRY PARATYPE REPRESENTATIVE. USNM 880380 (Fig. 20, ex MBI 488.41DP, 1 ad): 18°47°S, 44°47°E: Madagascar: N Bemaraha Reserve, 250 m: riverine scrub. 27-Jun-96. OTHER DRY PARATYPES. AMS C.203549-C.203554 (ex MBI 484-494DP; 6 lots; total 47 ad, 39 juv): 18°3’S, 44°31’E: Madagascar: N Bemaraha Reserve, 250 m: dry deciduous forest. 23-Jun-96. MNHN (ex MBI 494.41DP, 2 ad): 18°45’S, 44°45’E: Madagascar: N Bemaraha Reserve, 280 m: semideciduous forest. 29-Jun-96. ANSP 402003 (ex MBI 494.41DP, 2 ad): 18°45’S, 44°45’E: Madagascar: N Bemaraha Reserve, 280 m: semidecidu- ous forest. 29-Jun-96. USNM 880508-880514 (ex MBI 484- 495DP, 8 lots; total 44 ad, 42 juv): Madagascar: N Bemaraha Reserve. Jun-96. DESCRIPTION OF HOLOTYPE. Shell tall and bluntly pyramidal. Height 22.8 mm, diameter 9.2 mm, whorl count 8.2, coiling tightness (whorls/In height) 2.62. Body-whorl periphery gently rounded, slightly flattened; body-whorl shoulder narrow, sloped about 60 degrees; suture moderate- ly impressed, simple to weakly crenulate. Umbilicus a nar- row well with tear-drop-shaped periphery, about half cov- ered by the reflected columellar peristome; umbilicus maxi- mum diameter 0.8 mm. Aperture shape auriculate-oval. Apertural lip very narrowly reflected at the upper suture and to the basal area, whence it expands to a broadly trian- gular columellar insertion, within which is a narrow, trian- EMBERTON: EDENTULINA LAND SNAILS OF MADAGASCAR 105 Figs. 19-20. Edentulina florensi sp. nov.: Fig. 19 holotype in two views; Fig. 20 paratype. Scale bar 3 mm. gular, steeply sloping shelf. Aperture height 5.0 mm, width 5.1 mm. Preapertural deflection gradually upward, 0.2 whorl. Aperture side shape a reversed, very shallow comma. Embryonic whorl count 2.8; diameter of first 1.5 whorls 2.2 mm. Embryonic sculpture riblets. Post-embry- onic sculpture moderately strong, thin ribs, moderately spaced; a conspicuous, pre-apertural, duplicate peristome is present. Coloration ivory. VARIATION (Figs. 19-20, in part). Presumed adult whorls range 6.7-9.2, height 15.8-29.4; shell height/diameter 1.9- 3.0; coiling tightness (whorls/In height) 2.4-2.7; one or more pre-apertural, duplicate peristomes occur in about three-fourths of presumed adults. ETYMOLOGY. For Vincent Florens, in recognition of his outstanding contribution to the first Bemaraha expedition. Edentulina minor (Morelet, 1851) Piss. 1.2,3,4 NEW SYNONYMS. Edentulina alluaudi (Dautzenberg, 1895) (Fig. 3; Fischer-Piette et al., 1994:57, plate IV fig. 11). The slender shape of this variant from Montagne d’ Ambre (see below) falls within the morphological range of E. minor from other sites such as Namoroka Reserve (see below). Edentulina gaillardi Fischer-Piette and Bedoucha, 1964 (Fig. 4; Fischer-Piette et al., 1994:57, plate IV fig. 11). A wide-apertured, strong-ribbed variant of E. minor from Ankarafantsika Reserve and environs (including Ampijoroa Reserve, see below), which also occurs within E. minor’s range of variation within Namoroka Reserve (see below). Edentulina intermedia (Morelet, 1851) (Fischer- Piette et al., 1994:50, fig. 33). Fischer-Piette was unable to locate the holotype, so he reproduced Morelet’s inadequate figure and measurements; he also commissioned numerous collections at the sole known locality (Port-Léven), but without success (Fischer-Piette and Bedoucha, 1964). Fortunately, however, Tryon (1885:83; plate 17, fig. 20) had adequately illustrated and remeasured the holotype, which falls within E. minor’s known range of variation. Edentulina montis Fischer-Piette, F. Blanc, and Salvat, 1975 (Fischer-Piette ef al., 1994: 57, plate IV fig. 12-14). Despite its narrow umbilicus, this species seems to fall within the wide variation of E. minor, which occurs at the type locality of E. montis, Montagne des Francais (see below). Edentulina stumpfii Kobelt, 1904 (Fischer-Piette et al., 1994: 51, Fig. 34). On Nosy Be, the type locality of E. stumpfii, many E. minor were collected (see below), within whose range of variation E. stumpfii seems to fall. REPRESENTATIVES. USNM 894281 (Figs. | and 2, ex MBI 203.03DR, 2 ad): 12°44’°S, 49°30’E: Madagascar: Analamera Reserve, 285 m: bamboo-dry deciduous thicket; 16 July 1995. USNM 894282 (Fig. 3, ex MBI 184.01DR, | ad): 12°36’S, 49°9’E: Madagascar: Montagne d’ Ambre National Park, 1165 m: rainforest; 10 July 1995. USNM 106 AMER. MALAC. BULL. 15(1) (1999) 894283 (Fig. 4, ex MBI 61.03DR, | ad): 16°23’S, 45°18°E: Madagascar: Namoroka Reserve, 105 m: dry deciduous for- est; 25 May 1995. OTHER DRY VOUCHERS. AMS C.203514 (ex MBI 184.01D, 1 ad): 12°36°S, 49°9°E: Madagascar: Montagne d’ Ambre National Park, 1165 m: rainforest; 10 July 1995. AMS C.203555 (ex MBI 83.01D, | ad): 16°17°S, 46°49°E: Madagascar: Ampijoroa Reserve, 95 m: hardwood deciduous forest; 3 June 1995. AMS C.203558 (ex MBI 55.03D, 2 ad): 16°23’S, 45°21’E: Madagascar: Namoroka Reserve, 110 m: dry deciduous forest; 21 May 1995. AMS C.203559 (ex MBI 487.01D, 3 ad, 4 juv): 18°°S, 44°°E: Madagascar: N Bemaraha Reserve; 25 June 1996. AMS C.203560 (ex MBI 805.03D, 6 ad, 8 juv): 13°1’S, 49°O’E: Madagascar: Ankarana Reserve, 50 m; 8 Oct. 1994. MNHN (ex MBI 55.03D, 2 ad): 16°23°S, 45°21°E: Madagascar: Namoroka Reserve, 110 m: dry deciduous forest; 21 May 1995. MNHN (ex MBI 83.01D, | ad): 16°17°S, 46°49°E: Madagascar: Ampijoroa Reserve, 95 m: hardwood deciduous forest; 3 June 1995. MNHN (ex MBI 184.01D, 1 ad): 12°36°S, 49°9°E: Madagascar: Montagne d’Ambre National Park, 1165 m: rainforest; 10 July 1995. ANSP 402004 (ex MBI 55.03D, 2 ad): 16°23°S, 45°21E: Madagascar: Namoroka Reserve, 110 m: dry deciduous forest; 21 May 1995. ANSP 402005 (ex MBI 83.01D, 1 ad): 16°17°S, 46°49°E: Madagascar: Ampijoroa Reserve, 95 m: hardwood deciduous forest; 3 June 1995. ANSP 402006 (ex MBI 184.01D, 1 ad): 12°36°S, 49°9°E: Madagascar: Montagne d’Ambre National Park, 1165 m: rainforest; 10 July 1995. USNM 880515-880527 (ex MBI 55-74D, 13 lots; total 148 ad, 148 [no mistake] juv): Madagascar: Namoroka Reserve; May 1995. USNM 880528- 880531 (ex MBI 81-84D, 4 lots; total 28 ad, 57 juv): 16°17°S, 46°49°E: Madagascar: Ampijoroa Reserve; June 1995. USNM 880532 (ex MBI 86.01D, | juv): 16°8°S, 47°0°E: Madagascar: Ankarafantsika Reserve, 160 m: dry deciduous forest; 5 June 1995. USNM 880533-880549 (ex MBI 118-150D, 17 lots; total 14 ad, 42 juv): 13°S, 48°E: Madagascar: Nosy Be: Lokobe Reserve; June 1995. USNM 880550 (ex MBI 168.01D, 2 ad, 1 juv): 13°34°S, 48°45°E: Madagascar: Galoko Escarpment, 225 m: hardwood-palm rainforest; 4 July 1995. USNM 880551-880560 (ex MBI 177-194D, 10 lots; total 6 ad, 19 juv): 12°S, 49°E: Madagascar: Montagne d’Ambre National Park; July 1995. USNM 880561-880575 (ex MBI 199-214D, 15 lots; total 39 ad, 102 juv): 12°S, 49°E: Madagascar: Analamera Reserve; July 1995. USNM 880576-880577 (ex MBI 217-219D, 2 lots; total 6 ad, 7 juv): 12°23°S, 49°19°E: Madagascar: Montagne des Orchides; 20 July 1995. USNM 880578-880579 (ex MBI 220- 221D, 2 lots; total 1 ad, 10 juv): 12°19°S, 49°20°E: Madagascar: Montagne des Francais; 21 July 1995. USNM 880580-880591 (ex MBI 230-240D, 402-407D, 12 lots; total 25 ad, 80 juv): Madagascar: Cap d’Ambre; 1995. USNM 880592 (ex MBI 410.01D, 1 juv): 12°26’S, 49°12°E: Madagascar: W of Sakaramy, S of Diego Suarez, 470 m: dry deciduous viny forest; 26 Aug. 1995. USNM 880593-880600 (ex MBI 412-421D, 8 lots; total 10 ad, 29 juv): Madagascar: Andavakoera massif, N of Betsiaka; 1995. USNM 880601 (ex MBI 487.01D, 4 ad, 4 juv): 18°°S, 44°°E: Madagascar: N Bemaraha Reserve; 25 June 1996. USNM 880602 (ex MBI 548.01D, 9 juv): 13°28°S, 48°21°E: Madagascar: Nosy Komba, 622 m: cleared, former forest; 23 June 1995. USNM 880603-880605 (ex MBI 564-572D, 805D, 3 lots; total 9 ad, 13 juv): Madagascar: Ankarana Reserve; 1994, 1995. ALCOHOL VOUCHERS. USNM 880516-880523 (ex MBI 56- 70A, 2 lots; total 2 juv): 16°S, 45°E: Madagascar: Namoroka Reserve; May 1995. USNM 880530-880531 (ex MBI 83-84A, 2 lots; total 2 juv): 16°17°S, 46°49’E: Madagascar: Ampijoroa Reserve; June 1995. USNM 880538-880549 (ex MBI 125-150A, 7 lots; total 11 juv): Madagascar: Nosy Be: Lokobe Reserve; June 1995. USNM 880567-880574 (ex MBI 206-213A, 2 lots; total 1 ad, 2 juv): 12°S, 49°E: Madagascar: Analamera Reserve; July 1995. USNM 880585 (ex MBI 239.01A, 1 ad): 12°0’S, 49°17°E: Madagascar: Cap d’Ambre, near Ambatojanahary, 40 m: dry deciduous forest; 25 July 1995. USNM 880594-880600 (ex MBI 413-421A, 3 lots; total 3 juv): 13°S, 49°E: Madagascar: Andavakoera massif, N of Betsiaka; 1995. DESCRIPTION. Shell acute-oval to fusiform, aperture protruding moderately to greatly outside the oval-to- fusiform profile. Height 16.4-37.2 mm, height/diameter 1.6-2.0, whorl count 6.9-9.0, coiling tightness (whorls/In height) 2.4-2.5. Body-whorl periphery rounded; body- whorl shoulder a very narrow shelf defined by a subsutural cord; suture shallowly to moderately impressed, simple to faintly undulating to faintly crenulate. Subsutural spiral cord present. Umbilicus a slit-like crevice or crease to a tear-drop-shaped well, abruptly widened by rapid expan- sion in final 0.2-0.4 whorl; umbilicus maximum diameter/shell diameter 0.2. Aperture shape broad auricu- late-oval. Apertural lip unreflected to narrowly reflected at the upper suture, moderately and thickly relected thereafter, the columella broadened by a triangular shelf slanting inward within the aperture. Aperture width/shell diameter 0.5-0.6; apertural height/width 1.0-2.0. Preapertural deflec- tion weakly to moderately upward, 0.1-0.2 whorl. Aperture side shape a reversed, shallow comma, but straight and slightly recurved below. Embryonic whorl count 2.6-3.1; diameter of first 1.5 whorls 1.8-2.0 mm. Embryonic sculp- ture smooth, sometimes with very faint traces of low riblets in the final half whorl. Post-embryonic sculpture weak to moderately strong, closely spaced ribs. Coloration ivory to yellowish beige. Edentulina nitens (Dautzenberg, 1895) Fig. 5 REPRESENTATIVE. USNM 894284 (ex MBI 210.04DR, 1 ad): 12°44’S, 49°29°E: Madagascar: Analamera Reserve, 35 m: dry deciduous floodplain; 16 July 1995. DESCRIPTION OF REPRESENTATIVE (embryonic shell severely fractured and repaired during life). Shell blunt elongate-oval. Height 20.6 mm, diameter 10.4 mm, whorl count 6.7, coiling tightness (whorls/In height) 2.21. Body-whorl periphery flattened; suture well impressed, simple. Umbilicus a narrow well, scarcely widened in final 0.1 whorl; umbilicus maximum diameter 0.8 mm. Aperture EMBERTON: EDENTULINA LAND SNAILS OF MADAGASCAR 107 shape broadly auriculate-oval. Apertural lip nearly unre- flected at the upper suture, then narrow, widening greatly at the columella into a triangular insertion, with a narrow, steep shelf below the columella. Aperture height 6.1 mm, width 5.3 mm. No preapertural deflection. Aperture side shape a reversed, very shallow comma, but straight and just slightly recurved below. Embryonic whorl count 2.4; diam- eter of first 1.5 whorls 1.8 mm. Embryonic sculpture smooth. Post-embryonic sculpture smooth, with faint, very regularly spaced growth lines. Coloration ivory. Edentulina rugosa sp. nov. Fig. 11 HOLOTYPE. USNM 880381 (ex MBI 407.04DH, | ad): 12°15’°S, 49°15°E: Madagascar: Cap d’Ambre, Ambongoabo, 290 m: dry deciduous forest; 26 Aug. 1995. DRY PARATYPES. AMS C.203561 (ex MBI 406.01DP, | ad): 12°15’S, 49°15’E: Madagascar: Cap d’Ambre, Ambongoabo, 310 m: baobab deciduous forest; 25 Aug. 1995. MNHN (ex MBI 405.03DP, 1 ad): 12°15’S, 49°15’E: Madagascar: Cap d’ Ambre, Ambongoabo, 320 m: baobab deciduous forest; 25 Aug. 1995. ANSP 402007 (ex MBI 405.03DP, 1 ad): 12°15°S, 49°1S’E: Madagascar: Cap d’ Ambre, Ambongoabo, 320 m: baobab decidu- ous forest; 25 Aug. 1995. USNM 880606-880607 (ex MBI 405- 407DP, 2 lots; total 4 juv): 12°15’S, 49°15’E: Madagascar: Cap d’Ambre, Ambongoabo; Aug. 1995. USNM 880608-880609 (ex MBI 410-411DP, 2 lots; total 1 ad, 5 juv): 12°26’S, 49°12’E: Madagascar: W of Sakaramy, S of Diego Suarez; Aug. 1995. DESCRIPTION OF HOLOTYPE. Shell fusiform-oval. Height 20.2 mm, diameter 10.7 mm, whorl count 7.4, coil- ing tightness (whorls/In height) 2.46. Body-whorl periphery slightly flattened; suture mildly impressed, strongly crenu- late from ribbed sculpture. Umbilicus a moderately wide well, rapidly expanded in last 0.1 whorl; umbilicus maxi- mum diameter 1.5 mm. Aperture shape broadly auriculate. Apertural lip narrowly reflected at the upper suture, broad- ening to widely reflected at the columella, which continues inward as a triangular sloping shelf. Aperture height 5.8 mm, width 5.8 mm. Preapertural deflection gradually upward, 0.2 whorl. Aperture side shape a reversed, extremely shallow comma, straightened and very slightly recurved at the bottom. Embryonic whorl count 2.6. Embryonic sculpture smooth, with a hint of dense spiral lines visible at 40x magnification. Post-embryonic sculp- ture very strongly ribbed. Coloration light beige. VARIATION. Adult height ranges from 16.7 to 22.8 mm. ETYMOLOGY. For the conspicuous sculpture that is strongly wrinkled (Latin “rugosa’’) or ribbed. DISCUSSION Fischer-Piette et al.’s (1994) faunal list tentatively included two extra-Madagascan Edentulina: E. ovoidea (Bruguiere, 1792) of East Africa and the Comores, and E. dussumieri (Dufo, 1840) of the Seychelles. Neither of those species was encountered in the 1992-1995 survey. E. ovoidea should remain tentatively on the faunal list, because it reportedly was introduced (misguidedly) as a biological control agent to Madagascar (Fischer-Piette ef al., 1994:49-50). It seems safe, however, to remove E. dus- sumieri, because it has never been collected in Madagascar since Morelet’s old and non-vouchered report. Based on collections made in 1992-1995 from at or near their type localities, four species are synonymized above under FE. minor (Morelet, 1851): FE. alluaudi (Dautzenberg, 1895); E. gaillardi Fischer-Piette and Bedoucha, 1964; E. montis Fischer-Piette, F. Blanc, and Salvat, 1975; and E. stumpfii Kobelt, 1904. E. intermedia (Morelet, 1851) is also synonymized under E. minor, based on Tryon’s (1885) redescription of the holotype, and based on numerous barren collections at the type locality (Fischer-Piette and Bedoucha, 1964). Thus, of Fischer-Piette et al.’s (1994) list of 13 native species of Madagascan Edentulina, this paper deletes one, transfers three, and synonymizes five. This leaves four, all of which were collected in 1992-1995 and are described (from representatives) above. With the addition of 7 new species, also described above, Madagascar’s list of Edentulina now totals 11 native species and, possibly, the one introduced species. Many promising regions of Madagascar remain uncollected, so exploration should yield additional new species of Edentulina. ACKNOWLEDGMENTS Funded by the U.S. National Science Foundation (DEB 9201060), with some additional funding provided by Owen Griffiths. Permits were issued by the Madagascar government agencies DEF and ANGAP. Ranomafana National Park Project gave logistic support. Many people collected or otherwise assisted, but particular thanks go to Owen Griffiths, who conducted both Bemaraha expeditions and an Ankarana expedition; Dr. Tim Pearce, who conducted the bulk of the Montagne d’Ambre and a second Ankarana expedition; and Felix Rakotomalala (requiescat in pace), Tim Pearce, Jean Rakotoarison, and Roger Randalana, who were primary assistants in both the field and the sorting lab. LITERATURE CITED Emberton K. C. 1985. Seasonal changes in the reproductive gross anatomy of the land snail Triodopsis tridentata (Say) (Pulmonata: Polygyridae). Malacologia 26:225-239. Emberton K. C. 1989. Retraction/extension and measurement error in a land snail: effects on systematic characters. Malacologia 31:157- 173. 108 AMER. MALAC. BULL. 15(1) (1999) Emberton K. C. and T. A. Pearce. in press. Small high-spired land pul- monates from Mounts Mahermana, Ilapiry, and Vasiha, S.E. Madagascar, with description of a new genus and with conserva- tion statuses of 15 streptaxids. The Veliger. Fischer-Piette, E., and J. Bedoucha. 1964. Mollusques terrestres de Madagascar. Famille Streptaxidae. Bulletin du Muséum national d Histoire Naturelle, Paris (2) 36:502-505. Fischer-Piette, E., C. P. Blanc, F. Blanc, and F. Salvat. 1993. Gastéropodes terrestres prosobranches. Faune de Madagascar 80:1-281. Fischer-Piette, E., C. P. Blanc, F. Blanc, and F. Salvat. 1994. Gastéropodes terrestres pulmonés. Faune de Madagascar 83:1- 551. Richardson, C. L. 1988. Streptaxacea: catalog of species, Part I, Streptaxidae. Tryonia 16:1-326. Tryon, G. W. 1885. Testacellidae, Oleacinidae, Streptaxidae, Helicoidea, Vitrinidae, Arionidae. Manual of Conchology (Philadelphia) 2:1- 364, pl. 1-60. Vaught, K. C. 1989. A Classification of the Living Mollusca. American Malacologists Inc., Melbourne, Florida, 189 pp. Zilch, A. 1959-1960. Gastropoda, Teil 2, Euthyneura, Band 6. In: Handbuch der Paldozoologie, O. H. Schindewolf, ed. pp. 1-834. Gebriider Borntrager, Berlin. Date of manuscript acceptance: 15 January 1999 AMERICAN MALACOLOGICAL SOCIETY FINANCIAL REPORT General Accounts 1998 Income and Expenses MIRA SASS SEs S51) 119) Giscerssae a coneouteerannes pnsmecretirns ee se dea tesanditeas age sxuseataibesstestenscrdanecnosMensksvyersvsevatptceetantetsouss $131,991.62 MIN OE oars ceases cnc dss pecan cn zcaine hese ga ce na ey ers vaceeat nau unsu as asnauani acesunusveuuouses acter uben pasogias anevanstesuesnencvuy aes $46,954.96 Membership Dies 41996, 1997, 1998) ic csuxccprseccscnsers cenectusasasscassiet evnesoranvanganecseactaantisns 12,818.00 Menibersinp Dies C1999 ) vs siscccasetee 25 seta cienakecsnsattue iw tavesoaunsaseccsanssoneet sauntnnttvednesesvesonsinnes 1,515.00 NITES SPAMG DIV IG CIGS cae creneeanconnsadbo-sspieivacteteseneearnedunocetntasueuruacsraytassnnse pacenmeenccaniorsnasaers 3,966.35 Money Market Account Interest va .scssisicactunnesonssaiieceneaniotsuccgdsrniandanasanys 1,503.06 Lite Membership Endowment Pun .ij.ccciscssscccesteenssoacseccenssneosoevanesvereverin 236.88 SyMPOSiilm. ENAGWIMENCFUMAS: cas cjsceceapsexecacnsncascanscssecececuescuswonateosanscaeies 2,226.41 Publications InCOme ciorecstress.ce-ncecctatecsesetectuncecdssstessecssstareé seuceeteustesedeacovsescecsextertaestetsooaeas 7,413.40 AME Poreign SUDSCHIptl ms coc: cancccyeczvavestiedori ius tecentencettnaiaccmmpeonnnvasviane 1,696.00 AMB DOMeESUC SUDSCHPIIONS va siimsshenscatesiersraecdsrseveli sates tnmsantuvearernosnvey) 2,994.00 PIM PASE CAR GES correctsoscaczceecssententspesanonsdubiaeesnsveasaateiuetsantoenensavnansaaniaees 1,716.00 PUIVETS FRET CAR COS ccescs seca, tia tbe soreese yeaa ese scatdaactasas a }etevcaveeniatsnaingacntes 680.00 Sales Of Other PUD MCAMON Ss vciessivesssasansccensdassianinedlncsncesy Ustecsesdmsectantnnescoace 327.40 PAM AVIC CL 2S cts comers ater coc certheot ex aces eaces veces cede ts pao eo yecsativatosene re tac osg cat UR vouddh De neceoedesh Sons 16,910.33 MOS 7 CCU SUT PWS sc tetwacto-austsunsatusnsrinceneteroncandyeet tence yeireaisuedsaescturabecpnotesa 97.13 *ET OOS: Meeting: SULPIUS vovccccsesncevessesgesvenneatecsveencavetotedesidesoceteaevonssteedcenseey 12,279.50 OOF ACUI OCS SS cre esi ye aecatenciraartan sca dec aust yocunpnncicaonertedisucadgen@anosabioals 978.00 1998 Auction ProGeeds. sivc.c20, cscs: desascuceeec eset cecaiceeasiessionrecesseesvevee coeaveeees 3,555.70 DOM Ath ONS soe os sect ce cescan Svsanceavevessstereocereanes cseessoetunderaaniesunovasavestavsstevsanngsesansteetacddevsnceussntser destnes 4,299.38 SyMPOsilin HP AGOWMGME WWM 5, ccs cgsycceseeceeceaascececeancceneecseoraienaeanianvescce 440.00 StudentGrant Endowment PUNd in ccccicctecevsusvtdcescshgueetisvcanestcomwenesoianuens 939.00 IMISCElIAMCOUSUNCOME eacetee-ceseaeeceenceta sconces cctes caceereceeces anectstaceerin sreersts cokers sassesaanececes coreesee ss 32.50 | BD, 2 SUN Oy Be meee ec eC ROD DEN eC NE ee RST ere POP ee ee ee eee $24,731.48 DECTCLATY EXPENSES) cconssasasnssbinsasveesssteatorreideanedessiccsssercis Munsubsansvasesanpvicesyeotneotesraseteansdanavencisnene 208.00 Members ips Clairim atl EX Dense sai 0e ss decvacuacerertececosinduts omssseairysstssesuadseeseseneesnecarncxesaseseact 133.15 OTS AS UROL ex POMS cece trees ceetet atest (tea ga rents tacscrsn ce pecatea sue cai aren aan a etnesetacsacaecuewese ese pasar! 683.43 PEfiltate: Wie berslips siaspsasisathasasacseasstecsnssstsesa yneecabn-ronnasasesacsusu sess tones onrocad voccenvenananyanteuaeaiaes 15.00 | BUSEY UT il got ok arpa a ameter aoe ory anette er arate eke MOI EATnre ees ee ene nee Pe eR ee rere 30.00 WEG eV ES ACG sta cates sans edancatetace Gey cat agvavo-paveestey ances incase tase us we beonauace esd hcies cmeanen ay toEsniters 585.10 BITC OPO CICA’ Sncett act alts cue nacsaeine. comasnenneaseaeessreselaahiechades tanerehenaysot clea. faetenneseswa nose eRu deren ta 20.00 IS UrAN Ce OMA MCCS teary, eter sur Uinstreancenenpetaciettensenssvaerdayesgataieual aver vayteareran set nseniteaysanioceesys 600.00 Be Dials oyna ce cates e coca egestas racen ecoet eva crater waa toratey esc ha sonii ag ounsticeeaneas treaty eseansy ase nacntars@earteesess 250.00 Managing and Bulletin Bator EX pensesivveice cecevecscssassossserssannsersstebdennerasensactaal vanivensddbiavecesns 101.40 PUB UTC AEN OT OSS scales yos aay veo ey tess as catcey oe tae vena sacence as cahaerayee apts ee) 9,885.17 SAD 2 cats Seo phage rea oa ete pa gia gece eee ads Sac enas noes arora eed eA gL Reamer 7,467.87 POVES IN CW SLETECY wrercctcican tliat uytracteataeceuastiubesenes ead vevducriucsiaabasanvusnpuniustessgeinilietone ikages 12,417.30 BS MIGeNt IES ar CH CIDANILS yest career iearsai¥ reese aeashausaumer er souieiesaaser vex ir oouaecemanrcarona amin asare 1,957.00 Be NS ECAC UL AMMAN teeter tec eg ent atcee rcs, wehciecste eyes ets aethea eet eas naoneseeaesxasemtereenreseie rere 3,017.07 1997 council Mec ting: Expenses a2.cacdsntssstssiae ae or nc vasersn ss saw ieelacusiaaandnasnsenedann edits 353.43 ROOT AUCH ONE RPeM SOS ass soph eecta cae aaclianioet vescvusedvecacsvesiaaeaseaianiouaserseteeayy ses Vessivaveasegecaastivevataret 188.00 1998 Meeting Expenses - Separate Account 19D NICE Lum eat KISS 1 SOS 55 as vsicsaeeesckancacnceeinns mn uesesinvareras een taae uni eae ED Sneed 560.02 Officer Travel MANCnses (tO MAGEINGS) 1) present (1) plesiomorphic Body wall musculature present (1) absent (0) present (1) plesiomorphic Intercrossing d-v muscles many (0) few (1) lacking (?) equivocal Specific circumoral field present (1) present (1) present (1) uninformative Foot reduced (1) present (0) lacking (?) equivocal Spiculoblasts unicellular (1) uni- and multicellular (0) lacking (?) equivocal Shell-field lacking (0) present (1) lacking (0) plesiomorphic Coelomoducts one pair (1) several or one pair (0/1) many, one or no pair (0/1) uninformative Gametes through pericardium present (1) absent (0) No pericardium (?) equivocal Mantle cavity posterior (1) circumpedal (0) no mantle cavity (?) equivocal Osphradium extrapallial (1) intrapallial (0) no osphradium (?) equivocal Cerebral ganglia present (1) present or absent (0/1) present (1) uninformative Radula basically distichous (1) basically many teeth (0) no radula (?) equivocal HASZPRUNAR: IS APLACOPHORA MONOPHYLETIC? 117 were specifically listed and directly evaluated by means of outgroup comparison (Table 1). Doing this the “red-blue- tail” problem becomes crucial (see below). Red-blue-tail problem. Concerning the coding of certain characters, the so-called “red-blue-tail” problem, i. e., the problem of coding of inapplicable characters (Maddison, 1993; Nelson and Ladiges, 1993; Wilkinson, 1995; Hawkins et al., 1997, Kitching et al., 1998, Lee, 1999) turned out to be of crucial importance, in particular with respect to the outgroups and thus directly associated with (so far nearly exclusively done) a priori estimation of the polarity of characters within the Mollusca. The problem concerns many significant characters in molluscan evolution, namely the type of body cuticle (#2), presence of the periostracal groove (#6), type of man- tle cavity (#9), number of ctenidia (#11), number of inter- crossing dorso-ventral muscles (#15), details of the pericar- dioducts (#25, #26) and gamete release (#34), various aspects of the radular apparatus (#38, #39, #40), pedal gan- glia (#49), and position of the osphradium (#56). The spe- cific point concerning the Aplacophora-question is outlined in the following with the example of the various aspects of the radular apparatus (#38, #39, #40). There is large agreement that the two aplacophoran taxa share a similar radular type, regardless whether it is regarded as a strictly distichous or a monoserial dicuspid radula (see Salvini-Plawen, 1988:355-359 for detailed dis- cussion). In contrast, Polyplacophora and the conchiferan taxa share a polystichous rasping tongue for grazing with several to many teeth per row. None of the outgroups pos- sess a radula, thus it would be indeed nonsense to ask which type of radula would have been present, if there would be a radula at all. Ontogenetic patterns of ingroups also do not solve the problem: there is an early stage in the ontogeny of the chiton radula (Sirenko and Minichev, 1975; Salvini-Plawen, 1988: 365) resembling the aplacophoran type. However, one could interpret this similarity either as a recapitulation of the aplacophoran type in the chiton onto- genesis, or assume with equal a priori likelihood a paedo- morphic event for the aplacophoran taxa from a chiton-like predecessor. Thus, in this analysis it remains open which type, the dicuspid/distichous or the rasping type, is the ple- siomorphic condition for Mollusca. If the rasping type is plesiomorphic, the dicuspid / distichous one would be a synapomorphy of Aplacophora; in the opposite case the rasping tongue would serve as a synapomorphy for Testaria (Polyplacophora and Conchifera). In the given situation it is impossible to infer character polarity a priori to the phylo- genetic reconstruction. Nevertheless, the character is not useless for the Aplacophora-problem, because the given distribution makes many possible arrangements of the four- taxa problem such as [(Caudofoveata and Polyplacophora) (Solenogastres and Conchifera)] being less parsimonious than others such as {Caudofoveata [Solenogastres (Polyplacophora and Conchifera)]} or {Conchifera [Polyplacophora [Caudofoveata and Solenogastres)] }. Moreover, if either Aplacophora or Testaria turn out to be paraphyletic based on other characters, reasonable inference of character polarity is possible a posteriori of tree calcula- tion. The various authors recommend different ways out of the problem of inapplicable characters. One possibility is to calculate trees with all possible character states in all cases of inapplicable characters. In the present analysis this would result in trillions of calculations (15 characters, three outgroups, i. e., 2!5 *3 tree calculations) which were far beyond the computer capacity available. To omit such characters from the analysis would strongly influence the most parsimonious tree topology. Herein, inapplicable states are scored with an “x,” which is treated by PAUP as equal to “?” (unknown) (cf. Strong and Lipscomb, 1999). Software and options. The inference of the most parsimonious tree followed the standard procedure: PAUP 4.0 (Beta-Version; Swofford and Begle, 1999) was applied as the parsimony software, input files were done in the NEXUS format. The ACCTRAN option was selected for tree calculation. Tree calculation was followed by a thor- ough a posteriori analysis of characters (see Discussion Part). CHARACTER ANALYSIS A summary of character coding and the matrix are presented in Table 2. Taxa are abbreviated as follows: Ann - Annelida; Biv - Bivalvia, Cau - Caudofoveata, Cep - Cephalopoda, Gas - Gastropoda, Kam - Kamptozoa (= Entoprocta), Pol - Polyplacophora, Sca - Scaphopoda, Sip - Sipuncula, Sol - Solenogastres, Try - Tryblidia. Taxon/p means “taxon partim.” #1. Cuticle: (0) = absent: Try, Biv, Sca, Gas, Cep; (1) = present: Ann, Sip, Kam, Sol, Cau, Pol.. The condition of the unspecialized dorsal epider- mis of the adult is scored. #2. Type of cuticle: (0) = chitinous: Kam, Sol, Cau, Pol; (1) = collagenous: Ann, Sip; (x) = cuticle absent: Try, Biv, Sca, Gas, Cep. Coding of this character has been outlined by Haszprunar (1996a). #3. Aragonitic scales or spicules: (0) = absent: Ann, Sip, Kam, Try, Biv, Sca, Gas, Cep; (1) = present: Sol, Cau, Pol. All three aculiferan taxa have solitary scale- or spicule-building cells; in addition, the Polyplacophora also have multicellular spiculoblasts. #4. Shell: (0) = absent: Ann, Sip, Kam, Sol, Cau; 118 AMER. MALAC. BULL. 15(2) (2000) Table 2. Character coding and data matrix. #1: Cuticle [0 = absent, 1 = present] #2: Cuticle type [0 = chitinous, | = collagenous, x = cuticle absent] #3: Aculiferan condition (0 = absent, | = present] #4: Shell [0 = absent, 1 = shell plates, 2 = shell by shell gland] #5: Periostracum [0 = absent, | = present] #6: Periostracal groove [0 = absent, 1 = present, x = no periostracum] #7: Mantle papillae (0 = absent, | = present] #8: Mantle cavity [0 = absent, 1 = present] #9: Position of mantle cavity [0 = circumpedal, | = posterior, x = no man- tle cavity] #10: Ctenidia [0 = absent, 1 = present] #11: Number of ctenidia [0 = 1 pair,1 = 2,2 =3 to6,3 => 6, x = no cteni- dia] #12: Body wall musculature [0 = ring/diagonal/longitudinal, 1 = other- wise] #13: Longitudinal body muscles [0 = smooth, | = striated, x = no longitu- dinal muscles] #14: Intercrossing dorsoventral muscles (IDVM) [0 = absent, 1 = present] #15: Number of IDVMs [0 = many, | = eight, 2 = less than eight, 3= less than three, x = no IDVM] ORDERED #16: Hydrostatic muscle system [0 = absent, | = present] #17. Specific head retractor: [0 = absent, 1 = present] #18: Pedal digging by hemolymph pressure [0 = absent, | = present] #19: Pedal gland (0 = absent, | = present] #20: Pedal cirri [0 = absent, 1 = present] #21: Coelomatic cavity (histological sense) [0 = absent, 1 = present] #22: Eucoelomate condition [0 = absent, 1 = present] #23: Heart with pericardium [0 = absent, | = present] #24: Circulatory system [0 = pseudovessels, 1 = endothelial, 2 = sinusial] #25: Pericardioducts [0 = absent, 1 = present, x = no pericardium] #26: Formation of Coelomoduct [0 = ingrowth, | = outgrowth, x = no coelomoduct] #27: Number of Coelomoducts [0 = one pair, 1 = two pairs, 2 = more than two pairs, x = no coelomoduct] #28: Podocytes [0 = absent, 1 = present] #29: Protonephridia [0 = absent, 1 = present] #30: Rhogocytes (0 = absent, | = present] #31: Number of gonads [0 = single “right” (pretorsional left), 1 = single right, 2 = one pair, 3 = two pairs, 4 = more than two pairs] #32: Position of gonad [0 = dorsal to gut; 1 = ventral to gut; ? = not clear] #33: Urinogenital [0 = absent, 1 = present] #34: Gamete release through pericardium [0 = absent, | = present] #35: Molluscan Cross [0 = absent, | = present] #36: Jaws [0 = absent, | = present] #37: Radula [0 = absent, 1 = present] #38: Radular membrane [0 = absent, | = present, x = radula lacking] #39: Rasping tongue [0 = absent, 1 = present, x = radula lacking] #40: Buccal cartilage [0 = absent, 1 = present, x = no radula] #41: Oesophageal pouches [0 = absent, 1 = present] #42: Highly glandular midgut [0 = no, | = yes] #43: Subdivided midgut for sorting and uptake of food [0 = absent, 1 = present] #44: Bilobed midgut gland [0 = absent, | = present, x = no midgut gland] #45: Crystalline style [0 = absent, 1 = present] #46: Intestinal loops [0 = absent, 1 = longitudinal, 2 = unidirectional, 3 = bidirectional] #47: Position of anus [0 = opposite of oral opening, 1= near mouth open- ing at dorsal side, 2 = near dorsal opening at ventral side] #48: Tetraneury [0 = absent, 1 = present] #49: Precerebral ganglia [0 = absent, 1 = present] #50 Pedal ganglia [0 = absent, | = present, x = no pedal nervous system] #51: Visceral loop and IDVM [0 = between, | = outside, 2 = inside, x = no IDVM] #52: Visceral commissure [0 = suprarectal, 1 = subrectal, x = homology unclear] #53: Innervation of the shell(-plate) margin. [0 = also visceral, 1 = only cerebropleural, x = no shell-plate] #54: Cerebral eyes [0 = absent, 1 = present] #55: Statocysts [0 = absent, 1 = present] #56: Osphradia [0 = absent, 1 = present] #57: Position of osphradia [0 = pallial, 1 = extrapallial, x = no osphradi- um] #58: Subradular organ [0 = absent, | = present] (1) = shell plates: Pol; (2) = shell by shell gland: Try, Biv, Sca, Gas, Cep. The homology between the polyplacophoran shell- plates and the conchiferan shell is doubtful (e. g. Haas, 1981), therefore both stages are coded. #5. Periostracum: (0) = absent: Ann, Sip, Kam, Sol, Cau; (1) = present: Pol, Try, Biv, Sca, Gas, Cep. Periostracum is meant here in a very general way, namely as a purely organic layer covering a shell or shell plate. #6. Periostracal groove: (0) = absent: Pol; (1) = present: Try, Biv, Sca, Gas, Cep; (x) = no periostracum: Ann, Sip, Kam, Sol, Cau. Recent fine-structural investigations (Schaefer and Haszprunar, 1997b) revealed significant differences of the organization of the mantle margin between Neopilinidae and other conchiferans. Nevertheless the periostracal groove itself is regarded as homologous throughout the Conchifera. #7. Mantle papillae: (0) = absent: Ann, Sip, Kam, Try, Biv, Sca, Gas/p, Cep; (1) = present: Sol, Cau, Pol, Gas/p. Reind! and Haszprunar (1996a,b; Reindl et al., 1995, 1997) investigated the fine-structure and immuno- cytochemistry of shell pore contents (so-called papillae and caeca) of various molluscan groups and compared them among each other and with the caeca of articulate Brachiopoda. There is a striking similarity between poly- placophoran aesthetes and brachiopod caeca, whereas the bivalve caeca show entirely different structure and mode of formation. The available data (Hoffmann 1949, Fischer et al. 1980, 1988; Scheltema et al. 1994) and personal, unpub- lished fine-structural (TEM) studies on the mantle papillae of Solenogastres and Caudofoveata cannot exclude a possi- ble homology between these papillae and the polypla- HASZPRUNAR: IS APLACOPHORA MONOPHYLETIC? Table 2. Continued MATRIX 1 2 character 12345 67890 1 23 4 5 6 7 8 90 12345 Annelida 11000 x00x0 x 01 0 x 0 0 O OO 1100x Sipuncula 11000 x00x0 x O1 0 %*. O x O OO 1100x Kamptozoa 10000 x00x0 x 1x 0 x O x O 01 0001x Solenogastres 10100 x1110 x 00 1 0 0 0 O 11 10111 Caudofoveata 10100 x11l1l 0 00{0,1}0 0 O Q OO 10111 Polyplacophora 10111 01101 3 1x 1 1 #0 0 0 10 10111 Tryblidia 0x021 10101 2 1x 1 1 #0 0 O OO 10110 Bivalvia OxO021. TOOT “0 dx 2 22. 0) “Oesd - 00. 10171 Scaphopoda 0x021 10100 x 1x 1 3{0,1})0{0,1}00 10011 Gastropoda Ox021 101211 0 dx 2 3 21 1 0 410 10111 Cephalopoda Qx021 10111{0,1)1x ? 3 1 #1 O OO 10121 cophoran macroaesthetes, although there are no distinct similarities as in the case of the Brachiopoda. However, if homology is assumed, then the fissurellid (but not bival- vian) caeca need to be included, accordingly the Gastropoda are scored by {0,1}. #8. Mantle cavity: (0) = absent: Ann, Sip, Kam; (1) = present: Sol, Cau, Pol, Try, Biv, Sca, Gas, Cep. The reduction and loss of the mantle cavity in vari- ous gastropod and certain bivalve taxa is considered as a secondary matter of multiple convergence, therefore both are coded (1). #9. Position of mantle cavity: (0) = circumpedal: Pol, Try, Biv, Sca; (1) = posterior: Sol, Cau, Gas, Cep; (x) = no mantle cavity: Ann, Sip, Kam. Hoffmann (1949) and Salvini-Plawen (e. g. 1981, 1991) considered the mantle cavity of the Solenogastres as reduced and narrowed, yet of the circumpedal type found in Polyplacophora, Tryblidia, Bivalvia, and Scaphopoda. In all the latter groups the visceral nerve cord surrounds the dorsoventral (shell-) muscle bundles and directly innervates the mantle epithelium and the various organs. In contrast, the visceral cords of Solenogastres run between the two pairs of dorsoventral muscles (see Haszprunar, 1989), and innervation of the peripedal groove has never been shown. Thus, there is no reason to homologize the peripedal groove between the foot-sole and the cuticularized mantle with a mantle cavity. The mantle cavity of Solenogastres is there- fore coded as “posterior” (1). The neural condition of the Caudofoveata is less clear because of the reduction of the dorsoventral muscles. E19 3 4 5 6 8 9 0 1 2 3 4 +5 67890 1234 5 678 9 O 12345 678 0 1{0,1}0 4 ? O »% O 10xxx 000x O 000 O O xxx10 0x0 0 1 0 0 2 ? O %* 1 OOxxx 000x O 110 0 O xxxl0 0x0 x 0 1 0 2 0 1 x +O OOxxx 010x O O01? O x xxx10 0x0 ? 1 ? 1 2 +O O{0,1}1 01000 010x O 001{0,1)0 00x00 110 ? 1? 2 2.0.0. 2 2.02100: 0110{0;2)001, 1. 0 00x00 110 1 hook 22 O12 Or A OLLI, ATLL. 0 301. 0° 0 20000. 101 2 1? 612,371. OO 0: “FP 21ITd Tid 1° .<202 0 °0 TLO0T “OxL 1 To 1. 20 00. | 1 0Oxxx, TTI “1 301 30 cL) 110021100 2 Lod 2-2 Or 0 OF Ty TITIT VAT. Oy (324, 0) 1 LLOOL- Oxi al 12 2 2000, 2)0° 0 2 DLATE ATT 0). 1) 327. 0: 00 “2L1TFY 107 1.(0;,1}:2) “Oo tls 3}0. 210-0) LITT 0211. 0° -321.-0° “A 221. LOL However, there is no doubt that the caudofoveate mantle cavity 1s purely posterior. The anlage of the gastropod mantle cavity is situat- ed posteriorly, therefore gastropods are scored by (1). Because the character is inapplicable in all out- groups the coding of the circumpedal and posterior stage with (0) or (1) does not imply any decision concerning polarity. #10. Ctenidia: (0) = absent: Ann, Sip, Kam, Sol, Sca; (1) = present: Cau, Pol, Try, Biv, Gas, Cep. In recent times Morton (1988), Lindberg (1989), and also Ponder and Lindberg (1997: 112-113) expressed doubts about the homology of the ctenidia between the molluscan classes. Whereas there is high probability that the respiratory surfaces of the ctenidium evolved several times in molluscan evolution, the common ancestry of this originally ventilatory organ (i. e., causing water current; see Haszprunar, 1992) is well supported by shared position, structure, and innervation (see Haszprunar, 1987a: fig. 5). All authors agree that lack of ctenidia is a secondary phenomenon in the Gastropoda, therefore they are coded by Ch). #11. Number of ctenidial pairs: (0) = | pair: Cau, Biv, Gas, Cep/p; (1) = 2 pairs: Cep/p; (2) = 3 to 6 pairs: Try; (3) = more than 6 pairs: Pol; (x) = no ctenidia: Ann, Sip, Kam, Sca. Recent, unpublished observation on Micropilina minuta revealed the presence of four pairs of ctenidia, so that the Tryblidia exhibit a continuous range from 3 to 6 ctenidial pairs. 120 AMER. MALAC. BULL. 15(2) (2000) As outlined by Yonge (1939) the condition of the mantle cavity and ctenidial arrangement differ significantly between the Lepidopleurida and the remaining Polyplacophora (Chitonida). Whereas in the former group the ctenidial number increases with size towards the anal opening, in the Chitonida the ctenidia multiply forwards when becoming larger. Accordingly the multiplication of ctenidia is an independent matter in Lepidopleurida and Chitonida and is not used as a synapomorphy of the Polyplacophora. Because of these circumstances all charac- ter states are coded as unordered. #12. Body wall musculature: (0) = circular/diago- nal/longitudinal: Ann, Sip, Sol, Cau; 1 = otherwise: Kam, Pol, Try, Biv, Sca, Gas, Cep. This character replaces the more obscure “worm- like shape” by an observable character, i. e., the presence of a distinct body wall musculature, which is composed of outer circular, intermediate diagonal, and inner longitudinal muscle fibers. This condition is typical for Sipuncula and Polychaeta; among the Mollusca it exists solely in the apla- cophoran taxa. Similar conditions in various groups of opisthobranch or pulmonate slugs are without doubt sec- ondary conditions, because all earlier (5S to 10) gastropod clades lack this condition (Haszprunar, 1988; Ponder and Lindberg, 1997). #13. Structure of the longitudinal muscles of body wall: (0) = smooth: Sol, Cau; (1) = striated: Ann, Sip; (x) = no longitudinal muscles: Kam, Pol, Try, Biv, Sca, Gas, Cep. Annelids and sipunculans are known to have obliquely striated longitudinal muscles, whereas those in the aplacophoran taxa are smooth. For all other taxa the character is inapplicable. Contrary to Salvini-Plawen (1981, 1991) I regard the homology of the longitudinal enrolling muscles in Polyplacophora and Solenogastres as doubtful. The enrolling muscles of Solenogastres and Caudofoveata are a specialized, latero-ventral part of the longitudinal layer of the body wall musculature. Recent studies on the myogene- sis of the chiton Mopalia muscosa revealed that the enrolling muscle is in principle a ring-system independent of the original body wail muscle grid, and that the enrolling function is provided by the transverse stiffness of the shell- plates (Haszprunar and Wanninger, 2000). Purely pedal position and innervation also exclude homology of both lat- erally situated and innervated enrolling muscles with the suctorial muscle of extant Monoplacophora. #14. Intercrossing of the inner dorsoventral mus- culature (IDVM): (0) = absent: Ann, Sip, Kam; (1) = pre- sent: Sol, Cau, Pol, Try, Biv, Sca, Gas, Cep. In contrast to other “worms” the Mollusca are char- acterized by a ventral intercrossing of the inner muscle bundles of their dorsoventral (shell-) musculature. Because of the lack of a pedal sole this character is missing in most Caudofoveata, but the genus Scutopus shows the basic con- dition (Salvini-Plawen, 1972: fig. 16), therefore coding is {0,1}. Contrary to the statement by Voltzow (1988), Patellogastropoda also show intercrossing dorsoventral muscles (pers. obs.). The conditions in the Cephalopoda are unclear (?) because of the major reconstruction of the foot. #15. Number of dorsoventral muscle pairs: (0) = many: Sol, Cau; (1) = eight: Pol, Try; (2) = less than eight: Biv; (3) = less than three: Sca, Gas, Cep; (x) = no IDVM: Ann, Sip, Kam. There is a general tendency to reduce the number of dorso-ventral muscles in the Mollusca (Haszprunar and Wanninger, 2000), particularly exemplified in the Bivalvia, in which extant Protobranchia and Pteriomorpha have seven to three pairs, and Heterodonta usually three. Scaphopoda show one or two pairs (Steiner, 1992). The anterior pair of the cephalopod “‘depressores infundibuli” is the head retractor (see below), so that Cephalopoda have a single pair. Recent ontogenetic data on the myogenesis of Gastropoda confirmed the presence of a single pair of shell muscles even in cases of secondary splittings such as in Patella (Wanninger et al., 1999). Because this is a continuous series of reductions, it makes sense to code this multistate character as “ordered.” Indeed, this option is crucial for the resolution of the conchiferan taxa (see discussion). #16. Hydrostatic muscular system: (0) = absent: Ann, Sip, Kam, Sol, Cau, Pol, Try, Biv, Sac/p; (1) = pre- sent: Sca/p, Gas, Cep. As outlined by Haszprunar (1988: 405) cephalopods and gastropods share a “hydrostatic muscular system” meaning that extension of body parts or tentacles is caused by muscle contraction analogous to the vertebrate tongue rather than by hemolymphatic pressure. According to Shimek and Steiner (1997) the same is true for the foot of dentaliidan scaphopods explaining the ability of rapid extension and burrowing of the latter organ. #17. Specific head retractor: (0) = absent: Ann, Sol, Cau, Pol, Try, Biv, Sca; (1) = present: Gas, Cep; (x) = no head: Sip, Kam. Gastropoda and Cephalopoda share a free head, which is separately retractable by a specific head retractor. Limpets in particular show often a distinct insertion scar of this head retractor, in the Cephalopoda these are the anteri- or pair of the “depressores infundibuli.” Contrary to the statement that also Scaphopoda have a free head (e. g. Waller, 1998) the latter have a free buccal cone alone, whereas the remaining head (cerebral and buccal mass) is fixed (Shimek and Steiner, 1997). #18. Pedal digging by hemolymph pressure: (0) = absent: Ann, Sip, Kam, Sol, Cau, Pol, Try, Sca/p, Gas, Cep; (1) = present: Biv, Sca/p. HASZPRUNAR: IS APLACOPHORA MONOPHYLETIC? 12] Bivalves and gadilidan Scaphopoda (Shimek and Steiner, 1997) use their foot for digging by means of hemolymph pressure in a soft sediment. Caudofoveata show a similar feature but use the cerebrally innervated head-region for digging as secondarily achieved by e. g. the naticid caenogastropods (head and foot) or bullomorph opisthobranchs. To the contrary many bivalve taxa secon- darily settle on hard substrates by means of a byssus. #19. Pedal gland: (0) = absent: Ann, Sip, Kam, Cau, Biv, Sca, Cep; (1) = present: Sol, Pol, Gas. A true pedal gland is herein defined as a subepithe- lial gland at or near the anterior margin of the foot sole. Accordingly the “pedal glands” of Tryblidia and the “‘fun- nel gland” of coleoid cephalopods (Nautilus lacks a funnel gland) do not fit this definition, since both consist of purely epithelial mucous cells. Polyplacophora are scored by (1), since a true pedal gland occurs in early juveniles. The so-called “lip gland” of the sipunculid pelagosphaera larva is a cerebrally innervated structure (Rice, 1993) and thus is not homologous (Scheltema, 1993) to the pedally innervated molluscan gland (Gerould, 1907). #20. Pedal cirri: (0) = absent: Ann, Sip, Cau, Pol, Try, Biv, Sca, Gas, Cep; (1) = present: Kam, Sol. Haszprunar (1986; see also Scheltema et al. 1994) described the ultrastructure of pedal cirri in the “pedal pit” of certain Solenogastres. Very similar structures occur at the anterior margin of the gliding sole of kamptozoan lar- vae (Nielsen, 1971; Haszprunar et al., 1995), although fine- structural studies are still lacking. #21. Coelomatic cavities: (0) = absent: Kam; (1) = present: Ann, Sip, Sol, Cau, Pol, Try, Biv, Sca, Gas, Cep. There is no doubt that molluscs are coelomate in the histological and embryological sense, i. e., there are meso- dermal epithelial cavities, namely the gonopericardial sys- tem, out of the 4d-blastomere. #22. Eucoelomate condition: (0) = absent: Kam, Sol, Cau, Pol, Try, Biv, Sca, Gas, Cep; (1) = present: Ann, Sip. Bartolomaeus (1993, 1994, Salvini-Plawen and Bartolomaeus, 1995, Haszprunar, 1996a, b) defined the eucoelomatic condition (#22) by the checkable feature that the inner wall of the coelomic cavity forms the epithelio- muscular layer of the gut. This condition is not present in any mollusc, but is (among many other phyla) found in Sipuncula and Annelida. #23. Heart in pericardium: (0) = absent: Ann, Sip, Kam, Sca; (1) = present: Sol, Cau, Pol, Try, Biv, Gas, Cep. The specific structures of the molluscan heart are unique and constitute a synapomorphy for the phylum. One known tryblidian genus (Micropilina), the Scaphopoda, and certain opisthobranchs (e. g. Alderia modesta, Rhodope spp.) have lost the heart secondarily. Scaphopoda are unique in retaining at least the pericardial cavity and its function concerning ultrafiltration (Reynolds, 1990). #24. Circulatory system: (0) = pseudovessels: Ann, Sip; (1) = mainly sinusial: Kam, Sol, Cau, Pol, Try, Biv, Sca, Gas; (2) = mainly endothelial: Cep. As outlined by Bartolomaeus (1993) and Haszprunar (1996a), Kamptozoa and Mollusca share a cir- culatory system of sinuses. Pseudovessels, i. e., a lining by outwards orientated epithelia, are found in most eucoelo- mates (here Sipuncula and Annelida), whereas true, endothelial vessels occur in certain stylommatophorans (capillaries; cf. Luchtel et al. 1997) and cephalopods (all except capillaries, cf’ Budelmann et al., 1997). #25. Pericardioduct: (0) = absent: Try; (1) = pre- sent: Sol, Cau, Pol, Biv, Sca, Gas, Cep; (x) = no pericardi- um: Ann, Sip, Kam. With the notable exception of Tryblidia (Haszprunar and Schaefer, 1997a; Schaefer and Haszprunar, 1997a), all molluscs that possess a pericardial cavity also have a peri- cardioduct releasing the modified primary ultrafiltration product into the mantle cavity. Because a true pericardium is lacking in all outgroups, they are scored by (x). #26. Formation of coelomoducts: (0) = ingrowth: Ann, Sip; (1) = outgrowth: Pol, Biv, Gas, Cep; (x) = no coelomoduct: Kam; (?) = unknown: Sol, Cau, Try, Sca. As outlined in detail by Bartolomaeus (1993, 1994) and Salvini-Plawen and Bartolomaeus (1995) the nephridial ducts in molluscs (Polyplacophora, Bivalvia, Gastropoda, and Cephalopoda; no data on the remaining classes) are formed by outgrowth of the coelomic cavity, whereas the (gono-)nephridial ducts in eucoelomates are formed by epi- dermal ingrowth. According to Baba (1938), in the solenogastre Epimenia “there arise on the neck of proctodaeum a pair of short diverticula which may develop into gonoducts [l. e., pericardial = urinogenital ducts].’”’ However, the procto- daeum itself is an outgrowth of the endodermal mass and not an epidermal infolding. I still code the solenogastre condition as unknown. #27. Number of coelomoducts: (0) = one: Sip, Sol, Cau, Biv, Sca, Cep/p; (1) = two: Cep/p; (2) = more than two: Ann, Try; (x) = no coelomoduct: Kam. Among the Mollusca only Nautilus shows two pairs of coelomoducts. If (as is done here) the excretory organs are also considered as coelomic cavities (see above), the extant Tryblidia are characterized by several (three to seven) coelomoducts. #28. Podocytes: (0) = absent: Ann/p, Kam; (1) = present: Ann/p, Sip, Sol, Cau, Pol, Try, Biv, Sca, Gas, Cep. Podocytes have been described in all molluscan classes. Adults of certain polychaete taxa have solenocytes instead of podocytes (e. g. the recent review by Bartolomaeus, 1999). #29, Protonephridia: (0) = absent: Sip; Cep; (1) = 122 AMER. MALAC. BULL. 15(2) (2000) present: Ann, Kam, Pol, Biv, Sca, Gas; (?) = unknown: Sol, Cau, Try. Based on the discovery of protonephridial cyrto- cytes in the larva of a chiton, Bartolomaeus (1989) postu- lated larval protonephridia as a character of the molluscan ground pattern. Recent personal investigations on larvae of Chiton olivaceus confirmed the presence of prominent pro- tonephridia also for that species. Moreover, the recent dis- covery of protonephridia in larvae of the primitive limpet Patella caerulea (Haszprunar and Ruthensteiner, 2000) and in the larva of the scaphopod Antalis vulgatum (Haszprunar et al., 2000), close former significant gaps of occurrence. However, fine-structural data on aplacophoran larvae are badly needed to determine whether protonephridia are also present in Solenogastres and Caudofoveata. #30. Rhogocytes (pore cells): (0) = absent: Ann, Sip, Kam; (1) = present: Sol, Cau, Pol, Try, Biv, Sca, Gas, Cep. Haszprunar (1996a,b) has outlined in detail the sig- nificance of the diagnostic molluscan rhogocyte (often called pore-cell) for general nephridial evolution. In addi- tion, the presence of rhogocytes has been confirmed for both aplacophoran taxa (pers. obs.) and scaphopods (G. Steiner, pers. comm.). #31. Number of Gonads: (0) = single “right” (pre- torsional left): Gas; (1) = single right: Sca, Cep/p; (2) = one pair: Sip, Kam, Sol, Cau, Pol, Try/p, Biv, Cep/p; (3) = two pairs: Try/p, Cep/p; (4) = more than two pairs: Ann. The majority of molluscan classes show a single pair of gonads, but there is gonadal asymmetry in Cephalopoda in part (left side), Scaphopoda in general (right side), and Gastropoda in general (posttorsional right = pretorsional left side) show gonadal asymmetry. Nautilus and Neopilinidae (Micropilina arntzi has a single pair of gonads: Haszprunar and Schaefer, 1997b) have multiple gonads. #32. Position of Gonad: (0) = dorsal of gut: Kam, Sol, Cau, Pol, Biv, Sca, Gas/p, Cep; (1) = ventral of gut: Try, Gas/p; (?) = equivocal: Ann, Sip. Both aplacophoran taxa have dorsal gonads. According to Ponder and Lindberg (1997: 128) “the gonad lies dorsally in polyplacophorans, scaphopods, bivalves, cephalopods, and in all gastropods except patellogas- tropods, in which it is ventral as in monoplacophorans.” Sipuncula and Annelida are equivocal in this character, Kamptozoa clearly have a dorsal gonad. #33. True gonoducts: (0) = absent: Ann, Sip, Sol, Cau, Try, Biv, Sca, Gas; (1) = present: Kam, Pol, Cep. Among molluscs true gonoducts are restricted to Polyplacophora and Cephalopoda, although this condition also occurs secondarily in the Bivalvia (e. g. Mackie, 1984) and Gastropoda (e. g. Haszprunar, 1988; Ponder and Lindberg, 1997). In Solenogastres only the genus Phyllomenia possesses true gonoducts (see below). #34. Release of gametes through the pericardi- um: (0) = absent: Sol/p, Pol, Try, Biv, Sca, Gas, Cep; (1) = present: Sol/p, Cau; (x) = no pericardium: Ann, Sip, Kam. Both Salvini-Plawen (1972, 1981, 1985, but see 1991:17) and Scheltema (1996) agree in regarding the pas- sage of gametes through the pericardium as derived. However, whereas Scheltema regards this character as a synapomorphy of Aplacophora, Salvini-Plawen argued in favour of a convergent evolution of this character because of (secondary) elongation and lateral enrolling of the body. The latter view is based on the exceptional condition of the solenogastre genus Phyllomenia, in which true gonoducts do exist as well as true pericardioducts (Salvini-Plawen, 1970). As outlined in the general character analysis, the lack of a heart in all potential outgroups hinders the direct application of the outgroup-criterion for evaluating the polarity of this character. Concerning ingroup comparison and patterning there are three a priori possibilities: (1) The conditions in Phyllomenia (and in part Dorymenia) repre- sent a secondary atavistic reversal to the ancestral mollus- can features. This would be the only explanation in regard- ing “‘pericardial release of gametes” as a synapomorphy for Aplacophora. However, because of the highly complex genital system of Phyllomenia (Salvini-Plawen, 1970), this assumption is very unlikely. (2) Phyllomenia conditions represent a retained primitive feature, implying convergent evolution of heart passage at least once in the remaining Solenogastre (if Phyllomenia is the sister taxon of them) and in the Caudofoveata. (3) The general aplacophoran condition is plesiomorphic for Mollusca and Phyllomenia represents a parallel, derived condition (see Salvini-Plawen, 1991: 17 tor similar ideas). If so, the condition “evolution of true gonoducts” is paralleled in the Polyplacophora as a whole and within several lineages of Bivalvia and Gastropoda (see above). I regard this assumption as the most probable view. For the cladistic analysis Solenogastres were coded {0,1}, since both conditions occur within the taxon and (contrary to Bivalvia and Gastropoda) the plesiomorphic one is not clear. #35. Cleavage with “molluscan cross”: (0) = absent: Ann, Kam, Cep; (1) = present: Sip, Sol, Pol, Biv, Sca, Gas; (?) = unknown: Cau, Try. For discussion see Haszprunar (1996a). #36. Jaws: (0) = absent: Sip, Kam, Sol, Cau, Pol, Biv; (1) = present: Ann, Try, Sca, Gas, Cep. There is agreement that the presence of jaws is a conchiferan character being secondarily lost in Bivalvia and several gastropod taxa. #37. Radula: (0) = absent: Ann, Sip, Kam, Biv; (1) = present: Sol, Cau, Pol, Try, Sca, Gas, Cep. HASZPRUNAR: IS APLACOPHORA MONOPHYLETIC? 123 There are no convincing arguments in regarding any buccal structure in any annelidan taxon as a homolog of the molluscan radula. #38. Radular membrane: (0) = absent: Sol; (1) = present: Cau, Pol, Try, Sca, Gas, Cep; (x) = radula lacking: Ann, Sip, Kam, Biv. A true radular membrane, i.e., a distinct layer below the radular teeth proper (see Scheltema et al., 1994: fig. 19; Eernisse and Reynolds, 1994: fig.11A), is present in all molluscan classes with a radula except the Solenogastres, where TEM-studies (Haszprunar in Salvini-Plawen, 1988; Wolter, 1992) suggest a kind of “pre-ribbon” (Wolter, 1992). It should be mentioned that Scheltema (pers. comm.) still insists in the presence of a true radula mem- brane at least in certain solenogastre species, unfortunately there is no TEM-evidence for this point of view. #39. Radular type: (0) = basically distichous/bifid: Sol, Cau; (1) = basically rasping: Pol, Try, Sca, Gas, Cep; (x) = radula lacking: Ann, Sip, Kam, Biv. For discussion see under general methodology. #40. Buccal cartilages: (0) = absent: Sol, Cau; (1) = present: Pol, Try, Sca, Gas, Cep; (x) = radula lacking: Ann, Sip, Kam, Biv. There is a long-lasting and continuing equivocal use of the term “radular bolster’ which may be formed either by more or less vacuolized muscle cells or by true carti- lages. The latter type is restricted to Polyplacophora and Conchifera, whereas the aplacophoran taxa show the mus- cular type. Heterobranch Gastropoda secondarily show the purely muscular type again (Haszprunar, 1988; Ponder and Lindberg, 1997). #41. Oesophageal pouches: (0) = absent: Ann, Sip, Kam, Sol, Cau, Cep; (1) = present: Pol, Try, Biv, Sca, Gas. As outlined in detail by Salvini-Plawen (1988) the presence of broad and glandular oesophageal pouches with a ciliated, dorsal food channel is typical for Polyplacophora and most conchiferan classes at least in their primitive rep- resentatives. The only exception is the Cephalopoda. #42. Highly glandular midgut: (0) = no: Ann, Sip; (1) = yes: Kam, Sol, Cau, Pol, Try, Biv, Sca, Gas, Cep. Whereas the midgut of Sipuncula and Annelida is a more or less simple tube, those of Kamptozoa and Mollusca have large glandular areas. #43. Subdivided midgut fo: sorting and uptake of food: (0) = absent: Ann, Sip, Kam, Sol; (1) = present: Cau, Pol, Try, Biv, Sca, Gas, Cep. All Mollusca except the Solenogastres show distinct functional subdivisions of the midgut into a sorting area (stomach), digestion area (midgut sac or gland), and trans- port tube (intestine). Since none of the outgroups shows this subdivision, it is likely a (syn-?) apomorphic condition of Caudofoveata and all remaining Mollusca. Salvini-Plawen’s (1981, 1988) thorough reviews on the evolution of the molluscan alimentary tract emphasized the differences between the midgut in Caudofoveata and in Testaria. However, having in mind the variability of this region between and within the various conchiferan classes, this seems not a valid argument versus an a priori assump- tion of homology testable by parsimony. To avoid multiple, directly correlated characters, this complex has been taken as a single character, although multiple coding (e. g. for intestine, midgut gland, stomach) would be theoretically possible. #44. Bilobed midgut gland: (0) = absent: Cau; (1) = present: Pol, Try, Biv, Sca, Gas, Cep; (x) = no midgut gland: Ann, Sip, Kam, Sol. Whereas the midgut sac of the Caudofoveata is a solitary structure, the midgut gland of all other groups Is at least bilobed, although secondary asymmetry or multiplica- tion occurs repeatedly among many conchiferan taxa. #45. Crystalline style: (0) = absent: Ann, Sip, Kam, Sol, Cau/p, Pol, Sca, Gas/p, Cep; (1) = present: Cau/p, Try, Biv, Gas/p. Chaetodermatid caudofoveates have a gastric shield and a protostyle, whereas the remaining Caudofoveata lack these features. Tryblidia, Bivalvia and many gastropod taxa have so-called crystalline styles or protostyles in their stomach (Salvini-Plawen, 1981, 1988). #46. Intestinal loops: (0) = absent: Ann, Kam, Sol, Cau; (1) = along longitudinal axis: Sip; (2) unidirectional: Try; (3) true bidirectional looping: Pol, Biv, Sca, Gas, Cep. Whereas the gut 1s straight or simply U-shaped in Annelida, Kamptozoa, Solenogastres, and Caudofoveata, intestinal looping is present in the remaining taxa of this study. However, looping is caused by coiling around a lon- gitudinal muscle in Sipuncula, which is not the case in any molluscan taxon. Moreover, the Tryblidia uniquely show unidirectional looping, whereas Polyplacophora, Bivalvia, Scaphopoda, Gastropoda, and Cephalopoda exhibit bidirec- tional loops. #47. Position of anus: (0) = opposite of oral open- ing: Ann, Sol, Cau, Pol, Try, Biv; (1) near mouth opening at dorsal side: Sip, Kam; (2) = near dorsal opening at ven- tral side: Sca, Gas, Cep. Ponder and Lindberg (1997) and Waller (1998) have pointed out that Scaphopoda, Gastropoda and Cephalopoda share a so-called “‘ano-pedal flexure’, whereas the remain- ing molluscan classes the anterio-posterior axis is predom1- nate. #48. Tetraneury: (0) = absent: Ann, Sip; (1) = pre- sent: Sol, Cau, Pol, Try, Biv, Sca, Gas, Cep; (?) = unknown: Kam. Tetraneury 1s a typical molluscan character, whereas Sipuncula and Annelida show a single pair of longitudinal main cords. The conditions in the adult Kamptozoa are not relevant, and those of the larva are largely unknown, there- 124 AMER. MALAC. BULL. 15(2) (2000) fore Kamptozoa are coded by (?). Scheltema (1993, 1996, Scheltema ez al., 1994) considered “‘ganglionated tetraneury” as a synapomorphy of Aplacophora. However, according to personal observations on several genera of Caudofoveata and Solenogastres the so-called “ganglia” are in many species (but not all; Salvini-Plawen, pers. comm.) just thickenings in the lateral or pedal cords and do not fit the usual definition of ganglia as being interconnected by axons only. #49. Precerebral Ganglia: (0) = absent: Ann, Sip, Kam, Sol/p, Pol, Try, Biv, Sca, Gas, Cep; (1) = present: Sol/p, Cau. Precerebral ganglia are generally present in Caudofoveata, whereas they occur only in certain Solenogastres; the latter are therefore coded as {0,1}. Convergent precerebral ganglia are known from a number of interstitial opisthobranchs, but this 1s clearly a secondary condition in gastropods. #50. Pedal ganglia: (0) = absent: Ann, Sip, Sol, Cau, Pol, Try, Gas; (1) = present: Biv, Sca, Cep; (x) = no pedal nervous system: Kam. Among molluscs, true pedal ganglia (versus elon- gated, pedal cords) are restricted to Bivalvia, Scaphopoda, and Cephalopoda, and they occur as a secondary condition in many gastropod groups. #51. Position of visceral loop: (0) = between DVM: Sol, Cau; (1) = outwards DVM: Pol, Try, Biv, Sca; (2) = inwards DVM: Gas, Cep; (x) = no DVM: Ann, Sip, Kam. As outlined earlier (Haszprunar, 1985) the position of the visceral loop (= lateral cord) with respect to the posi- tion of the dorsoventral muscles differs significantly between the molluscan classes. In Caudofoveata and Solenogastres the visceral loop runs between the dorsoven- tral muscle fibers, whereas the visceral loop runs around the shell muscles in the Polyplacophora, Tryblidia, Bivalvia, and Scaphopoda. Gastropoda and Cephalopoda are characterized by a visceral loop lying between the shell muscles enabling these groups to concentrate their nervous system to a great extent. Outgroups lack the specific dorsoventral muscles and are therefore coded as inapplica- ble (x). #52. Position of visceral commissure: (0) = suprarectal: Sol, Cau, Pol; (1) = subrectal: Try, Biv, Sca, Gas, Cep; (x) = homology unclear: Ann, Sip, Kam. The homology of the lateral (visceral) and pedal cord to those of the outgroups is very questionable. Reisinger (1972) regarded the main cords of Sipuncula and Annelida as homologs to the molluscan pedal cord. Considering the suprarectal commissure in aculiferan mol- luscs and the eucoelomates, it is more likely that the viscer- al cord is the common one. Because of these uncertainties the outgroups are here coded by (x). #53. Innervation of the shell(-plate) margin: (0) = cerebropleural and visceral: Pol, Try, Biv, Sca; (1) = only cerebropleural: Gas, Cep; (x) = no shell-plate: Ann, Sip, Kam, Sol, Cau. This is a new character that has not been considered previously for molluscan relationships. The margin (or growth-zone) of the shell(-plates) of Polyplacophora (Eernisse and Reynolds, 1994), Tryblidia (Lemche and Wingstrand, 1959), Bivalvia (Haas, 1935), and Scaphopoda (Shimek and Steiner, 1997) are innervated by the lateral cords and by the cerebropleural region. In contrast, the shell-margin of Gastropoda (e. g. Fretter and Graham, 1962) and Cephalopoda (Young, 1965) are innervated sole- ly by the pleural region or ganglia, but not by the visceral loop. Since the cerebropleural region of gastropods is not involved in the torsion process, the orientation of the adult gastropod shell (teleoconch) is identical to those of the pre- torsional ancestor or to the sister group Cephalopoda. #54. Cerebral (pretrochal) eyes: (0) = absent: Sol, Cau, Pol, Try, Biv, Sca; (1) = present: Ann, Sip, Kam, Gas, Cep. The homology of metazoan cephalic eyes is still a matter of considerable debate, although there is increasing agreement in favour of multiple convergences despite a common epigenetic “master’’-basis (e. g. Salvini-Plawen and Mayr, 1977; Salvini-Plawen, 1982; Zuker, 1994; Nilsson, 1996; Gehring and [keo, 1999; Meyer-Rochow, 2000). In the Mollusca only Gastropoda and Cephalopoda show cerebrally innervated (i. e., pretrochal) eyes, whereas superficially similar, but pleurally-laterally innervated (i. e., posttrochal) photoreceptive organs in polyplacophoran and bivalve larvae or juveniles do not fulfill the criterion “cere- bral”. All outgroups possess cerebral (pretrochal) eyes in the given definition either as adults or at least in the larval stage, although homology is doubtful because of significant differences in their fine-structure (e. g. Woollacott and Eakin, 1973; Verger-Bocquet, 1992; Bartolomaeus, 1992; Blumer, 1997). #55. Paired statocysts: (0) = absent: Ann, Sip, Kam, Sol, Cau, Pol; (1) = present: Try, Biv, Sca, Gas, Cep. Among molluscs paired statocysts are restricted to the conchiferan classes, although analogous, unpaired grav- ity receptors have been described in some Solenogastres (Haszprunar, 1986, Scheltema et al., 1994). #56. Osphradia: (0) = absent: Ann, Sip, Kam, Try, Sca; (1) = present: Sol, Cau, Pol, Biv, Gas, Cep. Because of the correlation in position and innerva- tion Haszprunar (1987a, b) pointed out the homology of chemoreceptive sensory organs known as “dorsoterminal ” “Tacaze’s Organ,” or 99 66 sense organ”, “Geruchsorgan, “osphradium.” #57. Position of osphradia: (0) = pallial: Pol, Biv, HASZPRUNAR: IS APLACOPHORA MONOPHYLETIC? 125 Gas, Cep; (1) = extrapallial: Sol, Cau; (x) = no osphradium: Ann, Sip, Kam, Try, Sca. Both aplacophoran taxa are characterized by extra- pallial osphradia (dorsoterminal sense organs), whereas the osphradium of the remaining classes, if it is present, is pal- lial. #58. Subradular Sense Organ: (0) = absent: Ann, Sip, Kam, Sol, Cau, Biv; (1) = present: Pol, Try, Sca, Gas, Cep. Despite the repeated claim of Heath (1904, 1911) that a subradular organ is present in aplacophorans, it is lacking there (e. g. Salvini-Plawen, 1978, 1985; Scheltema et al., 1994). The presence of a subradular organ is restrict- ed to Polyplacophora, Tryblidia, Scaphopoda, Cephalopoda (Nautilus), and Gastropoda, where it is lost independently in several subclades. CHARACTERS BEING EXCLUDED FROM THIS ANALYSIS Cephalic appendages: Although a general cerebral innervation is present, the probability of homology of cephalic appendages between the molluscan taxa and the selected outgroups is extremely low, because their anlagen are placed anteriorly (gastropod tentacles) versus posterior- ly (bivalve oral lappets, scaphopod captaculae) of the pro- totroch. Therefore the present analysis does not consider this character. Oral lappets: Waller (1998) proposed the lack of oral lappets as a synapomorphy of Scaphopoda, Gastropoda, and Cephalopoda. However, many gastropods have prominent oral lappets, and cephalopods are equipped with oral tentacles. Therefore this character is not scored herein. Epipodial projections: Homologization of the scaphopod pedal flaps with the originally posterior epipodi- al tentacles and the cephalopod funnel (Waller, 1998) is extremely doubtful. Whereas the pedal flaps and the funnel are locomotory organs, gastropod epipodial tentacles are sensory structures. In Patellogastropoda the epipodial tenta- cles are present already prior to metamorphosis (Wanninger et al., 1999). Chitinized gill support: Waller (1998) proposed this as a synapomorphy for all Conchifera except Tryblidia. However, basal clades of Gastropoda (Patellogastropoda, Neritimorpha) lack ctenidial skeletal rods. In cephalopods the ctenidial skeleton is situated in the afferent axis, where- as gastropods and bivalves have efferent rods. Therefore homology of the ctenidial skeleton is very improbably and the character is not coded herein. Position of ctenidia or secondary gills: A posteri- or (versus lateral and posterior) position of ctenidia or sec- ondary gills has been considered as a synapomorphy for Scaphopoda, Gastropoda, and Cephalopoda by Waller (1998). However, lepidopleuran Polyplacophora and proto- branch Bivalvia also have purely posteriorly situated cteni- dia, and certain Patellogastropoda (e. g. Patella) exhibit also lateral gills. Therefore this character is not coded in this study. RESULTS The parsimony algorithm revealed a single most parsimonious tree with 95 steps, CI (excluding uninforma- tive characters) = 0.701; RI = 0.780 and RC = 0.566 (Fig. 2). Various re-arrangements of outgroups always show (1) that Kamptozoa and Mollusca as well as Annelida and Sipuncula are sister taxa. (2) The *Aplacophora* and *Aculifera* appear as basal and paraphyletic Mollusca, with the Solenogastres, Caudofoveata, and Polyplacophora as subsequent offshoots. (3) Thus, the monophyly of Testaria (Polyplacophora and Conchifera), further of Conchifera and Cyrtosoma (or better Visceroconcha: Gastropoda and Cephalopoda) is confirmed. Coding of character #15 as unordered results in three most parsimonious trees again with 95 steps. Tree #2 of these is identical to that of Fig. 2, the (majority rule and strict) consensus of all most parsimonious trees shows a polytomy: Tryblidia, Bivalvia, [Scaphopoda (Gastropoda & Cephalopoda)]. DISCUSSION Solenogastres as the first molluscan offshoot and the Hepagastralia-concept It was not before the recent cladistic study by Salvini-Plawen and Steiner (1996) that anyone regarded the Solenogastres as the earliest molluscan offshoot and con- sidered the monophyly of the remaining classes. Whereas Ivanov (1996), Scheltema (1993, 1996) and Waller (1998) favoured the Aplacophora-concept, also earlier papers by Salvini-Plawen (mainly 1972, 1981, 1985, 1991) argued for Caudofoveata versus the remaining classes (Adenopoda- concept). Despite some differences in character coding (see above), the present result is in full accordance with that of Salvini-Plawen and Steiner (1996). Herein I name the monophyletic clade consisting of Caudofoveata and Testaria as Hepagastralia! reflecting the main synapomor- phy, the highly distinct and complex subdivision of the midgut. A clade Hepagastralia is additionally supported by I regard it as nonsense to argue about the rank of Hepagastralia as a “‘sub- or infraphylum” or as a “mega- or gigaclass.” 126 AMER. MALAC. BULL. 15(2) (2000) ANNELIDA EUCOELOMATA Out- SIPUNCULA Groups KAMPTOZOA — SOLENOGASTRES = *Aplacophora* *Aculi- . CAUDOFOVEATA L fera* 5 8 POLYPLACOPHORA — U E Ss Pp T TRYBLIDIA C A E | A G S BIVALVIA A ak *DIASOMA* CONCHI- Ss A SCAPHOPODA T R FERA R I GASTROPODA A A VISCEROCONCHA L CEPHALOPODA I A - nov. Fig. 2. Single, most parsimonious tree of the current parsimony analysis revealed by coding character #15 as ordered. It is identical to tree #2 (of three), if #15 is coded unordered. *Taxon* means a paraphyletic group. further characters, the presence of ctenidia, loss of pedal cirri, and presence of a true radular membrane. None of the latter characters is unequivocal, however (see character analysis and below). Solenogastres show high variability in the principal configuration (anatomy, histology) of the foregut glands and in the number and arrangement of the osphradia (e .g. Salvini-Plawen, 1978). Considering the basal position of the Solenogastres in the molluscan framework, one gets the impression that in the Solenogastres these characters are not yet fuily constrained, in contrast to the remaining class- es. Moreover, the Solenogastres seem to be the only extant molluscan group in which the original, solely ciliary type of locomotion has been retained. In particular the anteriorly placed ciliary “pit”, which consists of compound cilia (Haszprunar, 1986) is strikingly similar to the anterior part of the foot sole of nearly all benthic larvae of Kamptozoa (Nielsen, 1971; see also Haszprunar et al., 1995), although fine-structural details are still lacking in the latter case. Position of Tryblidia Recent microanatomical and ultrastructural investi- gations (Haszprunar and Schaefer, 1997a,b; Schaefer and Haszprunar, 1997a,b) provided evidence that the Tryblidia are not “living fossils” or even “Archi-Mollusca”. Herein the Tryblidia appear as an early but not necessarily earliest (see above) conchiferan offshoot. As outlined above the monophyly of the remaining Character state distribution in Fig. 2. h = homoplasy, r = reversal Eucoelomata: #2:0->1,; #13:0->1; #20: 1->0(h,r); #22:0->1; NN (Sinusoida = Kamptozoa & Mollusca): ?#20:0->1(r); #24:0->1; #42:0- >1; Mollusca: #3:0->1(r); #7:0->1(r); #8:0->1; #9:x->1; #14:0->1; #23:0- >I(r); #25:x->1(r); #26:0->1; #30:0->1; #35:0->1(h); #37:0->1(r); #48:0- >1; #51:x->0; #56:0->1(r); Hepagastralia: #10:0->1(r); ?#20:1->0; #38:0->1; #43:0->1; Testaria: #4:0->1; #5:0->1; #9:1->0; #12:0->1(h); #15:0->1; 2#20:1- >O(h,r); #34:1->0; #39:0->1; #40:0->1; #41:0->1(r); #44:0->1: #46:0->3; #51:0->1; #53:x->0; #57:1->0; #58:0->1(r); Conchifera: #1:1->0; #3:1->O(r); #4:1->2; #6:0->1; #7:1->0(r); #9:1->0(r); #36:0->1(h,r); #45:0->1(r); #52:0->1; #55:0->1; NN (Biv, Sca, Gas, Cep): #15:1->2; #50:0->1(r); NN (Sca, Gas, Cep): #15:2->3; 2#16:0->1(r in Sca and Gas); #45:1->0(r); #47:0->2; Visceroconcha: #9:0->1(r);#17:0->1; #51:1->2; #53:0->1; #54:0->1(h); Annelida (Ann): #27:0->2(h); #31:2->4; #36:0->1(h,r); Sipuncula (Sip): #29:1->0(h); #35:0->1(h); #46:0->1; #47:0->1(h), Kamptozoa (Kam): #12:0->1(h); #21:1->0; #28:1->0; #33:0->1(h); #47:0- >1(h); Solenogastres (Sol): #19:0->1(h); Caudofoveata (Cau): #34:0->1(h in Sol); #49:0->1(h in Sol); Polyplacophora (Pol): #11:0->3; #19:0->1(h); #33:0->1(h); Tryblidia (Try): #11:0->2; #25:1->O(r); #27:0->2(h); #32:0->1; #46:3->2; #56:1->0(r); Bivalvia (Biv): #18:0->1(h in Sca); #36:1->0(r); #37: 1->0(r); #58: 1->0(r); Scaphopoda (Sca): #10:1->0(r); #15:1->4(h); #23:1->0(r); #31:2->1,; #45:1->0(h,r); #56: 1->0(r); Gastropoda (Gas): #15:3->4; #19:0->1(h); #31:2->0; ?#50:1->0(r); Cephalopoda (Cep): #24:1->2; #29:1->O(h); #33:0->1(h); #35:1->0(n); #41:1->0(1); HASZPRUNAR: IS APLACOPHORA MONOPHYLETIC? 127 conchiferan classes depends on the coding of character #15 (number of shell muscles) clade. In any case there 1s no unequivocal support for this clade, therefore this clade is not named. The proposed term “Ganglioneura” (Lauterbach, 1983) is in any case inappropriate, since prim- itive gastropods, bivalves, and cephalopods do not show true ganglia. Position of Scaphopoda Applying the Hennigian method Waller (1998) has strongly argued against the Diasoma-concept (Rostrochoncha, Scaphopoda and Bivalvia monophyletic). Instead he favoured a clade “Gastropoda (Scaphopoda and Cephalopoda).” As outlined in the character analysis many of his proposed synapomorphies for this arrangement are not accepted herein. Nevertheless, the present analysis again contradicts the Diasoma-concept and argues for monophyly of Scaphopoda, Gastropoda and Cephalopoda. However, contrary to Waller (1998) the sister-group rela- tionship of Gastropoda and Cephalopoda is very well sup- ported by no less than four non-homoplastic synapomor- phies. On the other hand, none of the proposed synapomor- phies is unequivocal and all show homoplasies within the terminal taxa. Therefore this clade is not named again. Implications for the groundplan of the Mollusca The consideration of the *Aplacophora* as the basic molluscan level of organization has major implica- tions for the understanding and reconstruction of the mol- luscan stem species (HAM = “Hypothetical Ancestral Mollusc”). Many shared aplacophoran features can now reasonably be considered as characters of HAM (see also Salvini-Plawen, 1972, 1981, 1985, 1991; Haszprunar, 1992; Salvini-Plawen and Steiner, 1996): worm-shaped body with chitinous cuticles covered with aragonitic spicules or scales; distinct body wall musculature; posterior mantle cavity with extrapallial osphradium; distichous radula and carnivory; urinogenital ducts and openings. Up to now all reconstructions of the molluscan archaetype show ctenidia. The lack of ctenidia in the Solenogastres has been regarded as a secondary loss com- parable to those in Scaphopoda and many gastropod taxa, and this scenario is still possible. However, Ockham¥s razor (minimalization of assumptions) favours the assump- tion of a plesiomorphic lack of ctenidia in the Solenogastres. Indeed, there are many larger species of Solenogastres with respiratory folds in the mantle cavity to increase the respiratory capacity, but none of these gill- leaflets shows the specific ctenidial structure. Outlook Is this the final word or the final decision concern- ing the *Aplacophora*? Of course it is not, since phyloge- netics never is a dogma, but is a matter of probabilities depending on the current state of data, character selection, and basic assumptions. In particular ontogenetic data on the aplacophoran taxa are badly needed to clear up certain points such as the possible presence of a pedal sole and gland in larval or early juvenile Caudofoveata. The same is true in particular for the larval charac- ters of the Kamptozoa, which appear again crucial for improving our understanding of molluscan origins and early evolution. Important questions concern presence or absence of locomotory cirri at the anterior end of the foot sole comparable to Solenogastres or the details of the (tetraneural?) nervous system of the larva. I do have some personal experience with these high- ly enigmatic and interesting taxa, but true experts may have a better chance to improve or modify the provided data- matrix based on their long expertise. Nevertheless, it is the argument and the data and not the author who decides the matter. Thus, the rejection of *Aplacophora* as a mono- phyletic taxon and accordingly the alternative Hepagastralia-concept is the most parsimonious assumption and thus the most probable solution at the present stage of knowledge. ACKNOWLEDGMENTS I am deeply indebted to my friend and teacher Dr. Luitfried v. Salvini-Plawen (University of Vienna), who not only opened the apla- cophoran world to me, but also provided many useful data and comments to the draft of this contribution. Thanks also to Amelie H. Scheltema (Woods Hole Oceanographic Institution) for several fruitful discussions on the subject. I am very grateful for many detailed and helpful contributions provided by two anonymous referees. I also thank Drs. Ellis L.Yochelson and Harold B. Rollins, who initiated this contribution and who kindly invited me to present these results at the AMS meeting in Pittsburgh, July 1999. LITERATURE CITED Adamkewicz, S. J., M. G. Harasewych, J. Blake, D. Saudek, and C. J. Bult. 1997. A molecular phylogeny of the bivalve mollusks. Molecular Biology and Evolution 14:19-629. Adoutte, A., G. Bavaloine, N. Lartillot, O. Lespinet, B. Prud’homme, and R. De Rosa. 2000. The new animal phylogeny: reliability and implications. Proceedings of the National Academy of Sciences of the U. S. A. 97:4453-4456. Baba, K. 1938. The later development of a solenogastre, Epimenia verru- cosa (Nierstrasz). Journal of the Department of Agriculture, Kyusyu Imperial University 6:21-40, pl. 1. Bartolomaeus, T. 1989. Larvale Nierenorgane bei Lepidochiton cinereus (Polyplacophora) und Aeolidia papillosa (Gastropoda). Zoomorphology 108:297-307. Bartolomaeus, T. 1992. Ultrastructure of the photoreceptors in certain lar- vae of the Annelida. Microfauna Marina 7:191-214. 128 AMER. MALAC. Bartolomaeus, T. 1993. Die Leibeshdhlenverhaltnisse und Nephridialorgane der Bilateria - Ultrastruktur, Entwicklung und Evolution. Habilitationsschrift at the University of Gottingen, Germany, Summary in Verhandlungen der Deutschen Zoologischen Gesellschaft 86, 1:42. Bartolomaeus, T. 1994. On the ultrastructure of the coelomic lining in the Annelida, Sipuncula and Echiura. Microfauna marina 9:171-220. Bartolomaeus, T. 1999. Structure, function and development of segmented organs in Annelida. Hydrobiologia 402:21-37. Blumer, M. J. F. 1997. The larval ocelli of Golfingia misakiana (Sipuncula, Golfingiidae) and of a pelagosphaera of another unidentified species. Zoomorphology 11:115-120. Budelmann, B. U., R. Schipp, and S. v. Boletzky. 1997. Cephalopoda. In: Microscopic Anatomy of Invertebrates, Vol. 6A: Mollusca II, F. W. Harrison and A. J. Kohn, eds. pp. 119-414. John Wiley & Sons, New York. Eernisse, D. L. and P. D. Reynolds. 1994. Polyplacophora. In: Microscopic Anatomy of Invertebrates, Vol. 5: Mollusca I, F. W. Harrison and A. J. Kohn, eds. pp. 55-110. John Wiley & Sons, New York. Fischer, F. P., W. Maile, and M. Renner. 1980. Die Mantelpapillen und Stacheln von Acanthochiton fascicularis L. (Mollusca, Polyplacophora). Zoomorphologie 94:121-131. Fischer, F. P., B. Eisenhammer, C. Militz, and I. Singer. 1988. Sense organs in the girdle of Chiton olivaceus (Mollusca, Polyplacophora). Helgoldnder wissenschaftliche Meeresunter- suchungen 31:425-443. Fretter, V. and A. Graham. 1962. British Prosobranch Molluscs. Ray Society, London. Gehring, W. J. and K. Ikeo. 1999. Pax6 mastering eye morphogenesis and eye evolution. Trends in Genetics 15:371-377. Gerould, J. H. 1907. Studies on the embryology of the Sipunculidae. II: The development of Phascolosoma. Zoologische Jahrbiicher, Abteilung Anatomie 23:77-162. Ghiselin, M. T. 1988. The origin of molluscs in the light of molecular evi- dence. In: Oxford Surveys in Evolutionary Biology, P. H. Harvey and L. Patridge, eds. Vol. 5:66-95. Oxford University Press, Oxford. Giribet, G. and S. Carranza. 1999. What can 18S rDNA do for bivalve phylogeny? Journal of Molecular Evolution 48:256-258. Gotting, K. -J. 1980a. Origin and relationships of the Mollusca. Zeitschrift fiir zoologische Systematik und Evolutionsforschung 18:24-27. G6tting, K. -J. 1980b. Argumente fiir eine Deszendenz der Mollusken von metameren Antezedenten. Zoologische Jahrbiicher, Abteilung Anatomie 103:211-218. Gustafson, R. G. 1987. Phylogenetic implications of the molluscan perica- lymma larva. American Zoologist 27(4):59A (abstract). Haas, F. 1935. Bivalvia, Teil I. In: Dr. H. G. Bronns Klassen und Ordnungen des Tierreichs. 3. Band: Mollusca. Akademische Verlagsgesellschaft, Leipzig. pp. 1-984. Haas, W. 1981. Evolution of calcareous hardparts in primitive molluscs. Malacologia 21: 403-418. Haszprunar, G. 1985. On the innervation of gastropod shell muscles. Journal of Molluscan Studies 51:309-314. Haszprunar, G. 1986. Feinmorphologische Untersuchungen an Sinnesstrukturen urspriinglicher Solenogastres. Zoologischer Anzeiger 217:345-362. Haszprunar, G. 1987a. The fine morphology of the osphradial sense organs of the Mollusca. III. Placophora and Bivalvia. Philosophical Transactions of the Royal Society of London B317:37-61, 3 plates. Haszprunar, G. 1987b. The fine morphology of the osphradial sense organs of the Mollusca. IV. Caudofoveata and Solenogastres. BULL. 15(2) (2000) Philosophical Transactions of the Royal Society of London B317:63-73, 3 plates. Haszprunar, G. 1988. On the origin and evolution of major gastropod groups, with special reference to the Streptoneura. Journal of Molluscan Studies 54:367-441. Haszprunar, G. 1989. Die Torsion der Gastropoda - ein biomechanischer ProzeB? Zeitschrift fiir zoologische Systematik und Evolutionsforschung 27:1-7. Haszprunar, G. 1992. The first molluscs - small animals. Bollettino di Zoologia 59:1-16. Haszprunar, G. 1996a. The Mollusca: Coelomate turbellarians or mes- enchymate annelids? In: Origin and Evolutionary Radiation of the Mollusca, J. Taylor, ed. pp. 1-28. Oxford University Press, Oxford. Haszprunar, G. 1996b. The molluscan rhogocyte (pore-cell, Blasenzelle, cellule nucale), and its significance for ideas on nephridial evolu- tion. Journal of Molluscan Studies 62:185-211. Haszprunar, G. 1998. Parsimony analysis as a specific kind of homology estimation, and the implications for character weighting. Molecular Phylogeny and Evolution 9:333-339. Haszprunar, G. and B. Ruthensteiner. 2000. Microanatomy and ultrastruc- ture of the protonephridial system in the larva of the limpet, Patella vulgata L. (Mollusca, Patellogastropoda). Journal of Submicroscopical Cytology and Pathology 32:59-67. Haszprunar, G., L. v. Salvini-Plawen, and R. M. Rieger. 1995. Larval planktotrophy - A primitive trait in the Bilateria? Acta Zoologica (Stockholm) 76:141-154. Haszprunar, G. and K. Schaefer. 1997a. Monoplacophora. In: Microscopic Anatomy of Invertebrates, Vol. 6B: Mollusca II, F. W. Harrison and A. J. Kohn, eds. pp. 415-457. John Wiley & Sons, New York. Haszprunar, G. and K. Schaefer. 1997b. Anatomy and phylogenetic signif- icance of Micropilina arntzi (Mollusca, Monoplacophora, Micropilinidae fam. nov.). Acta Zoologica (Stockholm) 77:315- 334. Haszprunar, G. and A. Wanninger. 2000. Molluscan muscle systems in development and evolution. Journal of Zoological Systematics and Evolutionary Research (in press). Haszprunar, G., A. Wanninger, and B. Ruthensteiner. 2000. Comparative fine-structure of the larval protonephridia of primitive Mollusca (Polyplacophora, Scaphopoda, Patellogastropoda). Abstracts of the 66th AMS and 33rd WSM Joint Congress, San Franciso, July 2000: 6. Hawkins, J. A., C. E. Hughes, and R. W. Scotland.1997. Primary homolo- gy assessment, characters and character states. Cladistics 13:275- 283. Heath, H. 1904. The nervous system and subradular organ in two genera of Solenogastres. Zoologische Jahrbiicher, Abteilung Anatomie 20:399-408, plate 27. Heath, H. 1911. Reports on the scientific results of the expedition to the tropical Pacific. XIV. The Solenogastres. Memoirs of the Museum of Comparative Zoology (Harvard) 45(1):17-179. Hofmann, S. 1949. Studien tiber das Integument der Solenogastren, nebst Bemerkungen iiber die Verwandtschaft zwischen den Solenogastres und den Placophoren. Zoologiska Bidrag fran Uppsala 27:293-427. Ivanov, D. L. 1996. Origin of Aculifera and problems of monophyly of higher taxa in molluscs. /n: Origin and Evolutionary Radiation of the Mollusca, J. Taylor, ed. pp. 59-65. Oxford University Press, Oxford. Jenner, R. A. and F. R. Schram. 1999. The grand game of metazoan phy- logeny: rules and strategies. Biological Reviews 74:121-142. Kitching, I. L., P. L. Forey, C. J. Humphries, and D. M. Williams. 1998. Cladistics - The Theory and Practice of Parsimony Analysis. The HASZPRUNAR: IS APLACOPHORA MONOPHYLETIC? 129 Systematics Association Publication No. 11. Oxford University Press, Oxford. Lauterbach, K. -E. 1983. Erérterungen zur Stammesgeschichte der Mollusca, insbesondere der Conchifera. Zeitschrift fiir Zoologische Systematik und Evolutionsforschung 21:201-216. Lee, D. -C. 1999. A reconsideration of the coding of inapplicable charac- ters: Assumptions and problems. Cladistics 15:373-378. Lemche, H. and K. G. Wingstrand. 1959. The anatomy of Neopilina galathea Lemche, 1957 (Mollusca, Tryblidiacea). Galathea Report 3:9-71, pls. 1-56. Lindberg, D. R. 1989. The evolution of respiratory structures in the Conchifera. Abstracts of the 10th International Malacologcal Congress, Tiibingen 1989: 147. Luchtel, D. L., A. W. Martin, I. Deyrup-Olsen, and H. H. Boer. 1997. Gastropoda: Pulmonata. In: Microscopic Anatomy of Invertebrates, Vol. 6B: Mollusca II, F. W. Harrison and A. J. Kohn, eds. pp. 459-718. John Wiley & Sons, New York. Mackie, G. L. 1984. Bivalves. In: The Mollusca. Vol. 7: Reproduction, A. Tompa, N. H. Verdonk, and J. A. M. van den Biggelaar, eds. pp. 351-418. Academic Press, New York. Maddison, W. P. 1993. Missing data versus missing characters in phyloge- netic analysis. Systematic Biology 42:576-581. Meyer-Rochow, V. B. 2000. The eye: monophyletic, polyphyletic or per- haps biphyletic? [includes a reply by W. Gehring] Trends in Genetics 16: 244-245. Morton, B. A. 1988. The pallial cavity. In: The Mollusca. Vol. 12: Form and Function, E. R. Trueman and M. R. Clarke, eds. pp. 253-286. Academic Press, San Diego. Nelson, G. and P. Y. Ladiges. 1993. Missing data and three-item analysis. Cladistics 9:111-114. Nilsson, D. -E. 1996. Eye-ancestry: old genes for new eyes. Current Biology 6:39-42. Nielsen, C. 1971. Entoproct life cycles and entoproct/ectoproct relation- ship. Ophelia 9:209-341. Ponder, W. F. and D. R. Lindberg. 1997. Towards a phylogeny of gastro- pod molluscs: An analysis using morphological characters. Zoological Journal of Linnean Society 119:83-265. Reindl, S. and G. Haszprunar. 1996a. Shell pores (caeca, aesthetes) of Mollusca: a case of polyphyly. In: Origin and Evolutionary Radiation of the Mollusca, J. D. Taylor, ed. pp. 115-118. Oxford University Press, Oxford. Reindl, S. and G. Haszprunar. 1996b. Fine structure of caeca and mantle of arcoid and limopsoid bivalves (Mollusca, Pteriomorpha). Veliger 39:101-116. Reindl, S., W. Salvenmoser, and G. Haszprunar. 1995. Fine structural and immunocytochemical investigations on caeca of Argyrotheca cor- data and A. cuneata (Brachiopoda, Terebratulida). Journal of Submicroscopic Cytology and Pathology 27:543-556. Reindl, S., W. Salvenmoser, and G. Haszprunar. 1997. Fine structural and immunocytochemical studies on the eyeless aesthetes of Leptochiton algesirensis, with comparison to Leptochiton cancel- latus (Mollusca, Polyplacophora). Journal of Submicroscopic Cytology and Pathology 29:135-151. Reisinger, E. 1972. Die Evolution des Orthogons der Spiralier und das ArchicOlomatenproblem. Zeitschrift fiir zoologische Systematik und Evolutionsforschung 10:1-43. Reynolds, P. D. 1990. Functional morphology of the perianal sinus and pericardium of Dentalium rectius (Mollusca: Scaphopoda) with a reinterpretation of the scaphopod heart. American Malacological Bulletin 8:137-146. Rice, M. E. 1993. Sipuncula. In: Microscopic Anatomy of Invertebrates, 12: Onychophora, Chilopoda, and lesser Protostomia, F. W. Harrison and M. E. Rice, eds. pp. 237-325. John Wiley and Sons, New York. Salvini-Plawen, L. v. 1970. Phyllomenia austrina, ein phylogenetisch bedeutsamer Solenogastre (Mollusca, Aculifera). Zeitschrift fiir Zoologische Systematik und Evolutionsforschung 8:297-309. Salvini-Plawen, L. v. 1972. Zur Morphologie und Phylogenie der Mollusken: Die Beziehung der Caudofoveata und Solenogastres als Aculifera, als Mollusca und als Spiralia. Zeitschrift fiir wis- senschaftliche Zoologie 184:205-394. Salvini-Plawen, L. v. 1978. Antarktische und subantarktische Solenogastres (eine Monographie 1898-1974). Zoologica (Stuttgart) 128(44):1-315. Salvini-Plawen, L. v. 1980. A reconsideration of systematics in the Mollusca (phylogeny and higher classification). Malacologia 19:249-278. Salvini-Plawen, L. v. 1981. On the origin and evolution of the Mollusca. In: Origine dei Grande Phyla dei Metazoi. Atti dei Convegni Lincei (Roma) 49:235-293. Salvini-Plawen, L. v. 1982. On the polyphyletic origin of photoreceptors. In: Visual Cells in Evolution, J. A. Westfall, ed. pp. 137-154, Raven Press, New York. Salvini-Plawen, L. v. 1985. Early evolution and the primitive groups. Jn: The Mollusca, Vol. 10: Evolution, E. R. Trueman and M. R. Clarke, eds. pp. 59-150. Academic Press, London. Salvini-Plawen, L. v. 1988. The structure and function of molluscan diges- tive systems. In: The Mollusca, Vol. 11: Form and Function, E. R. Trueman and M. R. Clarke, eds. pp. 301-379. Academic Press, London. Salvini-Plawen, L. v. 1991. Origin, phylogeny and classification of the phylum Mollusca. /berus “1990” 9:1-33. Salvini-Plawen, L. v. and T. Bartolomaeus. 1995. Mollusca: mesenchyma- ta with a “coelom.” Jn: Body cavities: Function and Phylogeny, G. Lanzavecchia, R. Valvassori, and M. D. Candia-Carrnevali, eds. pp. 75-92. Mucci Editore, Modena. Salvini-Plawen, L. v. and E. Mayr. 1977. On the evolution of photorecep- tors and eyes. Evolutionary Biology 10:207-263. Salvini-Plawen, L. v. and G. Steiner. 1996. Synapomorphies and ple- siomorphies in higher classification of Mollusca. In: Origin and Evolutionary Radiation of the Mollusca, J. Taylor, ed. pp. 29-51. Oxford University Press, Oxford. Schaefer, K. and G. Haszprunar.1997a. Anatomy of Laevipilina antarctica, a monoplacophoran limpet (Mollusca) from Antarctic waters. Acta Zoologica (Stockholm) 77:295-314. Schaefer, K. and G. Haszprunar.1997b. Organisation and fine structure of the mantle of Laevipilina antarctica (Mollusca, Mono- placophora). Zoologischer Anzeiger 236: 13-23. Scheltema, A. H. 1988. Ancestors and descendents: Relationships of the Aplacophora and Polyplacophora. American Malacological Bulletin 6:57-61. Scheltema, A. H. 1993. Aplacophora as progenetic aculiferans and the coelomatic origin of mollusks as the sister taxon of Sipuncula. Biological Bulletin, 184:57-78. Scheltema, A. H. 1996. Phylogenetic position of Sipuncula, Mollusca and the progenetic Aplacophora. Jn: Origin and Evolutionary Radiation of the Mollusca, J. Taylor, ed. pp. 53-58. Oxford University Press, Oxford. Scheltema, A. H., M. Tscherkassky, and A. M. Kuzirian. 1994. Aplacophora. In: Microscopic Anatomy of Invertebrates, Vol. 5: Mollusca I, F. W. Harrison and A. J. Kohn, eds. pp. 13-54. John Wiley and Sons, New York. Shimek, R. L. and G. Steiner. 1997. Scaphopoda. In: Microscopic Anatomy of Invertebrates, Vol. 6B: Mollusca 2, F. W. Harrison and A. J. Kohn, eds. pp. 719-781. John Wiley and Sons, New York. 130 AMER. MALAC Sirenko, B. I. and Yu. S. Minichev. 1975. Developpement ontogénétique de la radula chez les Polyplacophores. Cahiers de Biologie Marine 16:425-433. Steiner, G. 1992. The organisation of the pedal musculature and its con- nection to the dorso-ventral musculature in Scaphopoda. Journal of Molluscan Studies 58:181-197. Steiner, G. 1999. What can 18S rDNA do for bivalve phylogeny? - Response. Journal of Molecular Evolution 48:258-261. Steiner, G. and M. Miiller. 1996. What can 18S rDNA do for bivalve phy- logeny? Journal of Molecular Evolution 43:58-70. Strong, E. E. and D. Lipscomb. 1999. Character coding and inapplicable characters. Cladistics 15:363-371. Swofford, D. L. and D. P. Begle. 1999. PAUP: Phylogenetic Analysis Using Parsimony, Version 4.0. Computer Program distributed by MacMillan Inc. USA. Verger-Bocquet, M. 1992. Polychaeta: Sensory structures. In: Microscopic Anatomy of Invertebrates, 7: Annelida, F. W. Harrison and S. L. Gardiner, eds. pp. 181-196. John Wiley & Sons, New York. Voltzow, J. 1988. The organization of limpet pedal musculature and its evolutionary implications for the Gastropoda. In: Prosobranch Phylogeny, W. F. Ponder, ed. Malacological Review, Supplement 4:273-283. Waller, T. R. 1998. Origin of the molluscan class Bivalvia and a phyloge- . BULL. 15(2) (2000) ny of major groups. In: Bivalves: An Eon of Evolution, P. A. Johnston and J. W. Haggard, eds. pp. 1-45. University of Calgary Press, Calgary. Wanninger, A., B. Ruthensteiner, S. Lobenwein, W. Salvenmoser, W. J. A. G. Dictus, and G. Haszprunar. 1999. Development of the mus- culature in the limpet Patella (Mollusca, Patellogastropoda). Development, Genes and Evolution 209:226-238. Wilkinson, M. 1995. Coping with abundant missing entries in phylogenet- ic inference using parsimony. Systematic Biology 44:501-514. Wolter, K. 1992. Ultrastructure of the radula apparatus in some species of aplacophoran molluscs. Journal of Molluscan Studies 58:245-256. Woollacott, R. M. and R. M. Eakin. 1973. Ultrastructure of a potential photoreceptor organ in the larva of an entoproct. Journal of Ultrastructure Research 43:412-425. Yonge, C. M. 1939. On the mantle cavity and its contained organs in the Loricata (Placophora). Quarterly Journal of Microscopic Sciences 81:367-390. Young, J. Z. 1965. The central nervous system of Nautilus. Philosophical Transactions of the Royal Society of London B249:1-25. Zuker, C. S. 1994. On the evolution of eyes: would you like it simple or compound? Science 265:742-643. Date of manuscript acceptance: 23 March 2000 Considerations on Paleozoic Polyplacophora including the description of Plasiochiton curiosus n. gen. and sp. Richard D. Hoare Department of Geology, Bowling Green State University, Bowling Green, Ohio 43403-0218, U.S. A. Abstract: Numerous occurrences of undescribed Paleozoic polyplacophorans have been located, and new taxa of Silurian, Mississippian, and Permain specimens have been recently described. Problems commonly exist in classifying isolated plates at the species level. Sizes and patterns of aesthetes show some differences in plate microstructure between Paleozoic and modern taxa. Recently proposed classifications indicate differences of opinion concerning relationships and inclusion of non-polyplacophoran taxa. A possible, but incomplete, phylogenetic scheme at the family level is presented. Plastochiton curiosus n. gen. and sp. is described from the Devonian of Pennsylvania. Key Words: Mollusca, Polyplacophora, Paleozoic, classification, phylogeny Paleozoic polyplacophorans are not as rare strati- graphically and numerically as one might be led to believe from the published literature, largely as a result of collec- tions remaining undescribed. Collections by the author and associates in Pennsylvanian age strata of the Appalachian basin have found over 5O localities containing one to hun- dreds of plates of polyplacophorans. In many cases, speci- mens are placed in repositories with little or no reference to their existence. Inquiries to museums, universities, and individuals commonly turn up specimens, many of which are important in our understanding of the classificiation and phylogeny of the class. It is likely that individuals process- ing samples for ostracodes and foraminifers, have found, but did not recognize polyplacophoran plates, especially if the plates were fragmented. It is important to illustrate and specify in the literature the location of specimens to make known their existence. Illustrated specimens herein are in the repositories of the National Museum of Natural History (USNM), Ohio University Zoological Collections (OUZC), and New York State Museum (NYSM). RECENT STUDIES Since I began work on this study a number of unde- scribed specimens have been found. These include: two Ordovician plates of Gotlandochiton from Minnesota and two plates of Chelodes from the Ordovician of Alabama (Fig. 1); Silurian specimens from California containing new species of Paleochiton and Thairoplax, only the second known Silurian fauna from North America (Hoare, 2000a); an articulated partial specimen of a new genus from the Devonian of Pennsylvania (described herein); Devonian plates of Arcochiton? and Gotlandochiton and Mississippian plates of Gryphochiton trom Australia (Hoare and Cook, 2000); an abundant Mississippian fauna from Iowa containing new species of Gryphochiton and Euleptochiton and three new genera (Hoare, 2000c), a fauna quite different from that of the Salem Limestone in Indiana described by Kues (1978); two Mississippian plates of Gryphochiton from Kentucky, an undetermined Mississippian plate from Morocco, and a Permian plate of Cymatochiton from Sicily (Fig. 2); and Permian specimens from Argentina (Hoare and Sabattini, 2000), Malaysia (Hoare, 2000b), and Oregon (Hanger et al., 2000), contain- ing a number of new genera and species. A new Devonian genus from Germany and a new Permian genus from the U. S. are in press (Hoare, 2001). The Malaysian specimens are particulary interesting in their very large size with thick shell material. Recent studies by Cherns (1998a, b) have added significantly to our knowledge of Silurian polypla- cophorans. Undoubtedly there are still numerous taxa that have not been found or described. A significant collection of Lower Carboniferous specimens from Belgium has also been discovered. A study of this collection plus known collections at Harvard University and the Smithsonian Institution will allow a American Malacological Bulletin, Vol. 15(2) (2000):131-137 131] 132 AMER. MALAC. BULL. 15(2) (2000) Fig. 1. 1-8, Gotlandochiton sp., Oneata Fm. (Ordovician), Minnesota. 1-4, dorsal, ventral, anterior, and right lateral views of a tail plate, USNM 501824 (bar scale = 1 cm); 5-8, dorsal, posterior, ventral, and right lateral views of an intermediate plate, USNM 501825. 9-15, Chelodes sp., Odenville Fm. (Ordovician), Alabama. 9-12, dorsal, ventral, anterior, and left lateral views of a tail plate, USNM 501826; 13-15, posterior, right lateral, and dorsal views of an intermediate plate, USNM 501827 (bar scale = 1 cm). review of their taxonomic relationships which, as described and illustrated by de Koninck (1842, 1883), Miinster (1839, 1843), and de Ryckholt (1845, 1852), appear to contain a number of synonymous taxa. PROBLEMS The study of fossil polyplacophorans is challenging because the dorsal plates usually become disarticulated and are most commonly found as isolated entities. Complete articulated specimens are rare, although the occurrence of numerous specimens of Glaphurochiton concinnus (Richardson, 1956) in the Pennsylvanian Francis Creek Shale in Illinois is a notable exception. These specimens also show the presence of radula, girdle, and spines (Yochelson and Richardson, 1979). Incomplete articulated specimens are also helpful in determining the correct asso- ciation of isolated plates. Isolated plates, when described and named, have led to taxonomic confusion. As an example, Dunlop (1922), by careful collecting at one locality, was able to show that Chiton cordatus Kirkby, 1859, a head plate, C. armstron- gianus Etheridge, 1882, an intermediate plate, and C. gem- matus Kirkby, 1862, a tail plate, were all parts of the same species, C. cordatus Kirkby, 1859 [=Lekiskochiton cordatus (Kirkby, 1859)]. Because intermediate plates are usually more commonly found, a number of species have been erected on the basis of a single plate. Tail plates of most Paleozoic taxa are more diagnostic than head or intermedi- ate plates and distinction between genera and species can often be made on these plates alone. Tail plates also pro- vide much of the ontogenetic evidence of a taxon (Fig. 3). HOARE: PALEOZOIC POLYPLACOPHORA 133 Fig. 2. 1-4, Cymatochiton sp., Socio Ls. (Permian), Sicily. Dorsal, right lateral, anterior, and ventral views of an intermediate plate, USNM acc. No. 179,820 (bar scale = 1 cm). Note large apical area on ventral surface. 5, Undetermined sp., Mississippian formation, Morocco. Oblique right lateral view of an inter- mediate plate, OUZC 1633 (bar scale = 1 mm). 6-8, Gryphochiton parvus (Stevens), Tribune Ls. (Mississippian), Ky. Dorsal and posterior views of an inter- mediate plate, USNM 508371, and a dorsal view of an intermediate plate, USNM 508372, (bar scale = 1 mm). PLATE MICROSTRUCTURE Previous studies of plate microstructure (e. g. von Knorre, 1925; Bergenhayn, 1930; Haas, 1972) have normal- ly been made by cutting the plates vertically showing dispo- sition of shell layers and the occasional aesthete exposed in the section. Such sections do not show the pattern of the aesthetes. In cutting a piece of a plate parallel to and just below the dorsal surface of the tegmentum the size, pattern, and distribution of the aesthetes can be determined. Haas (1972) cut Recent material parallel to the plate surface. Specimens of nine species representing six genera of Mississippian and Pennsylvanian taxa were sectioned. Preservation of the aesthetes ranged from poor to excellent (Fig. 4) and measurements of diameters and spacing were made. Several preliminary conclusions were reached: 1. No size distinction for micraesthetes and megaesthetes was present in any specimen; 2. No pattern of micraesthetes encircling megaesthetes or any other circular pattern was seen. Such patterns must have developed post-Paleozoic; 3. For coarser surface ornamentation of pustules, the diame- ter of the aesthetes was larger, with one aesthete per pustule; 4. Aesthete diameter ranged in size from 19.3 um to 31.2 um in the specimens; 5. Several specimens showed aesthetes centered in polygo- nal structures. Whether this is an artifact of preservation or a structure related to plate development is unknown. CLASSIFICATION A number of multiplated organisms other than poly- placophorans are known to occur in the Paleozoic. The opportunity to study specimens of several of these taxa has led the author to reevaluate what should be or should not be included in the Polyplacophora. Some of these (e. g. Diadeloplax, Strobilepis, Aenigmatectus) from the upper Paleozoic have some characteristics in common with poly- placophorans including mucros on tail plates, insertion plates, and apical areas. However, the shapes, arrange- ments, and ornamentation of the plates are usually quite dif- ferent and the canal system in the plates is different from the aesthetes in polyplacophorans (Hoare and Mapes, 1995, 1996). Bischoff (1981) included Ordovician-Silurian phos- phatic multiplated organisms from Australia in the Polyplacophora. Besides being of a different composition these plates show structural differences from polypla- cophorans such as plate growth patterns, microstructure, and supposed insertion plates. Yu (1984, 1987) has illustrat- ed and described a number of multiplated taxa from the 134 AMER. MALAC. BULL. 15(2) (2000) Fig. 3. Acutichiton allynsmithi Hoare, Mapes, and Atwater. Gene Autry Fm. (Pennsylvanian), Oklahoma. 1-9, dorsal views of a series of tail plates showing increase in size and change in shape during ontogeny, OUZC 1621-1629 (bar scale = 0.5 cm). Lower Cambrian of China placing them in the Polyplacophora under five new families. These are small specimens, mostly less than 0.5 mm in width. Sirenko (1997) questioned the inclusion of these in the Polyplacophora, a conclusion with which I agree. The family Septemchitonidae Bergenhayn, 1955, contains multiplated organisms that developed plates that curve inward ventrally almost closing off the area for the position of the foot as illustrated by Dzik (1994, fig. 29). These plates are totally unlike polyplacophoran plate shape and the family is here omitted from the class. It is expected that many other types of multiplated organisms will eventu- ally be discovered in Paleozoic strata. Most recently, Sirenko’s (1997) classification of polyplacophorans recognized four orders, five suborders, and 14 families, including one new family, in the Paleozoic. Changes from his scheme suggested here (Fig. 5) include the recognition of the family Matthevidae Walcott (1886) in place of Chelodidae Bergenhayn (1943), the elimination of the family Septemchitonidae Bergenhayn (1955), and the addition of the family Choriplacidae Cotton and Weeding (1939). The latter is included based upon the new Devonian genus and species described herein, which has many char- acters comparable to the genus Choriplax Pilsbry, 1894. PHYLOGENY As with the classification, there are more questions than answers related to the phylogeny of the Polyplacophora at this time. This situation is true even to the extent of determining the origin of the class and rela- tionship with other mollusks. It is not the purpose of this study to discuss the origin of the phylum or that of the polyplacophorans. However, polyplacophorans are so dif- ferent from other mollusks in terms of their skeletal and anatomical structures that it seems likely that the phylum is polyphyletic. Yu (1989) proposed a multiplated ancestral stem leading to the polyplacophorans from which other mollus- can classes developed. The Polyplacophora undoubtedly developed from a segmented ancestral form but it seems very unlikely that the single or double plated classes of other mollusks arose from a similar ancestor (Runnegar and Pojeta, 1974; Pojeta, 1980). Figure 5 shows a possible phylogenetic relationship at the family level. It is based in part on the classification given by Sirenko (1997). There will be additional families added to this scheme based upon the studies of Cherns (1998a, b) in the Silurian, possibly in the Permian by the author, and by the discovery of additional specimens in the future. SYSTEMATICS Class Polyplacophora de Blainville, 1816 Order Neoloricata Bergenhayn, 1955 Suborder Lepidopleurina Thiele, 1910 Family Choriplacidae Cotton and Weeding, 1939 Genus Plasiochiton n. gen. Type species. Plasiochiton curiosus n. sp. Diagnosis. Intermediate plates subrectangular, wider than long, moderately arched, broadly convex; semicircular raised areas marked by concentric ridges on anterior jugal areas. Description. See under Plasiochiton curiosus n. sp. Discussion. This taxon is based upon a single incomplete articulated specimen. Plasiochiton differs from Glyptochiton de Koninck, 1883, in the shape of the plates and the shape and position of the raised jugal portion. The Recent genus Choriplax Pilsbry, 1894, is similar to Plasiochiton in its subrectangular plates with a raised and restricted tegmentum region located on the median portion HOARE: PALEOZOIC POLYPLACOPHORA 135 Fig. 4. Thin sections showing size and distribution of aesthetes on intermediate plates. 1, Acutichiton pyrmidalus Hoare, Sturgeon, and Hoare, Vanport Ls., Ohio; 2, Elachychiton juxaterminus Hoare and Mapes, Imo Fm., Ark.; 3, Euleptochiton spatulatus (Hoare, Sturgeon, and Hoare), Vanport Ls., Ohio; 4, Gryphochiton simplex (Raymond), Vanport Ls., Ohio; 5-6, Glaphurochiton carbonarius (Stevens), Putnam Hill Sh. and Washingtonville Sh., Ohio, OUZC 1633-1638 (bar scale = 100 tm). Note the differences in size and distribution of the aesthetes and the polygonal structure in 4 and 6. (Sections prepared by B. R. Hoare). of the intermediate plates. Plasiochiton is questionably assigned to the family Choriplacidae on this basis. Occurrence. Known only from the type locality of P. curiosus, Devonian (Erian). Etymology. Greek, plasion, oblong body. Plasiochiton curiosus Nn. sp. Fig. 6 Polyplacophoran Petzold, Clark, Makin, Owens, and Perry, 1992, p. 342, Fig. 1. Diagnosis. As for the genus. Description. Incomplete, articulated specimen of five plates preserved as mold of ventral surface; intermediate plates subrectangular, moderately arched, broadly convex, wider than long; posterior margins straight meeting lateral margins at right angles; anterior margin not observed; raised anterior jugal areas marked by concentric ridges in U-shaped pattern; raised area smaller towards posterior of specimen; shell material thin; head and tail plates unknown. Measurements. Size of exposed specimen 17 mm long, 4 mm wide; exposed plates range from 1.8 to 2.5 mm in length and from 4.0 to 4.6 mm in width; estimated approxi- mate total length 30 mm, total width 9.0 mm, and total height 4.0 mm. Etymology. Latin, curiosus, odd, strange. Type. Holotype, USNM 456242. Type locality. Devonian Sherman Ridge Member of the Mahantango Formation exposed in an abandoned shale pit at the intersection of U.S. 11-15 and Pennsylvania 104, Perry Co., Pennsylvania, 40°36°52”N, 76°57°21°W, Millersburg 7.5 minute quadrangle. Discussion. Devonian polyplacophorans are little known in 136 _ _N ~ Se et ee AMER. MALAC. BULL. 15(2) (2000) CAMBRIAN ORDOVICIAN SILURIAN DEVONIAN MISSISSIPPIAN JPENN SYLVANIAN PERMIAN Gryphochitonidae Choriplacidae ? er ee ee ee _ aN Yo fo Chip, a g, ae Fig. 5. A schematic phylogenetic diagram showing possible relationships at the family level. Fig. 6. Plasiochiton curiosus n. gen. and sp. 1, uncoated and 2, coated dor- sally oblique views of a partial articulated ventral mold, Mahantango Fm. (Devonian), Pennsylvania, USNM 456242 (bar scale = | cm). Note the subrectangular shape of the plates and the medial anterior restriction of the tegmentum layer. (Fig. | from Petzold ef al., 1992, p. 342, published with the permission of the publisher and author). North America and this specimen is important to our under- standing of the development of the class in the Paleozoic. Unfortunately the shell material is lacking, but the mold is diagnostic in the shape of the plates and the semicircular structure on the anterior portion of the jugal area. In general the raised structure is similar to that of Glyptochiton de Koninck, 1883, from the Lower Carboniferous of Belgium and the United Kingdom but differs in position on the plates and the shape of the plates of the latter is different. The complete size and shape of the plates of Plasiochiton curiosus cannot be determined without destroying a portion of the specimen. LITERATURE CITED Bergenhayn, J. R. M. 1930. Kurze Bemerkungen zur Kenntnis der Schalenstrucktur und Systematik der Loricaten. Kungliga Svenska Ventenskapsakademien, Handlingar, series 3, 9(3): 1-54. Bergenhayn, J. R. M. 1943. Preliminary notes on the fossil polypla- cophoras from Sweden. Geologiska Foreningen, Stockholm, Foérhandlingar 65(3):297-303. Bergenhayn, J. R. M. 1955. Die Fossilen Schwedischen Loricaten nebst einer vorlaufigen Revision des Systems der ganzen Klasse Loricata. Lunds Universitets Arsskrift, n.f., Avd. 2. 51(8), Kungliga Fysiografiska Sdllskapets, Handlingar, n.f., Avd. 2, 66(8):1-41. Bischoff, G. C. O. 1981. Cobcrephora n. g., representative of a new poly- placophoran order Phosphatoloricata with calciumphosphatic HOARE: PALEOZOIC POLYPLACOPHORA 137 shells. Senkenbergiana Lethaea 61:173-215. Blainville, H. M. D. de. 1816. Prodrome d’une nouvelle distribution systé- matique du regne animal. Société Philomathique, Paris, Nouveau Bulletin, p. 51-53, 93-97. Cherns, L. 1998a. Chelodes and closely related Polyplacophora (Mollusca) from the Silurian of Gotland, Sweden. Palaeontology 41 (3):545-573. Cherns, L. 1998b. Silurian polyplacophoran molluscs from Gotland, Sweden. Palaeontology 41(3):939-974. Cotton, B. C., and B. J. Weeding. 1939. Flindersian Loricates. Transactions of the Royal Society of South Australia 63(2):180- 199. Dunlop. R. 1922. Notes on the chitons of Woodmill. Transactions of the Geological Society of Glasgow 17:75-76. Dzik, J. 1994. Evolution of ‘small shelly fossils’ assemblages. Acta Palaeontologica Polonica 39(3):247-313. Etheridge, R., Jr. 1882. A contribution to the study of the British Carboniferous Chitonidae. Proceedings of the Natural History Society of Glasgow 5:84-107. Haas, W. 1972. Untersuchungen uber dis Mikro- und Ultrastruktur der Polyplacophorenschale. Akademie der Wissenschaften und der Literatur, Mainz. Forschungsberichte Biomineralisation; Forschungsberichte Kommission fuer Biokristallit- Biomineralization, Research Reports 5:3-52. Hanger, R. A., R. D. Hoare, and E. E. Strong. 2000. Permian Polyplacophora, Rostroconchia, and Problematica from Oregon. Journal of Paleontology 74(2):192-198. Hoare, R. D. 2000a. Silurian Polyplacophora and Rostroconchia (Mollusca) from northern California. Proceedings of the California Academy of Sciences 52(3):23-31. Hoare, R. D. 2000b. New Permian Polyplacophora (Mollusca) from Malaysia. Journal of Paleontology 74(4):571-574. Hoare, R. D. 2000c. Early Mississippian Polyplacophora (Mollusca) from Iowa. Journal of Paleontology 75(1):66-74. Hoare, R. D. 2001. New genera of Paleozoic Polyplacophora (Mollusca). Journal of Paleontology (in press). Hoare, R. D. and A. G. Cook. 2000. Devonian and Mississippian Polyplacophora from Western Australia. Memoirs of the Queensland Museum 45(2):395-403. Hoare, R. D. and R. H. Mapes. 1995. Relationships of the Devonian Strobilepis and related Pennsylvanian problematica. Acta Palaeontologica Polonica 40(2):111-128. Hoare, R. D. and R. H. Mapes. 1996. Late Paleozoic problematic sclerites of hercolepadid affinites. Journal of Paleontology 70(3):341-347. Hoare, R. D. and N. Sabattini. 2000. Lower Permian Polyplacophora (Mollusca) from Argentina. Journal of Paleontology 74(2):189- 191. Kirkby, J. W. 1859. On the Permian Chitonidae. Quarterly Journal of the Geological Society of London 15:607-626. Kirkby, J. W. 1862. On some remains of Chiton from the Mountain- Limestone of Yorkshire. Quarterly Journal of the Geological Society of London 18:233-237. Knorre, H., von. 1925. Duslia insignis Jahn - ein angeblich 11 oder 12 Schaleonplatten tragender Chiton der untersilurischen Schiten Bohmens. Jenaische Zeitschrift fiir Naturwissenschaft 61:497- 499. Koninck, L. G. de. 1842. Description des animaux fossiles, qui se trouvent dans le terrain carbonifére de Belgique. H. Dessain, Leige, 650 p. Koninck, L. G. de. 1883. Faune du calcaire carbonifere de la Belgique, Pt. 4, Gasteropodes. Annales Musee Royale d'Histoire Naturelle de Belgique, series Paleontologie 8(4):198-213. Kues, B. S. 1978. Polyplacophora from the Salem Limestone (Mississippian) in central Indiana. Journal of Paleontology 52(2):300-3 10. Miinster, G. G. zu. 1839. Der Chiton priscus und einige andere seltens Versteinerungen aus der Uebergangsformation. Beitrdge zur Petrefactenkunde, \st edition, 1:38; 1843, 2nd edition, 1:60-61. Buchner’ schen Buchhandlung, Bayreuth. Petzold, D. D., C. Clark, M. Malin, R. Owens, and E. Perry. 1992. A par- tial articulated polyplacophoran from the Devonian of Pennylvania. Journal of Paleontology 66(2):340. Pilsbry, H. A. 1892-94. Monograph of the Polyplacophora. Manual of Conchology vol. 14, 350 pp., vol. 15, 133 pp. Pojeta, J., Jr. 1980. Molluscan Phylogeny. Tulane Studies in Geology and Paleontology 16(2):55-80. Richardson, E. S., Jr. 1956. Pennsylvanian invertebrates of the Mazon Creek area, Illinois. Fieldiana: Geology, Chicago Natural History Museum 12(3):61-65. Runnegar, B. and J. Pojeta, Jr. 1974. Molluscan Phylogeny: The Paleontological Viewpoint. Science 186(4161):311-317. Ryckholt, P. de. 1845. Résumé geologique sur le genre Chiton Lin. Bulletin de l'AcadEmie Royale des Sciences et des Belles-Lettres de Bruxelles 12(7):36-62. Ryckholt, P. de. 1852. Mélanges Paléontologiques - Chitonidae. Mémoires Couronnes et Mémoires des Savants étrangers. l'Académie Royale des Sciences et des Beux-Arts de Belgique 24:63-65. Sirenko, B. I. 1997. The importance of the development of articulamentum for taxonomy of chitons (Mollusca, Polyplacophora). Ruthenica 7(1):1-24. Thiele, J. 1910. Revision des Systems der Chitonen. Zoologica 22(56):71- 132. Walcott, C. D. 1886. Studies on the Cambrian faunas of North America. U. S. Geological Survey Bulletin 30:223-225. Yochelson, E. L. and E. S. Richardson, Jr. 1979. Polyplacophoran mol- luscs of the Essex fauna (Middle Pennsylvanian, Illinois). In: Mazon Creek Fossils, H. Nitecki, ed. pp. 321-332. Academic Press, New York. Yu, Wen. 1984. Early Cambrian molluscan faunas of Meischucun Stage with special reference to Precambrian-Cambrian boundary. A Special Paper on the Academic Sinica Developments in Geoscience, Contribution to the 27th International Geological Congress, 1984, Moscow, pp. 21-35. Science Press, Beijing. Yu, Wen. 1987. Yangtze micromolluscan fauna in Yangtze region of China with notes on Precambrian-Cambrian boundary. Nanjing Institute of Geology and Palaeontology, Academica Sinica 8:19- 219; Yu, Wen. 1989. Did the shelled mollusks evolve from univalved to multi- valved forms or vice versa? Contribution to the 28th International Geological Congress, 1989, Washington, D. C. pp. 235-244. Science Press, Beijing. Date of manuscript acceptance: 17 May 2000 Cambrian monoplacophoran molluscs (Class Helcionelloida) Alexander P. Gubanov and John S. Peel Department of Earth Sciences (Historical Geology and Palaeontology), Uppsala University, Norbyvagen 22, SE-752 36 Uppsala, Sweden Abstract: Cambrian monoplacophoran molluscs (Class Helcionelloida) are morphologically diverse, but also show considerable interspecific variation. While usually small (1-2 mm), they may achieve a size of 30 mm or more. Shell form ranges from low limpets to high cones and these may be straight or curved; coiled shells may be bilaterally symmetrical, dextral or sinistral. This diversity provides a fertile ground for speculation about early molluscan evolu- tion and almost all extant classes of Mollusca have been recognised here. Certainly, molluscs can be traced back to the late Vendian (late Precambrian), but not all classes appeared during the Vendian-Cambrian transition. Small-sized late Wendian-Middle Cambrian molluscs (helcionelloids) show little resem- blance to crown molluscan groups, including the living monoplacophorans (tryblidiid Tergomya). Key Words: Cambrian, Mollusca, monoplacophoran, Helcionelloida, evolution Most Cambrian monoplacophoran molluscs are morphologically far removed from the familiar present day tryblidiid tergomyan Neopilina Lemche, 1957 or its Early Palaeozoic antecedents Pilina Koken, 1925 and Tryblidium Lindstrém, 1880. Placed in the Class Helcionelloida, these Cambrian univalves display a variety of shell forms, bear- ing witness to the first great diversification of molluscs near the Precambrian-Cambrian boundary (Peel, 1991a, b). Indeed, helcionelloids are represented in the first shelly faunas appearing in the late Precambrian in Siberia (Khomentovsky et al., 1990). The helcionelloid shell is typ- ically bilaterally symmetrical (isostrophic), rapidly expand- ing and coiled through about one whorl, but limpet-like shells, tall cones and strongly laterally compressed mor- phologies are also conspicuous. Aldanella Vostokova, 1962 is a multi-whorled anisostrophic helcionelloid from the ear- liest Cambrian, seductively reminiscent of the gastropods that first appeared in the latest Cambrian. Anisostrophic shells of Pelagiella Matthew, 1895 can be conspicuous throughout the Cambrian and are conveniently placed with- in the Helcionelloida, even if their precise relationship remains uncertain. Helcionelloids may be abundant in strata of Early and Middle Cambrian age. A variety of problematic mol- luscs are known from the Late Cambrian (e. g., Webers et al., 1992) but the helcionelloid record is uncertain. However, we here report undoubted helcionelloids from the latest Cambrian to early Ordovician of Kazakhstan, repre- senting the youngest known members of the class. Thus, the geological record of helcionelloids overlaps with that of the earliest tryblidiid tergomyans such as Proplina Kobayashi, 1933 and most other molluscan classes. Helcionelloids have been interpreted as torted (gas- tropods, e. g. Knight et al., 1960) and un-torted (monopla- cophorans) and this latter hypothesis is currently widely supported (Runnegar and Jell, 1976; Peel, 1991a, b; Geyer, 1994; Runnegar, 1996). Their relationship to tergomyans such as Neopilina is disputed, with some authors (Runnegar, 1996 and previous works cited therein) uniting all these untorted univalves in a single Class Monoplacophora. Peel (1991a, b; see also Geyer, 1994) reconstructed helcionelloids with the apex at the posterior and the shell expanding anteriorly (endogastric), as distinct from the anterior apex and posteriorly expanding (exogas- tric) shell of Neopilina and other tergomyans. Following the recommendation of Wingstrand (1985) and others, Peel (1991a, b) abandoned Monoplacophora as a formal taxon, employing Class Helcionelloida for the endogastric branch and Class Tergomya for the exogastric trybilidiid stock. Recognition of helcionelloids as the molluscan ancestral group focuses attention in the search for intermediate forms to molluscan classes such as Tergomya, Gastropoda and Cephalopoda to the Late Cambrian. In contrast, the Rostroconchia, as a direct descendant of the Helcionelloida, and the Bivalvia appear to be represented already in the Early Cambrian (Runnegar and Pojeta, 1985). THE OLDEST MOLLUSCS Helcionelloids are present in the oldest shelly fau- nas from Siberia of late Vendian (Nemakit-Daldynian) age American Malacological Bulletin, Vol. 15(2) (2000):139-145 139 140 AMER. MALAC. BULL. 15(2) (2000) (Khomentovsky et al., 1990; Gubanov and Peel, 1999). These, and the slightly younger, earliest Cambrian, species Oelandiella korobkovi Vostokova, 1962 have been referred to the genus Latouchella Cobbold, 1921 by most authors, following the unexplained synonymisation of Oelandiella Vostokova, 1962 with Latouchella by Missarzhevsky (in Rozanov et al., 1969; translated into English as Raaben, 1981). Recent re-description of the type materials of both genera by Gubanov and Peel (1998, 1999) has demonstrated that Oelandiella is a valid and recognisable genus whereas the widely reported Latouchella is uncommon. Oelandiella is characterised by comarginal rugae that cross the dorsum whereas the median dorsal area in Latouchella is smooth with rugae restricted to the lateral areas. Latouchella has also been confused with Oelandia Westergard, 1936 from the Middle Cambrian of Sweden, China and Bohemia. The latter genus, however, is not bilaterally symmetrical since the lateral rugae interdigitate at the dorsum (Peel and Yochelson, 1987; Gubanov and Peel, 1998). Inflated and strongly rugose shells of Oelandiella korobkovi occur contemporaneously with the laterally com- pressed and generally smooth-shelled Anabarella Vostokova, 1962, also originally described from northern Siberia. The close relationship between these two genera is demonstrated by the recent description of rugose Anabarella from Spain (Vidal et al., 1999). Recently, we have also iden- tified Anabarella from the Lontova Formation (Early Cambrian) of Estonia where it occurs together with Aldanella kunda (Opik, 1926) described by Posti (1978). SIZE, SYMMETRY, AND VARIATION Most described Cambrian helcionelloids are small, generally 1-2 mm in length (Runnegar and Pojeta, 1985). This reported size range sometimes reflects environmental, preservational and preparational biases because recovery of material suitable for description is commonly accomplished by digestion of phosphatised carbonate rocks in weak acids. Even in the Tommotian (earliest Cambrian), however, there are helcionelloids more than 30 mm long (Dzik, 1991 and personnal observations) (Fig. 1) while undescribed late Early Cambrian material from Sweden can be of similar size. The taxonomic significance of the size differences is uncertain but the importance of considering size in func- tional morphological interpretations of early molluscs was stressed by Peel (1991b). An unfulfilled need for ontoge- netic studies is also apparent in the taxonomy of helcionel- loids. Archaeospira Yu, 1979 from the earliest Cambrian of China closely resembles Oelandiella korobkovi (Fig. 2) and the latter name has been applied by earlier authors to material included within the synonymy of the type species Archaeospira ornata Yu, 1979 in the detailed re-description given by Qian and Bengtson (1989). They pointed out that O. korobkovi is not known to be asymmetrically coiled while the shell of A. ornata is slightly sinistral. Furthermore, they commented that a second Chinese species, Hubeispira nitida Yu, 1981, is dextral. Gubanov and Peel (1999) noted that some specimens of O. korobkovi were slightly dextral but recent study of material from the earliest Cambrian (Tommotian) of the Selinde River, south- ern Siberia, demonstrates the presence of sinistral, dextral and planispiral bilaterally symmetrical morphs in the same sample of O. korobkovi. The sinistral Archaeospira may thus prove to be a synonym of Oelandiella. The same sam- ples also show dextral morphs of O. korobkovi that approach the morphology of the anisostrophic Aldanella (Fig. 3). Aldanella itself also shows a similar spectrum of Fig. 1. Large limpet-shaped (A-C) and coiled (D, E) helcionelloid molluscs from the Lower Cambrian (Tommotian, Dokidocyathus regularis Zone) of the middle Lena River (southeastern Siberia). A, Institute of Palaeobiology, Polish Academy of Sciences, Warsaw (IPB PAN) Ga VIII-3, the largest specimen with broken apex, lateral view. B, C, IPB PAN Ga VIII-13, example with well-preserved shell, lateral (B) and apical (C) views. D, E, IPB PAN Ga VIll-7, tightly coiled mollusc, lateral (D) and anterio-lateral (E) views. GUBANOV AND PEEL: CAMBRIAN MONOPLACOPHORAN MOLLUSCS Fig. 2. Oelandiella korobkovi Vostokova, 1962 showing variation in comarginal ornamentation and rate of expansion. A, B, Museum of Evolution (Palaeontology), Uppsala University (PMU) SIB 1001, 1002, Pestrosvet Formation (early Tommotian) of the Dzhanda River, southeastern Siberia C, D, PMU CH 1, 2, Dahai Member (Meisuchunian) of southern China. Fig. 3. Oelandiella korobkovi Vostokova, 1962 from the lower Cambrian (Tommotian) of the Selinde River, southeastern Siberia. A, B, PMU SIB 1003, isostrophic morph in apertural (A) and lateral (apical) (B) views. C, D, PMU SIB 1004, dextral morph with comarginal ribs, apertural (C) and apical (D) views. E, F, PMU SIB 1005, dextral Aldanella-like morph with smooth shell surface, apertural (E) and apical (F) views. 142 AMER. MALAC. BULL. 15(2) (2000) variation occurring within the same sample, and with sub- stantial variation in such characters as rate of whorl expan- sion, rate of translation along the axis of coiling, tumidity, and tightness of coiling (Fig. 4). While early Cambrian helcionelloids, including Aldanella, appear to be taxonomically diverse, it is now evident from this substantial variation within samples that considerable taxonomic inflation has occurred. Recognition of this variation strengthens interpretation of Aldanella as an anisostrophic helcionelloid, refuting suggestions that it could be a gastropod (Runnegar and Pojeta, 1985). DIVERSITY Helcionelloids show a remarkable diversity in the Early and Middle Cambrian. The wide aperture and strong comarginal ornamentation suggest that Oelandiella was a benthic micro-herbivore or deposit feeder, a mode of life inferred in a variety of other taxa, such as Latouchella. Two adaptive morphological trends are clearly visible already in the Tommotian. Expansion of the shell to produce a low cone with a broad to near equidimensional aperture in plan view characterises Bemella Missarzhevsky, 1969 and Helcionella Grabau and Shimer, 1909, interpreted as benth- ic deposit feeders adapting to a hardening substratum (Linsley, 1977; Gubanov, 1984). Loss of ornamentation and strong lateral compression in Anabarella indicate adapta- tion to a softer substrate and a semi-infaunal life. This trend from Oelandiella to Anabarella is continued into Watsonella Grabau, 1900 (= Heraultipegma Runnegar and Pojeta, 1976), considered to be the oldest member of the Class Rostroconchia Runnegar and Pojeta, 1974 (Pojeta and Fig. 4. Aldanella sp. from the Lower Cambrian (early Tommotian) of the Anabar River, northeastern Siberia showing the spectrum of shell form variation, PMU SIB 1006-1012, apertural views of all specimens. GUBANOV AND PEEL: CAMBRIAN MONOPLACOPHORAN MOLLUSCS 143 Fig. 5. The hydrothermal vent mollusc Thermoconus shadlunae Little, Maslennikov, Morris, and Gubanov, 2000, from the Silurian volcanic rocks of Yaman-Kasy, southern Urals, Russia. A, C, lateral views of two specimens in the British Museum of Natural History (BMNH) VF 16 and BMNH VF 18 to show different rates of shell expansion. B, paratype, BMNH VF 23, oblique apical view showing the internal septa. Runnegar, 1976; Runnegar and Pojeta, 1985; Gubanov et al., 1999). Both of these morphological trends are repeated through the Early and Middle Cambrian (Peel, 1991b). Species referred to Scenella Billings, 1872 have widely expanded, limpet-like shells, although these are not univer- sally accepted as molluscs (Yochelson and Gil-Cid, 1984; Landing and Narbonne, 1992). Marocella Geyer, 1986 includes several species formerly assigned to Scenella (S. morenensis Yochelson and Gil-Cid, 1984 and S. antiqua Kiaer, 1916), but here, too, there is uncertainty about affini- ty (Geyer, 1986; Evans, 1992). Morphologically similar shells of uncertain systematic position also occur in the Late Cambrian and Early Ordovician (Stinchcomb, 1986; Webers et al., 1992; Webers and Yochelson, 1999). Lateral compression characterises the Early Cambrian Stenotheca Salter, 1859, some species of Yochelcionella Runnegar and Pojeta, 1974, and the Middle Cambrian genera Mellopegma Runnegar and Jell, 1976 and Eurekapegma MacKinnon, 1985. Peel (1991a, b) described a morphological series in the Early-Middle Cambrian Eotebenna Runnegar and Jell, 1976 in terms of increased penetration of soft bottom sedi- ments. Culminating in the elongate Eotebenna viviannae Peel, 1991, in which convergence of the lateral areas pro- duces separate anterior and posterior openings, this series can be directly compared with morphological adaptations in rostroconchs (Pojeta and Runnegar, 1976; Runnegar and Pojeta, 1985). THE LAST HELCIONELLOIDS Recently identified helcionelloids from the earliest Ordovician (Tremadoc) of Kazakhstan are the youngest undoubted members of the class. The specimens are typi- cally helcionelloid in form, coiled through about one whorl. They are about 1-2 mm in length, laterally compressed so that width is about 30 % of length and with ornamentation of comarginal growth lines. The recent recognition of these Kazakhstan helcionelloids requires re-examination of the variety of Late Cambrian-Ordovician problematic isostrophic shells generally placed in the Bellerophontoidea (Gastropoda) or the cyrtonellid Tergomya (Runnegar and Jell, 1976; Peel, 1991b). A recently described monoplacophoran mollusc from Silurian thermal vent deposits in the Ural Mountains of Russia (Little et al., 1999) provides a challenge to cur- rent understanding of tergomyan or helcionelloid evolution. 144 AMER. MALAC. BULL. 15(2) (2000) Thermoconus shadlunae Little, Maslennikov, Morris, and Gubanov, 1999 (Fig. 5) is an unusually large (5-8 cm tall), straight-sided orthoconical shell with a sub-circular cross- section. A specimen with a broken apex reveals internal septa, which are concave towards the aperture (Fig. 5B). The external surface shows two distinct sculpture patterns, with rounded costae in the early stages and reticulate ornamentation with regularly spaced flanges in the late stage. The shell form of Thermoconus is unlike other described tergomyans, but superficially similar tall cones are known in the Late Cambrian-Early Ordovician Hypseloconus Berkey, 1898 and Knightoconus Yochelson, Flower and Webers, 1973 (Webers ef al., 1992). Interpreted as tergomyan, Thermoconus represents the first evidence of the migration of this group from the shallow water environ- ments in which it occurs from the Late Cambrian to the Devonian, to the deep-water environments that characterise most of its species at the present day. While Thermoconus is quite reasonably interpret- ed as a tergomyan mollusc by Little et al. (1999), it lacks any muscle scars to substantiate the determination. In addi- tion to their occurrence in the tergomyan Knightoconus, septa of the type found in Thermoconus are very common in some euomphaloidean gastropods (Gubanov et al., 1995) and among other fossil and recent gastropod taxa (Yochelson et al., 1973). Septation is also a characteristic feature of some helcionelloid molluscs such as Helcionella chinensis Walcott, 1905, H.? insulcata Rasetti, 1957 and H. coreanica Kobayashi, 1958. Thermoconus is unlike known Lower Palaeozoic gastropod limpets (e. g. Peel and Horny, 1999), but deep water or hydrothermal vent limpets are lacking in the geological record. The array of exotic gastro- pod limpets described from these habitats at the present day by McLean (1981, 1990) and later papers surely indicates that atypical gastropod limpets could have occurred in vent communities already in the Palaeozoic. Thermoconus is much larger than any described helcionelloid, but this is a feature of Silurian molluscs in general when compared to their Cambrian progenitors (Runnegar and Pojeta, 1985), although some Tommotian helcionelloid molluscs have already attained a rather large size (see discussion above). Recently one of us (APG) collected possible helcionelloid molluscs, tentative- ly determined as Helcionella sp. with a size up to 8 cm, in the Upper Cambrian-Lower Ordovician Kruzhilikha Formation of the Severnaya Zemlya Archipelago (Arctic Siberia). If we interpret Thermoconus as a helcionelloid, it represents a relic of a group of mol- luscs that were abundant in the Early-Middle Cambrian but survived into the Silurian by taking refuge in the deep ocean. The scenario parallels the more recent history of the Tergomya. ACKNOWLEDGMENTS We thank Prof. J. Dzik, Institute of Palaeobiology, Warsaw for lending the large helcionelloids illustrated in Fig. 1. Our research is sup- ported by grants from the Swedish Natural Science Research Council (NFR) and The Swedish Royal Academy of Science (KVA). LITERATURE CITED Dzik, J., 1991. Is fossil evidence consistent with traditional views of the early metazoan phylogeny? In: The early evolution of Metazoa and the significance of problematic taxa, A. M. Simonetta and S. Conway Morris, eds. pp. 47-56. Cambridge University Press and University of Camerino. Evans, K. R. 1992. Marocella: Antarctic specimens of an enigmatic Cambrian animal. Journal of Paleontology 66:558-562. Geyer, G. 1986. Mittelkambrische Mollusken aus Marokko und Spanien. Senckenbergiana lethaea 67:55-118. Geyer, G., 1994. Middle Cambrian molluscs from Idaho and early conchiferan evolution. New York State Museum Bulletin 481:69- 86. Gubanov, A. P. 1984. Forma rakoviny gastropod i ee svyaz’ so sredoj obi- taniya 1 obrazom zhisni. In: Geologo-geofisicheskie issledovaniya v Sibiri. IGiG SO AN SSSR, Novosibirsk, pp. 53-57. [In Russian]. Gubanov, A. P., A. V. Kouchinsky, and J. S. Peel. 1999. First evolution- ary-adaptive lineage within the earliest molluscs. Lethaia 32:155- 157. Gubanov, A. P. and J. S. Peel. 1998. Redescription of the type species of Latouchella Cobbold, 1921 (Mollusca) from the Lower Cambrian of Comley (England). GFF 120:17-20. Gubanov, A. P. and J. S. Peel. 1999. Oelandiella, the earliest Cambrian helcionelloid mollusc from Siberia. Palaeontology 42(2):211-222. Gubanoy, A. P., J. S. Peel, and I. A. Pianovskaya. 1995. Soft-sediment adaptations in a new Silurian gastropod from central Asia. Palaeontology 38(4):83 1-842. Khomentovsky, V. V., A. K. Val’kov, and G. A. Karlova. 1990. Novye dannye po biostratigraphii perekhodnykh sloev v bassejne sred- nego techeniya r. Aldan. In: Pozdnij dokembrij i rannij paleozoj Sibiri. Voprosy regional’noj stratigrafii, V. V. Khomentovsky and A. S. Gibsher, eds. pp. 3-63. Institut Geologii i Geofiziki SO AN SSSR, Novosibirsk. [in Russian] Knight, J. B., L. R. Cox, A. M. Keen, R. L. Batten, E. L. Yochelson, and R. Robertson. 1960. Systematic descriptions. In: Treatise on invertebrate paleontology, I, Mollusca 1, R. S. Moore, ed. pp. 1169-11310. University of Kansas Press and Geological Society of America. Landing, E. and G. M. Narbonne. 1992. Scenella and a chondrophorine (medusoid hydrozoan) from the basal Cambrian (Placentian) of Newfoundland. Journal of Paleontology 66:338. Linsley R. M. 1977. Some ‘laws’ of gastropod shell form. Paleobiology 3:196-206. Little, C., V. Maslennikov, N. Morris, and A. P. Gubanov. 1999. Two Palaeozoic hydrothermal vent communities from the southern Urals, Russia. Palaeontology 42:1043-1078. McLean, J. H. 1981. The Galapagos Rift limpet Neomphalus: relevance to understanding the evolution of a major Paleozoic-Mesozoic radia- tion. Malacologia 25:3-20. McLean, J. H. 1990. Neolepetopsidae, a new doccoglossate limpet family from hydrothermal vents and its relevance to patellogastropod evolution. Journal of Zoology 222:485-528. GUBANOV AND PEEL: CAMBRIAN MONOPLACOPHORAN MOLLUSCS 145 Opik, A. 1926. Uber den estlandischen Blauen Ton. Publications of the Geological Institution of the University of Tartu 6:39-46. Peel, J. S. 1991a. Functional morphology of the Class Helcionelloida nov., and the early evolution of the Mollusca. In: The early evolution of Metazoa and the significance of problematic taxa, A. Simonetta and S. Conway Morris, eds. pp. 157-177. Cambridge University Press and University of Camerino. Peel, J. S. 1991b. The Classes Tergomya and Helcionelloida, and early molluscan evolution. Grgnlands Geologiske Undersggelse Bulletin 161:11-65. Peel, J. S. and R. J. Horny. 1999. Muscle scars and systematic position of the Lower Palaeozoic limpets Archinacella and Barrandicella gen. n. (Mollusca). Journal of the Czech Geological Society 44:97-115. Peel, J. S. and E. L. Yochelson. 1987. New information on Oelandia (Mollusca) from the Middle Cambrian of Sweden. Bulletin of the Geological Society of Denmark 36:263-273. Pojeta, J., Jr. and B. Runnegar. 1976. The paleontology of rostroconch molluscs and the early history of the phylum Mollusca. /U. S./ Geological Survey Professional Paper 969:88 pp. Posti, E. 1978. New finds of platysolenitids and gastropods from the Lontova stage of Estonia. ENSV TA Toimetised Geoloogia 27(3): 103-107. [in Russian] Qian, Yi and S. Bengtson. 1989. Palaeontology and biostratigraphy of the Early Cambrian Meishucunian Stage in Yunnan Province, South China. Fossils and Strata 24:156 pp. Raaben, M. E. (ed.) 1981. The Tommotian Stage and the Cambrian lower boundary problem. Amerind Publishing Company, New Delhi, 359 pp. Rozanov, A. Yu., V. V. Missarzhevsky, N. A. Volkova, L. G. Voronova, I. N. Krylov, B. M. Keller, I. K. Korolyuk, K. Lendzion, R. Michniak, N. G. Pykhova, and A. D. Sidorov. 1969. Tommotskij yarus i problema nizhnej granitsy kembriya. Publishing office Nauka, Moscow, 380 pp. Runnegar, B. 1996. Early evolution of the Mollusca: the fossil record. In: Origin and Evolutionary radiation of the Mollusca, J. D. Taylor, ed. pp. 77-87. Oxford Science Publications. Runnegar, B. and P. A. Jell. 1976. Australian Middle Cambrian Molluscs, their bearing on early molluscan evolution. Alcheringa 1(2):109- 138. Runnegar, B. and J. Pojeta, Jr. 1985. Origin and diversification of the Mollusca. Jn: The Mollusca 10, Evolution, E. R. Trueman and M. R. Clarke, eds. pp. 1-57. Academic Press, Orlando. Stinchcomb, B. L. 1986. New Monoplacophora (Mollusca) from late Cambrian and early Ordovician of Missouri. Journal of Paleontology 60:606-626. Vidal, G., T. Palacios, M. Moczydlowska, and A. Gubanov. 1999. Age constraints from small shelly fossils on the Early Cambrian termi- nal Codomian Phase in Iberia. GFF 121:137-143. Webers, G. F., J. Pojeta, Jr., and E. L. Yochelson. 1992. Cambrian Mollusca from the Minaret Formation, Ellsworth Mountains, West Antarctica. Geological Society of America Memoir 170:181- 248. Webers, G. F. and E. L. Yochelson. 1999. A revision of Palaeacmea (Upper Cambrian) (?Cnidaria). Journal of Paleontology 73(4):598-607. Wingstrand, K. G. 1985. On the anatomy and relationships of Recent Monoplacophora. Galathea Report 16:7-94. Yochelson, E. L. and D. Gil-Cid. 1984. Reevaluation of the systematic position of Scenella. Lethaia 17:331-340. Yochelson, E. L., R. H. Flower, and G. F. Webers. 1973. The bearing of the new Late Cambrian monopi. -ophoran genus Knightoconus on the origin of the Cephalopoda. Lethaia 6:275-310. Date of manuscript acceptance: 21 March 2000 The Bellerophont Controversy Revisited John A. Harper! and Harold B. Rollins” 'Pennsylvania Geological Survey, Pittsburgh, Pennsylvania 15222-4745, U.S. A. 2Department of Geology and Planetary Sciences, University of Pittsburgh, Pittsburgh, Pennsylvania 15260, U.S. A. Abstract: An old controversy reestablished itself in the late 1970s and early 1980s that focused on the systematic placement of the enigmatic Bellerophontoidea (informally, “bellerophonts”’), a group of planispirally coiled, wholly fossil molluscs. The controversy embraced three fundamental con- cepts that are based on different philosophical interpretations of shell form, muscle scar patterns, and other preserved shell features: 1) all bellerophonts were monoplacophorans; 2) all bellerophonts were gastropods; and 3) some bellerophonts were monoplacophorans and some were gastropods. A review of the main issues appearing in the literature since the early 1980s indicates that these three philosophical divisions still exist and, indeed, have become entrenched. An examination of the relevant anatomical and shell features of recent gastropods and monoplacophorans, and comparison with preserved fea- tures in enigmatic fossil forms, convinces us that the bellerophontoideans and the coiled and high-domed “monoplacophorans” (Cyclomya) were gastropods Only the flattened, spoon- and cap-shaped monoplacophorans (Tergomya) were true monoplacophorans. We present a hypothetical scheme for the morpho- logical diversification of gastropods from early monoplacophorans that could account for Cyclomya, Belleropkontoidea, Patellogastropoda, and Prosobranchia. Key Words: Gastropoda, Monoplacophora., Bellerophontoidea, functional morphology, phylogeny One of the more polarizing arguments in malacology eral authors (Salvini-Plawen, 1980; Wingstrand, 1985; in the latter half of the 20th century has been termed the Peel, 1991; and Geyer, 1994) recently urged abandonment “bellerophont controversy”. At the heart of this controversy of that formal name. N. H. Ohdner (/n: Wenz, 1940) intro- is the question of whether the Bellerophontoidea, an extinct duced “Monoplacophora” with the intent that it would be group of bilaterally symmetrical univalved molluscs, of an informal term separating the superfamily Tryblidioidea which Bellerophon Montfort (Fig. 1) 1s a typical example, from the Polyplacophora. Knight (1952) formalized the were gastropods or monoplacophorans. If they were mono- name by including Monoplacophora as a gastropod order placophorans, they were untorted and exogastric (with the (Knight included the order Polyplacophora with the gas- shell coiled over the head). If they were gastropods, they tropods as well!). The fossil tryblidioids had long been con- were torted and endogastric (with the shell coiled over the sidered patelliform gastropods with multiple muscle scars foot as in the majority of extant gastropods). The history of until Wenz hypothesized that they had been, in fact, untort- the controversy has been summarized many times and need ed animals. The discovery of Neopilina Lemche (Lemche, not be repeated here. Readers interested in exploring the 1957; also Lemche and Wingstrand, 1959) gave credence to details might begin with summaries published by Yochelson that hypothesis. (1967 — an excellent introduction to the natural history and Horny (1965a) established two new monopla- evolution of thought on the bellerophonts), Harper and cophoran subclasses, Tergomya and Cyclomya. Tergomya Rollins (1982), Peel (1985b), and WahIman (1992). includes the cap-shaped or spoon-shaped monopla- cophorans such as Tryblidium Lindstr6m in which the plane TERMINOLOGY of the muscle field lies outside the apical axis (a curved line marking the exact center of the shell during ontogeny) (Fig. We use the term “bellerophont” in an informal sense 2A). Cyclomya includes a wide variety of coiled and to designate both the Bellerophontoidea and the coiled uncoiled forms in which the plane of the muscle field inter- Cyclomya; that is, any planispirally coiled, univalved mol- sects the apical axis (Fig. 2B). This appears to represent a lusc that definitely is not a cephalopod. We also use the distinct and natural morphological separation among shells term ““monoplacophoran” in an informal sense because sev- exhibiting multiple pairs of muscle scars. American Malacological Bulletin, Vol. 15(2) (2000):147-156 147 148 AMER. MALAC. BULL. 15(2) (2000) Fig. 1. Bellerophon Montfort, the isostrophically coiled mollusc that lent its name to the bellerophont controversy. A - Apertural view. B - Lateral view. C - Antero-dorsal view. Notice the similarity of the shell form to an ancient Greek helmet - D THE CONTROVERSY AND ITS EFFECTS By the early 1980s many of those working with bellerophonts and/or molluscan phylogeny had encountered the puzzle of what to do with planispirally coiled, non- cephalopod, univalved molluscs bearing evidence of sym- metrical, often multiple, muscles. As a result of quite dif- ferent philosophies, these workers became polarized into three camps: group | — those who considered all bellerophonts to be monoplacophorans (Runnegar and Pojeta, 1974; Pojeta and Runnegar, 1976; Runnegar and Jell, 1976; Salvini-Plawen, 1980; Runnegar, 1981); group 2 — those who considered all bellerophonts to be gastropods (the majority of those who worked primarily on gastropod systematics and biostratigraphy); and group 3 — those who considered the bellerophonts a polyphyletic group that included both monoplacophorans and gastropods (Horny, 1963a, b, 1965a, b; Peel, 1972, 1974, 1976, 1980; Linsley, 1977, 1978a, b; Berg-Madsen and Peel, 1978). This third group considered that those bellerophonts having multiple sets of muscle scars were monoplacophorans whereas those with a single set of “columellar” scars were gastropods. Harper and Rollins (1982) reviewed the controversy and critically examined the significance of shell structure, apertura] re-entrants, parietal deposits, and muscle scars, all of which had been used at times by previous workers in dis- tinguishing gastropods and monoplacophorans among the bellerophonts. All of these features have limitations in sys- tematics, and at the time we felt muscle scars in particular were probably the least reliable single criterion on which to base a phylogeny. We concluded that placing bellerophonts and monoplacophorans in a single class based simply on shell form and muscle scar patterns was “tantamount to classifying bats, birds, and insects together because they all have bilateral symmetry and wings” (Harper and Rollins, 1982, p. 229), Opinion in the paleontological community has changed little since the early 1980s. A review of the litera- ture since mid-1982 indicates that group | still considers all bellerophonts to be monoplacophorans (Stanley, 1982; Runnegar, 1985; Runnegar and Pojeta, 1985; Signor, 1985; Geyer, 1994). Those in group 2 still consider the bellerophonts to be gastropods (McLean, 1984; Harper and Rollins, 1985; Kase and Nishida, 1986; Boucot et al., 1986; Rohr and Yochelson, 1990; Fryda and Guitierrez-Marco, 1996; Ebbestad, 1999; and many others). Those in group 3 still consider the bellerophonts to be divided among the gastropods and monoplacophorans (Linsley and Peel, 1983; Peel, 1985a, b, 1993; Horny, 1986, 1993; Edlinger, 1988; WahlIman, 1992; Berg-Madsen and Peel, 1994). The primary questionable contention that is driving the bellerophont controversy is that multiple pairs of sym- metrical muscles in conjunction with the bilateral symme- try of the shell indicate a monoplacophoran affinity for all bellerophonts (Wenz, 1940; Runnegar and Jell, 1976; Salvini-Plawen, 1980; Geyer, 1994). This argument is based on the assumption that asymmetry is a necessary consequence of torsion, involving all internal and external organs, including the shell muscle. In effect, this argument ignores the innumerable biological investigations done over the past 200 years showing post-torsional symmetry of many gastropod species and their shell muscles, and even of organs such as the ctenidia, osphradia, and hypo- branchial glands (as in many fissurelloideans). Yet the exis- tence of multiple, symmetrical muscles has been used time and again as a valid systematic character allying such dis- parate molluscs as Neopilina Lemche and Bellerophon Montfort. For example, Rollins and Batten (1968), when confronted with conflicting morphological evidence, were persuaded to assign the sinus-bearing bellerophont Sinuitopsis acutilira (Hall) to the monoplacophorans based solely upon multiple symmetrical pairs of muscle scars. TORSION AND ITS CONSEQUENCES The argument that bellerophonts were untorted because asymmetry, as represented by helical coiling as well as asymmetry of soft parts, must be a necessary conse- quence of torsion has been reiterated by many authors (Ghiselin, 1966; Runnegar and Pojeta, 1974; Stanley, 1982; HARPER AND ROLLINS: BELLEROPHONT CONTROVERSY 149 cm © B Q. E C — cm cm — 0}. cm F Fig. 2. Diagrams of representative tergomyan and cyclomyan monoplacophorans and patelliform gastropods, showing lateral and dorsal views of muscle scar patterns (black), including the discrete cephalic muscle (cm). A - Silurian tergomyan Tryblidium Lindstrém, showing the muscle field lying outside the apical axis (dashed line); B - Devonian cyclomyan Cyrtonella Hall, showing the muscle field intersecting the apical axis; C - Ordovician archinacelloidean Floripatella Yochelson; D - Recent fissurelloidean Diodora Gray; E - Recent acmaecidean Lottia Gray; and F - Recent cocculinoidean Cocculina Dall. Geyer, 1994, just to name a few). Linsley and Kier (1984), on the other hand, have argued that torsion and asymmetry are separate events in gastropod ontogeny; this has been shown to be the case by Bandel (1982). In fact, both the fossil record and the modern seas are full of prosobranch gastropods (torted) with symmetrical shells and opistho- branch gastropods and other molluscs (untorted) with heli- cally coiled shells. Haszprunar (1988a, b) also found no pri- mary correlation between torsion and helical coiling. He suggested that the occurrence of hyperstrophy in the proto- conchs of higher gastropods argues for two independent processes, demonstrating that the direction of shell coiling is not correlated with the direction of torsion. We therefore reject the notion that bellerophonts must be untorted because they were isostrophically coiled. APERTURAL RE-ENTRANTS One of the most frequently used shell features for separating gastropods from monoplacophorans are slits, sinuses, and other re-entrants on the margins of the shell aperture. For example, the Pleurotomarioidea are best- known for having a deep, narrow slit near the middle of the whorl that is used to channel the exhalant current away from the inhalant currents. Linsley’s fourth “law” of 150 AMER. MALAC. BULL. 15(2) (2000) gastropod shell form states, simply, “Angulations or re- entrants on the aperture are usually indicative of inhalant or exhalant areas; inhalant areas will be directed as anteriorly as possible” (Linsley, 1977:200). This law makes sense only in motile animals — the gastropod has the advantage of sensing its environment in advance of its direction of motion. Because some bellerophonts have re-entrants situ- ated laterally or postero-laterally, close to the shell coil, Starobogotov (1970), Berg-Madsen and Peel (1978), and others considered these forms monoplacophorans. However, Linsley (1977) admitted that he did not consider all gastropods when formulating this “law.” Harper and Rollins (1982) argued that the fourth “law” was not neces- sarily applicable to limpets and other bottom clampers. Many limpet gastropods have “re-entrants” all around the shell, related to coarse radial ribs that end at the shell mar- gin in small concave, v-shaped emarginations. Also, the inhalant and exhalant currents in limpets are not necessarily positioned anterior and posterior as they are in the helically coiled gastropods. In Patina, for example, the inhalant cur- rent enters the pallial groove along the length of the gill skirt and exits anteriorly through the right side of the nuchal cavity (Fretter and Graham, 1962). Patella similarly draws water in along the whole margin of the mantle, but expels it ventrally. In the caenogastropod limpet Crepidula, the current enters the mantle cavity on the left and leaves on the right — both inhalant and exhalant currents are lateral. We agree that apertural reentrants should be a good indication of whether a mollusc is a monoplacophoran or a gastropod, but only if the animal is truly motile. However, of all the coiled “monoplacophorans” in which the muscle scars are known, none is what we would call a truly motile animal. We envision most of the cyclomyans as limpets or shell clampers (as opposed to those that retract the body into the shell like most prosobranchs). These forms com- monly have large, tangential apertures and loosely coiled shells, as opposed to the tight coils and radial apertures of forms such as Bellerophon. They are also often covered with epibionts. This suggests to us that these forms were functionally limpet-like (although some very motile gas- tropods harbor shell-covering epibionts). The most func- tionally advantageous place for inhalant currents would be close to the gills and relatively far from the exhalant cur- rents. As in Patina, Patella, and Crepidula, this could easi- ly have been just about anywhere on the shell. We conclude that presence of lateral re-entrants is not a definitive criteri- on for separating limpet-like monoplacophorans and gas- tropods. It should be noted that certain cyclomyans have small re-entrants on the aperture beneath the shell coil (e. g. Neocyrtolites Horny — see Horny, 1993, Pl. 1, Figs. 7, 8). If these functioned as inhalant currents, the animal almost certainly was exogastric. However, these features common- ly are very small and probably would not have been very effective for channeling currents into a large aperture. Their function is uncertain, but may have been associated with an operculum or some other feature. COMPARISONS OF MUSCLE SCARS The suggestion that multiple muscles are a hallmark only of the monoplacophorans has no basis in fact. Most non-helically coiled limpet gastropods (fissurelloideans, patellogastropods, cocculinoideans, etc.) have horseshoe- shaped muscle fields with a distinctive break at the anterior end that defines the position of the mantle cavity. In the Patellogastropoda, the muscle field appears to be continu- ous, but in fact it is arranged as a series of discrete muscle bundles separated by narrow clefts. In the Cocculinoidea the muscle bundles display a widely varying degree of sep- aration (Haszprunar, 1987, 1988a). Thiem (1917) showed that the clefts allow space for afferent vessels to transfer blood from the foot and visceral mass into spaces in the mantle skirt where oxygenation takes place. The degree of separation appears to be directly related to the efficiency of the gills (Fretter and Graham, 1962). The Fissurelloidea have two very efficient ctenidia and no pallial gills, so the muscle bundles have little division (Fig. 2D). The Patellidae and Lottiidae exhibit well-developed pallial gills, and the Lottiidae have one ctenidium, so that oxygenation occurs easily. There are few afferent vessels and few mus- cle divisions (Fig. 2E). The Acmaeidae and Lepetidae have one ctenidium that is not very effective, and no pallial gills, so that pallial respiration via afferent vessels is necessary for adequate blood oxygenation. Species of these families commonly exhibit numerous muscle clefts. The separation of the individual muscle bundles is especially noticeable in the Cocculinoidea (Haszprunar, 1987, 1988a) (Fig. 2F). Based on innervation of shell muscles in limpet gas- tropods, Haszprunar (1985) showed that there is only a sin- gle pair of muscles. However, he also indicated the possi- bility that a single pair of shell muscles divided into dis- crete bundles might be the primitive condition shared by monoplacophorans (Haszprunar, 1988b, p. 374). It seems likely that multiplicity of muscles (regardless of division) is a normal condition in animals that pull the shell down to the substrate, as opposed to those that retract into their shells. It is also likely, based on those animals with discrete muscle bundles and those with seem- ingly solid horseshoe-shaped muscle masses, that environ- mental conditions and habitat diversity played a large part in the origination of the particular muscular condition. More muscle mass (the “solid” horseshoe) probably was an adaptation either to ward off predation or to the vagaries of near-shore conditions (storms, rough seas, desiccation HARPER AND ROLLINS: BELLEROPHONT CONTROVERSY 151 episodes, etc.). In either case, it would allow the animal to clamp the shell more tightly around the body and foot. If this is the case, then discrete, multiple muscles must indi- cate a lack of need, or less need, for such muscle mass. Indeed, modern monoplacophorans, which live in deep, quiet water, tend to have weakly developed muscles that leave no scars on their shell interiors. The Cocculinidae, which probably have the best set of discrete muscles among Recent limpet gastropods, also occur in deep water (McLean, 1987; Haszprunar, 1987). On the other hand, the fossil tryblidiids, which lived in shallow water and proba- bly functioned in much the same fashion as modern limpet gastropods, had strong muscles inserted well into the shell. The largest of the paired muscle bundles in modern limpet gastropods appears at the anterior end of the muscle field and is associated with retraction of the head. In at least two Recent limpet groups, Fissurelloidea and Lepetelloidea, this muscle pair has hook-like protrusions on the interior side of the bundle. McLean (1984) suggested these hook-like protrusions in the fissurelloideans mark the position of muscles that control the ctenidia. Although this is possible, we feel it is unlikely because similar hook-like protrusions occur on the cephalic muscles of many other limpets that have only one, or no, ctenidium. The Lepetelloidea, for example, have several pallial leaflets of secondary origin restricted to the right post-torsional side of the mantle skirt, outside the shell muscle field (Haszprunar, 1988b). The hook-shaped muscles also occur in Polyplacophora and in many Paleozoic monoplacophorans such as Tryblidium (cm in Fig. 2A). For McLean’s hypothe- sis to be correct, the ctenidia in Tryblidium and other mono- placophorans would have to be situated in an anterior man- tle cavity which, of course, does not exist in untorted mol- luscs. Alternatives include: 1) the hook-shaped muscles in fissurelloideans had distinctly different functions in mono- placophorans and, following torsion, the gastropod limpets adapted the muscle for use with the ctenidia; 2) the hook- shaped muscles are convergent features in the two lineages with no homologous functions; and 3) the hook-shaped muscles are homologous features but have a different func- tion than suggested by McLean (1984). We prefer the third option, but are uncertain of their function. The hook-shaped muscles are likely related to retraction of the head and, pos- sibly, anchoring of the buccal muscles. Purchon (1977, p. 467), in discussing the muscles of Recent monopla- cophorans, described a “complex series of muscles inserted into the shell anteriorly on either side of the mouth [that] serve to move the lips of the mouth, the velar lobes, the post-oral tentacles, and the radular apparatus.” Although the cephalic muscle area of Neopilina Lemche does not look much like that of Tryblidium Lindstrém, in all likeli- hood they are homologous, as well as analogous, structures. It may be that the change from a shallow water, presumably herbivorous mode of life in Paleozoic monoplacophorans to a deep water, deposit feeding mode of life in Recent forms accounts for the differences. We should, therefore, be able to use the relatively larger cephalic muscles in cap-shaped or spoon-shaped monoplacophoran and gastropod limpets to recognize the anterior ends of the shells. In 7ryblidium Lemche (Fig. 2A) the cephalic muscles occur near the shell apex, indicating the animal’s exogastric nature. In the gastropod limpets (Figs. 2C-F) the cephalic muscles should occur on the end opposite the shell apex (endogastric). The same is true of the archinacelloideans (Fig. 2C) which, despite their appar- ent tergomyan appearance, generally are considered to be gastropods (Starobogotov, 1970; Harper and Rollins, 1982; Yochelson, 1988; Mazaev, 1998). (Note, however, that in many limpet gastropods, the “coil” is anterior or central- ized. Therefore, the terms “endogastric” and “‘exogastric”’ are essentially meaningless without reference to soft anatomy.) We suggest these cephalic muscles also occur in the cyclomyans as the pair of subcircular to oblong muscle scars on the shell dorsum, farthest from the shell coil (Figs. 2B, 3A). Dzik (1981) reconstructed the cyclomyan Sinuitopsis Perner with these muscles acting as retractors for an operculum on the trailing foot of the animal. We find this highly unlikely, not because there is no evidence for an operculum in the cyclomyans, but simply because the mus- cle pads appear to be homologous with the cephalic mus- cles in tergomyans, archinacelloideans, and modern limpet gastropods. If the cyclomyans were exogastric, like the ter- gomyans, the cephalic muscles of these animals then would have been attached to the back of the shell above the viscer- al hump and well away from the head, which makes no sense. The head must have been situated below the cephalic muscles and on the opposite end of the body whorl from the coil, indicating the cyclomyans were endogastric. While this is not an impossible situation in monoplacophorans, in view of most efficient muscle function a more likely expla- nation is that the cyclomyans were gastropods. We envision a complete range of shell muscles arranged in a functional sequence (Fig. 4) based on coiling parameters and apertural characters, which in turn were very probably based on mode of life and physical environ- ment. According to figure 4, the uncoiled cyclomyans like Cyrtonella Hall were the most limpet-like forms, grading (functionally, but not necessarily phylogenetically) through the cyrtolitiform (e. g. Cyrtolites Conrad) and sinuitiform (e. g. Sinuitopsis Perner which has an anal emargination like the pleurotomarioideans) cyclomyans to the bellerophontoideans (e. g. Bellerophon Montfort). Forms such as Sinuites Koken, Syvestrosphaera Peel, and Carcassonella Horny and Peel were intermediaries, retain- ing the cephalic muscles of the limpets while accomplish- By AMER. MALAC. BULL. 15(2) (2000) ing retraction through application of lateral retractor mus- cles similar to the helically-coiled gastropods. These forms may have had a completely different mode of life from both the limpets and bellerophontoideans. In fact, Horny (1996) has described Sinuites Koken as a semi-infaunal gastropod based on secondary shell layers similar to those in the bellerophontoidean Euphemites Warthin (see Harper and Rollins, 1985). Pronounced variation in shell muscle posi- tion and pattern correlated with degree of shell coiling is also recognized in the Paleozoic platyceratids (Rollins and Brezinski, 1988). PARIETAL DEPOSITS Harper and Rollins (1982) and WahlIman (1992) were convinced that the most reliable criterion for assign- ment of any isostrophically-coiled mollusc to the Gastropoda was the presence of a massively developed parietal deposit. Such deposits are most easily viewed as functionally enhancing shell stability by placing the bulk of shell weight over the posterior foot, and less reasonable if elevated over the head of the animal. Rollins (1966) noted that the parietal pad of the bellerophont Promatis also dis- played a central cleft that most likely rested directly upon the foot. Massive parietal deposits, however, are variously developed in bellerophonts. When present, they appear to confirm torsion and justify ready assignment to the Gastropoda. Their absence, or even the presence of a thin inductural wash, is inconclusive. As noted by Pojeta and Runnegar (1976), thin secondary shell deposits might have been secreted by epithelial tissue near the head of the animal. DISCUSSION It has become obvious that the monoplacophorans, at least those represented by Neopilina Lemche, do not rep- resent the molluscan archetype (hypothetical ancestral mol- lusc). All of those forms we know or suspect to be mono- placophorans are too highly specialized, with their serial pairs of organs and muscles, the mantle covering the entire animal, and a pallial groove that completely encircles the body substituting for a centralized mantle cavity. However, given the close anatomical relationship of monopla- cophorans and polyplacophorans, this must be the ple- siomorphic condition from which the gastropods arose. Haszprunar (1988b) suggested that a shallow mantle cavity, with a concomitant degree of respiration in the mantle roof and margin, is the primitive condition in gastropods. In addition, Haszprunar believes the subpallial cavity was probably used for respiration until the gastropods evolved Fig. 3. Proposed configurations of head, foot, visceral mass, and various organs for: A - Cyrtolites, a representative cyclomyan; and B - Bellerophon, a representative bellerophontoidean. a - anus; cm - cephalic muscle; ct - ctenidia; e - eye; es - esophagus; f - foot; mc - mantle cavity; me - mantle edge; pg - pallial gill; rm - retractor muscle; sh - shell; sm - shell muscles; sn - snout; st - stomach; t - tentacle; vm - visceral mass. secondary gills. The fossil record suggests otherwise, as shown below. If, in fact, the fossil record is good enough, we should be able to see many of the morphological changes that evolved throughout the Phanerozoic in the tergomyans, cyclomyans, and bellerophontoideans. And indeed, we can. In figure 4 we show a series of hypothetical monopla- cophoran and gastropod forms that are simply generalized models based on actual fossils such as Tryblidium Lindstrom, Cyrtolites Conrad, and Bellerophon Montfort. We envision a series of anatomical changes that occurred relatively rapidly in geologic time. Following torsion (Fig. 4, fundamental change A) the archetypal gastropod was a torted monoplacophoran having serially arranged muscles and organs, including multiple gills in a pallial groove sur- rounding the body. At some later time the population of archetypal gastropods gave rise to forms that increased the HARPER AND ROLLINS: BELLEROPHONT CONTROVERSY Extinction in the Triassic Patellogastropoda Prosobranchia Neopilininae Morphological Change ——-———_® Fig. 4. Diagram illustrating a hypothetical generalized evolutionary scheme for the gastropods and monoplacophorans. Muscle scars are shown in black, stippled where hidden behind shell. Dotted lines indicate proposed extent of the mantle/pallial cavity. Letters designate fundamental changes in anatomy. A - torsion; B - vertical expansion of the shell; C - development of an anterior mantle cavity; D - decrease in afferent vessels, resulting in fusion of the muscles to a horseshoe; E - helical coiling; F - isostrophic coiling; G - development of an anterior mantle cavity; and H - reduction of shell muscles to two retractors. 154 AMER. MALAC. BULL. 15(2) (2000) height of the shell (B), allowing for enlargement of the pal- lial groove. Certain of these high-domed gastropods experi- mented with centralizing the mantle cavity and reducing the number of gills (C). This allowed a decrease in the number of afferent vessels and a consolidation of the discrete shell muscles into a horseshoe (D). These gastropods eventually gave rise to the Patellogastropoda, which are distinct from all other archaeogastropods (see Lindberg, 1988 for a review of this gastropod order). Whether or not there was a coiled intermediate in this lineage, as some authors have suggested, is beyond the scope of this discussion. High doming and centralization of the mantle cavity also allowed a separate lineage to experiment with coiling, particularly helical coiling. These forms gave rise to the most successful group of molluscs, the prosobranchs. Meanwhile, the original group of high-domed archetypal gastropods also were “experimenting” with coil- ing, probably as a means to increase the amount of visceral mass without making the shell unstable and/or to afford greater protection of soft organs within the shell, perhaps concomitant with a more motile existence. These gas- tropods, which we have been calling cyclomyans, retained the plesiomorphic pallial groove and multiple gills of the ancestral gastropods. Through time these molluscs enlarged the pallial groove and concomitantly moved the shell mus- cles farther back into the shell. This provided a morpholog- ical setting that would lead, eventually, to the independent development of the centralized mantle cavity. Once this important feature was attained, the cyclomyans went through an adaptive radiation that resulted in an explosion of forms within the morphospace of the isostrophically- coiled shell. This radiation, linked to probable predation pressures and adaptive changes to the physical environ- ment, forced modification in the shell from evolute, open- coiled, limpets to tightly involute, compact animals that spread out into a variety of adaptive niches - the Bellerophontoidea. Unfortunately, as successful as they were, the bellerophont lineage decreased substantially in diversity toward the end of the Paleozoic and ultimately became extinct, but well into the Triassic, not at the biotic crisis of the Permian extinction. SUMMARY AND CONCLUSIONS During the 1970s and early 1980s, a controversy concerning the systematic placement of the Bellerophontoidea (‘‘bellerophonts”) — a group of planispi- rally coiled marine molluscs known only from the fossil record and commonly considered to be primitive gas- tropods - emerged in the paleontologic literature. Although some points of contention had been raised as early as the 1940s, the primary concern now focused upon interpreta- tion of multiple, symmetrical muscle scars preserved on internal molds (steinkerns) of certain Bellerophontoidea. One camp insisted that, because multiple, symmetrical muscles suggest metamery, the Bellerophontoidea must have been untorted and, therefore, were monoplacophorans rather than gastropods. Another group insisted that “bellerophonts” were gastropods. Yet a third camp argued that single pairs of muscles and various gastropod-like fea- tures of some “bellerophonts” indicate that many of these animals were gastropods. They conceded that other Bellerophontoidea were monoplacophorans and, although distinguishing between the separate lineages would be diffi- cult, it could be done by analyzing all the shell features and not just the muscle scars (which were notoriously rare). In 1982 we presented evidence that the various arguments for “bellerophonts-as-monoplacophorans” were specious and concluded that all “bellerophonts” were gastropods. A reexamination of that evidence and consideration of research on these molluscs since 1982 has failed to con- vince us otherwise. Following torsion, the archetypal gastropod was a torted monoplacophoran having serially arranged muscles and organs, including multiple gills in a pallial groove sur- rounding the body. At some later time the population of archetypal gastropods gave rise to forms that increased the height of the shell, allowing for enlargement of the pallial groove and centralization of the mantle cavity, and eventual development of the Patellogastropoda. High doming and centralization of the mantle cavity also allowed a separate lineage to experiment with helical coiling leading to the very successful Prosobranchia. The original group of high- domed archetypal gastropods also “experimented” with coiling, and this group, which we have been calling cyclomyans, retained the plesiomorphic pallial groove and multiple gills of the ancestral gastropods. This provided a morphological setting that would lead, eventually, to the independent development of the centralized mantle cavity and the adaptive radiation of the Bellerophontoidea. The evolutionary scenario we described is specula- tive, of course, but it does have its basis in the fossil record and in comparative anatomy. We conclude, as we did previ- ously (Harper and Rollins, 1982), that the Cyclomya and Bellerophontoidea were gastropods rather than monopla- cophorans, and Cyclomya should be retained as a suborder of Bellerophontina. Only the Tergomya were, and are, true monoplacophorans. It should be apparent that there is no reason to aban- don the term Monoplacophora. Even though it originated as an informal name, and recently became a “garbage can” term for an unlikely grouping of exotic molluscs and, perhaps, non-molluscs, Monoplacophora should be retained as a formal class name. The taxonomic scope of this term has come full circle - originally proposed to include only HARPER AND ROLLINS: BELLEROPHONT CONTROVERSY 155 the Tryblidiidae, it once again can be defined as the Tryblidiidae and associated Recent neopilinids. The term Tergomya then is synonymous with Monoplacophora and should be abandoned. Cyclomya should be retained as a suborder of Bellerophontina. LITERATURE CITED Bandel, K. 1982. Morphologie und Bildung der friihontogenetischen Gehiuse bei conchiferen Mollusken. Facies 7:1-198. Berg-Madsen, V. and J. S. Peel. 1978. Middle Cambrian monopla- cophorans from Bornholm and Australia, and the systematic posi- tion of the bellerophontiform molluscs. Lethaia 11:113-125. Berg-Madsen, V. and J. S. Peel. 1994. A tergomyan mollusc from the Upper Cambrian of Wales. Palaeontology 37:505-512. Boucot, A. J., D. M. Rohr, J. Gray, A. de Faria, and G. K. Colbath. 1986. Plectonotus and Plectonotoides, new subgenus of Plectonotus (Bellerophontacea: Gastropoda) and their biogeographic signifi- cance. Neues Jahrbuch fiir Geologie und Palaoéntologie, Abhandlungen 173:167-180. Dzik, J. 1981. Larval development, musculature, and relationships of Sinuitopsis and related Baltic bellerophonts. Norsk Geologisk Tidsskrift 61:111-121. Ebbestad, J. O. R. 1999. Bucaniid gastropods from the Upper Ordovician of Baltica, with a discussion of the Bucaniinae. Palaeontology 42(1):149-169. Edlinger, K. 1988. Torsion in gastropods: A phylogenetic model. Jn: Prosobranch Phylogeny, W. F. Ponder, D. J. Eernisse, and J. H. Waterhouse, eds. pp. 239-248. Malacological Review Supplement 4, Ann Arbor, Michigan. Fretter, V. and A. Graham. 1962. British Prosobranch Gastropods. Ray Society, London, 755 pp. Fryda, J. and J. C. Guitierrez-Marco. 1996. An unusual new sinuitid mol- lusc (Bellerophontoidea, Gastropoda) from the Ordovician of Spain. Journal of Paleontology 70:602-609. Geyer, G. 1994. Middle Cambrian mollusks from Idaho and early conchiferan evolution. /n: Studies in Stratigraphy and Paleontology in Honor of Donald W. Fisher, E. Landing, ed. pp. 69-86. New York State Museum Bulletin 481, Albany, New York. Ghiselin, M. T. 1966. The adaptive significance of gastropod torsion. Evolution 20:337-348. Harper, J. A. and H. B. Rollins. 1982. Recognition of Monoplacophora and Gastropoda in the fossil record: A functional morphological look at the bellerophont controversy. Proceedings of the Third North American Paleontological Convention 1:227-232. Harper, J. A., and H. B. Rollins. 1985. Infaunal or semi-infaunal bellerophont gastropods: Analysis of Euphemites and functionally related taxa. Lethaia 18:21-37. Haszprunar, G. 1985. On the innervation of gastropod shell muscles. Journal of Molluscan Studies 51:309-314. Haszprunar, G. 1987. Anatomy and affinities of cocculinid limpets (Mollusca, Archaeogastropoda). Zoologica Scripta 16:305-324. Haszprunar, G. 1988a. Comparative anatomy of cocculiniform gastropods and its bearing on archaeogastropod systematics. Jn: Prosobranch Phylogeny, W. F. Ponder, D. J. Eernisse, and J. H. Waterhouse, eds. pp. 64-84. Malacological Review Supplement 4, Ann Arbor, Michigan. Haszprunar, G. 1988b. On the origin and evolution of major gastropod groups, with special reference to the Streptoneura. Journal of Molluscan Studies 54:367-441 Horny, R. 1963a. On the systematic position of the cyrtonelloids (Mollusca). Casopis Narodniho Muzea, Oddil prirodovedna 132:90-91. Horny, R. 1963b. Archaeopraga, a new problematic genus of monopla- cophoran mollusks from the Silurian of Bohemia. Journal of Paleontology 55:1071-1073 Horny, R. 1965a. On the systematic position of Cyrrolites Conrad, 1838 (Mollusca). Casopis Narodniho Muzea, Oddil prirodovedna 134:8-10. Horny, R. 1965b. Cyrtolites Conrad, 1838 and its position among the Monoplacophora (Mollusca). Sbornik Narodntho Muzea v Praza, ser. B 21:57-70. Hormy, R. 1986. Muscle scars in Sinuites (Strangulites) (Mollusca) from the Ordovician of Bohemia. Casopis Narodniho Muzea, Oddil prirodovedna 155:109-118. Horny, R. J. 1993. Shell morphology and mode of life of the Lower Devonian cyclomyan Neocyrtolites (Mollusca, Tergomya). Casopis Narodniho Muzea, Oddil prirodovedna 162:57-66. Horny, R. J. 1996. Secondary shell deposits and presumed mode of life in Sinuites (Mollusca, Gastropoda). Acta musei Nationalis Pragae, ser. B., Historia Naturalis 51:89-103. Kase, T. and T. Nishida. 1986. A Middle Devonian gastropod faunule from the Nakazato Formation of Kitakami, Northeast Japan. Bulletin of the National Science Museum, ser. C, Geology and Paleontology 12:73-89. Knight, J. B. 1952. Primitive fossil gastropods and their bearing on gastro- pod classification. Smithsonian Miscellaneous Collections 117:1- 56. Lemche, H. 1957. A new living deep-sea mollusc of the Cambro- Devonian class Monoplacophora. Nature 179:413-416. Lemche, H. and K. G. Wingstrand. 1959. The comparative anatomy of Neopilina galatheae Lemche, 1957 (Mollusca, Tryblidiacea). Galathaea Report 3:9-72. Lindberg, D. R. 1988. The Patellogastropoda. In: Prosobranch Phylogeny, W. F. Ponder, ed. pp. 35-63. Malacological Review, Supplement 4. Linsley, R. M. 1977. Some “laws” of gastropod shell form. Paleobiology 3:196-206. Linsley, R. M. 1978a. Shell form and the evolution of gastropods. American Scientist 66:432-441. Linsley, R. M. 1978b. Locomotion rates and shell form in the Gastropoda. Malacologia 17:193-206. Linsley, R. M. and W. M. Kier. 1984. The Paragastropoda: A proposal for a new class of Paleozoic Mollusca. Malacologia 25:241-254. Linsley, R. M. and J. S. Peel. 1983. Gonad pouches in a new clam-like monoplacophoran from the Silurian of Sweden. Lethaia 16:273- 280, 5 fig. Mazaev, A. V. 1998. A new genus of Patelliformes (Gastropoda) and some systematic questions on Paleozoic limpets. In: Abstracts, World Congress of Malacology, R. Bieler and P. M. Mikkelsen, eds. p. 213. Unitas Malacologica, Washington, D.C. McLean, J. H. 1984. A case for derivation of the Fissurellidae from the Bellerophontacea. Malacologia 25:3-20. McLean, J. H. 1987. Taxonomic description of cocculinid limpets (Mollusca, Archaeogastropoda): Two new species and three redis- covered species. Zoologica Scripta 16:325-333. Peel, J. S. 1972. Observation on some Lower Paleozoic tremanotiform Bellerophontacea (Gastropoda) from North America. Palaeontology 15:412-422. Peel, J. S. 1974. Systematics, ontogeny and functional morphology of Silurian trilobed bellerophontacean gastropods. Bulletin of the Geological Society of Denmark 23:231-264. 156 AMER. MALAC. BULL. 15(2) (2000) Peel, J. S. 1976. Musculature and systematic position of Megalomphala taenia (Bellerophontacea; Gastropoda) from the Silurian of Gotland. Bulletin of the Geological Society of Denmark 25-49-55. Peel, J. S. 1980. A new Silurian retractile monoplacophoran and the origin of the gastropods. Proceedings of the Geologists Association 91:91-97. Peel, J. S. 1985a. 6.1 Class Monoplacophora. /n: Atlas of Invertebrate Macrofossils, J. W. Murray, ed. pp. 79. The Palaeontological Association, London, England. Peel, J. S. 1985b. Autecology of Silurian gastropods and monopla- cophorans. Special Papers in Paleontology 32:65-182. Peel, J. S. 1991. Functional morphology of the Class Helcionelloida nov., and the early evolution of the Mollusca. /n: The Early Evolution of Metazoa and the Significance of Problematic Taxa, A. M. Simonetta and S. Conway Morris, eds. pp. 157-177. Cambridge University Press, Cambridge, England. Peel, J. S. 1993. Muscle scars and mode of life of Carinaropsis (Bellerophontoidea, Gastropoda) from the Ordovician of Tennessee. Journal of Paleontology 67:528-534. Pojeta, J. and B. Runnegar. 1976. The paleontology of rostroconch mol- lusks and the early history of the phylum Mollusca. U. S. Geological Survey Professional Paper 968: 1-88. Purchon, R. D. 1977. The Biology of the Mollusca, 2nd ed. Pergamon Press, Oxford, 560 pp. Rohr, D. M. and E. L. Yochelson. 1990. An unusual new Bellerophontacean gastropod from the Ordovician (Whiterockian) of Nevada. Journal of Paleontology 64:956-960. Rollins, H. B. 1966. Morphological observations on the bellerophont Ptomatis patulus (Hall) (Gastropoda, Bellerophontacea). American Museum of Natural History, Novitates 2242:1-7. Rollins, H. B. and R. L. Batten. 1968. A sinus-bearing monoplacophoran and its role in the classification of primitive molluscs. Palaeontology 11(1):132-140. Rollins, H. B. and D. K. Brezinski. 1988. Reinterpretation of crinoid- platyceratid interaction. Lethaia 21:207-217. Runnegar, B. 1981. Muscle scars, shell form and torsion in Cambrian and Ordovician univalved molluscs. Lethaia 14:31 1-322. Runnegar, B. 1985. Origin and early history of mollusks. Jn: Mollusks: Notes for a Short Course, T. W. Broadhead, ed. pp. 17-32. University of Tennessee, Department of Geological Sciences, Studies in Geology 13, 306 pp. Runnegar, B. and P. A. Jell. 1976. Australian Middle Cambrian molluscs and their bearing on early molluscan evolution. Alcheringa 1:109- 138. Runnegar, B. and J. Pojeta. 1974. Molluscan phylogeny; the paleontologi- cal viewpoint. Science 186(4161):311-317. Runnegar, B. and J. Pojeta. 1985. Origin and diversification of the Mollusca. In: The Mollusca 10, Evolution, E. R. Trueman and M. R. Clarke, eds. pp. 1-57. Academic Press, Orlando, Florida, 491 p. Salvini-Plawen, L. V. 1980. A reconsideration of systematics in the Mollusca (phylogeny and higher classification). Malacologia 19:249-278. Signor, P. W. 1985. Gastropod evolutionary history. Jn: Mollusks: Notes for a Short Course, T. W. Broadhead, ed. pp. 157-173. University of Tennessee, Department of Geological Sciences, Studies in Geology 13, 306 pp. Stanley, S. M. 1982. Torsion: A new twist. Neues Jahrbuch fiir Geologie und Paladntology 1807:95-107. Starobogotov, Y. I. 1970. K sistematike rannepaleozoyskikh Monoplacophora [Systematics of Early Paleozoic Monoplacophora]. Paleontologicheski Zhurnal 4:6-17. Thiem, H. 1917. Beitrage zur Anatomie und Phylogenie der Docoglossen. II. Die Anatomie und Phylogenie der Monobranchen (Akmaiden und Scurniden nach der Sammlung Plates). Jenaische Zeitschrif fiir Naturwissenschaft 54:405-630. Wahlman, G. P. 1992. Middle and Upper Ordovician symmetrical uni- valved mollusks (Monoplacophora and Bellerophontina) of the Cincinnati arch region. U. S. Geological Survey Professional Paper 1066-O, pp. 01-0213. Wenz, W. 1940. Ursprung und friihe Stammesgeschichte der Gastropoden. Archiv fiir Molluskenkunde, Frankfort 72:1-10. Wingstrand, K. G. 1985. On the anatomy and relationships of Recent Monoplacophora. Galathaea Report 16:7-94. Yochelson, E. L. 1967. Quo vadis, Bellerophon? In: Essays in Paleontology and Stratigraphy—-R. C. Moore Commemorative Volume, C. Teichert and E. L. Yochelson, eds. pp. 141-161. University of Kansas Press, Manhattan, Kansas. Yochelson, E. L. 1988. A new genus of Patellacea (Gastropoda) from the Middle Ordovician of Utah: The oldest known example of the superfamily. New Mexico Bureau of Mines & Mineral Resources Memoir 44:195-200. Date of manuscript acceptance: 23 June 2000 Cambrian Pelecypoda (Mollusca) John Pojeta, Jr. U.S. Geological Survey & Smithsonian Institution, MRC 137, Museum of Natural History, Washington, D.C. 20560, U. S. A. Abstract: The record of Cambrian bivalved organisms that have been placed in the Mollusca includes undoubted pelecypods and other bivalved shells. This record is reviewed herein. Additional topotype material of Delgado’s (1904, Communicacoes da Commissao do Servico Geologico de Portugal 5:307- 374) Lower Cambrian specimens from Portugal shows that they are distorted brachiopods. Restudy of Zhang’s (1980a, Bulletin Chinese Academy of Geological Sciences, Series 8, 1:1-19; 1980b, 26th International Geological Congress 4:121-129) type specimens from the Lower Cambrian of China shows that most of them are also brachiopods. New specimens of Tuarangia MacKinnon indicate that this genus is a Middle-Upper Cambrian pelecypod. Key Words: Cambrian, Pelecypoda, Brachiopoda, Fordilla, Pojetaia, Mollusca, bivalves Over the past generation, a great deal of new infor- mation and data have been added to our knowledge of Cambrian and Ordovician pelecypods; much of this new information is summarized by Babin in this volume. My paper concentrates on the expanding knowledge of Cambrian pelecypods. Undoubted Cambrian pelecypods are now known from North America, Europe, Greenland, Asia, Africa, and Australia. However, the stratigraphic range of the class in the Cambrian, which bivalved taxa are pelecypods and which ones are not, and how the Cambrian pelecypods are related to Ordovician taxa, are all questions under study by various workers who have reached differing conclusions. Much recent work in Europe and South America has shown that the great Ordovician radiation of pelecy- pods began in Arenigian time about 10 million years sooner than previously thought. Thus, it is reasonable to assume that Cambrian rocks hold part of the key to understanding Ordovician pelecypods. The Ordovician radiation causes us to debate the number of subclasses of pelecypods that should be recog- nized. However, by the end of Ordovician time almost all modes of life exploited by pelecypods were present — deposit feeding, filter feeding, epifaunal byssal attachment, burrowing, semi-infaunal byssal attachment, nestling, bor- ing and, probably, swimming; ligament types and dentition were also highly varied. EARLY CAMBRIAN TAXA The known Cambrian stratigraphic range of pelecy- pods is being filled in with records now ranging from the Early Cambrian through the early Late Cambrian. In a previous summary of information about Cambrian pelecypods to that date (Pojeta, 1975) I came to the conclusion that the only undoubted Cambrian pelecy- pod was Fordilla troyensis Barrande, 1881 (Fig. 1A,C,D). I noted that Lamellodonta simplex Vogel, 1962, probably was an obolellid brachiopod, a decision later confirmed by Havlicek and Kriz (1978). In the 1975 paper, I gave the age of Fordilla as Early Cambrian. More recently, Geyer and Streng (1998:87) indicated the age of Fordilla as “early late Early Cambrian to the late Early Cambrian (probably equivalents of the Siberian Atdabanian and Boto- man{ian]...”. Runnegar and Pojeta (1992:117) indicated Fordilla was known from rocks as old as Tommotian in Siberia (Fig. 2). My 1975 summary noted that Fordilla was known from Lower Cambrian rocks in New York State, Newfoundland, Denmark, and Greenland. The genus has subsequently been found in Lower Cambrian rocks of the Siberian platform (Krasilova, 1977, 1987; Jermak, 1986) and eastern Germany (Elicki, 1994); Geyer and Streng, (1998:88) suggested that the German Fordilla could be bet- ter assigned to Pojetaia. In addition, undescribed material is known from Labrador. The interpretation of Fordilla as a pelecypod was based on articulated material and internal molds that show pelecypod muscle scars. Subsequently, teeth and ligament groove fillings have been found (Pojeta, 1978). Subsequent workers accepted Fordilla as a pelecypod, documented the constancy of the muscle scar patterns, provided data about American Malacological Bulletin, Vol. 15(2) (2000):157-166 157 158 AMER. MALAC. BULL. 15(2) (2000) Fordilla’s shell microstructure, and added information about its geographic distribution in Early Cambrian time (Runnegar and Pojeta, 1992; Hinz-Schallreuter, 1995; Cope, 1996; Geyer and Streng, 1998; Cope and Babin, 1999). Yochelson (1978) treated Fordilla as having origi- nated from a hypothetical unknown ancestral mollusk sepa- rately from the origin of the Pelecypoda from the same ancestor. Yochelson (1981) preferred to treat Fordilla as molluscan incertae sedis, until more information about the genus became known. The authors cited herein have pro- vided that new information. The other undoubted Early Cambrian pelecypod is Pojetaia runnegari Jell, 1980 (Fig. 1E-H), first described Fig. 1. Lower Cambrian pelecypods Fordilla troyensis Barrande (A, C, D), F. siberica Krasilova (B), and Pojetaia runnegari Jell (E-H). All scale bars equal one millimeter. A, right valve composite drawing of the muscle scars as seen on internal molds. Drawing by B. Runnegar. B, dorsal view of internal mold showing trace of one tooth in each valve and the filling of the ligament space, Lower Cambrian, Siberia. C, right valve internal mold showing most of the muscle scars, Lower Cambrian rocks, New York. D, left valve of partially shelled specimen showing growth lines, Lower Cambrian, Greenland. E, composite reconstruction of the muscle scars, modified from Runnegar and Bentley (1983); published with permission of The Journal of Paleontology. F, dorsal part of left valve showing ligament area and teeth, specimen | mm long, Lower Cambrian, Australia. G, dorsal view of internal mold showing trace of one tooth in each valve, anterior to left, specimen 1.2 mm long, Lower Cambrian, Australia. H, right lateral view of internal mold. The apparent texture on the surface was interpreted (Runnegar and Bentley, 1983) as impressions of crystals on the inner surface of the shell, specimen 0.9 mm long, Lower Cambrian, Australia. POJETA: CAMBRIAN PELECYPODA 159 from the Lower Cambrian rocks of Australia. This species was intensively studied by Runnegar and Bentley (1983), who dealt with all aspects of its morphology from prodisso- conch IJ to shell microstructure. Runnegar and Pojeta (1992) added some information about comparative shell microstructure, stratigraphic distribution, and probable syn- onyms of both Pojetaia and Fordilla. Pojetaia has been widely accepted as the second well-documented Early Cambrian pelecypod. The genus is now known from Lower Cambrian rocks in South Australia; New South Wales, Australia; Anhui Province, China; Henan Province, China; Bornholm, Denmark; Morocco (Geyer and Streng, 1998); and, questionably from Newfoundland (Landing and Westrop, 1998). Geyer and Streng (1998:89) provided a table comparing the six named species of Pojetaia. Runnegar and Pojeta (1992:117) noted that Fordilla and Pojetaia are about coeval in their known first occurrences. The relationships of Pojetaia and Fordilla to each other and to Ordovician pelecypods have been under fre- quent discussion. Making comparisons to Ordovician pele- cypods, Pojeta and Runnegar (1985) regarded Pojetaia as a palaeotaxodont and Fordilla as an isofilibranch. Runnegar and Pojeta (1992) placed the two genera in a monophyletic grouping, Fordillidae Pojeta, 1975, based on their similar shell microstructure. Hinz-Schallreuter (1995) followed Runnegar and Pojeta (1992) and placed Pojetaia in the Fordillidae. Geyer and Streng (1998) and Cope (1996) treated Pojetaia as a palaeotaxodont. Cope (1996) noted that most authorities regard Fordilla as a pelecypod, but he doubted its assignment to the Isofilibranchia, as did Waller (1998). Cope and Babin (1999:175) noted: “Prolonged debate on the affinities of the genus Fordilla...has finally been settled and it is now regarded as one of the two known early Cambrian bivalve genera.” Runnegar and Pojeta (1992) and Geyer and Streng (1998) discussed the various Lower Cambrian taxa that have been assigned to the Pelecypoda. They noted probable synonyms and taxa that are probably not pelecypods. Herein, new information is provided on two occurrences, in Portugal and China, indicating that certain Lower Cambrian bivalves once regarded as pelecypods should be assigned elsewhere. From the Lower Cambrian rocks of Portugal, Delgado (1904) described nine species, placed in six gen- era, of what he regarded as pelecypods. For years since his publication, Delgado’s material was little studied or noticed. Teixeira (1952) restudied Delgado’s material, placed three of the species defined by Delgado in “Lamellibranchia?,” and the remainder in the single species “Modiolopsis” bocagei Delgado, which Teixeira regarded as a pelecypod. In 1904, Delgado sent C. D. Walcott identified topo- type specimens of Modiolopsis bocagei and Davidia doll- fusi Delgado; these are figured herein for the first time (Fig. 3G, H). These specimens are probably deformed bra- chiopods; they clearly show cardinalia in the presumed brachial valves at the midlength of the hinge margin indi- cating a possible articulate brachiopod (identified by J. T. Dutro, Jr., who suggested that the cardinalia could be called socket ridges in these specimens). Thus, it seems likely that all of Delgado’s “species” of Early Cambrian pelecypods are deformed brachiopods; the other elements of the fauna described by Delgado are also clearly deformed. Zhang (1980a, b) described four new genera, seven new species, and two new families of bivalved creatures; these were first proposed as nomina nuda (Zhang, 1979). His material is from the Tianheban Formation, a Redlichia trilobite-bearing Lower Cambrian unit at Zhongbao Village, Xianfeng County, Hubei Province, China; Zhang treated the bivalves from Hubei as pelecypods. Zhang (pers. comm., June 1996) noted that the Hubei fauna may correlate with Ordovician Late Cambrian Middle Cambrian uarangia Ti Toyonian Stage Arhouriella | Botomian Stage = = ‘= 5 = a oO Pojetaia Atdabanian Stage Fordilla Tommotian Stage y Nemakit- Daldynian Stage Neoproterozoic Fig. 2. Stratigraphic distribution of Cambrian pelecypods. The numbers in the left hand column indicate millions of years ago. The Early-Middle Cambrian boundary is queried because there is debate about where to place it (Landing ef al., 1998; Geyer, 1998). 160 AMER. MALAC. BULL. 15(2) (2000) Fig. 3. Probable Lower Cambrian brachiopods that have been described as pelecypods. Hubeinella formosa Zhang (A-F), “Davidia” dollfusi Delgado (G), “Modiolopsis” bocagei Delgado (H), and Xianfengoconcha minuta Zhang (I-K). All scale bars represent | mm. A, paratype, figured by Zhang (1980a, pl. 3, fig. 3). B, paratype, figured by Zhang (1980a, pl. 3, fig. 4). C, paratype, figured by Zhang (1980a. pl. 3, fig. 2). D, holotype, figured by Zhang (1980a, pl. 3, fig. 1). E, previously unfigured specimen. F, paratype, figured by Zhang (1980a, pl. 3, fig. 5). A-F all from Lower Cambrian, China. G, H, symmetri- cal internal molds. Arrow points to apex of shell, which to either side has socket ridges. G and H are from Lower Cambrian, Portugal. I, holotype, figured by Zhang (1980a, pl. 2, figs. 12, 13). J, paratype, figured by Zhang (1980a, pl. 2, fig. 14). K, paratype, figured by Zhang (1980a, pl. 2, fig. 15). I-K are from Lower Cambrian rocks, China. the Siberian Botomian Stage. Subsequent to Zhang’s work, Runnegar and Pojeta (1992) treated his taxa as stenothecoids, a group of uncer- tain affinities, and Geyer and Streng (1998:87) in dis- cussing Zhang’s taxa noted: “...the interpretation of the fol- lowing taxa, all from the Early Cambrian of Hubei Province on the Cambrian Yangtze Platform, is refused: ... All of them are based on slightly or considerably distorted valves of inarticulate brachiopods rather than pelecypods.” A difficulty in using Zhang’s work is the low print- ed quality of the photographic illustrations. I have had the opportunity to examine and photograph many of his speci- mens (Figs. 3A-F, I-K, 4, 5). Hubeinella formosa Zhang (1980a, pl. 3, figs. 1-5; POJETA: CAMBRIAN PELECYPODA 161 1980b pl. 1, figs. 20-21; Fig. 3A-F herein), the type and only known species of Hubeinella, has a near teardrop shape and prominent comarginal ornament similar to stenothecoids (Horny, 1957; Koneva, 1976, 1979). However, such a shape also occurs in obolellid inarticulate brachiopods. Some specimens of Hubeinella (Fig. 3B, D, E) show faint radial external ornament, a feature that occurs in obolellids but is absent in stenothecoids (Yochelson, 1969). Unfortunately, nothing is known of the internal fea- tures of Hubeinella. Of Xianfengoconcha Zhang, | was able to restudy and photograph three specimens of X. minuta Zhang (1980a, pl. 2, figs. 13-15; 1980b, Plate 1, figs. 13, 14; Figs. 3I-K, 4A herein). All three specimens are poorly pre- served molds that show nothing beyond an inequilateral shape and faint coarse and fine comarginal ornament. X. minuta is one of three species placed in Xianfengoconcha by Zhang; his figures of X. elliptica Zhang, type species of the genus, and X. rotunda Zhang are no more informative than the specimens of X. minuta. Zhang (1980a:10) provided a drawing of the hinge of X. elliptica that shows an elongate toothlike structure on each side of the beak. Such structures occur in obolellid brachiopods (Rowell, 1965:292-293) including Lamel- lodonta Vogel (Havlicek and Kriz, 1978). In any case, these teeth do not show well on Zhang’s plate, and most of the specimens of X. elliptica and X. rotunda figured by him are essentially equilateral in shape. The most that can be said about Xianfengoconcha is that the known specimens are of a bivalved creature that shows no diagnostic pelecypod features, and which proba- bly is an inarticulate brachiopod. I have been able to examine and photograph four specimens (including the holotype) of Praelamellodonta elegansa Zhang, the type species of Praelamellodonta (Zhang, 1980a, pl. 1, figs. 1, 3, 8, 12, 13; 1980b, pl. 1, figs. 1, 3, 5; Fig. 4B-H herein). The specimens range from a composite mold to one that is mostly covered with shell. The shape of the specimens varies as does the strength of the posterior umbonal slope, both of which suggest defor- mation of the original shape; the specimen shown in Figure 4H is clearly deformed. The holotype of Praelamellodonta elegansa (Fig. 4D, E) has a prominent beak to either side of which is a bladelike structure, and at least three specimens preserve some radial ornament (Fig. 4C, G, H). The bladelike structures to either side of the beak and the presence of radial ornament suggest that the specimens of P. elegansa are distorted inarticulate brachiopods. Thus, restudy of Zhang’s specimens of Hubeinella, Xianfengoconcha, and Praelamellodonta reinforces the suggestion of Geyer and Streng (1998) that they are distort- ed valves of inarticulate brachiopods. In addition to the above three genera, Zhang described the genus Cycloconchoides (Fig. SA-D). He defined two species, C. venustus Zhang, the type species of the genus, and C. elongatus Zhang. I have examined and photographed four specimens including the holotypes of both species. All the known specimens of Cycloconchoides pre- serve the shell and have a gross ornament of comarginal growth lines and radial ribs. No internal features are known. The growth lines are prominently raised and seem to truncate the radial ribs before the next set of ribs is formed. The general appearance of the specimens is that of a series of stacked shells such as occur in some arthropods that do not completely molt the old carapace. This is a prominent feature of conchostracan branchiopods some of which have radial ribs between the growth increments (Tasch, 1969). This type of ornament is not common in pelecypods and is not known to occur in any Cambrian or Ordovician species. However, branchiopods are not known from Lower Paleozoic rocks, and most branchiopods occur in fresh or brackish waters. Other bivalved arthropods are known from the Cambrian, such as Canadaspis perfecta (Walcott) (Conway Morris ef al., 1982); although Canadaspis does not stack its molts. Determining the affinities of Cycloconchoides will depend on finding new material that shows diagnostic features; specimens in which the shell is partially decorticated would be the most useful. It is unlikely that Cycloconchoides was either a pelecypod or a brachiopod. MIDDLE AND UPPER CAMBRIAN TAXA Recently, Middle Cambrian species of Pojetaia were described from Bornholm, Denmark (Hinz- Schallreuter, 1995) and Morocco (Geyer and Streng, 1998) (Fig. 2). Geyer and Streng (1998) described the genus Arhouriella from the early Middle Cambrian Jbel Wawrmast Formation of Morocco (Fig. 6C). A. opheodon- toides Geyer and Streng is known from two silicified incomplete presumed right valves. The species is roughly equilateral. Except for the hinge, internal features are poor- ly preserved; the authors (Geyer and Streng, 1998:93) noted that: “*...the presence of a faint imprint interpreted as posterior adductor muscle scar [called anterior muscle scar in their fig. 7H]...Faint pallial [sic] commences at adductor muscle scar.” These features are not readily apparent on their photographic figures and are not included on their reconstruction. One specimen of A. opheodontoides preserves a well-developed slightly concave hinge plate and some of the presumed posterior ligament area, which partly covers 162 AMER. MALAC. BULL. 15(2) (2000) Fig. 4. Probable Lower Cambrian brachiopods from China that have been described as pelecypods. All bar scales represent 1 mm. Xianfengoconcha minuta Zhang (A) and Praelamellodonta elegansa Zhang (B-H). A, enlargement of apical area of specimen shown on Fig. 31. B, C, paratype, C shows some radial ribbing, figured by Zhang (1980a, pl. 1, fig. 8). D, E, holotype, E is an enlargement of the apical area, arrows point to bladelike structures to either side of the apex, figured by Zhang (1980a, pl. 1, fig. 1). F, G, paratype, G shows some radial ribbing, figured by Zhang (1980a, pl. 1, figs. 12, 13). H, distorted paratype showing radial ribbing, figured by Zhang (1980a, pl. 2, fig. 3). the hinge plate (Fig. 6C). The hinge plate undercuts the two pronglike central teeth. Anterior to the teeth there is an elongate fossette that Geyer and Streng interpret as an addi- tional ligament area; they reckon that the ligament extended over the entire hinge plate in an early ontogenetic stage. The anterior of the two teeth originates directly on the dor- sal margin and not on the hinge plate. If Arhouriella opheodontoides is a pelecypod, and if the amphidetic interpretation of its ligament is correct, it is clearly very different from Fordilla and Pojetaia which have opisthodetic ligaments. John C. W. Cope (pers. comm., September 1999) suggested that the teeth of Arhouriella opheodontoides could best be compared to those of the Upper Cretaceous solemyoid Nucinella sohli Pojeta (1988, pl. 5). Leaving aside the fact that the oldest known nucinellid is Early Jurassic in age, a major dentition difference between the two species is that in N. sohli the peglike teeth are mounted POJETA: CAMBRIAN PELECYPODA 163 Fig. 5. Incertae sedis Lower Cambrian bivalves from China; possible arthropods. Cycloconchoides elongatus Zhang (A), and C. venustus Zhang (B-D). All scale bars represent | mm. A, holotype, figured by Zhang (1980a, pl. 3, fig. 14). B, paratype, figured by Zhang (1980a, pl. 3, fig. 8). C, paratype, figured by Zhang (1980a, pl. 3, fig. 9). D, holotype, figured by Zhang (1980a, pl. 3, fig. 6). directly on the hinge plate, and, as in other solemyoids, N. sohii is anteriorly elongated. Clearly, Geyer and Streng have a very unusual bivalve and additional specimens will help determine its systematic position. Tuarangia paparua MacKinnon (1982) was first described from the Middle Cambrian Tasman Formation of New Zealand; the species is based on at least one hundred specimens. Subsequently, Berg-Madsen (1987) described T. gravgaerdensis from the Middle Cambrian Andrarum Limestone of Bornholm, Denmark. Tuarangia (Fig. 6A, D-E) has been the subject of much discussion as to whether or not it is a pelecypod, and, if it is, how it is related to other pelecypods. MacKinnon (1982) placed Tuarangia in the Pteriomorphia with ques- tion, even though the genus possesses numerous taxodont teeth. He interpreted its shell microstructure as foliated cal- cite, which he noted in pelecypods occurs only in the Pteriomorphia. He also noted that the genus has an amphidetic ligament that shows as a raised ridge, between the two rows of taxodont teeth on internal molds. Runnegar (1983) regarded Tuarangia as being relat- ed to the “quasirostroconch” Pseudomyona queenslandica (Runnegar and Jell), because both have foliated calcite shell microstructure, and he interpreted the presumed amphidetic ligament of Tuarangia as representing the place where a univalved protoconch had broken off from the rest of the shell. Pseudomyona is known to have a univalved protoconch and a bivalved adult shell. Runnegar thought of Tuarangia and Pseudomyona as bivalved monopla- cophorans. Carter (1990:179) noted: “The order Tuarangioida is, therefore, presently regarded as a mono- placophoran derivative which is convergent toward the Bivalvia.” Waller (1998:10) regarded Tuarangia as being poorly known and noted its various taxonomic placements. With conviction, MacKinnon (1985) placed Tuarangia firmly in the Pteriomorphia and disassociated it from Pseudomyona by noting that the middorsal ridge on internal molds was the underside of an amphidetic liga- ment. The ridge between the rows of taxodont teeth is pre- served in the same way on all known specimens, and he noted that there is no sign that it represents a broken-off 164 AMER. MALAC Fig. 6. Middle and Upper Cambrian pelecypods. Tuarangia gravgaerden- sis tenuiumbonata (A), Camya asy (B), Arhouriella opheodontoides (C), and Tuarangia paparua (D, E). A, reconstruction of muscle scars from Hinz-Schallreuter (1995:75); VS = anterior adductor, HS = posterior adductor, PM = pallial line. B, drawing of shape and dentition from Hinz- Schallreuter (1995:75). C, drawing of hinge from Geyer and Streng (1998:93); af = anterior fossette, u = umbo, pl = posterior ligament area, hp = hinge plate. D, E, dorsal and left lateral views, specimen 0.9 mm long. Figures A and B reproduced with permission of Geschiebekunde aktuell. Figure C reproduced with permission of Revista Espanola de Paleontologia. structure. Berg-Madsen (1987) maintained the placement of Tuarangia in the Pteriomorphia, but with reservations. She also figured Tuarangia sp. from an erratic boulder from Poland; the boulder contained Upper Cambrian conodonts, which indicate that Tuarangia ranged into lower Upper . BULL. 15(2) (2000) Cambrian. Runnegar and Pojeta (1992) maintained the close association of Tuarangia and Pseudomyona, based on similar shell microstructure, and separated them from the lineage including Fordilla and Pojetaia. Hinz-Schallreuter (1995) described a pelecypod faunule from the Middle Cambrian Exsulans Limestone of Bornholm, Denmark. The faunule contains Pojetaia ost- seensis, Tuarangia gravgaerdensis tenuiumbonata, and Camya asy, all named by Hinz-Schallreuter in her 1995 paper. In discussing Tuarangia gravgaerdensis tenuium- bonata, Hinz-Schallreuter noted that she had complete shells showing both the interior and exterior and that they have a space for an amphidetic ligament as interpreted by MacKinnon (1982, 1985). Her description of this species noted the presence of pallial muscles and an anterior adduc- tor muscle (Fig. 6A). These features do not show well where she notes their presence on her figure 5.3. The mus- cles are shown on her reconstruction (Hinz-Schallreuter, 1995:75) where the position of the anterior adductor muscle scar is in black and queried. She indicated the possible position of the posterior adductor muscle with a dashed open circle within which is a question mark; in her descrip- tion she noted that the posterior adductor muscle was not preserved on her material. If Hinz-Schallreuter’s interpre- tation of the musculature of Tuarangia proves correct, there is no doubt that it is a Cambrian pelecypod and is not relat- ed to Pseudomyona, because that genus has a single sub- central muscle. Cope and Babin (1999) accepted Hinz- Schallreuter’s reconstruction of Tuarangia. Hinz-Schallreuter also defined the Middle Cambrian pelecypod genus Camya, type species C. asy. This species (Fig. 6B) is like Tuarangia in its anterior-pos- terior elongation and long dorsal margin, but it is more con- stricted anteriorly; the beaks are in an anterior position, and there is a single tooth to either side of the beak. The species is known from two left valves that show comarginal growth lines. Other features of the species are unknown. Babin (1993) has shown that the supposed Middle Cambrian pelecypod from Belgium, Modiolopsis ? malaisii Fraipont, 1910, is a pseudofossil. Pojeta (1980:77) figured an unnamed possible Late Cambrian pelecypod from the Frederick Limestone of Maryland. This specimen shows only shell shape; no other specimens have been found, and its status as a pelecypod remains dubious. ACKNOWLEDGMENTS I thank Zhang Renjie for arranging for my study of the Hubei specimens. My thanks to David MacKinnon for arranging for me to photo- graph the specimen of Tuarangia paparua and to John Peel for the loan of POJETA: CAMBRIAN PELECYPODA 165 the specimen of Fordilla from Greenland. I congratulate the organizers of the symposium for bringing together paleomalacologists and neomalacolo- gists in a common forum. LITERATURE CITED Babin, C. 1993. A propos d’un pretendu mollusque bivalve du Cambrien de Belgique. Annales de la Societe geologique de Belgique 116:13-14. Barrande, J. 1881. Systéme Silurien du centre Bohéme. v.6, Acéphalés. Paris and Prague, 342. Berg-Madsen, V. 1987. Tuarangia from Bornholm (Denmark) and simi- larities in Baltoscandian and Australasian late Middle Cambrian faunas. Alcheringa 11:245-259. Carter, J. G. 1990. Evolutionary significance of shell microstructure in the Palaeotaxodonta, Pteriomorphia, and Isofilibranchia (Bivalvia: Mollusca), In: Skeletal Biomineralization: Patterns, Processes, and Evolutionary Trends, v. 1, J. G. Carter, ed. pp. 135-296. Van Nostrand Reinhold, New York. Conway Morris, S., H. B. Whittington, D. E. G. Briggs, C. P. Hughes, and D. L. Bruton. 1982. Atlas of the Burgess Shale. Palaeontological Association. pp. 1-32. Cope, J. C. W. 1996. The early evolution of the Bivalvia. Jn: Origin and evolutionary radiation of the Mollusca, J. D. Taylor, ed. pp. 361- 370. Oxford University Press. Cope, J. C. W. and C. Babin. 1999. Diversification of bivalves in the Ordovician. Geobios 32:175-185. Delgado, J. F. N. 1904. Faune Cambrienne du Haut-Alemtejo (Portugal). Communicacoes da Commissao do Servico Geologico de Portugal 5:307-374. Elicki, O. 1994. Lower Cambrian carbonates from eastern Germany. Neues Jahrbuch fiir Geologie und Palaontologie 191:69-93. Fraipont, C. 1910. Modiolopsis ?? Malaisii Ch. Fraip. Lamellibranche nouveau du Revinien Belge (Cambrien moyen). Annales Société Géologique Belgique 37:MS-8. Geyer, G. 1998. Intercontinental, trilobite-based correlation of the Moroccan early Middle Cambrian. Canadian Journal of Earth Sciences 35:374-401. Geyer, G. and M. Streng. 1998. Middle Cambrian pelecypods from the Anti-Atlas, Morocco. Revista Espanola de Paleontologia, No extrordinario, Homenaje al Prof. Gonzalo Vidal. pp. 83-96. Havlicek, V. and J. Kriz. 1978. Middle Cambrian Lamellodonta simplex Vogel: ‘bivalve’ turned brachiopod Trematobolus simplex (Vogel). Journal of Paleontology 52:972-975. Hinz-Schallreuter, I. 1995. Muscheln (Pelecypoda) aus dem Mittelkanbrium von Bornholm. Geschiebekunde Aktuell, Mitteilungen der Gesellschaft fiir Geschlebekunde 11:71-84. Horny, R. 1957. Problematic molluscs (?Amphineura) from the Lower Cambrian of south and east Siberia (U.S.S.R.). Sbornik Ustredniho Ustavu Geologickeho 1956 23:397-432. [English Summary 423-432] Jell, P. 1980. Earliest known pelecypod on Earth—a new Early Cambrian genus from South Australia. Alcheringa 4:233-239. Jermak, V. 1986. Early Cambrian Fordillidae (Bivalvia) from the northern Siberian Platform. Trudy Sibirskoe Otdelenie Institut Geoiogogti i Geophysiki Akademiya Nauk SSSR 669: 183-188. [In Russian. ] Koneva, C. 1976. New Members of the Class Stenothecoida from the Lower Cambrian of central Kazakhstan. Paleontologicheskiy Zhurnal 10:125-128. [In Russian] Koneva, C. 1979. Stenothecoids and inarticulate brachiopods from the Lower and Middle Cambrian of Central Kazakhstan. Academy of Sciences, Kazakhstan SSR, Satpaev Institute of Geological Sciences, Nauka Kazakhstan SSR, Alma-Ata. pp. 123. [In Russian] Krasilova, I. 1977. Fordillidae (Bivalvia) from the lower Paleozoic of the Siberian Platform. Paleontologicheskiy Zhurnal 11:42-48. [In Russian] Krasilova, I. 1987. The oldest representatives of the bivalve mollusks. Paleontologicheskiy Zhurnal 21:24-30. [In Russian] Landing, E, S. A. Bowring, K. L. Davidek, S. R. Westrop, G. Geyer, and W. Heldmaier. 1998. Duration of the Early Cambrian: U-Pb ages of volcanic ashes from Avalon and Gondwana. Canadian Journal of Earth Sciences 35:329-338. Landing, E. and S. R. Westrop. 1998. Cambrian faunal sequence and depositional history of Avalonian Newfoundland and New Brunswick. Bulletin New York State Museum 492:5-70. MacKinnon, D. 1982. Tuarangia paparua n. gen. and n. sp., a late Middle Cambrian pelecypod from New Zealand. Journal of Paleontology 56:589-598. MacKinnon, D. 1985. New Zealand late Middle Cambrian molluscs and the origin of Rostroconchia and Bivalvia. Alcheringa 9:65-81. Pojeta, J. 1975. Fordilla troyensis Barrande and early pelecypod phyloge- ny. Bulletins of American Paleontology 67:363-384. Pojeta, J. 1978. The origin and early taxonomic diversification of pelecy- pods. Philosophical Transactions of the Royal Society of London B284:225-246. Pojeta, J. 1980. Molluscan Phylogeny. Tulane University Studies in Geology and Paleontology 16:55-80. Pojeta, J., 1988. The origin and Paleozoic diversification of solemyoid pelecypods. Memoir New Mexico Bureau of Mines and Mineral Resources 44:201-271. Pojeta, J. and B. Runnegar. 1985. The early evolution of diasome mol- luscs. In: The Mollusca, v. 10, Evolution, E. R. Trueman and M. R. Clarke, eds. pp. 295-336. Academic Press, Orlando. Rowell, A. 1965. Obolellida. In: Treatise on Invertebrate Paleontology, part H, Brachiopoda I. R. C. Moore, ed. pp. 292-293. Geological Society of America and University of Kansas, New York and Lawrence. Runnegar, B. 1983. Molluscan phylogeny revisited. Memoir Association of Australasian Palaeontologists 1:121-144. Runnegar, B. and C. Bentley. 1983. Anatomy, ecology, and affinities of the Australian Early Cambrian bivalve Pojetaia runnegari Jell. Journal of Paleontology 57:73-92. Runnegar, B. and P. A. Jell 1976. Australian Middle Cambrian molluscs and their bearing on early molluscan evolution. Alcheringa 1:109- 138. Runnegar, B. and J. Pojeta. 1992. The earliest bivalves and their Ordovician descendants. American Malacological Bulletin 9:117- 122. Tasch, P. 1969. Branchiopoda. Jn: Treatise on Invertebrate Paleontology, part R, Arthropoda 4, R. C. Moore, ed. pp. 128-191. Geological Society of America and University of Kansas, Boulder and Lawrence. Teixeira, C. 1952. La faune Cambrienne de vila boin au Portugal. Boletim da Sociedade Geologica de Portugal 10:169-188. Vogel, K. 1962. Muscheln mit Schlosszahnen aus dem spanischen Kambrium und ihre bedeutung fiir die Evolution der Lamellibranchiaten. Akademie der Wissenschaften und der Literatur in Mainz, Abhandlungen der Mathematisch- Naturwissenschaftlichen Klasse 4:197-244. Waller, T. R. 1998. Origin of the molluscan Class Bivalvia and a Phylogeny of major groups, /n: Bivalves: An eon of evolution, P. A. Johnston and J. W. Haggart, eds. pp. 1-45. University of Calgary Press, Calgary, Canada. 166 AMER. MALAC. BULL. 15(2) (2000) Yochelson, E. 1969. Stenothecoida, a proposed new class of Cambrian Mollusca. Lethaia 2:49-62. Yochelson, E. 1978. An alternative approach to the interpretation of the phylogeny of ancient mollusks. Malacologia 17:165-191. Yochelson, E. 1981. Fordilla troyensis Barrande: “The oldest known pele- cypod” may not be a pelecypod. Journal of Paleontology 55:113- 125. Zhang, R. 1979. Bivalves from Lower Cambrian Tianheban Formation in Xianfeng of Hubei. Abstracts 12th Annual Conference and 3rd National Congress, Palaeontological Society of China. pp. 24. [In Chinese] Zhang, R. 1980a. On the earliest bivalve fauna—bivalves from Lower Cambrian Tianheban Formation, Xianfeng, Hubei. Bulletin Chinese Academy of Geological Sciences, Series 8, 1:1-19. [In Chinese] Zhang, R. 1980b. On the earliest bivalv[e] fauna—bivalves from Lower Cambrian Tianheban Formation, Xianfeng, Hubei. 26th International Geological Congress 4:121-129. [In Chinese] Date of manuscript acceptance: 8 May 2000 Ordovician to Devonian diversification of the Bivalvia Claude Babin UFR Sciences de la Terre, Université Claude Bernard-Lyon I, France Abstract: Studies of the Paleozoic Bivalvia have accelerated in the last three decades and we have numerous new data dealing with the diversification of this Class of mollusks. After the extreme scarcity of the Cambrian data, the abruptness of the diversity of the bivalve faunas in the fossil record during the Early Ordovician is an outstanding event. It is also noteworthy that both this first explosion and the succeeding diversification during the Middle Ordovician were located primarily on the Gondwanan and Avalonian shelves in shallow clastic facies. By contrast, from the late Ordovician, Baltica and Laurentia were more propitious for the further diversifications of bivalves, notably epibenthic ones. Thus before the close of the Period, all the subclasses of bivalves were established and the Class was dispersed throughout the world oceans. After the uppermost Ordovician extinctions, an important replacement at familial and generic levels occurred during the Silurian. The Pteriomorphia, many of them adapted to an epibyssate mode of life, underwent an explosive evolution, par- ticularly during the Ludlow, while many free-burrowing suspension-feeding genera were adapted to the broad expanses of soft muds. During the Devonian, bivalves continued their diversification at both familial and generic levels, and for the first time, some of them colonized fresh-water. The continuing paleo- geographic changes favored faunal exchanges, for example between Appalachian and western European areas, and led to the establishment of cosmopolitan faunas. Key Words: bivalves, diversification, Paleozoic, Ordovician, Silurian, Devonian “The evolutionary radiation of the bivalves is one of Families the success stories of invertebrate evolution; few other 100 marine taxa have shown such a steady and consistent increase in diversity” (Gould, 1977:273). At present, the 80 Class Bivalvia constitutes a major part of modern marine benthic faunas even without consideration of the freshwater ss representatives. This diversity of the bivalves (about 9000 40 living species according to Nicol, 1989) is the successful conclusion of a long history (Fig. 1), and it is interesting to 20 attempt to track the key stages marking the beginning of this odyssey, especially the initial diversifications of the Class. The difficulties for such a process are numerous and any reconstruction is largely provisional. However in the three last decades, numerous new investigations have brought important data dealing with the Lower Paleozoic Fig. 1. The continuous increase of the diversity of the bivalves during the Phanerozoic. (after Paul, 1989). faunas of bivalves. Nevertheless, to estimate the value of (Babin, 1993a, 1995) I used genera and regrouped them at the study of early diversifications we have to enumerate the subclass level. I think that the results were conclusive main sources of difficulty. enough but undoubtedly that procedure is not without prob- lems. In the literature, the genus concept of authors is MAIN DIFFICULTIES OF THE STUDY probably only a little less insecure than the species one. The latter is the result of diverse splittings and regroupings. ¢ Systematics. Identifying the times of diversifica- For instance, comparing Devonian faunas from Europe and tion requires expressing the succession of taxonomic diver- North America, Bailey (1983) proposed a regrouping of 99 sities. For that, it is fitting to choose adequate taxonomic fossil species into only 16 species. Nevertheless, the notion levels, obviously a subjective process. In previous papers of the genus is also sometimes the subject of dramatic American Malacological Bulletin, Vol. 15(2) (2000):167-178 167 168 AMER. MALAC. BULL. 15(2) (2000) changes. For instance, two classical genera in the old litera- ture, the “praecardioids” Cardiola from the Silurian and Buchiola from the middle and upper Devonian have been split, the former into twelve genera (Kriz, 1979a), and the latter into six genera (Grimm, 1998). Regarding the subclasses, there are also some dif- ferences of opinion amongst authors. The neontologists generally recognize two subclasses of Bivalvia, Protobranchia and Autobranchia, while paleontologists, basing their classification on the dentition, make use of more subclasses, considering them as evolutionary units. The latter are regarded rather as superorders in the former classification. Pojeta (1987) uses five subclasses: Palaeotaxodonta, Isofilibranchia, Heteroconchia, Pteriomorphia and Anomalodesmata. Cope (1995) defines seven subclasses: Palaeotaxodonta, Lipodonta (for the sin- gle solemyoids included in the Cryptodonta by Newell, 1969 and in the Palaeotaxodonta by Pojeta, 1988), Palaeoheterodonta (including the mytiloids which are Isofilibranchia for Pojeta), Heterodonta (these two units are included in Heteroconchia by Pojeta), Neotaxodonta, Pteriomorphia and Anomalodesmata. We retained Cope’s classification in a recent paper about the Ordovician diver- sification (Cope and Babin, 1999); nevertheless, Cope (1998) recently proposed that “some long-accepted palaeo- heterodonts are in fact heterodonts, and that even in the early Ordovician heterodonts were already present.” In these works all extant bivalve subclasses were represented from the early Ordovician. However, another subclass, never mentioned in all these papers, is the Cryptodonta, which included (Newell, 1969) two orders, solemyoids and praecardioids. The former are possibly paleotaxodonts or lipodonts (see above) but the latter raise a true problem. They were abundant during the Paleozoic, particularly dur- ing the Silurian; Kriz (1965, 1979) has put a large part of them into the isofilibranchs and the pteriomorph arcoids (which would be neotaxodonts for Cope) but he has lately kept Cryptodonta for the praecardiids and antipleurids (Kriz, 1996). More recently, Johnston and Collom (1998) have used afresh the subclass Cryptodonta for numerous bivalves. It is not an unimportant choice regarding the evo- lution of the bivalves because Cryptodonta, if they include only the praecardioids, would be the single subclass of Bivalvia without modern representatives. Last but not the least, it is evident that these prob- lems of classification also reflect different conceptions about the phylogeny of the bivalves (see for instance, Cope, 1995; Waller, 1998). Notwithstanding their importance, these points are beyond the present enquiry and they are not really constraining for my purpose. So, despite these unde- niable divergences and difficulties, it remains possible to appraise the properties of the initial bivalve diversifica- tions. ¢ Fossil determinations correspond to another prob- lem of taxonomy. Paleontological material is often of poor preservation. Paleozoic bivalve fossils are often poor molds showing few characters without the dentition for instance. So many determinations have been based solely on shell shape explaining many “wastebasket” taxa; in the older lit- erature genera such as Ctenodonta or Modiomorpha can appear to have had a very large spatio-temporal distribu- tion. It is another cause of weakness for the inventory in addition to the splitting or regrouping cited above. ¢ Field investigations and study of the new material are of course among the most important factors initiating change in diversity estimates. The comparison of recent graphs of the Ordovician bivalve diversification, for exam- ple, shows an important amplification of the generic num- bers for South Gondwana and Avalonia (Fig. 2). It is the result of new data from Argentina on the one hand (Sanchez, 1999), and from Wales on the other (Cope, 1996). Another example may be given. There is no paleo- taxodont in my database for the middle Ordovician from Baltica. However, in a recent paper Dzik (1994, Fig. 24, A- C) figured a minute specimen of a juvenile paleotaxodont bivalve (Praenucula?) from the lower Llanvirn of Baltica. ¢ Chronology. Identifying the times of diversifica- tion requires having a sufficient resolution of the geologic time scale and precise chronologic correlations between the different parts of the Earth. The stratigraphic information concerning the lower Paleozoic is indeed various and of irregular quality according to the regions. In addition there are several regional scales and they often remain without clear correlation. In this way, I have underlined (Babin, 1995), for example, the equivocal use of middle Ordovician in the North American literature compared with the stan- dard scale of Great Britain. The American Blackriveran stage, containing bivalves, considered as middle Ordovician corresponds really to the lower part of the Caradoc Series from the upper Ordovician in the standard scale. On that basis, we have not known until now bivalves from the middle Ordovician in North America. Elsewhere, in regions where stratigraphical investi- gations are incomplete, we can sometimes read Ordovician, Silurian or Devonian without more precision and we are hazy about the times. In the Treatise on Invertebrate Paleontology (1969), such vagueness is common enough too, even though it has been the reference for several pub- lished compilations. To establish the database one must evidently consult the old fundamental publications of the last century. Unfortunately, the stratigraphical location of the fossils, which are described and figured, is usually imprecise. Barrande (1881) published the volume VI of his monumen- tal “Systéme silurien du centre de la Bohéme.” Devoted to BABIN: ORDOVICIAN TO DEVONIAN BIVALVIA [_] Palaeotaxodonta WA, Lipodonta Palaeoheterodonta Neotaxodonta ‘fae Pteriomorphia Anomalodesmata 169 Genera 24 21 18 15 BY x 12 ‘ ~ ‘ x ‘ ‘ ay ~ x 7 BY PA COS » et a AY s,s NNN NN SSS NSN SSNS NN NS ff # #4 4 # # EO MO LO. EO MOLO £0 MO LO SOUTHERN NORTHERN GONDWANA GONDWANA al Genera 24 21 18 15 Pr a A A a a A 12 ff # £4 ee Fe .N NNN NNN sy NNN N NNN SNS NSS SNS SN SN Pr a a a a a 2 a a a A A A i‘ EO MO LO EO MO LO EO MO LO SOUTHERN NORTHERN GONDWANA GONDWANA AuenOTn EO = Early Ordovician (Tremadoc + Arenig) MO = Middle Ordovician (Llanvim + Llandeilo) LO = Late Ordovician (Caradoc + Ashgill) Fig. 2. Comparison of two successive compilations of data concerning the Ordovician bivalves. It shows the effect of the recent investigations in the field. Left: data from 1993 (after Babin). Right: data from 1999. the “Acéphalés” (i.e. to the bivalves), it comprises 342 pages and 361 plates to describe 1269 species distributed among 58 genera, of which 20 are new. However, they are not all from the Silurian, as the title mentions, but they are distributed from the lower Ordovician (Llanvirn) to the middle Devonian (Givetian) because Barrande always used Murchison’s original definition (1835) of the Silurian Period. Thus, to utilize Barrande’s data, one needs to refer to more recent documents of Czech geologists for the tem- poral correlations. Concerning the Devonian, the important books of Hall (1884, 1885) with the description of 69 gen- era, of which 49 were established in North America, and of Beushausen (1895) with the description of 46 genera from the Rhineland, excluding the pteriomorphs studied by Frech (1891), are easier to consult regarding the stratigraphical assignment. Nevertheless, the utilization of the data for the database requires the establishing of synonymies between all these works. 170 AMER. MALAC. BULL. 15(2) (2000) These non-exhaustive remarks go to show that every database is imperfect, provisional and could be con- tinually improved. ¢ Paleogeography. To locate the paleodiversifica- tions in space we have at our disposal numerous and more or less different paleogeographic reconstructions. Despite great strides over three decades in this field following the plate tectonics model, we know that these reconstructions remain difficult for the Paleozoic. However, a general enough agreement exists today about the broad outlines of the relative position of the large paleocontinents and it is sufficient for the present matter. Taking all these restrictions into consideration, I will consider the available data to draw the main events regarding bivalve paleodiversifications from the Ordovician to the Devonian. Nevertheless, before examin- ing the rough data, diversifications need some considera- tion. Actually, when the radiations are put in evidence, the question remains of their explanation. Two categories of causes need consideration. Are the diversifications the result of intrinsic factors or of inductive extrinsic condi- tions? The former may particularly correspond to key inno- vations and it is understandable that many new structures, innovative ones, are not preserved in the fossil record. The interpretation of the material, more often than not hypothet- ical, will be an inescapable way to reconstruct the soft parts. As for the extrinsic factors, some of them are given by the sediments like the grain-size or some information regarding the bathymetry, the oxygenation, etc., but many other factors will remain hypothetical too. THE GREAT ORDOVICIAN RADIATION The need to look at the Ordovician diversification has been renewed for some years because this Period encompasses one of the most important evolutionary radia- tions. This is particularly clear regarding the bivalves, which offer in the fossil record a true explosion during the Arenig (late Lower Ordovician). In reality, there is a very poor record and important gaps in the story of Bivalvia before the Arenig and one may consider that this apparent explosive diversification is an artefact. The extremely scarce representatives of Cambrian bivalves, small shells of perhaps meiofauna and “the abysmally poor record of bivalves from the Tremadoc Series (early Lower Ordovician) of the Ordovician” (Cope, 1995) suggest such an artefact. Diverse reasons can be offered to explain this lack of fossil bivalves, such as their particular fragility, unfavorable environmental conditions for their preserva- tion, or geodynamical causes such as the rifting of the near- shore shelves that they inhabited, or even deficiency of the investigations, among others. No one consideration is really convincing. There are fossils of other marine benthic groups during the upper Cambrian and Tremadoc. Finally, as the Ordovician was a period of important radiation for many groups, it is not surprising to observe the same for bivalves. However, its explosive aspect with the simultane- ous appearance of representatives of each subclass is very peculiar. Unfortunately, the paucity of information before the Arenig remains a serious handicap to understanding the temporal relations between the subclasses. Considering the known fossil record, it is noticeable that these Arenig bivalve faunas were located on the peri- Gondwanan shelves (Babin, 1993b), Avalonia being con- sidered as a neighboring area (Fig. 3). Possible reasons to explain this spatial restriction have been considered (Babin, 1995, Cope and Babin, 1999). Clearly the temperature of the sea water was not a deciding factor since the faunas were settled from very high latitudes (Morocco, Montagne Noire) to very low latitudes (Argentina, Australia), in spite of some obvious preferences such as the deposit-feeding paleotaxodonts, for instance, which are more numerous and varied in the high latitudes. On the other hand, a condition shared by the whole peri-Gondwanan shelves is siliciclastic sedimentation, and it is clear that all these first bivalves inhabited clastic sea-beds wherein they had infaunal or semi-infaunal modes of life. This relationship was probably linked with food supply because terrigenous sediments are considered more favorable from this point of view. Moreover, another distinctive feature of the Early Ordovician bivalve settings is their shallowness. The deep- § 1 genus Fig. 3. Peri-Gondwanan location of the bivalves during the late Tremadoc- Arenig, south polar view. BABIN: ORDOVICIAN TO DEVONIAN BIVALVIA 171 est environments were perhaps those of the Montagne Noire (France), but it was certainly less than fifty meters in depth. So, it is clear that the early bivalve diversification was globally located in relatively shallow waters and is in accordance with Jablonski and Bottjer’s model (1990) regarding the role of the onshore settings for the appear- ance of the major groups. In these environments, the empty ecospace hypothesis (Erwin, 1994) may help explain the burst of the endobenthic and semi-infaunal bivalves, which were the first representatives of shelly burrowers. Finally, it seems that this preference for siliciclastic sedimentation may have been the restrictive factor against a geographic expansion towards other plate margins like Baltica or Laurentia [the recent assertion of Droser et al. (1995) and of Droser et al. (1996) that “bivalves occur in Ibexian strata” in shell beds is probably erroneous. The old- est North American described bivalves are from the upper- most Whiterockian or very lowest Blackriveran (pers. comm. of R. Ross, 4 Oct. 1999)]. It is also plausible that sheer ocean width was a barrier to the migration of the planktic larvae. Bivalve larvae live about six weeks before settlement; however Baltica, for instance, was not so far from Gondwana-Avalonia (Fig. 3). Prevailing surface cur- rents could have been unfavorable but other benthic Arenig mollusks like rostroconchs were nevertheless cosmopolitan. The other aspect to consider about the initial diver- sification concerns the possible impact of some intrinsic factor. I have said that it is difficult with the fossil material, that is with the hard parts (shells) only, to discern one or several key characters that induced the adaptive radiation. Stanley (1975, 1977) suggested that “the delayed radiation (since the first Cambrian bivalves) was triggered by the ori- gin of the byssus” and that “retention of the byssus into the adult stage was an important aspect of the initial adaptive radiation of the class.” However, this consideration seems based on North American bivalves with emphasis on the upper Ordovician faunas with pteriomorphs. Tevesz and McCall (1976, 1985) who proposed an epifaunal mode of life for Cambrian bivalves, suggested “that the Ordovician expansion of pelecypods is due to their invasion of the infauna habitat” and “that the delayed radiation of pelecy- pods was due to the Ordovician evolution of some other non-preserved structure that promoted survival of the group, such as, for example, some essential pedal muscula- ture or a more efficient circulatory system.” One fundamental anatomical change could have been realized with the development of gills involved in the feeding process, that is with the appearance of the fili- branch gill (Yonge 1947). Waller (1988) suggested that the Tironucula group, small paleotaxodonts from the Llanvirn (Middle Ordovician; see below), “probably had already evolved filibranchiate gills”, and since 1995 Cope has cor- Spain. This paleotaxodont shows two separate series of teeth anterior and posterior; Cope (1995) assumed that this arrangement could be functional- ly related with filibranch gills. x6. related this change of the gills with the early diversification of the bivalves. Cope also looked for paleotaxodonts that could have acquired this important novelty. By analysis of the dentition allowing rapid fluttering of the valves to void the pseudofaeces, Cope (1995) proposed that it could have been the Cardiolaria group (Fig. 4). The advantages and changes of the mode of life resulting from this key innova- tion have been discussed elsewhere (Cope and Babin, 1999). However that may be, it appears that the Arenig bivalve faunas, located in Gondwana and Avalonia, con- tained representatives of all subclasses except Cryptodonta, if that is restricted to the praecardioids. The causes of this explosive radiation remain unclear, probably an interfer- ence of intrinsic and extrinsic factors, but the first unknown divergences took place probably before the Arenig (five or six genera are known from the latest Tremadoc of Gondwana). During the Middle Ordovician, that is the Llanvirn of the standard stratigraphical scale, bivalve faunas remain located in essentially the same areas, often with a great abundance and variety such as in the Ibero-Armorican mas- sif (Babin and Gutiérrez-Marco, 1991). Seven genera are also known in Avalonia (Cope, 1999). The Llanvirn was a period of transgression, and the flooding of the Gondwana craton also stimulated the burst of other groups like the trilobites. However, the geodynamic activity simultaneous- ly modified the relations between the continental plates; thus, the progressive reduction of the Iapetus Ocean allows some communication between both its sides (Fortey and Cocks, 1988). For the first time we find some scarce bivalves in Baltica with Babinka and in Laurentia (Svalbard) with Tironucula. A large dispersal of the bivalves over an extensive area (Baltica, Laurentia, Siberia) took place during the Upper Ordovician (Fig. 3) probably favored by the mid- Caradoc transgression, perhaps the most extensive in Phanerozoic history. The Chinese Ordovician bivalves probably proceed too from the upper Ordovician. It is no doubt the case for some of the genera in Yunnan cited by V72 AMER. MALAC. BULL. 15(2) (2000) Guo (1985) and assigned possibly to the Arenig-Llanvirn. Then, the dispersal in the earliest Late Ordovician provides bivalves firstly in the inshore siliciclastic environments such as the St. Peter Sandstone (Minnesota) but it corre- sponds rapidly to a second important diversification explor- ing particularly the epibenthic mode of life on the favorable carbonate platforms of these low latitudes. The pteri- omorphs, for instance, at the time rapidly radiate while some modiomorphoids (Modiolopsis, Corallidomus, and Whiteavesia) are the first gregarious genera. Are Cryptodonta also present in the Late Ordovician of Laurentia? A species of Viasta (V. americana) was cited by Fritz (1951) in the Richmondian of Canada but Kriz (1998) considers it is not a Vlasta. Elsewhere, other genera of the Upper Ordovician, such as Shanina from China and Shaninopsis from Sweden are not Vlastidae after Kriz (1998). We do not know precisely whether there were Cryptodonta in the Upper Ordovician. The origin of this subclass is obscure. Simultaneously, in the high latitudes of south Gondwana (Ibero-Armorican area, Bohemia) bivalves became scarce and we do not know the reason. It is long before the cooling of the latest Ordovician and in any case some bivalves continue during the Ashgill in Morocco near the Gondwanan cap-ice. Thus, the major paleodiversification of bivalves took place on the Gondwanan and neighboring Avalonian shelves as early as Late Tremadoc-Early Arenig. The dis- persal and migration during the Late Ordovician favored particularly the radiation of epifaunal bivalves. The first peak of the diversity of the suspension feeding bivalves took place throughout the Upper Ordovician (Bretsky, 1973). It is noticeable also that the sizes of the first bivalve representatives were small or moderate, the largest known form is a Modiolopsis from the Arenig of Wales of almost 50 mm length (Cope, 1996). The increase of size in the Middle and Upper Ordovician affected especially the pteri- omorphs and paleoheterodonts but the largest bivalve known from the Ordovician is the problematic “Vlasta” americana quoted above with a length of about 120 mm (Fritz, 1951). THE SILURIAN - DOWNTURN FOR THE BIVALVES The mass extinction of the latest Ordovician, proba- bly one of the most drastic crises, evidently affected the bivalves; nevertheless all the subclasses crossed the thresh- old of the Silurian. Moreover it is noteworthy that the Bivalvia have stood up to all the mass extinctions (Fig. 1), and their better success than other benthic groups (bra- chiopods for instance) is often explained by this remarkable adaptive resistance to major crises. In a fundamental paper Kriz (1984) gave a quantita- tive overview of changes in bivalve faunas between the Ordovician and Silurian and of the new Silurian radiation of the Class. He worked with data taken from the Treatise on Invertebrate Paleontology (1969) modified by his own taxonomic concepts classifying, for instance, a large part of the Cryptodonta of the Treatise in the pteriomorphs. In spite of the new data gathered since 1984, this analysis is still probably conclusive enough. It shows that about 30% of the Ordovician genera survived into the Silurian; 35% of them were free-burrowers and 58% were endobyssate infaunal and semi-infaunal. Obviously, the epifaunal bivalves were less resistant against this crisis from which the causes (climatic deterioration, regression then trans- gression with anoxic conditions) are probably many and are poorly established. The widespread anoxic conditions per- sisted during the Early Silurian (Llandovery) and it was only a period of survival for the bivalves without clear recovery (Fig. 5). Among the fauna, the paleotaxodonts included several surviving Ordovician genera (about eleven after Liljedahl, 1994), but their diversity had distinctly decreased; they were widely distributed but with a possible preference for the equatorial areas (Laurentia, Baltica, South China, Australia). However, they were sometimes present at higher latitudes (South America; Sanchez et al., 1995). Locally, as in Gotland (Baltica) some endemic gen- era appeared (Liljedahl, 1984). However, the subclass did not experience Silurian expansion. The paleoheterodonts, anomalodesmatans, some pteriomorphs, endobyssate infau- nal and semi-infaunal suspension feeders, were the major participants of a very weak diversity increase while the epi- faunal pteriomorphs remained inconspicuous after an almost total extinction at the end of the Ordovician (Fig. 5). The environmental changes of the Middle Silurian (Wenlock) induced a new bivalve diversification which continued during the Ludlow (early Late Silurian). In sev- eral basins (Bohemia, Montagne Noire, Morocco, etc.), the development of alternating soft muds and biodetrital deposits favored the abrupt diversification of shallow free- burrowing suspension-feeders. Following Kriz (1979), the new specialized genera, such as Cardiola (Fig. 9) and oth- ers that he considers as pteriomorphs (see above), proceed- ed from epi- or endobyssate ancestors by a reversion neote- ny. In these environments endobyssate infaunal and endobyssate semi-infaunal suspension-feeders also showed a rapid development with, for instance, Modiomorphidae and some (?) Pterineidae [Stanley (1972) and Bailey (1983) admitted that some Paleozoic pterineids were endobyssate; for Johnston (1993), these views are poorly supported, and all the pterineids were epibyssate]. Lastly, the epibyssate BABIN: ORDOVICIAN TO DEVONIAN BIVALVIA 173 INFAUNAL FREE-BURROWING DEPOSIT FEEDERS ENDOBYSSATE SUSPENSION FEEDERS INFAUNAL OR SEMI-INFAUNAL SUSPENSION FEEDERS LLANDOVERY EPIBYSSATE RECLINING SUSPENSION = SUSPENSION FEEDERS FEEDERS PRIDOLI | LUDLOW WENLOCK ed = 10 genera Fig. 5. Silurian diversification of the bivalves with their life habits (after Kitz, 1984, modified). bivalves increased rapidly after the Wenlock and during the Late Silurian, and Ktiz (1984) considers that this extensive adaptation is “one of the most important characteristics of the evolution of the Bivalvia in the Period.” The distribu- tion of the Bohemian-like fauna, linked with low oxygen conditions on the sea-bed is explained by the surface cur- rent circulation (Bogolepova and Kriz, 1995). Johnston and Collom’s new ideas on cryptodonts as chemoautotrophs (1998) are supported by the existence of these low oxygen conditions. It is interesting to note that in their paper, Bogolepova and Kriz (1995) indicate the presence of such fossils as early as the lower Llandovery (basal Silurian) in Siberia, considering them as ancestral forms of the Bohemian-type bivalves. Can we regard the Siberian area as the cradle of these forms? It is not possible to reply with certainty to such a question. To locate precisely the cradles of the groups remains another problem generally insoluble for the paleodiversifications. During the Wenlock and Ludlow, some rare elements of the Bohemian-type fauna could have reached other areas, Slava is known in Laurentia (Pojeta and Norford, 1987), Dualina is cited in South America (Sanchez, 1991) and Florida (Pojeta et al., 1976). The fauna considered in these above-mentioned basins is different from that of the other areas dominated by paleotaxodonts, paleoheterodonts, and anomalodesmatans (Fig. 6). In these associations occurring in both clastic and carbonate platform muds, deposit-feeding species (paleo- taxodonts) generally formed a large majority of the bivalves. Lastly, in the reef communities developed in sev- eral areas the bivalves constituted a minor group (Watkins, 1997). They were represented especially by epibyssate sus- pension-feeders like ambonychiids, pterineids, with genera known in other level-bottom associations with a lower species diversity and with larger specimen sizes. Apart from the megalodont Megalomoidea known in some reefs, there were no bivalves specialized for reef environments. Finally, succeeding the Late Ordovician mass extinction, the Silurian was a Period of progressive recov- ery for the bivalves. Two major kinds of environments induced different adaptive responses. On the platforms, with normally oxygenated waters and at all latitudes, repre- sentatives of the main subclasses are found. In the Bohemian-type basins, environmental conditions with a low oxygenation induced an explosive evolution of the Pteriomorphia and Cryptodonta giving an odd cachet to the faunas. A slight decrease in bivalve diversity occurred dur- ing the Pridoli (Late Silurian); however, there was not a cri- sis like that of the Upper Ordovician. THE DEVONIAN - DIVERSIFICATION AND DISPERSAL OF BIVALVE FAUNAS The Devonian was another period of important diversification for the bivalves. According to Ktiz (1979b), basing his argument on the Treatise on Invertebrate Paleontology (1969), the Devonian diversity in comparison with that of the Silurian increased 59% for genera and 36% for families. Amongst the latter it is noticeable that a dozen 174 AMER. MALAC. BULL. 15(2) (2000) % Bohemian type faunas © Paleotaxodonts - Paleoheterodonts dominated faunas & Faunas with one or two Bohemian bivalves Fig. 6. Distribution of Wenlockian bivalve faunas (after Babin, 1993a, updated). of them are extant families, four of which are new families of Heterodonta. Thus the Devonian bivalve fauna acquired a more modern aspect. The increase in diversity was progressive during the Early Devonian and attained a peak in the Emsian. A first decrease took place during the Middle Devonian (Fig. 7) with different aspects according to the regions. After the Givetian, the global decrease became more pronounced until the end of the Devonian. It may be the result of “plate- tectonic regulation” (Kriz, 1979b) but it was also probably related to the development of regional environmental dete- rioration, especially in regard to benthic oxygen. The Early Devonian bivalve faunas were particular- ly numerous on the terrigenous shallow sea-floors. The paleotaxodonts remained abundant in such environments and sometimes associations of Palaeoneilo or Nuculites were almost monospecific. Nevertheless, the pteriomorphs and the anomalodesmatans were particularly varied. Some of the former were possibly attached to flexible algae (Johnston, 1993) though the largest of them were fixed on elements on the sea-bed. In Europe, the Lower Devonian bivalves have been described and listed since the nineteenth century in the siliciclastic sediments of the Rhenish area (Frech, 1891; Beushausen, 1895; Maillieux, 1937, etc.) where they were especially numerous, a proportion of them being endemic genera. Similarly, in the carbonate facies of Bohemia a large part of the bivalve genera were endemic such as the cryptodonts Antipleura or Kralovna. In North America the generic variety is lesser but recent investiga- tions (Johnston and Goodbody, 1988; Desbiens, 1994a, b) show, in siliciclastic sediments of Canada, the presence of several Rhenish genera. In South America in cool water the genera cited (Sanchez et al., 1995) were cosmopolitan. In North China, 28 genera are cited, all of which were cos- mopolitan, whereas 13 of the 56 Lower Devonian genera cited in South China were endemic, especially during the an genera 20 a Mee. Ar GS families \_ a B .” endemic genera 10 ‘ c ar Ged-Sieg. Emsian Eifel. Givet. Fras. Famen. Fig. 7. Curves showing variations of Devonian bivalves in China. The Y- axis is Number (after Zhang, 1988). BABIN: ORDOVICIAN TO DEVONIAN BIVALVIA 7S \ A Paleotaxodonts % Paleoheterodonts § ™@ Pteriomorphs O Anomalodesmatans Fig. 8. Characteristics of Emsian bivalve faunas. West-European faunas have been taken as reference, and for each region those groups are shown that have at least three genera in common with them (after Babin, 1993a, updated). Emsian (Zhang, 1988). Like the Rhenish area, the latter region was apparently an important center of diversification (Fig. 8). In southeastern Australia, Chapman (1908) described 27 genera and noted that they were known in North America or Europe. More recently, from a shallow marine carbonate shelf environment Johnston (1993) described a remarkably preserved silicified fauna of bivalves. Amongst 24 genera cited, two or three are endem- ee Fig. 9. Possible life habits of small praecardioids. Cardiola (below; Silurian) as semi-endobyssate according to Kfiz (1979a). Buchiola (above; upper Devonian) as epibyssate pseudoplankton according to Babin (1966) (after Vannier et al., 1995; modified). ic; the others show close affinities with the bivalves of cen- tral Europe. From the lower Devonian (probably Emsian) of New Zealand, Bradshaw (1999) lists 27 genera and she acknowledges only “a distant link with Europe” for her fauna, however at the generic level, at least 16 of these taxa are known also in Europe. Finally, bivalves experienced a noteworthy diversification during Emsian times but they evolved only at family, genus and species levels. This radi- ation is probably related to the Emsian transgressions, and the most favorable environments seem have been shallow terrigenous shelves located at low latitudes such as central Europe or south China. After the Emsian acme a slight decline occurred during the Middle Devonian. However bivalve faunas were still flourishing in many regions but with little new radia- tion. For instance, the megalodontids show some adaptive diversification with new genera such as Megalodon and Eomegalodus, these having a thick shell related to coral environments. The most significant for Middle Devonian times is the continued dispersal of the bivalve faunas, resulting in increased cosmopolitanism. It is the result of the progressive restriction of some oceanic areas such as the Rheic between Europe and North America. Bailey (1983) has finely illustrated the similarities between the Rhenish and Appalachian communities; however, there was some diachronism, since the Middle Devonian faunas of North America include Lower Devonian forms from Western Europe. Therefore, the fauna migrated westwards giving a large expansion to the genera initiated primarily in the Rhineland. The decline of the bivalve faunas continued through 176 AMER. MALAC. BULL. 15(2) (2000) the Late Devonian, but the paleogeographic changes induced some new limited diversifications. For instance, the development of large continental areas such as the Red Sandstone continent was favorable to the first appareance of fresh-water bivalves with Archanodon. In the marine environments a new extension of dysaerobic facies induced the development of small praecardioids (Buchiola) and pte- rioids (Guerichia = Posidonia in previous papers like Babin 1966) which were probably attached to floating algae or wood fragments (Fig. 9) although Grimm (1998) argues that Buchiola was infaunal. After the drastic decrease by the end of the Frasnian some bivalve faunas of the latest Famennian (Strunian) dominated by pteriomorphs and located at the southern margin of Laurussia initiated a new diversification (Amler 1996). Nevertheless, a clear decline in bivalve diversity is related to the global crisis near the end of the Devonian. CONCLUSION After a very discrete appearance during the Lower Cambrian and an obscure history during the Cambrian and the earlier Ordovician (Tremadoc), bivalves showed during the Arenig a dramatic phase of radiation corresponding to the famous Ordovician diversification. It seems that in a few millions years all or nearly all, the subclasses appeared. This drastic paleodiversification coincided with key inno- vations like the filibranchiate gills and the retention of the postlarval byssus in the adult. It was located on shallow clastic sea-beds, apparently on the peri-Gondwanan and Avalonian shelves. After a major dispersal during the Upper Ordovician corresponding notably to the diversifica- tion of the semi-endobyssate and epibyssate forms, the bivalves underwent a first decline with the latest Ordovician mass extinction. New radiations occurred dur- ing the Middle and Upper Silurian especially with the development of the cryptodonts related to particular dysaer- obic environments, and with an explosive evolution of the pteriomorphs. The Devonian opened another period of paleodiversification at family, genus and species levels but with a progressive acquisition of a more modern cachet. The acme of this diversification took place during the late Early Devonian (Emsian), and the paleogeographic changes led to the establishment of cosmopolitan faunas. During this Period the bivalves dispersed into various marine envi- ronments and for the first time into fresh-water. A second decline coincided with the Late Devonian mass extinction. ACKNOWLEDGMENTS I acknowledge the American Malacological Society, which gave me support for its meeting in Pittsburgh. I am grateful to Dr. John C.W. Cope (Cardiff) and to an anonymous reviewer for their improvement of the English and for their comments and suggestions. LITERATURE CITED Amler, M. R. 1996. Die Bivalvenfauna des Oberen Famenniums West- Europas. 2. Evolution, Palaogeographie, Palaodkologie, Systematik 2. Palaeotaxodonta und Anomalodesmata. Geologica et Palaeontologica, 30:49-117. Babin, C. 1966. Mollusques Bivalves et Céphalopodes du Paléozoique armoricain. Imprimerie Commerciale et Administrative, Brest. 470 p., 18 pl. Babin, C. 1993a. La répartition géographique des mollusques bivalves du Paléozoique (Cambrien a Dévonien): une revue. Palaeo- geography, Palaeoclimatology, Palaeoecology 100:7-20. Babin, C. 1993b. Role des plates-formes gondwaniennes dans les diversi- fications des mollusques bivalves durant |’Ordovicien. Bulletin Société géologique de France 164:141-153. Babin, C. 1995. The initial Ordovician bivalve mollusc radiations on the western Gondwanan shelves. In: Ordovician Odyssey. J. D. Cooper, M. L. Droser, and S. C. Finney, eds. Pacific Section Society Sedimentary Geology 77:491-498. Babin, C. and J. C. Gutiérrez Marco. 1991. Middle Ordovician bivalves from Spain and their phyletic and palaeogeographic significance. Palaeontology 34:109-147. Bailey, J. B. 1983. Middle Devonian Bivalvia from the Solsville Member (Marcellus Formation), central New York State. Bulletin American Museum of Natural History 174:190-325. Barrande, J. 1881. Systeme Silurien du Centre de la Bohéme. Vol. V1. Acéphalés. Paris et Prague. 342 pp., 361 pl. Beushausen, L. 1895. Die Lamellibranchiaten des rheinischen Devon, mit Ausschluss der Aviculiden. Abhandlungen der Koniglich Preussischen geologischen Landesanstalt, N.F. 17:1-514, Atlas 38 pl. Bogolepova, O. K. and J. Kitz. 1995. Ancestral forms of Bohemian type Bivalvia from the lower Silurian of Siberia (Tungusskaja syneclise, Russia). Geobios 28:69 1-699. Bradshaw, M. A. 1999. Lower Devonian bivalves from the Reefton Group, New Zealand. Memoir 20 of the Association of Australasian Palaeontologists:1-171. Bretsky, P. W. 1973. Evolutionary patterns in the Paleozoic Bivalvia doc- umentation and some theoretical considerations. Geological Society of America Bulletin 84:2079-2096. Chapman, F. 1908. A monograph of the Silurian bivalved Mollusca of Victoria. Memoirs of the National Museum, Melbourne 2:1-62. Cope, J. C. W. 1995. The early evolution of the Bivalvia. In: Origin and evolutionary radiation of the Mollusca. J. D. Taylor, ed. pp. 361- 370. The Malacological Society London. Oxford University Press (Dated 1996 but published 1995). Cope, J. C. W. 1996. Early Ordovician (Arenig) bivalves from the Llangynog Inlier, South Wales. Palaeontology 39:979-1025. Cope, J. C. W. 1998. The origin of heterodont bivalves. 42nd Annual Meeting, The Palaeontological Association, Portsmouth, Abstracts. Cope, J. C. W. 1999. Middle Ordovician bivalves from Mid-Wales and the Welsh Borderland. Palaeontology 42:467-499. Cope, J. C. W. and C. Babin. 1999. Diversification of bivalves in the Ordovician. Geobios 32:175-185. Desbiens, S. 1994a. Le bivalve dévonien Prosocoelus Keferstein, 1857 dans le nord-est de 1’ Amérique du Nord. Geological Survey of Canada, Bulletin 479:49-66. BABIN: ORDOVICIAN TO DEVONIAN BIVALVIA 177 Desbiens, S. 1994b. Le bivalve Montanaria Spriestersbach, 1909: habitat et morphologie d’une nouvelle espéce emsienne de la Formation de York River de Gaspé, Québec. Canadian Journal of Earth Sciences 31:381-392. Droser, M. L., R. Fortey, and X. Li. 1996. The Ordovician Radiation. American Scientist 84:122-131. Droser, M. L., P. M. Sheehan, R. A. Fortey, and X. Li. 1995. The nature of diversification and paleoecology of the Ordovician radiation with evidence from the Great Basin. In: Ordovician Odyssey, J. D. Cooper, M. L. Droser, and S. C. Finney, eds. Pacific Section Society Sedimentary Geology 77:405-408. Dzik, J. 1994. Evolution of “small shelly fossils” assemblages. Acta Palaeontologica Polonica 39:247-313. Erwin, D. H. 1994. Early introduction of major morphological innova- tions. Acta Palaeontologica Polonica 38:281-294. Fortey, R. A. and L. R. M. Cocks. 1988. Arenig to Llandovery faunal dis- tributions in the Caledonides. In: The Caledonian-Appalachian Orogen, A. L. Harris and D. J. Fettes, eds. Geological Society Special Publication 38:233-246. Frech, F. 1891. Die Devonischen Aviculiden Deutschlands. Koniglich Preussischen geologischen Landesanstalt. Berlin. 253 pp. Atlas 18 pl. Fritz, M. A. 1951. The pelecypod genus Viasta in the Ordovician of North America. Journal of Paleontology 25:696-698. Gould, S. J. 1977. Ontogeny and phylogeny. Harvard University Press, Cambridge. 501 pp. Grimm, M. C. 1998. Systematik und Palaéodkologie der Buchiolinae noy. subfam. (Cardiolidae, Arcoida, Lamellibranchiata, Devon). Schweizerische Paldontologische Abhandlungen 118, 135 pp. Guo, F. 1985. Fossil bivalves of Yunnan. Yunnan Science and Technology Publishing House, Kunming. 319 pp. (in Chinese with Summary in English). Hall, J. 1884-1885. Palaeontology, vol. V. Lamellibranchiata. Part I Monomyaria (1884). Part II Dimyaria (1885). Geological Survey State New York, Albany. 561 pp., 96 pl. Jablonski, D. and D. J. Bottjer. 1990. Onshore-offshore trends in marine invertebrate evolution. In: Causes of Evolution: A Paleontological Perspective, R. M. Ross and W. D. Allmon, eds. pp. 21-75. University Chicago Press. Johnston, P. A. 1993. Lower Devonian pelecypoda from southeastern Australia. Association Australasian Palaeontologists. 134 pp. Johnston, P. A. and C. J. Collom. 1998. The Bivalve Heresies - Inoceramidae are Cryptodonta, not Pteriomorphia. Jn: Bivalves: An Eon of Evolution, P. A. Johnston and J. W. Haggart, eds. pp. 347-360. University Calgary Press. Johnston, P. A. and Q. H. Goodbody. 1988. Middle Devonian bivalves from Melville Island, Arctic Canada. In: Devonian of the world, N. J. McMillan, A. F. Embry, and D. J. Glass, eds. Canadian Society of Petroleum Geologists, Memoir 14:337-346. KZ, J. 1965. Genus Butovicella gen. n. in the Silurian (Bivalvia). Zvlastni Otisk Vestniku Ustredntho Ustavu Geologického 40:207-208. Kz, J. 1979a. Silurian Cardiolidae (Bivalvia). Sbornik Geologickych Ved. Paleontologie 22: 157 pp. KiiZ, J. 1979b. Devonian Bivalvia. In: The Devonian System. Special Papers in Palaeontology, Palaeontological Association 23:255- 257. Kriz, J. 1984. Autecology and ecogeny of Silurian Bivalvia. In: The Murchison Symposium, Special Papers in Palaeontology, Palaeontological Association 44:183-195. Kiiz, J. 1996. Silurian Bivalvia of Bohemian type from the Montagne Noire and Mouthoumet massif, France. Palaeontographica Abt A 240: 29-63. KiizZ, J., 1998. Taxonomy, functional morphology and autecology of the sinistrally twisted bivalve Viasta Barrande, 1881 from the lower Devonian of Bohemia, Morocco and Central Asia. Geobios 31:455-465. Liljedahl, L. 1984. Silurian silicified bivalves from Gotland. Sveriges Geologiska Undersokning 804, 78, 82 pp. Liljedahl, L. 1994. Silurian nuculoid and modiomorphid bivalves from Sweden. Fossils & Strata 33, 89 p. Maillieux, E. 1937. Les Lamellibranches du Dévonien inférieur de l’Ardenne. Mémoires Musée royal Histoire naturelle Belgique 81. 274 pp. Murchison, R. I. 1835. On the Silurian System of rocks. London and Edinburgh Philosophical Magazine and Journal of Science, Series 3, 7:46-52. Newell, N. D. 1969. Subclass Cryptodonta Neumayr, 1884. /n: Treatise on Invertebrate Paleontology, R. C. Moore, ed. pp. N241-N248. Part N. Mollusca 6. Bivalvia. Nicol, D. 1989. Diversity and adaptive radiation in marine and freshwater Bivalvia. American Conchologist 17:10-11. Paul, C. R. C. 1989. Patterns of evolution and extinction in invertebrates. In: Evolution and the fossil record, K. C. Allen and D. E. G. Briggs, eds. pp. 99-121. Belhaven Press. Pojeta, J. Jr. 1987. Class Pelecypoda. In: Fossil Invertebrates, R. S. Boardman, A. H. Cheetham, and R. J. Rowell, eds. pp. 386-435. Blackwell Scientific Publications. Palo Alto. Pojeta, J. Jr. 1988. The origin and Paleozoic diversification of solemyoid pelecypods. New Mexico Bureau of Mines and Mineral Resources Memoir 44:201-271. Pojeta, J. Jr., J. Kitz, and J. M. Berdan. 1976. Silurian-Devonian pelecy- pods and Paleozoic stratigraphy of subsurface rocks in Florida and Georgia and related Silurian pelecypods from Bolivia and Turkey. Geological Survey, Professional Paper 879, 32 p. Pojeta, J. Jr. and B. S. Norford. 1987. A Bohemian type Silurian (Wenlockian) pelecypod fauna from Arctic Canada. Journal of Paleontology 61:508-520. Sanchez, T. M. 1991. El género Dualina (Bivalvia, Praecardioida) en la Formacion Lipeon (Silurico), Sierra de Zapla, Provincia de Jujuy, Argentina. Ameghiniana 28:3 1-34. Sanchez, T. M. 1999. New late Ordovician (early Caradoc) bivalves from the Sierra de Villicum (Argentina Precordillera). Journal of Paleontology 73:66-76. Sanchez, T. M., B. G. Waisfeld, and B. A. Toro. 1995. Silurian and Devonian molluscan bivalves from Precordillera region, western Argentina. Journal of Paleontology 69:869-886. Sharpe, D. 1853. Description of the new species of Zoophyta and Mollusca. In: On the Carboniferous and Silurian formations in the neighbourhood of Bussaco, Portugal, C. Ribeiro, D. Sharpe, and T. R. Jones. Quarterly Journal of the Geological Society of London 9.3, 35:146-158. Stanley, S. M. 1972. Functional morphology and evolution of byssally attached bivalve mollusks. Journal of Paleontology 46:165-212. Stanley, S. M. 1975. Adaptive themes in the evolution of the Bivalvia (Mollusca). Earth and Planetary Sciences 3:361-385. Stanley, S. M. 1977. Trends, rates, and patterns of evolution in the Bivalvia. In: Patterns of evolution illustrated by the fossil record, A. Hallam, ed. pp. 209-250. Elsevier Scientific Publishing Company. Tevesz, M. J. S. and P. L. McCall. 1976. Primitive life habits and adaptive significance of the pelecypod form. Paleobiology 2:183-190. Tevesz, M. J. S. and P. L. McCall. 1985. Primitive life habits of Bivalvia reconsidered. Journal of Paleontology 59:1326-1330. Treatise on Invertebrate Paleontology. 1969. Part N. Mollusca 6. Bivalvia, vols. 1 and 2. R. C. Moore, ed. 952 pp. Vannier, J., C. Babin, and P. R. Racheboeuf. 1995. Le principe 178 AMER. MALAC. BULL. 15(2) (2000) d’actualisme appliqué aux faunes paléozoiques: un outil ou un leurre? Geobios Mémoire Spécial 18:395-407. Waller, T. R. 1988. A phylogeny of the Bivalvia: progress and problems. 54th Annual Meeting American Malacological Union, Charleston. Abstracts: 19. Waller, T. R. 1998. Origin of the Molluscan Class Bivalvia and a Phylogeny of Major Groups. Jn: Bivalves: An Eon of Evolution, P. A. Johnston and J. W. Haggart, eds. pp. 1-45. University Calgary Press. Watkins, R. 1997. Paleoecology of Silurian reef bivalves, Racine Formation, North America. Lethaia (1996) 29:171-180. Yonge, C. M. 1947. The pallial organs in the aspidobranch Gastropoda and their evolution throughout the Mollusca. Philosophical Transactions Royal Society London B, 232: 443-518. Zhang, R. 1988. Stratigraphic and paleobiogeographic aspects of Devonian bivalves of China. In: Devonian of the world. N. J. McMillan, A. F. Embry, and D. J. Glass, eds. pp. 347-356. Canadian Society of Petroleum Geologists III. Date of manuscript acceptance: 4 April 2000 Reconstruction of ancestral character states in neocoleoid cephalopods based on parsimony Michael Vecchione!, Richard E. Young?, and David B. Carlini National Marine Fisheries Service, Systematics Laboratory, National Museum of Natural History, Washington, DC 20560, U.S. A. 2Department of Oceanography, University of Hawaii, Honolulu, Hawaii 96822, U.S. A. 3Virginia Institute of Marine Science, School of Marine Science, College of William and Mary, Gloucester Point, Virginia 23062, U.S. A; current address: Department of Biology, University of Rochester, Rochester, New York 14627, U.S. A. Abstract: The Neocoleoidea, sister group to the Belemnoidea, includes all living cephalopod species except nautilids, as well as their immediate ances- tors. Several hypotheses have been published about the morphology of ancestral neocoleoids. Ancestral states are easily inferred from fossils for some char- acters, such as 10 arms and the presence of an ink sac in basal coleoids or the presence of fins in ancient octopods. Many inferences are less strongly sup- ported, though, and open to debate. We examine this problem using three cladograms resulting from analyses of morphology and DNA sequences (both mitochondrial and nuclear) from samples representing the full diversity of extant coleoids. Character states at three ancestral nodes (neocoleoid, octopodi- form, and decapodiform) are reconstructed for 51 morphological characters using cladistic parsimony. Strong or moderate agreement among the three trees was found for almost 3/4 of the character-at-node reconstructions. The level of agreement among the trees varied among nodes, with strongest agreement found at the ancestral octopodiform node. However, some of these reconstructions seem unlikely to be correct. Changes in subclade resolution can exert varying effects on inferences about basal nodes. Because several subclades within the neocoleoids are not yet adequately resolved, we cannot be very confi- dent in reconstructions of ancestral character states based solely on parsimony and we propose a provisional suite of character-state reconstructions including other sources of inference in addition to parsimony. Key Words: squid, octopod, cuttlefish, morphology, evolution, phylogeny, cladistic consensus The reconstruction of ancestral character states is an ny to infer character states at ancestral nodes of a clado- important step in unraveling the pathways of evolution and gram. Cladistic reconstruction is based on rules that define the changes that led to the present diversification in which character states are most likely to occur at a node cephalopods. Naef (1921-3) relied on inference based on (Cunningham et al., 1998). However, even with carefully his extensive knowledge of comparative anatomy, embryol- and accurately defined character states, the results are not ogy and paleontology to reconstruct generalized types from definitive. Nevertheless, inferences can be robust (Shultz et which all major coleoid groups could be derived. While not al., 1996). We present here the first attempt to reconstruct intending these to represent ancestral forms, he often felt ancestral character states for neocoleoid cephalopods that they did so. With the advent of more rigorous means of (defined by Young ef al., 1998 as the sister group to the reconstruction, Naef’s types are no longer acceptable for belemnoids and containing all extant coleoids) using parsi- determining primitive states even though many of his con- mony analyses of cladistic hypotheses of phylogeny based clusions ultimately may be supported. Subsequent efforts on both morphological and molecular characters. We previ- (e. g., Bandel and Leich, 1986; Haas, 1997) have been ously analyzed 50 morphological characters in order to based largely on inferences from fossils and the literature determine relationships among the major groups of extant on anatomy of living cephalopods. Coleoid fossils, although coleoid cephalopods (Young and Vecchione, 1996). These rare, have occasionally been of use in determining the pres- characters, for the most part, are those for which we try to ence of some anatomical structures, such as an ink sac and construct evolutionary histories, although several characters ten arms in early coleoids and fins in ancient octopods have been redefined or replaced. We use the cladistic rela- (Young ef al., 1998). However, the fossil record of coleoids tionships inferred from morphology by Young and is exceptionally poor (Foote and Sepkoski, 1999), limiting Vecchione (1996) as well as molecular phylogenies based its usefulness for character-state reconstruction. on mitochondrial (Carlini and Graves, 1999) and nuclear An alternative approach is to use cladistic parsimo- (Carlini, 1998) DNA sequences. American Malacological Bulletin, Vol. 15(2) (2000):179-193 179 180 AMER. MALAC. BULL. 15(2) (2000) MATERIALS AND METHODS Morphology The data matrix used for the reconstructions is mod- ified from Young and Vecchione (1996) and that paper can be consulted for much of the material examined. We have revised a few character definitions for the present analysis and have added or replaced a few characters. The character descriptions and data matrix are presented in Appendices 1- 2. The consensus cladogram that Young and Vecchione considered the best estimate of neocoleoid phylogeny based on morphology is reproduced in Fig. 1. For the present study, we added Belemnoidea, the sister group to the neo- coleoids, with character states derived from literature on fossil belemnoids. As is discussed below, reconstruction by parsimony does not always infer reasonable character states at ancestral nodes. We therefore offer a set of opinions in Appendix 3 about ancestral characters states. These opin- ions summarize the discussion of individual characters in Appendix |. The parsimony solution is accepted unless some other source of information, such as paleontology or ontogeny, indicates that the inference from parsimony is particularly questionable. Loliginidac Bathyteuthidac A Enoplotcuthidac Gonatidae Ommastrephidae Gi) Onychoteuthidae Thysanoteuthidae Sepiidae Sepiolidae Spirulidae Bolitaenidae ) Octopodidae co Ocythoidae Staurotecuthidae Molecular sequences Taxonomic sampling and details of extraction, amplification, cloning, and sequencing methods for DNA are presented by Carlini (1998). Many alternative phyloge- netic analyses, using different models for cladogenesis, substitution, weighting, and combinations of molecular data sets are presented there as well, along with cladogram para- meters such as tree lengths, consistency indices, bootstrap values, etc. We have selected here what we feel are the most reasonable phylogenetic hypotheses for separate mito- chondrial and nuclear DNA sequences. The cladograms for both molecular data sets are fully resolved and are the most parsimonious from unweighted parsimony analysis. For reconstruction analyses, the trees representing these hypotheses were “pruned” so that the nodal structure of the trees was retained but terminal taxa are those for which morphological character states were assessed. Although the resulting trees appear pectinate, many of the terminal branches actually represent clades on the original molecular trees. Tree statistics are presented in the captions to Figs. 2 and 3. Mitochondrial. The cytochrome c oxidase I (COI) cladogram (Fig. 2) is similar to that of Carlini and Graves @ Thysanoteuthidse {MH Onychoteuthidse [BOmmastrephidee WGonatidae Enoploteuthidee @ Bathyteuthidae IO Loliginidse lOSpirulidse Se piolidae ID Se piidse ‘OYampyroteuthidae fi Bolitsenidae M Oc ythoidae IM Octopodidee Opisthoteuthidae Intestine vs vena cava unordered Cirroteuthidae Vampyrotcuthidae J) ventral ‘ WEE oDorsa) or anterior Opisthoteuthidae Nautilidac ES equivocal bd YQ. “I belemnoid Nautilidae Fig. 1. A. Morphological consensus cladogram based on Young and Vecchione (1996); consensus of 14 trees, tree length 46, consistency index 0.98, reten- tion index 0.99. Numbers within circles are the number of unambiguous synapomorphies supporting the node at that location. B. Example of reconstruction cladogram based on A. In this and the following two figures, the same character (number 47, the position of the intestine relative to the vena cava) is used as an example of character-state reconstruction. VECCHIONE ET AL.: ANCESTRAL STATES IN NEOCOLEOID CEPHALOPODS 18] Chirareuthts Mastigoteuthts Phahdoteuthts Foubinureuthis Brachtoreuthis Octopoteuthis hysanoteutas Alluroteathis Spuula Archueuthts A Cryeloreuthis Lepidoteathts Disxcoteuthis Onvehoteuttin Moroteuthis Hisnioteuthts Powvchroteuthis Ommastrephes Sthenateuthis Chtenoplervx Bathyreuthis Cranchia Lioeranchta Ancistrocheirus Pyroteuthin L. opalescens L. pealet Sepioreuthis Lycoteushis Abralta Enoploteuthis OssIa Stoloteuthin Heteroteuthis S. offictualis S. opipara Idiosepius Septoluidea Octopus Hapalochlaena Argonauta Graneledone Verreledonetla Eledonellu Japatetla Srauroteahts Cirrothauna Gronpoteuthis Opisthoreuthis sp. Bathypolypus Vampyroteuthis Nautilus Katharina G.berry! G. ONYX Gonatopsis @ Thysanoteuthidse BGonatidse Spirulidae MB Onychoteuthidee @Ommastrephidae @ Bathyteuthidee (DD Loliginidae BW Enoploteuthidae IO Se piolidae (Se piidee OO Vamp yroteuthidae Mi Octopodidee Intestine ys vena cava 4 steps 1 Boliteenidae unordered CJ ventral BCirroteuthidae MB sCDorsal or anterior E35 equivocal MOpisthoteuthidae Nautilidae Fig. 2. A. Maximum parsimony cladogram based on DNA sequence of the mitochondrial gene for cytochrome c oxidase I; tree length 3763, consistency index 0.167, retention index 0.329. Branch lengths are proportional to changes in nucleotide sequence. B. Cladogram “pruned” for character-state recon- struction to retain nodal structure but reduce terminal taxa to those for which morphological data were assessed. (1999), but the data were reanalyzed after including addi- tional taxa. For the present paper, a 657 base-pair segment of the COI gene was analyzed for 55 species. Of these char- acters (base pairs), 297 were parsimony-informative. The result was a single most-parsimonious tree of length 3763 (Fig. 2). Nuclear.— Preliminary phylogenetic analysis by Carlini (1998) on cephalopod actin sequences suggested that at least three paralogous actin genes had been amplified and cloned. These were subsequently discriminated among using restriction endonucleases. This finding was supported subsequently by genomic southern blotting, a more rigorous means of assessing gene copy number (Southern, 1975; Carlini et al., 2000). Two of the three copies of the actin gene were analyzed in detail, and gene sequences from both paralogs were obtained from 26 taxa. The two genes, desig- nated actin I and actin II, were concatenated in each taxon for a total of 1568 base pairs in the analysis of the com- bined data sets. The total number of parsimony-informative characters was 376. The cladogram presented here (Fig. 3) represents the single most parsimonious tree (length = 1581) derived from cladistic analysis of the combined actin genes. Reconstructions Ancestral character states have been reconstructed using the computer program MacClade (Maddison and Maddison, 1992), which finds the most parsimonious 182 AMER. MALAC. BULL. 15(2) (2000) Spirula Loliga pealet Seproreuthis S. officinalis S. opipara A Chrenopreryx Bathyreuthis Thyvsanoteuthis Cranchia Sepwloidea Stoloteuthis Rossia Idiosepius Ommastrephes Sthenoteuthis Enaploteuthis Histioreuthis Gonatus onvx Cycloteuthis Discoreuthis Octopus Graneledone Efedonella Japatella Cirrothauma Vampyroteuthis Loliginidae Sepiidse Spirulidae @ Bathyteuthidee WB Thysanoteuthidse Sepiolidae WOmmastrephidae 1 Enoploteuthidee M@Gonatidae Yempyroteuthidae Intestine vs vena cava WCirroteuthidee 3 steps unordered Cc Ventral GB Dorsal or anterior MOctopodidse WB Bolitaenidse Neutilidse Fig. 3. A. Maximum parsimony cladogram based on combined DNA sequences of paralogs I and II for the nuclear gene for actin (Carlini, 1998; Carlini et al., 2000); tree length 1581, consistency index 0.461, retention index 0.549. Branch lengths are proportional to changes in nucleotide sequence. B. Cladogram “pruned” for character-state reconstruction to retain nodal structure but reduce terminal taxa to those for which morphological data were assessed. reconstructions. A characteristic of MacClade must be noted, however. In the case of polytomies, such as the deca- pod clade in the morphological tree, MacClade attempts to reconstruct character evolution as if the polytomy were resolved dichotomously by inventing an intercalated node between the polytomy and the basal dichotomous node out- side of the polytomy. This convention is used when a char- acter is polymorphic within a polytomy because the context of the basal node will vary with different resolutions of the phylogeny. However, for cases in which all taxa within a polytomy share a character state, we consider that state to be ancestral to the polytomy, rather than equivocal as would be reconstructed by MacClade. In the current pre- sentation, this convention does not apply to the molecular trees, which are fully resolved, but affects many reconstruc- tions based on the morphological tree, especially for the ancestral decapod node. Character states were considered to be unordered except for three characters for analysis on the morphologi- cal tree. Young and Vecchione (1996) treated characters number 8 (arms II), 35 (superior buccal lobe) and 36 (sub- frontal lobe) as ordered to infer the morphological phyloge- ny used here. Because these characters were considered to be ordered for development of the tree, we have treated VECCHIONE ET AL.: ANCESTRAL STATES IN NEOCOLEOID CEPHALOPODS 183 them as ordered for character-state reconstruction based on the morphological tree. However, rather than impose these morphological assumptions on the independent molecular analyses, we considered all character states to be unordered for reconstructions based on gene trees. To determine the degree of agreement among trees, we define a series of rules for the consensus column in Tables 1-3. (1) Strong agreement occurs when all three trees indicate the same character state at a node. (2) Moderate agreement is when two trees agree on a character state and the third reconstructs the character state as equiv- ocal. (3) If two trees agree on a character state and the third reconstructs the character as an alternative state (as in majority-rules consensus) we consider this to be weak agreement. (4) When the reconstructed state is equivocal on two or three trees, this is listed as lack of consensus. (5) Disagreement occurs when two trees reconstruct a charac- ter as having different states and the third indicates that the state is equivocal. RESULTS Reconstructions of the states of 51 characters on morphological, nuclear (actins I and II), and mitochondrial (COI) trees are summarized in Table | for the ancestral neocoleoid node, Table 2 for the ancestral octopodiform node, and Table 3 for the ancestral decapod node. Of the total of 153 characters-at-nodes, strong agreement (all three trees agree on a character state) was found at 97, or 63.4%. If moderate agreement is included, the number increases to 112 (73.2%). Overall, the amount of agreement among trees, including strong, moderate, and weak agreement, was high at the decapodiform (43 characters) and octopodiform nodes (41 characters) and much lower (30 characters) at the neocoleoid node. The maximum strong agreement (38 char- acters) and least disagreement (1 character) among trees was found at the ancestral octopodiform node. Although the ancestral decapod node had a higher number of charac- ters with strong agreement among trees (34) than did the ancestral neocoleoid node (25), the neocoleoid node had only two characters for which the trees disagreed; there were five such characters at the decapod node. For some characters, agreement among trees was consistently strong at all focus nodes. The inferred charac- ter state remained unchanged among nodes for some of these characters. Examples include character 5, fins, char- acter 10, suckers, and character 14, arm trabeculae, all pre- sent at all nodes. Other characters for which strong agree- ment was consistent within nodes exhibited changes in inferred state among nodes. Examples of these include character 7, the buccal crown (present in ancestral decapods but absent in octopodiforms), and character 19, the outer capsule of the statocyst (absent in ancestral neo- coleoids and decapods but present in octopodiforms). Several characters are otherwise noteworthy. Reconstructions of character 47, the position of the intes- tine relative to the vena cava, resulted in disagreement among trees at all three focus nodes (Figs. 1-3). Similarly, head width reconstruction, character 49, resulted in dis- agreement at two nodes and only weak agreement at the third. Assessment of states for both of these characters was particularly difficult (Appendix 1; Young and Vecchione, 1996). Character 1, the phragmocone, is an example of a problem that is perhaps more important. Parsimony recon- structed the phragmocone as either absent or equivocal at every focus node on every tree. As discussed below, the phragmocone almost certainly was present at least in ances- tral neocoleoids and decapods and possibly was present also in ancestral Octopodiformes. Therefore, these recon- structions of this character are almost certainly either incor- rect (absent in actins at these nodes) or uninformative (equivocal elsewhere). Similarly, the reconstruction of the digestive gland, character 44, as fused at all ancestral nodes can be questioned based on embryological evidence of a paired origin of this structure. Such problems raise ques- tions of reliability of the other reconstructions. We there- fore present in Appendix 3 a summary of hypothesized ancestral character states based on ontogenetic and paleon- tological evidence in addition to parsimony, as discussed in Appendix |. DISCUSSION Some neocoleoid morphological character states can be polarized by the extant outgroup, Nautilus, fossil cephalopods, or cephalopod embryology, but many cannot be polarized due to the lack of a living, closely-related out- group for neocoleoids. Sufficient information exists, how- ever, to make a first attempt at defining some of the basic character states (Appendix 1). The morphological hypothe- sis of relationships used here suffers from a lack of resolu- tion within decapod and incirrate octopod clades. As a result, inferences are equivocal for character-state changes of many characters. The molecular analyses do not have this problem but suffer from unsubstantiated (7. e., lack of bootstrap support) subclade relationships. Whereas the tree derived from the actin genes has the same major nodes as the morphological tree, the COI gene tree disagrees in the relationships within the Octopoda. Both molecular trees, however, are completely resolved within the Decapodi- formes, although showing considerable disagreement between them, while the morphological tree depicts the Decapodiformes as an unresolved (i. e., soft) polytomy. Although our analyses involve three quite different trees, 184 AMER. MALAC. BULL. 15(2) (2000) Table 1. Character states inferred using parsimony for the ancestral neocoleoid nodes of trees based on morphol- ogy, DNA sequence for combined actins I and II (of three paralogs), and DNA sequence for cytochrome oxidase subunit I. “Consensus” indicates agreement among trees with: --, indicating disagreement; -, indicating no con- sensus; a character state without an asterix indicating weak agreement; *, indicating moderate agreement; and ** indicating strong agreement. Character Morphology Actins 1 Phragmocone Equivocal Absent 2 Proostracum Present Present 3. Median Field Equivocal Equivocal 4 — Shell Number 1 1 5 Fins Present Present 6 Fin Cartilage Equivocal Equivocal 7 ~~ Buccal Crown Equivocal Equivocal 8 Arms II Unmodified Equivocal 9 Arms IV Unmodified Equivocal 10 Suckers Present Present 11 Sucker Stalks Equivocal Equivocal 12. Sucker symmetry Equivocal Equivocal 13. Sucker Rings Equivocal Equivocal 14. Arm Trabeculae Present Present 15 Arm Protective Membrane Equivocal Present 16 Arm ILI Sucker Series Equivocal Equivocal 17. Arm I Web Present Equivocal 18 Arm V Web Equivocal Equivocal 19 Statocyst Outer Capsule Absent Absent 20 Nephridial Coelom 2 2 21 Visceropericardial Coelom Extensive Extensive 22 Dorsal Mantle Cavity Absent Absent 23. Nidamental Glands Equivocal Present 24 Crop Present Present 25 Branchial Canal Equivocal Present 26 Mantle Septum Equivocal Equivocal 27. Mantle Adductor Present Equivocal 28 Funnel Valve Present Present 29° Nuchal Cartilage Present Present 30. Cornea Absent Absent 31 Right Oviduct Present Present 32 Oviducal Gland Symmetry Equivocal Equivocal 33 Oviducal Gland Position Terminal Terminal 34 Photosensitive Vesicle Equivocal Equivocal 35 Superior Buccal Lobe Broad Sep. Broad Sep. 36 = Subfrontal Lobes Equivocal Equivocal 37. Arm-Mantle Muscle Absent Absent 38 Horizontal Arm Septae Absent Absent 39 Arm IV (III) Hectocotylus Absent Absent 40 Arm V Hectocotylus Absent Equivocal 41 Spermatophores Present Present 42 DGDA, Number Equivocal Equivocal 43 DGDA, Location Equivocal Equivocal 44 Digestive Gland Fused Fused 45 Gonad: Coelom Coverage Mostly Cov. Mostly Cov. 46 Posterior Salivary Gland Post.to Post.to 47 Intestine vs Vena Cava Equivocal Dorsal 48 Gill Filaments Both Free Both Free 49 Head Width 1.00-1.49 Equivocal 50. Long. Mantle Muscle Present Present 51 Arm Orientation Anterior Anterior Col Equivocal Equivocal Equivocal 1 Present Equivocal Equivocal Equivocal Equivocal Present Equivocal Equivocal Equivocal Present Present Equivocal Present Equivocal Absent 2 Extensive Absent Present Present Present Equivocal Equivocal Present Present Absent Present Equivocal Terminal Equivocal Broad Sep. Equivocal Absent Absent Absent Equivocal Present Equivocal Equivocal Fused Mostly Cov. Post.to Ventral Both Free 0.50-0.99 Present Anterior consensus VECCHIONE ET AL.: ANCESTRAL STATES IN NEOCOLEOID CEPHALOPODS 185 Table 2. Character state inferences as in Table 1, but for the ancestral Octopodiformes nodes. Character Morphology Actins COI consensus 1 Phragmocone Absent Absent Absent sats 2 Proostracum Present Present Equivocal = 3. Median Field Equivocal Equivocal Equivocal - 4 Shell Number 1 1 1 re 5 Fins Present Present Present me 6 Fin Cartilage Base + Core Base + Core Base + Core em 7 ~~ Buccal Crown Absent Absent Absent + 8 Arms II Modified Equivocal Equivocal . 9 Arms IV Unmodified Unmodified Unmodified e 10 Suckers Present Present Present ae 11 Sucker Stalks Equivocal Equivocal Equivocal - 12. Sucker symmetry Radial Radial Radial ai 13 Sucker Rings Equivocal Equivocal Equivocal - 14. Arm Trabeculae Present Present Present — 15 Arm Prot. Membrane Equivocal Present Present es 16 Arm III Sucker Series | 1 1 “e 17 Arm I Web Present Present Present ne 18 Arm V Web Present Present Present a 19 Statocyst Outer Capsule Present Present Present ns 20 Nephridial Coelom 2 2 2 ss 21 Visceropericardial Coelom Extensive Extensive Extensive _ 22 Dorsal Mantle Cavity Absent Absent Absent ae 23 Nidamental Glands Absent Absent Absent ne 24 Crop Present Present Present a 25. Branchial Canal Present Present Present si 26 Mantle Septum Equivocal Equivocal Equivocal - 27 Mantle Adductor Present Equivocal Equivocal - 28 Funnel Valve Present Present Present a 29 Nuchal Cartilage Present Present Present a 30 Cornea Absent Absent Absent _ 31 Right Oviduct Present Present Present ss 32 Oviducal Gland Symmetry Radial Radial Radial nm 33 Oviducal Gland Position Terminal Terminal Terminal * 34 Photosensitive Vesicle Equivocal Equivocal Equivocal - 35 Superior Buccal Lobe Adjacent Equivocal Equivocal - 36 Subfrontal Lobes Incipient Equivocal Equivocal - 37 Arm-Mantle Muscle Absent Absent Absent a 38 Horizontal Arm Septae Absent Absent Absent iid 39 Arm IV (IIT) Hectocotylus Absent Absent Absent an 40 Arm V Hectocotylus Absent Absent Absent ae 41 Spermatophores Present Present Present ee 42 DGDA, Number Single Single Single te 43 DGDA, Location Not in coel Not in coel Not in coel aoe 44 Digestive Gland Fused Fused Fused oe 45 Gonad: Coel. Coverage Mostly Cov. Mostly Cov. Mostly Cov. si 46 Posterior Salivary Gland Post.to Post.to Post.to He 47 Intestine vs Vena Cava Equivocal Dorsal Ventral -- 48 Gill Filaments Both Free Both Free Both Free eh 49 Head Width 1.00-1.49 0.50-0.99 0.50-0.99 50. Long. Mantle Muscle Present Present Present a 51 Arm Orientation Lateral Lateral Lateral si agreement among them in reconstructing specific character states provides some confidence in the reconstruction. The Octopodiformes clade, which is most consis- tently resolved among the three trees, resulted in the most consistent reconstructions. The decapod clade, which was totally unresolved in the morphological tree considered here and was inconsistently resolved between the two mol- ecular trees, resulted in many disagreements in reconstruct- ed character states. The deeper node for ancestral neo- coleoids was most noteworthy because of the large number 186 AMER. MALAC. BULL. 15(2) (2000) Table 3. Character state inferences as in Table 1, but for the ancestral Decapodiformes nodes. Character Morphology Actins COI consensus 1 Phragmocone Equivocal Absent Equivocal - 2 Proostracum Present Present Equivocal ~ * 3. Median Field Equivocal Narrow Equivocal - 4 Shell Number l 1 1 4 5 Fins Present Present Present _ 6 Fin Cartilage Base Only Base Only Base Only ae 7 ~~ Buccal Crown Present Present Present An 8 Arms II Unmodified Unmodified Unmodified nm 9 Arms IV Tentacles Tentacles Tentacles as 10 Suckers Present Present Present ne 11 Sucker Stalks Base & Neck Base & Neck Base & Neck a 12. Sucker symmetry Bilateral Bilateral Bilateral i 13 Sucker Rings Horny Horny Horny a 14. Arm Trabeculae Present Present Present a 1S Arm Protective Membrane Equivocal Present Present - 16 Arm II Sucker Series Equivocal Equivocal Equivocal - 17 ArmI Web Present Absent Present 18 Arm V Web Absent Absent Absent am 19 Statocyst Outer Capsule Absent Absent Absent #% 20 Nephridial Coelom 1 l | 7s 21 Visceropericardial Coelom Extensive Extensive Extensive as 22 Dorsal Mantle Cavity Absent Absent Absent Ee 23 Nidamental Glands Equivocal Present Present * 24 Crop Absent Absent Absent ae 25 Branchial Canal Equivocal Present Absent -- 26 Mantle Septum Equivocal Continuous Continuous i 27 Mantle Adductor Present Absent Equivocal -- 28 Funnel Valve Present Present Present sates 29 Nuchal Cartilage Present Present Present aie 30 Cornea Absent Absent Equivocal = 31 Right Oviduct Present Present Equivocal = 32 Oviducal Gland Symmetry Bilateral Bilateral Bilateral = 33 Oviducal Gland Position Terminal Terminal Terminal *m 34 Photosensitive Vesicle Equivocal In Cart. In Cart. 7 35 Superior Buccal Lobe Br.Separate Br.Separate Br.Separate ae 36 Subfrontal Lobes Equivocal Absent Absent . 37 Arm-Mantle Muscle Absent Absent Absent sai 38 Horizontal Arm Septae Absent Absent Absent ae 39 Arm IV (III) Hectocotylus Absent Absent Absent am 40 Arm V Hectocotylus Absent Present Equivocal -- 41 Spermatophores Present Present Present * 42 DGDA, Number Paired Paired Paired oe 43 DGDA, Location Nephrocoel Nephrocoel Nephrocoel ree 44 Digestive Gland Fused Fused Fused ee 45 Gonad: Coelomic Coverage Covered Covered Covered ee 46 Posterior Salivary Gland Post.to Post. to Post. to a 47 Intestine vs Vena Cava Equivocal Dorsal Ventral -- 48 Gill Filaments Both Free Both Free Both Free ne 49 Head Width 1.00-1.49 0.00-0.49 Equivocal -- 50 Long. Mantle Muscle Present Present Present ae 51 Arm Orientation Anterior Anterior Anterior ae of characters (19) for which there was no consensus among reconstructions, a result of the many characters for which reconstructions on individual trees were equivocal. Reconstruction of ancestral character states using parsimony is a three-step process, including down-tree opti- mization, up-tree optimization, and final optimization rec- onciling the previous two steps (see Box | in Cunningham et al., 1998 for a more complete explanation). As a result, changes in resolution of a subclade can have repercussions at nodes much deeper in a cladogram. For example, if the VECCHIONE ET AL.: ANCESTRAL STATES IN NEOCOLEOID CEPHALOPODS 187 traditional, but controversial, suborder Sepioidea (= Sepiidae + Sepiadariidae + Sepiolidae + Idiosepiidae + Spirulidae) is considered to be a subclade within the Decapodiformes on the morphological cladogram used here, change in character states inferred at the ancestral decapod and neocoleoid nodes may (e. g., character 3, median field of the proostracum, changes from absent to equivocal) or may not (e. g., character 2, presence of the proostracum, is unchanged) occur, depending on the distri- bution of character states at the terminal branches. Parsimony-based reconstruction is also sensitive to assump- tions about rates of evolution and probabilities of gain and losses (Cunningham ef al., 1998), questions about which very little information exists for neocoleoid cephalopods. It is encouraging that some of these reconstructions are consistent with other sources of information. For instance, the reconstructed presence of fins, suckers, and arm trabeculae agree with inferences from the fossil record (Bandel and Leich, 1986). However, inability of parsimony to reconstruct the phragmocone as present based on the cur- rent phylogenetic hypotheses is troubling. A phragmocone is present in cuttlefish and Spirula, as well as in the extant outgroup, Nautilus. Additionally, phragmocones are known from fossil coleoids, including the belemnoid outgroup and early spirulids. It seems unlikely that such a complex struc- ture evolved independently in all of these groups. Therefore, ancestral decapods and neocoleoids almost cer- tainly possessed a phragmocone. Vampyroteuthis also has a structure of unknown function that could be a remnant of the siphuncle from a phragmocone (Young and Vecchione, 1996), indicating the possibility that ancestral octopodi- forms may also have retained a phragmocone. The recon- structed state of this character as either absent or equivocal, together with the reconstructed state of the digestive gland, which contradicts embryological evidence (Appendix 1), greatly reduced our confidence in parsimonious reconstruc- tions of ancestral character states based on our current knowledge of cephalopod phylogeny. Character reconstruction requires known phyloge- netic relationships. The analyses presented here, however, involve three trees depicting somewhat different phyloge- netic relationships. The reconstruction of morphology in Appendix 3 is based on the assumption that the following relationships are correct: (1) The Belemnoidea and Neocoleoidea are sister groups (Young ef al., 1998). The apomorphic character states of the Neocoleoidea are: A. presence of suckers; B. absence of a nacreous layer in the shell; C. presence of fins. This foundation, while presently convincing, requires confirmation. (2) The Octopodiformes and Decapodiformes are monophyletic sister groups. Monophyly of the decapods was supported by a morpho- logical cladistic study (Young and Vecchione, 1996) and molecular studies (Bonnaud ef al., 1997; Carlini, 1998; Carlini and Graves, 1999). Morphological support for monophyly was weak. The sole unambiguous morphologi- cal character found to unite the decapods was the modifica- tion of the fourth pair of arms into tentacles (belemnoids such as Jeletzkya and Belemnotheutis had 10 equal arms, presumably the primitive condition) although they share a variety of characters that could not be polarized. The mono- phyly of the Octopodiformes (Octopoda + Vampyro- morpha; see Young et al., 1998 for discussion of the proper name of this clade) was supported morphologically by: A. the shared outer capsule of the statocyst; B. modification of the second pair of arms; C. the position of the superior buc- cal lobe of the brain. Extant octopods have lost one pair of arms but the lost pair, apparently, is not the tentacles (arms IV), but rather arms II, which became retractile filaments in the Vampyromorpha; this problem is discussed in more detail by Young and Vecchione (1996; 1999) and Vecchione et al. (1999). (3) The Vampyromorpha and Octopoda are sister groups within the Octopodiformes. This genealogy has now been confirmed by separate cladistic studies of morphological and molecular data (Young and Vecchione, 1996; Bonnaud et al., 1997; Carlini and Graves, 1999). (4) The Cirrata and Incirrata are sister groups within the Octopoda. This relationship has been supported by mor- phology (Young and Vecchione, 1996; Voight, 1997). A sis- ter-group relationship betweeen the cirrates and incirrates is not supported by COI data (Carlini and Graves, 1999). This relationship was not adequately tested by the actin data in Carlini (1998) because few cirrates were included in the analysis due to difficulties in cloning cirrate actin DNA. However, the few cirrate taxa sampled for actin genes sug- gest a sister-group relationship between cirrates and incir- rates. Furthermore, monophyly of the Octopoda is support- ed by both actin and COI. Although reconstruction of ancestral character states is, of necessity, speculative (Frumhoff and Reeve, 1994), Shultz et al. (1996) concluded that such inferences can be remarkably robust. We have only begun the process of reconstruction here. Our understanding of coleoid evolu- tion needs: (1) addition of characters to the list presented here, (2) resolution of the phylogenetic relationships among the decapods, the cirrates, and the incirrates and (3) greater knowledge of the developmental history of these characters in the embryos of all families considered. The latter will greatly increase our ability to define, assess, and polarize characters and clarify reconstructions that are presently ambiguous. ACKNOWLEDGMENTS This research was funded in part by a grant/cooperative agree- ment from the National Oceanic and Atmospheric Administration, project 188 AMER. MALAC. BULL. 15(2) (2000) #R/MR-S51, which is sponsored by the University of Hawaii Sea Grant College Program, SOEST, under institutional grant No. NA86RG0041 from the NOAA Sea Grant Office, Department of Commerce. UNIHI- SEAGRANT-JC-99-03. LITERATURE CITED Bandel, K. and H. Leich. 1986. Jurassic Vampyromorpha (dibranchiate cephalopods). Neues Jahrbuch fur Geologie und Palaontologie Monatshefte 1986:129-148. Bonnaud, L., R. Boucher-Rodoni, and M. Monnerot. 1997. Phylogeny of cephalopods inferred from mitochondrial DNA sequences. Molecular Phylogeny and Evolution 7:44-54. Carlini, D. B. 1998. The phylogeny of coleoid cephalopods inferred from molecular evolutionary analyses of the cytochrome c oxidase I, muscle actin, and cytoplasmic actin genes. Doctoral Dissertation, College of William and Mary. 273 pp. Carlini, D. B. and J. E. Graves. 1999. Phylogenetic analysis of cytochrome c oxidase I sequences to determine higher-level relationships within the coleoid cephalopods. Bulletin of Marine Science 64:57-76. Carlini, D. B., K. S. Reece, and J. E. Graves. 2000. Actin family gene evo- lution and the phylogeny of coleoid cephalopods (Mollusca: Cephalopoda). Molecular Biology and Evolution 17:1353-1370. Cunningham, C. W., K. E. Omland, and T. H. Oakley. 1998. Reconstructing ancestral character states: a critical reappraisal. Trends in Ecology and Evolution 13:361-366. Donovan, D. T. and M. D. Crane. 1992. The type material of the Jurassic cephalopod Belemnotheutis. Palaeontology 35:273-296. Engeser, T. and K. Bandel. 1988. Phylogenetic classification of coleoid cephalopods. In: Cephalopods - Present and Past, J. Wiedman and J. Kullman, eds. pp. 105-116. Schweizerbart’ sche, Stuttgart. Foote, M. and J. Sepkoski. 1999. Absolute measures of the completeness of the fossil record. Nature 398:415-417. Frumhoff, P. C. and H. K. Reeve. 1994. Using phylogenies to test hypotheses of adaptation: A critique of some current proposals. Evolution 48:172-180. Haas, W. 1989. Suckers and arm hooks in Coleoidea (Cephalopoda, Mollusca) and their bearing for phylogenetic systematics. Abhandlungen des Naturwissenschaftlichen Vereins in Hamburg 28:165-185. Haas, W. 1997. Der ablauf der entwicklungsgeschichte der Decabrachia (Cephalopoda, Coleoidea). Palaeontographica B 245:63-81. Maddison, W. P. and D. R. Maddison. 1992. MacClade: Analysis of phy- logeny and character evolution. Version 3.02. Sinauer Associates, Sunderland, Massachusetts. 398 pp. Naef, A. 1921/1923. Fauna und Flora des Golfes von Neapel. Monograph, no. 35. Cephalopoda. Part I, Volume I, Fascicle I-II:1-917. Pickford, G. E. 1940. The Vampyromorpha, living-fossil Cephalopoda. New York Academy of Science, Series 2 2:169-181. Schultz, T. R., R. B. Cocroft, and G. A. Churchill. 1996. The reconstruc- tion of ancestral character states. Evolution 50:504-511. Southern, E. 1975. Detection of specific sequences among DNA fragments separated by gel electrophoresis. Journal of Molecular Biology 98:503-517. Vecchione, M., R. E. Young, D. T. Donovan, and P. G. Rodhouse. 1999. Reevaluations of coleoid cephalopod relationships based on modi- fied arms in the Jurassic coleoid Mastigophora. Lethaia 32:113- 118. Voight, J. R. 1997. Cladistic analysis of the octopods based on anatomical characters. Journal of Molluscan Studies 63:311-325. Young, J. Z. 1965. The centres for touch discrimination in Octopus. Philosophical Transactions of the Royal Society B 249:45-67. Young, J. Z. 1971. The Anatomy of the Nervous System of Octopus vul- garis. Claredon Press, Oxford. 690 pp. Young, J. Z. 1977. Brain, behavior and evolution of cephalopods. Symposia of the Zoological Society of London 38:377-434. Young, R. E. and M. Vecchione. 1996. Analysis of morphology to deter- mine primary sister taxon relationships within coleoid cephalopods. American Malacological Bulletin 12:91-112. Young, R. E. and M. Vecchione. 1999. Morphological observations on a hatchling and a paralarva of the vampire squid, Vampyroteuthis infernalis Chun (Mollusca, Cephalopoda). Proceedings of the Biological Society of Washington 112:661-666. Young R. E., M. Vecchione, and D. T. Donovan. 1998. The evolution of coleoid cephalopods and their present biodiversity and ecology. South African Journal of Marine Science 20:393-420. Date of manuscript acceptance: 18 May 2000 VECCHIONE ET AL.: ANCESTRAL STATES IN NEOCOLEOID CEPHALOPODS 189 APPENDIX 1. Comments on individual characters. 1. Phragmocone. States: 0- Present; 1- Absent. Young and Vecchione (1996) considered only the character “siphuncle.” We expand the definition here to “phragmocone,” which includes a siphuncle. The presence of a calcareous phragmocone in belemnoid fossils and Nautilus indicates that this structure was present in the ancestral coleoid. Among Neocoleoidea only sepiids and Spirula have a phragmocone. Since the highly complex phragmocone is unlikely to have arisen anew in neo- coleoid evolution, we consider this an irreversible character. Pickford (1940) and Young and Vecchione (1996) have described what could be a remnant of a siphuncle in Vampyroteuthis. If correct, this supports the presence of a phragmocone in the early octopodiform. This possibility, however, is uncertain so the reconstruction of the early octopodiform, for now, must be “absent.” 2. Proostracum. States: 0- Present; 1- Absent; 2- Entire. The coleoid proostracum (i. e., the dorsal remnant of the living chamber of the pre-coleoid) is considered to be homologous with the gladius, except conus, in decapods and Vampyroteuthis. A proostracum does not exist in sepiids or Spirula. Some sepiolids (e. g., Rossia) have a small anterior gladius that appears to represent the anterior end of the proostracum based on its embryonic development (Naef, 1921-1923). Other sepiolids and sepiadariids that lack a gladius obviously lack a proostracum. In octopods the internal shell, the apparent homologue of the gladius, is transversely elongate; it is single in cirrates and divided, when present, in incirrates. The structure of the octopod shell, however, cannot be related to the sub- divisions seen in the gladius and the state of the proostracum in octopods coded as a “?”. A proostracum is present in belemnoids. The complete living chamber of Nautilus is given the state “Entire.” A narrow proost- racum is present in the fossil sepioid Groenlandibelus and, therefore, absence in Sepia and Spirula can be interpreted as a loss. 3. Median field width of proostracum. States: O0- Broad; 1- Narrow; 2- Unrecognizable; 3- Absent; 4- Entire (Nautilus). In many extant coleoids, the gladius has a three-part structure: a medial field (= rhachis), lateral fields (= wings) and conus fields (these often form part of the terminal primary conus but may extend well anterior of the conus proper). A primary conus can be present or absent. The primary conus and conus fields are presumably remnants of the phragmocone while the rest of the gladius represents the proostracum. The anterior portion of the gladius of extant coleoids consists solely of the median field, which can be narrow (teuthoids and some sepioids) or broad in Vampyroteuthis. Belemnoids also have a broad proostracum, but its relationship to the median field is also uncertain so this has been coded as **?”. 4. Shell number. States: 0- None; 1- One; 2- Two. The shell can be single (i. e., teuthoid gladius) or double (incirrate shell) or absent (some sepiolids, incirrates). The double break in the mantle musculature in the position of stylets in the two incirrate families that lack stylets clearly indicates that the double state is ancestral to these families. The single condition as in cirrates clearly is the ancestral state as the double state in incirrates is derived from a single shell sac during embryology (Naef, 1921-1923). This interpretation contrasts with that of Voight (1997) that the double shell was ancestral to the Octopoda. We therefore consider this an ordered character and, as such, reconstruct the single con- dition as the ancestral incirrate condition. 5. Fins. States: 0- Present; 1- Absent. Fins are unknown in belemnoids (Young ef al., 1988), but are found throughout the decapods, as well as in Vampyroteuthis and the cirrate octopods, but are absent in incirrates. Vampyroteuthis has two pairs of fins. In the youngest speci- mens a juvenile fin is present. As the animal grows, a second pair of fins develops anteriorly, which persist into the adult stage; the juvenile fins are resorbed. Young and Vecchione (1996) presented evidence that the juvenile fin is the homologue of other cephalopod fins. MacClade recon- structs the ancestral incirrate condition as equivocal. However, fin folds are present in incirrate embryology (Naef, 1921-1923) suggesting that early incirrates had fins. In addition, fins are present in the fossil Palaeoctopus newboldi, which has been interpreted as an incirrate (Engeser and Bandel, 1988; Young ef al., 1998). 6. Fin cartilage. States: O- Base only; 1- Base and core; 2- Absent. Both Vampyroteuthis (juvenile fin) and the cirrates have an extensive core of flexible cartilage extending through half or more of the fin (Young and Vecchione, 1996). This is absent from decapods, which have a cartilage at the base of each fin that doesn’t penetrate the fin core. Incirrates lack fins and are coded as ‘“‘?”. Because the Decapodiformes are considered an unresolved bush in the morphological consensus tree, MacClade cannot determine whether the uniform decapod condition ocurred at the node of the bush or earlier. 7. Buccal crown. States: 0- Oral arms; 1- Present; 2- Absent. The buccal crown is present in decapods and absent in octopodiforms. It is thought to be homologous with the oral arms of Nautilus (Young and Vecchione, 1996). We therefore consider the states to be ordered: oral arms - buccal crown - absent. With this constraint, the ancestral neo- coleoid condition is reconstructed as “present.” The state in belemnoids is unknown. However, since the arm crown of belemnoids is known and oral arms are not present the options are either state | or 2. 8. Arms II. States: 0- Unmodified; 1- Modified; 2- Absent. Arms II are modified in Vampyroteuthis and lost in octopods. The likelihood that these assumptions are correct is discussed in Young and Vecchione (1996). On the assumption that modification preceeded loss, we order the states, which then reconstructs the ancestral octopodiform as “modified.” 9. Arms IV. States: 0- Unmodified; 1- Modified (tentacles). Arms IV are modified as tentacles in decapods but are unmodified in octopodiforms. In several decapod species adults have only eight arms due to the loss of the tentacles during ontogeny. The eight-armed condi- tion in these decapods is clearly secondary as tentacles are present in par- alarvae. 10. Suckers. States: 0- Absent; 1- Present. Nautilus lacks suckers as, apparently, did the belemnoids. Donovan and Crane (1992) reported possible suckers in Belemnotheutis, but these are more likely the muscular bases of hooks (Young ef al., 1998). Because all living coleoids have suckers, we consider the presence of suckers to be a neocoleoid synapo- morphy as suggested by Engeser and Bandel (1988) and as reconstructed by MacClade. 11. Sucker stalks. States: O- Base and neck; 1- Base and plug; 2- Cylinder. The suckers of decapods have stalks that are cone-like and ter- minate in a constricted, narrow neck. Octopods have broad, cylindrical sucker stalks. Vampyroteuthis has stalks that are unique (state 1), but in some ways intermediate between the decapod and octopod conditions. Polarity among these character states is presently undetermined. 12. Sucker symmetry. States: 0- Radial; 1- Bilateral. Decapods are characterized by having bilaterally symmetrical suckers. Octopods and Vampyroteuthis have suckers that are radially symmetrical, and reconstruction of the ancestral octopodiform is therefore with radial suck- ers. Polarity among these character states is presently undetermined. 13. Sucker rings. States: O- Cuticular; 1- Absent; 2- Horny. Octopods have cuticular sucker rings; decapods have horny sucker rings often modified into hooks, and Vampyroteuthis lacks sucker rings. Polarity among these character states is presently undertermined. 14. Arm trabeculae. States: O- Present; 1- Absent. Trabeculae (including their apparent homologues, cirri) are present in many decapods, Vampyroteuthis and the cirrate octopods. Within decapods, however, trabeculae are often reduced or absent. Belemnoids and Nautilus were coded as “?”; belemnoids because of uncertainty and Nautilus because the arms are so different that the character is not applicable. Haas (1989) has proposed that belemnoid arm hooks could be homologous with 190 AMER. MALAC. BULL. 15(2) (2000) neocoleoid trabeculae. 15. Arm protective membranes. States: 0- Present; 1- Absent. Although protective membranes are present between distal trabeculae on the arms of Vampyroteuthis, they are completely absent from octopods and from one decapod family (Sepiolidae); all other decapods examined possess protective membranes although they are often reduced. We con- sider that the absence from the sepiolids is likely due to secondary loss (as reconstructed in the molecular data sets). Under this constraint, the ances- tral decapod, octopodiform and neocoleoid are reconstructed as “present.” 16. Armature series on arms III. States: 0- One; 1- Two; 2- Four; 3- > Four. In Octopodiformes the suckers are either one or two series, or occasionally a combination of these states. However, many incirrates with two series of arm suckers have a single series as hatchlings. This supports a reconstruction of a single series in the ancestral octopodiform. Recent decapods have their armature in two, four or sometimes more than four series. In decapods, as in octopods with two series, the suckers are stag- gered, suggesting a sequence in which single series become double series and double series become quadruple series (as noted by Naef, 1921-1923). Unfortunately there are no concrete data to support this hypothesis and the condition in the ancestral decapod must be reconstructed as equivocal. In Vampyroteuthis and the cirrates, there are two trabeculae/sucker on the sucker-bearing portion of the arms. In decapods where two sucker series exist, there is one trabecula/sucker,; where four sucker series are present, there is one trabecula/two suckers. Trabeculae, therefore, appear to be pro- gressively lost as the number of sucker series increases. That is, the trabec- ula is lost from the side of the sucker that no longer is adjacent to the mar- gin of the arm. This is a scenario that would be expected as one series becomes multiple series. 17. Arm webs, dorsal. States: 0- Absent; 1- Present. Well-devel- oped webs between the dorsal pair of arms are present in most octopodi- forms and some decapods. We consider the dorsal sector to be representa- tive of web development between the dorsal six arms. 18. Arm webs, ventral. States: O- Absent; 1- Present. A web between the ventral arms is present in all major octopodiform lineages, but is uniformly absent in decapods. 19. Outer capsule of statocyst. States: O- Absent; 1- Present. An outer capsule is present in the statocyst of Vampyroteuthis and all octopods. It is absent in decapods and Nautilus. 20. Nephridial coelom. States: 0- Two; 1- One. Octopodiforms and Nautilus have two nephridial coeloms (one pair), whereas one (fusion) coelom is uniform in decapods. This character was not used by Young and Vecchione (1996) to support decapod monophyly due to difficulties in determining if fusion occurred more than once. 21. Visceropericardial coelom. States: O- Extensive; 1- Reduced. An extensive visceropericardial coelom is found in Nautilus, the decapods and Vampyroteuthis. Reduction in octopods is a synapomorphy in this clade. 22. Dorsal mantle cavity. States: 0- Absent; 1- Present. The dor- sal mantle cavity as defined by Young and Vecchione (1996) is an autapo- morphy in the Octopoda. This structure bears considerable resemblance to the large nuchal cavity of Spirula and we suspect that both were derived in the same manner. The nuchal cavity is a space where the dorsal mantle articulates with the head. In most decapods the gladius (= proostracum) extends to the anterior tip of the mantle where a cartilaginous reinforce- ment of the shell sac articulates with the nuchal cartilage. Spirula lacks a proostracum and the Recent octopod shell doesn’t reach the anterior man- tle margin, yet a proostracum was probably present in the ancestors of both groups (see character no. 2). Perhaps the proostracum was progres- sively reduced over evolutionary time, and as the proostracum receeded posteriorly, the nuchal cavity increased accordingly. The later anterior fusion of the head and mantle margin in the Octopoda formed the dorsal mantle cavity. This scenario is further supported by the reduced gladius found in the Idiosepiidae. Here the gladius is absent both anteriorly and posteriorly. Anteriorly, an expanded nuchal cavity is also present. The convergent condition in Spirula and Idiosepius occurred without the dorsal mantle and head fusing. For species in which the head-mantle fusion has occurred without a posterior regression of the gladius (e.g., Sepiola, Vampyroteuthis), the nuchal cavity has been obliterated. 23. Nidamental glands. States: 0- Absent; 1- Present. These glands produce some of the external coatings on eggs as they are spawned. The glands are found in Nautilus and nearly all decapods except the Enoploteuthidae. Because of their absence in the latter family, reconstruc- tions of the ancestral neocoleoid and decapod based on morphology are equivocal. However, we consider that the absence in the Enoploteuthidae, which spawn individual eggs strung together in a single gelatinous strand rather than gelatinous or encapsulated egg masses, is probably a secondary loss (as predicted by the molecular trees). It is unlikely that nidamental glands evolved twice (i. e., decapods and Nautilus). With this constraint we reconstruct the ancestral decapod as having nidamental glands. 24. Crop. States: 0- Present; 1- Absent. A crop, defined here as a swelling or diverticulum of the esophagus, is present in Nautilus, Vampyroteuthis and most octopods. The loss of the crop, therefore, is an apomorphy for the Decapodiformes 25. Branchial canal. States: 0- Absent; 1- Present; 2- Secondary loss. A branchial canal is present in teuthoids, Vampyroteuthis and in the incirrate octopods, but absent from Nautilus, some decapods (sepioids) and the cirrates. The cirrates, however, have highly modified gills, which likely resulted in the loss of the canal independent of the loss in some decapods; the cirrate condition, therefore, is coded as a different state. Because the condition in the Decapodiformes is polymorphic, the ancestral decapod state 1s equivocal. 26. Median mantle septum. States: 0- Absent; 1- Present and continuous; 2- Present but open posteriorly; 3- Present as a blood vessel only. The visceral mass of Spirula is highly distorted by the presence of a coiled phragmocone and the mantle septum is absent except for the pres- ence of the median mantle artery. Because this artery normally passes along the anterior margin of the septum, we consider the artery to repre- sent the mantle septum in Spirula. This makes the presence of the septum uniform within the decapods as it is in the octopods. It is absent in Vampyroteuthis. The septum is open in all octopods except the cirrate Grimpoteuthis glacialis. 27. Mantle adductor. States: 0- Absent; 1- Present. The mantle adductor 1s uniformly present in the octopods and in the sepiolid decapods. The sepiolids have a strong effect on the reconstructions because decapod relationships are unresolved. We consider that the mantle adductor in octopods and sepiolids is a result of convergence (as suggested by the molecular actin tree). Muscles are typically associated with the mantle septum and hypertrophy of these muscles into a mantle adductor in sepa- rate lineages could easily occur. Under this constraint, the ancestral deca- pod condition is reconstructed as “present.” 28. Funnel valve. States: 0- Present; 1- Absent. The funnel valve is present in Nautilus, Vampyroteuthis and most decapods. Although it is absent among decapods in some cranchiids and Planctoteuthis, the pres- ence of the funnel valve in some members of these families suggests that this is a secondary loss. 29. Nuchal cartilage. States: O- Present; 1- Absent. Although Nautilus doesn’t offer any information on this structure, the nuchal carti- lage is present in Vampyroteuthis and nearly all decapods (it is absent in some sepiolids and sepiadariids; its presence in some members of these families indicates that these are secondary losses). 30. Cornea. States: 0- Absent; 1- One part; 2- Two part. A cornea is absent in the cirrate octopods and Vampyroteuthis as well as many decapods. A two-part cornea is present in the incirrates. A one-part cornea is present in some decapods. These two types of corneas are considered to be independent derivatives of the eyelid. We did not find any corneas in the cirrates examined. However, Opisthoteuthis possesses inner, muscular, VECCHIONE ET AL.: ANCESTRAL STATES IN NEOCOLEOID CEPHALOPODS 19] pigmented eyelids in the form of convex, overlapping horizontal mem- branes. These have the same form as the clear, fixed corneas of the incir- rates and can be interpreted either as the forerunner or remnant of a cornea. 31. Right oviduct. States: 0- Absent; 1- Present. The right oviduct is present in many decapods (it is absent in loliginids and sepi- oids), Vampyroteuthis, incirrate octopods and Nautilus. This means that the oviducts were paired since the left oviduct is present in all neo- coleoids. The absence of the right oviduct in cirrate octopods and some decapods represents convergence. 32. Oviducal gland symmetry. States: O- Radial; 1- Bilateral; 2- Asymmetrical. The oviducal glands are radially symmetrical in the octopodiform lineage and bilaterally symmetrical in the decapod lineage but asymmetrical in Nautilus. The character therefore cannot be polarized. 33. Position of oviducal gland. States: 0- Terminal; 1- Subterminal. The oviducal gland is located in the terminal position in decapods, Vampyroteuthis and Nautilus, but is subterminal in the Octopoda. 34. Photosensitive vesicles. States: 0- Within cephalic cartilage; 1- Above funnel; 2- On stellate ganglia. In some decapods the photosensi- tive vesicles lie on the optic stalks of the brain, and in others they have moved off the stalks but lie mostly within the confines of the cephalic car- tilage with nerves running to the optic stalks. The photosensitive vesicles lie on the stellate ganglia in octopods and their nerves pass through this ganglion and into the pallial nerve (J. Z. Young, 1977), which leads to the brain. In Vampyroteuthis, they lie just dorsal to the funnel and their nerves follow the posterior funnel nerve toward the brain. Thus, the three major lineages have vesicles in different localities. We suspect that nerves from photosensitive vesicles of all cephalopods enter the brain in the region of the optic stalk and that the vesicles originally evolved at this location. If so, the decapod state would be the plesiomorphic state for the Neocoleoidea. Unfortunately, this cannot be confirmed at present and this reconstruction is equivocal. The reconstruction further assumes that the unknown state in Thysanoteuthis will conform to that of other decapods. 35. Superior buccal lobes. States: O- Broadly separated; 1|- Adjacent; 2- Fused. The superior buccal lobes are far removed from the brain in Nautilus and decapods but are adjacent to the brain in Vampyroteuthis and are fused with the brain in octopods, with the greatest compaction occurring in the cirrate octopods. The situation in Vampyroteuthis 1s actually somewhat more intimate than “adjacent”; the lateral edges of the superior buccal lobe and posterior buccal lobes lie within the same connective tissue covering. The state of this character strongly reflects the distance between the brain and the buccal mass. Young and Vecchione (1996) considered this to be an ordered character: separate - adjacent - fused. 36. Inferior frontal system of the brain. States: O- Absent; 1- Insipient; 2- Present. The inferior frontal system of incirrates deals with the use of chemotactile information from the arms (J. Z. Young, 1971). This system is composed of the posterior buccal, lateral inferior frontals, subfrontals and the median inferior frontal lobe (J. Z. Young, 1971). The system develops embryologically from the posterior buccal lobes (J. Z. Young, 1965) and is best developed in incirrate octopods, but is present in cirrates as well. In decapods only the posterior buccal lobes are present. In Vampyroteuthis, complexities of the posterior buccal lobes and their con- nections have been interpreted as an incipient inferior frontal system (J. Z. Young, 1977). We consider this to be an ordered character with the vampyromorph condition intermediate, as did Young and Vecchione (1996). The reconstruction further assumes that the unknown state in Thysanoteuthis will conform to that of other decapods. 37. Arm-mantle muscle. States: O- Present; 1- Absent. Special muscle bundles run between the bases of the dorsal arms and the dorsal, anterior end of the mantle in the octopods. This feature is a synapomorphy of the octopods and defines the dorsal head-mantle fusion peculiar to them. These muscles are not present in the head-mantle fusions of Vampyroteuthis or some decapods. 38. Horizontal arm septa. States: 0- Absent; 1- Present. Peculiar orally concave horizontal septa extend along the arms of all cirrate octopods and are found nowhere else. The arms of the incirrate bolitaenids have a somewhat similar arrangement but with different septal attach- ments; this condition was considered as a separate character state by Young and Vecchione (1996). Concave horizontal septa, therefore, is an apomorphy for the Cirrata. 39. Arm IV hectocotylizaton. States: O- Absent; 1- Present. Because of the loss of arm pair II in octopods (see character 8), arms IV are generally considered by students of neocoleoids to be the “third” pair of arms. Modification of one of the “third” arms (actually arms IV) for the transmission of spermatophores is a synapomorphy in the incirrate octopods. 40. Arm V hectocotylization. States: O- Absent; 1- Present. Modification of one of the ventral arms (arms V are often referred to as arms IV, not counting the tentacles in decapods as an arm pair) for the transmission of spermatophores occurs among many decapods, but not all, and is absent in other lineages. 41. Spermatophores. States: 0- Present; 1- Sperm packets; 2- Encapsulated coil. Typical spermatophores with an ejaculatory apparatus are found throughout the coleoids with the exception of the cirrate octopods. The presence of special sperm packets in cirrates, apparently a secondary simplification, is an apomorphy in this group. 42. Digestive-gland-duct appendages (DGDA), number. States: 0- Single; 1- Paired. In nearly all decapods the DGDA are spread along the ducts between the digestive gland and the caecum. In Vampyroteuthis and the octopods they are fused and compacted against the digestive gland. In a few genera of decapods (e. g., Batoteuthis, various cranchiids) com- paction exists but not fusion. Because the appendages are lacking in Nautilus, polarity is uncertain. 43. DGDA, location. States: O- In nephridial coelom; 1- Not in nephridial coelom. The DGDA in the decapods are located within (actual- ly surrounded by) the dorsal sac of the nephridial coelom but they are sep- arate from this coelom in the octopodiform lineage. The states cannot be polarized. 44. Digestive gland. States: 0- Many; I- Paired; 2- Fused. Digestive glands are paired only in Sepia, Spirula and Sepiadarium. Nautilus has numerous digestive glands. Parsimony indicates the single state to be ancestral. However, embryology clearly indicates the paired origin of this structure in some species having a single organ (e.g., Loligo, Octopus). We therefore consider that paired glands are the ancestral neo- coleoid state in spite of their reconstruction as fused. If this is correct, sec- ondary fusion to produce a single digestive gland has occurred in more than one lineage. 45. Gonad: coelomic covering. States: O- Mostly covered; 1- Less than 50% covered. In most neocoleoids the gonad lies suspended in the visceropericardial coelom (virtually 100% covered) although lined by the coelomic epithelium. In incirrates much less, but in excess of 50%, is covered. A synapomorphic condition exists in the cirrates in which less than 50% of the gonad lies within the coelom. 46. Posterior salivary glands. States: O- Absent; 1- Posterior to cephalic cartilage; 2- On or in buccal mass. The posterior salivary glands are usually found posterior to the brain and the cephalic cartilage. Only in the cirrate octopods are they found anterior to the brain and on or in the buccal mass. Therefore, this latter state is synapomorphic in cirrates. 47. Intestine: position relative to vena cava. States: 0- Ventral; 1- Dorsal/anterior. The intestine either runs dorsal/anterior to the vena cava (Vampyroteuthis, sepioids, loliginids) or ventral to it (oegopsids and octopods). This character exhibits homoplasy (Young and Vecchione, 1996) and outside the Octopoda ancestral states are equivocal. 48. Gill filaments. States: O- Both free; 1- Outer attached; 2- 192 Both attached. The tips of the gill filaments are free in some taxa (Nautilus, many decapods, Vampyroteuthis, some cirrates). Alternatively, one (Onychoteuthididae, Ocythoidae) or both (many incirrates, Opisthoteuthididae) filaments may be attached to the gill base. 49. Head width proportional to eye diameter. States: 0- 0-0.49; 1- 0.5-0.99; 2- 1.0-1.49; 3- 1.5-1.99; 4- 2.0-2.49; 5- 2.5-2.99; 6-3.0-3.49; 6- 3.5-3.99; 7- 4.0-4.49; 8- 4.5-4.99. Young and Vecchione (1996) attempted to quantify the head width by using the eye diameter as a size standard against which to measure head width. They compared the eye diameter to the head width measured between the extremities of the lenses and expressed it as a ratio. The method was only partially satisfactory because animals with dorsally tilted eyes added a complication and, in some, the eye size is simply a poor size standard for judging head width. Because of these problems, results of reconstruction must be taken cautiously. A gen- eral pattern, nevertheless, exists with many of the oegopsids (and Spirula) having narrow heads, most sepioids, loliginids and Vampyroteuthis having intermediate head widths and octopods having broad heads. The head of Vampyroteuthis is actually rather broad but since the eyes are especially large in this species our measure doesn’t reflect head size very well. Head width, in addition, seems to be a good, but not absolute, indicator of body width. 50. Longitudinal mantle muscles. States: O- Present; 1- Absent. Mantles of many decapods are composed mostly of circular and radial muscles but thin, discontinuous layers of longitudinal muscles are also AMER. MALAC. BULL. 15(2) (2000) present on the outer surface of the mantle especially near the anterior and posterior ends of the mantle. All groups examined, with the exception of a few families of decapods, had longitudinal muscles. 51. Arm orientation. States: O- Lateral; 1- Anterior. In the relaxed position, the arms of some cephalopods extend laterally away from the head while in others they extend anteriorly. The arms of all cephalopods, however, are very muscular and capable of moving through a wide range in orientation. We have searched for anatomical correlates (e. g., how the arms relate to the buccal mass) of the two basic orientations in order to quantify the character states. We have, unfortunately, been unsuccessful; as a result this character has not been adequately surveyed. Nevertheless, there seems to be little question that a basic difference in arm orientation exists between the octopodiform lineage and the decapods. When the arms of the former (typically oriented laterally) bend forward, their base near the buccal mass generally extends first laterally then anteri- orly as the arm curves forward. In contrast, one usually finds that, in decapods, arms are typically directed forward and when they are directed laterally the orientation at the base generally extends first anteriorly then laterally as the arm curves aborally. The lateral orientation is most obvious in Vampyroteuthis and the cirrate octopods. The difference is less obvious in some of the muscular pelagic octopods such as Ocythoe. Nautilus tenta- cles and the preserved arms of some belemnoids (e. g., Belemnotheutis) show anteriorly oriented arms. APPENDIX 2. Matrix of morphological character states used for reconstructions. Explanation of numerical designations for character states is presented in Appendix |. Most of the material examined is listed in Young and Vecchione (1996). 0 l 2 Character 1 2 3 4 56789012345 6 7 8&9 O Bathyteuthidae 1 O 1 1 OO1O1101I210 2 I OO 1 Enoploteuthidae 1 O l 1 00101101200 1 0 OO 1 Gonatidae 1 O l 1 00101101200 2 0 00 1 Loliginidae 1 0 1] 1 00101101200 1 O 00 1 Ommastrephidae 1 O 1 1 00101101200 1 0 OO 1 Onychoteuthidae 1 O 1 1 00101101210 1 0 00 1 Sepiidae QO | 3 1 00101101200 2 1 00 1 Sepiolidae 1 (OL) (13) (OL) OO1O1101211 (12)(01) 00 1 Spirulidae O 1 3 Jt 00101101210 2 1 OO 1 Thysanoteuthidae 1 O 1 1 00101101200 1 O 00 1 Bolitaenidae 1 ot 2? 0 12220120011 O | 11 0 Octopodidae 1? 2? 2 12220120011 (OL) 1 11 (OL) Ocythoidae 1 1 ? O | 17220120011 1 0 O1 0 Cirroteuthidae 1? 2? - 01220120001 0 1 ll O Opisthoteuthidae 1? 2? 1 01220120001 0 | 11 O Vampyroteuthidae 1 O O 1 01210110100 O 1 ll O Nautilidae O 2 4 1 12072072??? ? 2? 20 O belemnoid 0 O ? LT 12200072??? , 2 92? 123 001 000 001 001 001 O01 001 O01 001 001 110 110 110 110 110 000 001 22? 3 4 5 4 5 6 78 9 012345678901234567 8 9 Ol 1 tt 00 0 011000010000102011 0 0 Ol 1 1 1 00 0 011000010010102011 0 0 Ol 1 tf 1 00 O 011000010011?02011 0 0 il 1 tf 1 00 0 101000010010102010 O (12) Ol 1 1 ft 00 0 011000010010102011 (01) O 11 1 1 t 00 0 011000010000102011 1 0 Ol 1 0 t 00 0 -— 101000010010101010 0 2 Ol 1 0 tf 10 (O01) 101000010000102010 0 1 Ol 1 0 3 00 0 001000010010101010 0 0 O01 1 ot ft 00 O- 0110???10010102011 0 0 11 O 1 2 It ft 21012220010?012011 2 3 00 O tL 2 IL Ft 210122200100012011 2 5 00 O 1 2 Il 1 210122200100012011 1 4 Ol (Ol) 2. 2 It t 000122201001012121 0 4 00 O 2 (12) 11) 1 0001?2201001012121 2 2 0 O t O 20 O 010011110000012010 0 2 00 0 0 ? 7 ? 012070710797279000? 0 9 Ol ns met A an GCAO Se OS ee a te BLY ? TEU VECCHIONE ET AL.: ANCESTRAL STATES IN NEOCOLEOID CEPHALOPODS 193 APPENDIX 3. Provisional reconstructions of character states for ancestral nodes based subjectively on evidence from ontogeny and paleontology as well as on morphological parsimony. Character number 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Ancestral coleoid 0) 0 2 l I ? ? 0 0) 0) ? ? ? ? ? ? ? ? 0 0) Belemnoid 0 0 ? l 1 ? v4 0 0 0 ? ? ? ? ? ? ? ? 0 0 Ancestral neocoleoid ) 0 ? I 0 ? l 0) 0 | ? y ? 0 0 ? | ? 0) 0) Ancestral decapod 0 0 Hi l 0 0 1 0) l 1 0 | 2 0) 0 ? 1 0 0 l Decapodiformes O/L O/L 1/2 1 0 0 1 0 | 1 0 l 2 O/L O/L 1/2) OL OO 0 | Ancestral octopodiform | 0 ? | 0 | 2 | 0 | ? 0 ? 0 0 0 | | | 0 Vampyromorpha 1 0 0 l 0) | 2 l 0) | l 0 l 0 0 0 1 | 1 0 Ancestral octopod l ? ? | 0) I 2 2 0 1 2 0 0 0) l 0 1 | I 0) Ancestral cirrate l ? ? 1 0) | 2 2 0 I 2 0 0 0) | 0 l I l 0) Cirrata | ? 2 l 0 l 2 2 0 l 2 0 0 0 1 0 1 | l 0) Ancestral incirrate | ? ? l 0) l 2 2 0 l 2 0) 0 | I 0) ] 1 1 0) Incirrata ] /? ? 0/2 1 ? 2 2 0 1 2 0 0 1 | Ol O/L O/L I O/1 Character number 21 22 23 24 25 26 2 28 29 30, 31 32 33 34 35 36 37 38 39,40 Ancestral coleoid 0 0 | 0) y ? ? 0 ? 0) | ui; 0 ? 0 ? ! 0 ? ? Belemnoid 0 0 | 0 ? 2 ? 0 2 0 I ? 0) ? 0 ? l 0) ? ? Ancestral neocoleoid 0 0 I 0 ? ? ? 0) 0 0 l ? 0) ? 0 ? I 0) 0 0) Ancestral decapod 0 0 l l ? l 0 0 0 0 l 1 0 0 0 0 | 0 0 0 Decapodiformes 0 0 O/L 1 O/L 1/3) O/L OO OL O/L OL I 0 0 0 0 l 0 0 O/1 Ancestral octopodiform 0 0) 0 0 | ? ? 0 0 0 l ) 0) ? I l 1 0 0 0 Vampyromorpha 0 0 0 0 | 0 ? 0 0 0 | 0 0 | I l | 0) 0) 0 Ancestral octopod 1 l 0 0 1 | 1 | | 0 I 0 | 2 2 2 0 0 0 0 Ancestral cirrate I l 0 0 2 | 1 1 l 0) 0 0 | 2 2 2 0) l 0) 0 Cirrata l ] 0 O/1 2 1/2 1 l l 0) 0 0 l 2 2 2 0 1 0 0 Ancestral incirrate l l 0 0 l | l | | 2 l 0 l 2 2 2 0 0 I 0 Incirrata l l 0 0) l I 1 | l 2 1 0 l 2 2 2 0) 0) l 0) Character number 4] 42 43 44 45 46 47 48 49 50 5] Ancestral coleoid ? ? ? ? 0 ? ? 0) ? 0) l Belemnoid ? ? 2 ? 0 ? 4 0 ? 0) 1 Ancestral neocoleoid 0) ? ? l 0 l ? 0) 2 0 l Ancestral decapod 0 I 0) | ) | ? 0 2 0 l Decapodiformes 0) “1 O 1/2 0 l O/L O/L 0-2 O/L I Ancestral octopodiform 0 0 1 2 0 | ? 0 2 0 0 Vampyromorpha 0 0 | 2 0 l 0 0 2 0 0 Ancestral octopod 0 0 | 2 0 | l ? 4 0 0 Ancestral cirrate l 0) l 2 l 2 l 2 4 0 0 Cirrata | 0 l 2 l 2 | 0/2. 4 0 0 Ancestral incirrate 0 0) l 2 0 l l z 4 0) 0 Incirrata 0) 0 l 2 0 l l 1/2 3-5 O 0/1 Concerning the concept of extinct classes of Mollusca: or what may/may not be a class of mollusks Ellis L. Yochelson Research Associate, Department of Paleobiology, National Museum of Natural History, Washington, DC 20560-0121, U.S. A. Abstract: In the notion of classes, all of whose members are extinct, at least two fundamental issues are intertwined: (1) what criteria should be used to assign a fossil to the Mollusca; and (2) what level of morphological distinctiveness is needed to establish a fossil as a representative of an extinct class. One may also ask how far the morphologic limits of extant classes should be extended. Several current textbooks recognize at least one class of extinct mollusks. There is no consistency in the number of extinct or extant classes in use, nor need there be while the field is still vigorously under discussion. Key Words: classification, evolution, extinction, fossils, morphology Class proposals for extinct mollusks, or fossils thought by some paleontologists to be mollusks, have been in the literature for about a century. Discovery of living examples of the Class Monoplacophora stirred both mala- cology and paleomalacology and the last four decades have seen new proposals of both extant and extinct classes. Study of specimens is far more informative than examina- tion of photographs. Notwithstanding that point, sketches are included herein, but only for the purpose of making the text less obtuse. This essay repeats ideas written earlier (Yochelson, 1978, 1979) and, all my remarks should to be preceded with “in my view” or “in my opinion.” Those fun- damentally important phrases are eliminated. only to save space. Classification is a subjective activity and the higher one moves up in the Linnaean hierarchy, the greater the degree of subjectivity. ‘“Authoritarianism is dangerous, especially for scientists, and the reader should approach my comments with a skeptical attitude” (Yochelson, 1979:324). ISSUES The more that has been learned of living animals, the more complex has become their classification. In 1758, Linnaeus did not use the category of phylum, but his equiv- alent has increased ten-fold. Over centuries, the relative importance assigned to various features by different work- ers has changed. For example, Molluscoidea was split off from Vermes and disappeared when Brachiopoda and Bryozoa came into use. The latter is now two phyla and some workers would also divide the former in two. The present concept of Mollusca is not necessarily sacrosanct. During one-quarter of a millennium, data on fossils have also dramatically increased and there is equally good rea- son to expect the number of higher-level taxa of extinct forms should also increase. Defining a class as the taxon rank below a phylum is correct, though not particularly helpful. Much of accep- tance or rejection in high-level systematics is based on the opinion of textbook authors. The fundamental question of whether extinct classes of mollusks are recognized is answered in the affirmative by reviewing paleontology texts of the last three decades. Consensus is that at least one extinct molluscan class, Rostroconchia (Pojeta et al., 1972) is recognized in the Paleozoic; it features in a major paper (Waller, 1998). “One” is not large, but it constitutes a dra- matic philosophical change from viewing Mollusca as con- taining only extant classes. Which extinct class proposals are “reasonable” is another issue. The number of classes recognized among phyla varies widely. Within classes, the number of orders, families, or genera also varies widely. There are no rules to follow on the number and content of extinct and extant classes. How one decides whether a genus or a group of genera should be recognized as of class rank is yet a differ- ent issue. A digression, which appears simplistic, 1s appropri- American Malacological Bulletin, Vol. 15(2) (2000): 195-202 195 196 AMER. MALAC. BULL. 15(2) (2000) ate. The Class Gastropoda constitutes mollusks that have undergone torsion; all extant classes are based on features of soft anatomy. In the absence of torted soft parts, there are no fossils one can state with 100% confidence are Gastropoda. A paleontologist compares the shell of a living gastropod with a fossil shell; similarity of modern and fos- sil hard parts is at such a high level of probability, that the concept that it is a probability is forgotten. Classification is based on probability, not certainty. For fossil mollusks to be considered an extinct class-rank taxon the hard part morphology should differ dramatically from that of other classes of Mollusca. To determine what features are dramatically different is as much a matter of art as it is of science. After defining species as a community of individuals “whose distinctive characters, are in the opinion of a competent specialist, suf- ficiently definite to entitle . . . a specific name” a fish spe- cialist at the British Museum (Natural History), added “Groups of higher or lower rank . . . can be defined in the same way” (Regan, 1925:D1). In the game of classification, the only rule may be appeal to authority, a most unscientific approach. Cladistics and molecular systematics may appear more rigorous, but they too are ultimately grounded on cer- tain assumptions; change the assumption and one changes the end result. As a further complication in studying fossil mol- lusks that might have had more than one hard part, effects of taphonomy should not be ignored (Yochelson, 1984). Bizarre forms have been described, because modifications by sorting, transport, wear, and solution were ignored. Only after the taxon is satisfactorily defined by reference to the hard part(s) should one begin to speculate on soft parts and how they might have functioned. Listing interpretation of unknown anatomical features as the least significant criteri- on may not be popular, but it is the only way I see to under- stand molluscan fossils that do not closely resemble Recent mollusks. Three examples of possible extinct classes may help dispel some of this miasma. These are most familiar to me, but other proposals are equally worthy of discussion. CLASS MATTHEVA My proposal of this extinct class (Yochelson, 1966), which contained one genus, was based on the co-occur- rence of two triangular-shaped, narrow pieces, each of which contained two deep cavities (Fig. 1). I was in error in suggesting the theoretical possibility of intermediate plates, for these were based on the presence of worn scraps. An assumption is that the cavities were for muscles and that two forms were anterior and posterior. I did not know which piece is anterior and which is posterior, and still do not know, but that uncertainly does not in any way affect Fig. 1. Mattheva. A. One piece, possibly anterior. B. Another piece, pos- sibly posterior. About three times natural size. (Yochelson, 1966, figs. 1 and 2). the morphological uniqueness of this fossil. Specimens occur with algal domes. A paleoecological assumption is that the organisms did not cling to the substrate, but the weight of the two pieces helped maintain it in place during times the water was flowing. Runnegar and Pojeta (1974) presented a reconstruc- tion of the Late Cambrian Matthevia, and used a then unde- scribed intermediate (Runnegar er al., 1979) to link it to an Early Ordovician polyplacophoran, Chelodes. That genus has only a single cavity and is broad and low, like the pro- file of a roof. I did not agree with their drawing, which resembled a seven-parted hedgehog (Yochelson 1978), nor did the latter suggestion of eight parts improve the recon- struction (Runnegar, 1983). There is evidence from fossils of only two thick heavy pieces, not the various slight modi- fications shown in their diagram. Runnegar et al. (1979) and Pojeta (1980) refuted my views, but it would be pedantry, to attempt to refute them here. Those not com- pletely turned off by squabble should concordantly examine these papers. One could expand the Polyplacophora to include Matthevia, strikingly different from all others, as is the approach in current literature. Alternatively, one could rec- ognize a class that could have had more than one hard part, YOCHELSON: EXTINCT MOLLUSCAN CLASSES 197 but whose parts do not resemble polyplacophoran plates of living taxa. My idea of a class has fallen on sterile ground, yet I remain mumpsimus. Because a paleontologist deals with hard parts, it is better to recognize smaller groups of more or less uniform “bauplan” than to have larger unwieldy groups. Gould (1989) used “shoehorn” for the process of cramming the wrong fossils into a taxon based on living material. Polyplacophora have been reported from the Early Cambrian (Yu, 1984; 1990). If these fossils are part of the class — which I doubt — Matthevia 1s far off the main line of polyplacophoran evolution. If plates of Preacanthochiton, which most parties agree are polyplacophoran, are known from the earliest Early Ordovician, there is only a very short time to modify Matthevia from its bizarre shape to that approaching a modern chiton. There are multiple possibilities in classification. A century and a quarter ago, von Ihering proposed that the Amphineura (= Polyplacophora of current literature) consti- tuted a phylum. Most neomalacologists accept that polypla- cophorans are quite different from most other mollusks and they are often put in a separate subphylum. Perhaps a phy- lum Amphineura, containing two classes, Polyplacophora and Mattheva would be appropriate. Of course, it does not follow that either von Ihering, Yochelson, or Runnegar et al. had the correct answer. CLASS STENOTHECOIDA My proposal for this class (Yochelson, 1968, 1969) was based on Early and Middle Cambrian bivalved fossils, commonly found as elongate, isolated valves; probably five or six genera fall within the class (Fig. 2). The valves are asymmetrical and inequivalved and have a single apical tooth and socket holding them together. Internally, one valve has many closely spaced ridges. Little can be men- tioned about their paleoecology except that where they occur, they commonly are abundant. They are widespread in Asia (Yu, 1996) and North America. As regards interpretation of Stenothecoides, ‘““We offer the alternative suggestion that it may have been a bivalved monoplacophoran, with the lower (smaller?) valve formed by the sole of the foot” (Runnegar and Pojeta, 1974:316). This concept was repeated by Pojeta and Runnegar (1976:44), and seemingly there has been no fur- ther discussion. Members of Class Bivalvia (= Pelecypoda of older literature) have two valves and if the choices were between that class or Monoplacophora, I would have preferred the former. There are inquivalve bivalves and asymmetric bivalves, but both are uncommon. The tooth and socket is quite unlike a hinge line. To place Stenothecoida within the Fig. 2. Stenothecoida. A, dorsal, B, right side, C, ventral view of Stenothecoides. About ten times natural size. (Yochelson, 1969, fig. 3) Bivalvia would require such an expansion of the concept as to make it unpalatable. As an additional complication, a quarter of a century ago, the early Cambrian Fordilla (Pojeta et al., 1973; Pojeta and Runnegar, 1974) was con- sidered the ancestral member of Bivalvia. If that is true, then stenothecoids might be a dramatic morphologic expan- sion concurrently early in time with a conservative lineage. Adding this factor to the morphology, placing these aber- rant fossils in an extinct class seemed reasonable. Although it is popular to speak of classification based on ancestor-descendent relationships, there is no way to determine these relationships apart from which comes first in the fossil record. An additional complication is that evolution need not proceed in a straight line. If one wanted to concentrate on diversity, an alternative could be to define 198 AMER. MALAC. BULL. 15(2) (2000) Mollusca as univalves. That done, a phylum Bivalvia might be proposed, with Stenothecoidea and Pelecypoda as class- es within it. Bivalved gastropods are a complication, but one might make them another class of Bivalvia. At the least, this might stir debate on whether a classification based on hard part morphology of fossils should be sub- servient to one based on soft parts of living forms. Were such a phylum accepted, Fordilla could well form another class. My objections to its strange internal markings and my suggested reconstruction of soft parts different from typical Bivalvia has had no impact whatsoever (Yochelson, 1981) “There is a significant stratigraphic gap in the fossil record of the pelecypods, between the occurrence of Fordilla in the late Early Cambrian and the Early Ordovician (Tremadocian-Arenigian) when the pelecypods undergo a major radiation” (Pojeta, 1975:371). Time gaps in the fossil record are to be expected where one is dealing with rare forms, but long gaps where the forms are wide- spread and abundant below and above a gap ought to sug- gest something is not quite right. In reporting a presumed pelecypod slightly older than Fordilla, Jell (1980:239) noted “Yochelson’s other argument — the stratigraphic gap in the fossil record of pelecypods in Middle and Upper Cambrian — may disappear in the future as more finds are made.” The gap is still there. I do not really think that all bivalved mollusks will be placed in a separate phylum, but I mention this to show again that classification may be approached from different viewpoints. In turn that affects what is and what is not accepted as a class. CLASS HYOLITHA The hyoliths (Fig. 3) have been variously ancient “pteropods” (see Yochelson, 1979), a class of Mollusca (Marek and Yochelson, 1976), or a separate phylum (Runnegar et al., 1975a). These fossils are abundant in some places in the Cambrian, but taper off rapidly in younger strata and die out by the end of the Paleozoic. The most common hyolith order is the Hyolithida, bilaterally symmetrical closed tubes, uniformly expanding and generally with a triangular cross-section. The lower part of the aperture extends forward as a rounded shelf and the aperture is closed by an operculum; some opercula sug- gest a tooth and socket arrangement with the shell. Between the operculum and apertural margin are a pair of flattened curved pieces of calcium carbonate (= helens). The Orthothecida are generally smaller, have a cross-section that may be circular, oval, or kidney bean- shaped among other configurations. The aperture lacks an anterior shelf. The operculum is simpler than that of the Fig. 3. Hyolitha. A. Reconstruction of Orthotheca. Slightly enlarged from natural size (Marek, 1963, fig. 13). B. Reconstruction of Hyolithes. Slightly reduced from natural size (Marek, 1963, fig. 12). Hyolithida and does not show internal processes that would indicate hingement to the shell dorsum. The Class Xenoconchia Shimanskiy (1963) con- tained Mississippian and Permian forms found in the Soviet Union. The older forms may have been platyceratid gas- tropods, which clung to crinoid calyxes and developed a variety of shapes, including nearly symmetrical cones. Some Permian fossils from Greenland may have been the largest Paleozoic fossil invertebrates, apart from giant cephalopods (Peel and Yochelson, 1981). In our judge- ment, these Permian xenoconchids were better placed as a third order within the Class Hyolitha (Peel and Yochelson, 1984). So far as I know the name Xenoconchia has not since appeared in the literature, except one suggestion that they are internal shells of cephalopods, an unlikely interpre- tation. These fossils have not entered into discussions on the class/phylum rank for the hyoliths. The reconstruction of Runnegar et al. (1975) shows the closely folded gut of an orthothecid within the shell of a hyolithid. The peculiar “helens” projecting between oper- culum and apertural margin are fascinating, but they are an ordinal feature, neither a class characteristic nor a phylum characteristic. Others have objected to that reconstruction (Marek et al., 1997). However, “... it depends on one’s concept of the phylum Mollusca. If one believes that all molluscs are descended from forms that had developed a dorsal exoskeleton, it is possible to exclude the Hyolitha from the phylum. The known muscle insertions of hyoliths suggest that their skeleton was not primitively dorsal” (Runnegar, 1978:332). Runnegar (1980) again discussed his concept of a phylum Hyolitha and, once they were removed from the Mollusca, they were not again discussed with that phylum YOCHELSON: EXTINCT MOLLUSCAN CLASSES 199 (Runnegar, 1983). Since the Monoplacophora are judged by Runnegar and others to be the stem group from which all other Mollusca are derived, the Hyolitha cannot be Mollusca. Possibly the only objective way to form an opin- ion of this highly subjective matter is for an interested per- son to read the paper by Runneger er al. (1975) coordinate with that of Marek and Yochelson (1976). Hyolitha are dis- tinct from Gastropoda, or Rostroconchia, but I do not see that they are so vastly different as to be a separate phylum. A tangential point is a recent redefinition of Monoplacophora in which that class term is abandoned for Tergomya (Peel, 1991). These forms, “classical” monopla- cophorans if you will, begin in the Late Cambrian, not the Early Cambrian. Other small curved forms were placed in the extinct class Helcionelloida (Peel, 1991; Gubanov, 2000). It is too soon to claim that Helcionelloida has found acceptance as a molluscan extinct class, but its prospects are promising. WHAT IS A PALEOZOIC MOLLUSK? My involvement with Hyolitha brought the issue of how one defines a mollusk without considering soft parts. “Pragmatically defined a fossil mollusk is an organism whose hard parts show most of the following items: |. The shell is composed predominately of calcium carbonate and may contain both calcite and aragonite; 2. The shell is lay- ered and not pierced by holes; 3. The shell shows promi- nent growth lines; 4. The shell shows a logarithmic growth pattern; 5. The shell is basically a univalve, but may be modified to a bivalved condition; 6. The shell has basic bilateral symmetry, but may be modified to an asymmetri- cal condition; 7. The shell shows no trace of an apical attachment disk or foramen; 8. The shell may contain septa, either longitudinal or transverse; 9. The shell may have an operculum associated with it” (Yochelson, 1963:163). That 1963 paper discussed the Class Coniconchia Liashenko, and I suggested two unrelated groups were involved, Tentaculites in a broad sense and Hyolithes in a broad sense. My objection was that Coniconchia seemed to have too wide a span of morphology, even though both forms were elongate tubes closed at the apex. Before any- one notes that I suggested both Hyolithes and Tentaculites might be mollusks, this was more than three decades ago and [ knew even less than I know now. I have recanted on the latter. A considerable body of literature exists on both large and small tentaculitids. The small ones show more details of form and ornament than was anticipated half a century ago; possibly they were pelagic. Many of the large tubes are judged to have lived point down in sediment and presumably lived by filter feeding. An elaborate systematic scheme has been developed for the Class Tentaculita (Farsan, 1994) which has been assigned to the Mollusca. My opinion is that including these fossils stretches the phylum beyond reasonable limits. They do not seem to have one of the basic features of Mollusca. The tube-like shell contains a multiple number of very thin layers; it is laminated. The Tentaculita are unique at the highest level and an extinct phylum is appropriate. In a critique, “I proposed a list of features consid- ered common to all or most molluscs (Yochelson 1961) [sic] in an attempt to define that phylum without reference to soft parts; there has been no discussion of this approach, nor of the features listed. The late Cambrian molluscan class Mattheva has no relationship to the various asymmet- trical phosphatic sclerites discussed by Matthews and Missarzhevsky (1975:298-299): the latter separate these on morphological rather than chemical grounds. It does not follow, however, that all fossils with hard parts of calcium carbonate are molluscs, and there is no consensus among workers as to what fossils should be included in the phylum or excluded from it. Thus, Runnegar er al. (1975) suggest that Hyolithes and its allies constitute an extinct phylum, whereas Marek and Yochelson (1976) continue to place them as an extinct class of Mollusca” (Yochelson, 1975). There still has been essentially no more discussion of char- acterizing Mollusca from only the hard parts. It is not easy to define a mollusk, living or dead. “Because no single soft- or hard part character or combina- tion of a relatively few characters, is common to all mol- lusks, it is not possible to frame a succinct morphological definition of the phylum Mollusca as can be done for such phyla as the Echinodermata and Chordata. ... Mollusks are unified by morphological gradations between the different forms, by embryological similarities, and by information deduced from fossils of the various classes assigned to the phylum” (Pojeta, 1980:55). It seems to me that possession of the mineral arago- nite is basic to Mollusca and could be the reason that many fossils are not well preserved; reversion of the original shell to calcite may explain why most shells preserved in lime- stone exfoliate and leave only the internal mold. Immediately following in importance to shell mineralogy, I would list shell structure. A “thought experiment” is to imagine that the scaphopods are all extinct and wonder how they would be classified. Probably they would be consid- ered “worm tubes” until someone making thin-sections saw a similarity to the shell structure of gastropods. Similarly, one could imagine that Cephalopoda were known only from the living octopus in today’s seas. Where would paleontologists place all the straight, curved and coiled septate fossil shells. If they were satisfied by shell composition, and microstructure that these were mol- lusks, surely the fossils would have to be assigned to a class, all of whose members were extinct. Debate between 200 AMER. MALAC. BULL. 15(2) (2000) malacologists and paleontologists could break down as each group looks at different features for high-level classifi- cation within the Mollusca. Worm tubes can produce an amazing variety of shapes and it might be educational and enlightening to compare the presumed Ordovician and Silurian larval gas- tropods of Dzik (1994) with a Middle Ordovician popula- tion of highly variable worm tubes (Bockelie and Yochelson, 1979). Every author, including the present one, is selective in citation of references. For an excellent exam- ple of this process, Dzik illustrated Januspira as a monopla- cophoran, foliowing Runnegar (1977), whereas an interpre- tation on the following page (Yochelson, 1977) of it as a bizarre worm tube was ignored. In the past, a number of strange fossils were tossed into the Mollusca. A Paleozoic scaphopod may be a worm tube (Yochelson and Goodison, 1999) or a monopla- cophoran may be a medusoid (Webers and Yochelson, 1999); it depends on what features one considers to be most significant. To begin to make sense of the Mollusca, one should remove those fossils that ought not to be in the phy- lum. Unfortunately, by some workers ignoring composition of the integument, or great variation in shape, or concen- trating on internal molds, there is danger of further confus- ing the fossil record of the Mollusca. EVOLUTION Patterns in evolution is another subject with much discussion and little resolution. One notion that may have wide acceptance is that of adaptive radiation - a new form appears and diversifies rapidly, at least rapidly on the geo- logic time scale. The concept was first clearly enunciated in regard to early Cenozoic land mammals, but it seems to apply to other organisms, plant and animal, at other times and at various taxonomic levels. There is a near consensus that the early record of both Arthropoda and Echinodermata includes much high-level differentiation; in the Echinodermata, the number of extinct classes could be dou- ble those that are extant (Campbell and Marshall, 1987). If adaptive radiation is a general phenomenon, why does the phylum Mollusca have a different pattern? In one evolutionary scheme (Runnegar and Pojeta, 1974; Pojeta and Runnegar, 1976) there is room for only one extinct class. “At least six [of the eight] higher taxa of Mollusca originated in the Early Cambrian and did not begin to radi- ate in any substantial way until the late Cambrian or early Ordovician” (Runnegar, 1987:50). In this scheme, the Cephalopoda is the only class that appears in the Late Cambrian and immediately undergoes radiation. Scaphopoda appeared even later in the fossil record. In an alternative interpretation (Yochelson, 1978) most Cambrian mollusks superficially resemble those in extant classes, but are actually products of an early adaptive radiation, the extant classes come later. It is impossible to demonstrate that one of these two approaches is true and one is false, but I would be less than human if I did not pre- fer my view. With the latter approach one will seek and find fossils that belong to extinct classes, and with the former, one will not. Fossils do not have their systematic position inscribed on the shell and there is room for honest disagree- ment when unusual or poorly preserved material is at hand. Despite what has been written (Runnegar and Pojeta, 1974), there is no single paleontological viewpoint. I am not qualified to discuss the mechanism of evo- lution of classes. In 1978, I guessed at a non-shelled form beginning in the Precambrian and existing to at least mid- Paleozoic as the source for the various classes, but now think it was a bad guess. Changes that are recognized as being of class rank could have developed in the larval stage. I do not necessarily think that major steps happened instantly, but with the current resolution in stratigraphy, they appear to be instantaneous. The similarity of shell throughout the Mollusca may suggest that an organic template formed originally and then the mollusks diversified. The mechanism of formation of hard parts in the Phanerozoic is still a murky area. It can be made murkier, for did the hypothetical noncalcified mol- lusk form a calcified hard part or parts, or did having a plat- form on which to anchor musculature result in a fundamen- tally different organism? Put more starkly, did the mollusk make a shell or did the shell make the mollusk? SUMMARY “The poor Middle and Late Cambrian record of the Gastropoda, Pelecypoda and Rostroconchia . . . is difficult to explain. Possibly more fossils of these groups will be found when more microfossils are extracted from Cambrian rocks” (Runnegar, 1978:333). We are still waiting. If gaps in the record are real, there may be merit in considering that the later record of the mollusks consists of convergent forms rather than direct ancestors. As mentioned, a fundamental problem is what fos- sils should be placed in the Mollusca. A second problem is what is the level of distinctiveness that makes one fossil a representative of an extant class and another a representa- tive of an extinct one. Even if those points are resolved, others may be irreconcilable. In his view, Runnegar (1983) refuted my refutation of the Runnegar and Pojeta hypothe- sis. This “yes it is and no it is not” has hardened both posi- tions. If one side were to concede that hyoliths were not mollusks, would the other concede that the rostroconchs are simply peculiar pelecypods? In the long run, science might YOCHELSON: EXTINCT MOLLUSCAN CLASSES 201 not be better served, but we could save a lot of time and paper if we agree that there are no extinct classes of mollusks. For some specialists, “Higher taxa are recognized largely by hindsight, after sufficient evolution and diversifi- cation have produced a cohesive group of related organ- isms” (Runnegar, 1978:329). Likewise “If this symposium had been held in the Early Cambrian, it is probable that Fordilla and Pojetia would, at best, be ranked as a family or superfamily of the Monoplacophora” (Runnegar, 1987:49). It is a cheap shot to note the clarity of 20/20 hind- sight vision. In investigations of classification, should emphasis be on the product — morphology of the organism itself — or on the process — interpretation of ancestor- descendent relationships? My position is that one classifies on the basis of the features that are preserved. I admit this is very much an old-fashioned viewpoint. LITERATURE CITED Bockelie, Tove and E. L. Yochelson. 1979. Variation in a species of “worm” from the Ordovician of Spitsbergen. Norsk Polarinstitutt Skrifter 167:225-237. Dzik, Jerzy. 1994. Evolution of ‘small shelly fossils’ assemblages. Acta Palaeontologica Polonica 39:247-313. Campbell, K. W. S. and C. R. Marshall. 1987. Rates of evolution among Paleozoic echinoderms. /n: Rates of Evolution, K. W. S. Campbell and M. F. Day, eds. pp. 61-100. Allen and Unwin, London. Farsan, N. M. 1994. Tentaculiten: Ontogenese, Systematik, Phylogenese, Biostratonomie und Morphologie. Abhandlungen der Senckenbergischen Naturforschenden Gesellschaft 547:1-128. Gould, S. J. 1989. Wonderful Life. W. W. Norton, Inc. New York. Gubanov, A. P. 2000. Evolution of the earliest molluscs. Department of Earth Sciences-Historical geology and paleontology, Uppsala University. 28 pp. + 14 included papers. Jell, P. A. 1980. Earliest known pelecypod on Earth - a new Early Cambrian genus from South Australia. Alcheringa 4:233-239. Marek, L. 1963. New knowledge on the morphology of Hyolithes. Sbornick Geologickych Ved. Paleontologie, Prada P, sv 1: 53-73. Marek, L., R. L. Parsley, and A. Gallé. 1997. Functional morphology of hyoliths based on flume studies. Vestnik Ceskeho geologickeho ustavu 72:35 1-358. Marek, L. and E. L. Yochelson. 1976. Aspects of the biology of Hyolitha. Lethaia 9:65-82. Matthews, S. C. and V. V. Missarzhevsky. 1975. Small shelly fossils of late Precambrian and early Cambrian age: a review of recent work. Journal of the Geological Society of London 131:289-304. Peel, J. S. 1991. The classes Tergomya and Helcionelloida, and early mol- luscan evolution. Grgnlands Geologiske Undersggelse Bulletin 161:11-65. Peel, J. S. and E. L. Yochelson. 1981. Giant Mollusca (Hyolitha) from the Permian of East Greenland. Rapport Gronlands Geologiske Undersggelse Bulletin 101:65-67. Peel, J. S. and E. L. Yochelson. 1984. Permian Toxeumorphida from Greenland: an appraisal of the molluscan class Xenoconchia. Lethaia 17:211-221. Pojeta, J., Jr. 1975. Fordilla troyensis Barrande and early pelecypod phy- logeny. Bulletins of American Paleontology 67:363-384. Pojeta, J., Jr. 1980. Molluscan phylogeny. Tulane Studies in Geology and Paleontology 16:55-80. Pojeta, J., Jr. and B. Runnegar. 1974. Fordilla troyensis and the early his- tory of pelecypod mollusks. American Scientist 62:706-711. Pojeta, J., Jr. and B. Runnegar. 1976. The paleontology of rostroconch mollusks and the early history of the Phylum Mollusca. U. S. Geological Survey Professional Paper 968. Pojeta, J., Jr., B. Runnegar, and J. Kriz. 1973. Fordilla troyensis Barrande: the oldest known pelecypod. Science 180:866-868. Pojeta, J., Jr., B. Runnegar, N. J. Morris, and N. D. Newell. 1972. Rostroconchia: a new class of bivalved mollusks. Science 177:264-267. Regan, C. T. 1925. Organic evolution [address by sectional president]. British Association for the Advancement of Science, Southampton, Section D:D1-D12. Runnegar, B. 1977. Found — a phylum for Januspira. Lethaia 10:203. Runnegar, B. 1978. Origin and evolution of the Class Rostroconchia. Philosophical Transactions of the Royal Society of London series B 284:319-333. Runnegar, B. 1980. Hyolitha: status of the phylum. Lethaia 13:21-25. Runnegar, B. 1983. Molluscan phylogeny revisited. Association of Australian Palaeontologists Memoir 1:121-144. Runnegar, B. 1987. Rates and modes of evolution in the Mollusca. Jn: Rates of Evolution, K. S. W. Campbell and M. F. Day, eds. pp. 39-60. Allen and Unwin, London. Runnegar, B. and J. Pojeta, Jr. 1974. Molluscan phylogeny: the paleonto- logical viewpoint. Science 186:311-317. Runnegar, B., J. Pojeta, Jr., N. J. Morris, J. D. Taylor, M. E. Taylor, and G. McClung. 1975. Biology of Hyolitha. Lethaia 8:181-191. Runnegar, B., J. Pojeta, Jr, M. E. Taylor, and D. Collins. 1979. New species of Cambrian and Ordovician chitons Matthevia and Chelodes from Wisconsin and Queensland: evidence for the early history of polyplacophoran mollusks. Journal of Paleontology 53:374-394. Shimanskiy, V. N. 1963. [Systematic position and scope of Xenoconchia] Paleontologiske Zhurnal (4):53-63. [Translated by American Geological Institute]. Waller, T. R. 1998. Origin of the molluscan Class Bivalvia and a phyloge- ny of major groups. /n: Bivalves: an eon of evolution, P. A. Johnston and J. W. Haggart, eds. pp. 1-45. University of Calgary Press. Webers, G. F. and Yochelson, E. L. 1999. A revision of Palaeacmaea (Upper Cambrian) (?Cnidaria). Journal of Paleontology 73:598- 607. Yochelson, E. L. 1963. Notes on the class Coniconchia. Journal of Paleontology 35:162-167. Yochelson, E. L. 1966. Mattheva, a proposed new class of mollusks. U. S. Geological Survey Professtonal Paper 523:B1-B13. Yochelson, E. L. 1968. Stenothecoida, a proposed new class of Cambrian Mollusca. /nternational Palaeontological Union, Prague, Abstracts, p. 34. Yochelson, E. L. 1969. Stenothecoida, a proposed new class of Cambrian Mollusca. Lethaia 2:49-62. Yochelson, E. L. 1975. Discussion of early Cambrian “molluscs.” Journal of the Geological Society of London 131(6):661-662. Yochelson, E. L. 1977. Comments on Januspira. Lethaia 10:204 Yochelson, E. L. 1978. An alternative approach to the interpretation of the phylogeny of ancient mollusks. Malacologia 17(2):185-191. Yochelson, E. L. 1979. Early radiation of Mollusca and Mollusc-like groups. In: The Origin of Major Invertebrate Groups, M. R. House, ed. pp. 157-178. Systematics Association, special volume 202 AMER. MALAC. BULL. 15(2) (2000) 12. Academic Press, London, Oxford University Press, Oxford, Yu Wen. 1984. Early Cambrian molluscan faunas of Meischucan Stage, England. with special reference to Precambrian-Cambrian Boundary. Jn: Yochelson, E. L. 1981. Fordilla troyensis Barrande: “The oldest known Contributions to the 27th International Geological Congress, Sun pelecypod” may not be a pelecypod. Journal of Paleontology Shu, ed. pp. 21-25. 55:113-125. Yu Wen. 1990. The first radiation of the shelled molluscs. Palaeontologia Yochelson, E. L. 1984. Speculative functional morphology and morpholo- Cathayana 5:139-170. gy that could not function: The example of Hyolithes and Yu Wen. 1996. Early Cambrian stenothecoid molluscs from China. Biconulites. Malacologia 25(1):255-264. Records of the Western Australian Museum 18:209-217. Yochelson, E. L. and R. Goodison, R. 1999. Devonian Dentalium martini Whitfield, 1882, is not a mollusk but a worm. Journal of Date of manuscript acceptance: 23 June 2000 Paleontology 73:634-641. AMERICAN MALACOLOGICAL SOCIETY, INC. FINANCIAL REPORT General Accounts 1999 Income and Expenses BE De DOD eran eens rrr ces pacctdosstentanatic grap p aad aut dinn cesta nau le ear enite pins eatovaaeniaschoun tanh akeeeitaalqucecs $174,042.52 MOS oon arna tee earn Staataas eae ene Cea es eae eile e acts ces apter on GgMiona aa aera deaalaceigie te eget apa eb ane binanaaeanee $ 30,440.69 Wem Deis hip USS (19961 99 1; TOO) ccacsacsaeesbeasss tahcusece ntpnnncaadegsan edn wiwassataonansanegenniansaasass 1,962.00 PVC TIE YS EO TUES 1 9) Vc Sec yccine cance: as eaeaaccie natal aa vdontoen couininasaita ie rninb olesautid yotedeeaanens 12,047.00 Membersinip: Dues (2000) aces cansctes sass a eaahsatvaven aaioniavedsecindnitaaviv ana asiaxeieie aracsercnasi anatase 15.00 RTPCROS (AEE DIVING Foo win costae ovat cceceusivaacataiiee hc wtaa edb ta tbs bdudawacd sdaaotts bane cog Naadaearpsuadeanite 5,080.46 Money Market Account Interest ..0......0. ec eeeeeeeeereeeeereeeeeseeeeneeeeeeeneeeee 1,132.73 Lite Membership Endowment Fund x.sccctsecsiccesstieiccssevtinrsiasistscsancestamncsin 345.84 Symposium and Student Grant Endowment Funds..........0...0.ceeeee 3,601.89 PUNO AN NS TANT ore cnet Gace avant a ua ona uh c0VeNs praa atShatestenwedt casa esh enki eae cawad lis eoenaaatntieys 4,547.15 AMB Foreign SUDSCrptlOns ies isvinesssvaisassaansesspaseanicesetespendeseoncondorsnnayonds 1,372.00 ALES TOMIESEIC SUS CLI UEONS devices sasaices cesek shah once Racnnesedtonte rn ceasmnecstgncenes 2,565.00 AIM ES PAGS CIALOSS sais caseniavetisidanoseansusssadinnsanenibuniesssnsaaogadtestvsiiedaynassaiasnivanss 296.00 INAS AOS SUC Ga e's a aps ca geese ead scrote ste bine dnc uneentaateagiannel ay anateicanmiatete 10.15 AVES IROPIPUINY CAPS OS ga casscadviavasscrendyteieumnlaps heceabdnahsanntesbucacsbossecnetabbasnniieds 160.00 Sales of Other Publications..........0..ececceeeeeeneeeceeeeeeeeeeseenteeeeeseeeneeeeeeeee 144.00 URTEEDNL VU Se UAIA Soles cae chemo eoncea a tei ov eas oc ar ai ons sea Rea aa mead RatLd Sea dan semen ahaauapanacauuleaaaRe 2,248.50 POOOVAUCHON PLOCEEMS esscdorsccadeusectecatsaiessthibiisdadaabbbescuebebeccte ts jethavateodese 2,248.50 EQ OMatl OM Sirizes se 5 secession cdac cos sas tes ac eeeSER casa deka ux ceawaasa ga Naedanatanwersavs sndanaabseznues caeasemaeavesteleaeiae: 4,520.00 SYMPOSIUM: ENGOW MEME PUM 2 ao ccecscsSecovesussnesictiassaceutsiadeaunaiaasdanwatiyeueans 20.00 Student Grant Endowment Fund ........ eee eeeeeeeeeeeeeeeeeeeeeeeeneneeeee 1,400.00 Bernice Barbour Foundation.............ccccccccceceeeeeeeeeeeeeeeeeeeneeeeeeneeeeenseees 3,000.00 Bor Reprinting Revised “TOW TO COMEC( .o.ssncesessseseadsensnnsadancsasdssensxveenss 100.00 IEtS@ell ame GUS, ENCOUN e oc cr ccncscscnoitstocnsasuneeiavleseies pac dedagnannsinvacsdansidobstaabtuivebaaaps bikaocondanaaiyaidans 20.58 Be OTE ES ase crpot cca Gat aor sue aac eos aa nan Ga a ea a Raa eH aRaTLA awe Tae aaa eg a NeRR RENAE eaR ua dae $27,791.16 RS ASURREE, eC NS OS ow. ec cd Pyne cpran sash peta sx aise ed dnapdua une caeeasnsatinnaticanytenbenteimanenataaabbencvancats 1,293.11 PTET Aes Ny FO MIMD OES IVINS 25 CA ics ssl sateen aaa atch nian nea etcaaighinBcxdy se Nona conde Patnatenaatsida: 135.00 Pate eS ivigana viahcconndeuiteanccticuvlsetia sasansanaaiauig ea tepaeciuutceiunaieebennaniayereanaa adederaanaaaiecsorstaatbnetons 21.51 INES SV DS Sr bes ceatraicaen Gaddvacnsdpuls oasesaneMeanaaennd tadamstacnresseaddadetaciienieeadoneiedeiGiueaesniunhedens 725.85 IMEOLPOPAallOn: Pee cis cvsavisaveustvesausooarsudandssayoasyssiasedsstlaceday ipssaibanadonnsassabiuvis tuadinbenectinsumniseiouunntaes 40.00 TSA COPS Ot Th SS ict cece ca ts cause neteses alata seta aedawioaatel nate dinact tach auseemeanaAraea cena 605.00 Managing and Bulletin Editor Expenses............ccccececeeeeeeeeeeeeeeeeeseeeeeneeeeeecaeeeesneeeeetteeeseees 626.14 PDH Catton COStS sigste cg sk tecacicsdushwisauatanasavesinpherussantn adie cnaceoth uebunchenannsdtentdasaaniaulinesoewetsanens 13,872.20 FN 3 aan ee ee ne ee Reo Dy ee ne ee ee Re cera Econ ere nee 11,237.08 PUES IN CW STC LOR 2 a sraelics boi alsisu cas yop eauwaaasiiawalifauaiaeagpauenablanitaiucntayshaaaghsGncaiiase 932.12 PST OGINURG Sioa circisccaustiva veoh sta svazcas ea vnatuptaausdiesset dhnaavnaiennseecweruaneaetigassonieioonetesatbes 608.00 BES Or OMG te cine ae dete oahlev tana cesish caeiaeteeata fai rede elas ica maaan ana 1,095.00 Dissolving Council of Systematic Malacologists ...........ccccscccsscssseseesseeseseteeetssssssesseesseessees 02.00 LUC OME FRESE ACIN G3 LINCS saisas saa sokd wrevtunngettgaspau nieve ccote secured uanaandonka encaladstunastanesaetucensmanicannieas 2,000.00 POS 8 SACs MTR 20am inceas a ean es gi cose ceed oivgeeeciobs ca mbaan abae nai pcp aude hee danagin 137.36 1999 Meeting Expenses - Separate Account AV CIEE RPeusesi( DO NICS ie) oi mics Syesvsaceyaduespnnananeaueidenemeaten cochaelaugstuactensacteumdcebadeeees 3,776.99 | sui Deiat Mp S161 (c/a @l Aca: } Baer Renna Renee emer peo ereery cman cent oer oe are te Nearer ne 1,487.33 WEARS SE CIPOR C199 9 ) ond sat vsacasatnstrrsasssenaseduiner suas egionedegealdetaeanraanwen 294.40 PERG AGU to) anu rtacidln te cncenistetnlaneiidinsiial ovanmubacenemupensneianmaiGiaks 1,995.26 EUG CIN DG wy Stain oyacta ciysp cam cutsaasdaai un bles ict aslle Cavaumsieancteecasnandttanersaceyeendunisanraentsrmts 1,500.00 Bermice Barbour student Travel Awards... ciao cxet sccsucveseuhbiaueengadesasdoouisrieaanetvarcioniendis OF00000 OI END Oe pe scoee ea ey eater ep sate Una ieies iarctstaesEaeanclay Seu tguedcocisse aqua Rel aud evokes es ueaieudsouadccee tetcdwactidenar $2,649.53 DS Se 1) 9) aa ase tae reread te wanda ca ill 5 iP PU oxscmanuieh angered mem osaeomakuinunnade $179,908.98 **Includes an increase in capital investments. The above income and expenses do not include the annual meeting accounts. 67th ANNUAL MEETING THE AMERICAN MALACOLOGICAL SOCIETY WORLD CONGRESS OF MALACOLOGY VIENNA, AUSTRIA AUGUST 19 - 25, 2001 The American Malacological Society will be meeting jointly with Unitas Malacologica in Vienna, Austria, from 19-25 August 2001, at the second World Congress of Malacology. Scientific presentations will take place in the lecture rooms of the “Biozentrum” of the University of Vienna (founded in 1365!). A variety of housing facilities from student dorms to hotels will be available, all within 20 minutes of the meeting venue. Five symposia are currently being organized: * Evolution and Development in Molluscs (organizers Gerhard Haszprunar and W. A. G. Dictus) ¢ Chemosymbiosis (organizers Carole Hickman, Penelope Barnes, and Martin Zuschin) * Mollusca in Long-lived Lakes (organizers Frank Wesselingh and Elinor Michel) ¢ Molluscan Conversation (organizers Ian Killeen and Mary Seddon) ¢ Functional Morphology (organizers Dianna Padilla and Shirley Baker) English is the preferred language of scientific presentations. Oral contributions will be limited to 15 minutes, including discussion; poster presentations are strongly encouraged, as is the organization of workshops on spe- cial topics. There will be two poster sessions. Field trips to Schneeberg (alpine gastropods), the Danube alluvial forest (limnic and terrestrial molluscs), and to the tethyan and paratethyan fossil sites will provide a break in the scientific program. A Curator Session at the Naturhistorisches Museum Wien will be organized by Helmut Sattman and Anita Eschner. The Austrian capital offers a sparkling array of scientific, historical, architectural, and musical highlights, both within its borders and in its nearby surroundings. The social program will include a Congress Opening Reception at the Naturhistorisches Museum Wien and a Congress Dinner at the Rathaus (City Hall) under the patronage of the mayor of Vienna, Dr. M. Haupl, who is a zoologist (!). In addition, a mid-week dinner at a “Heurigen” (typi- cal Viennese open-air wine tavern) is planned. There will also be the traditional AMS auction led by Dick Petit. AMS will be providing up to six $500 travel awards for graduate students to attend the meeting. These awards will require a special early application and will be available on a competitive basis. Details will be posted on th AMS website (http://erato.acnatsci.org/ams/), the Mollusca listserver, and in the spring newsletter of the American Malacological Society. Preliminary registration is available at the meeting website: http://www.univie.ac.a/WCM2001/index.htm. For further information please contact: Janice Voltzow, President, AMS Department of Biology University of Scranton Scranton, PA 18510-4625 E-mail: voltzowj2 @scranton.edu 204 IN MEMORIAM Ruth D. Turner 205 Babin, C. 15:167 Carlini, D. B. 15: 179 Chiba, S. 15:75 Clark, R. N. 15:33 Emberton, K. C. 15:83, 97 Gubanov, A. P. 15:139 Harper, J. A. 15:147 Haszprunar, G. 15:115 Hayakaze, E. 15:75 INDEX TO VOLUME 15 AUTHOR INDEX Henley, W. F. 15:65 Hoare, R. D. 15:131 McMurtay, S. E. 15:57 Monroe, E. M. 15:51 Naimo, T. J. 15:51 Neves, R. J. 15:47, 65 Parker, B.C. 15:47 Patterson, M. A. 15:47 Peel, J.S. 15:139 PRIMARY MOLLUSCAN TAXA INDEX Pojeta, J., Jr. 15:157 Ramey, B. A. 15:57 Rollins, H. B. 15:147 Schuster, G. A. 15:57 Vecchione, M. 15:179 Wagner, P. 15:1 Yochelson, E. L. 15:195 Young, R. E. 15:179 [first occurrence in each paper recorded; new tax (including species) in bold face] Abralia 15:181 Acanthochitonina 15:34 Acavacea 15:83 Acavidae 15:83 Acavoidea 15:83 Acéphalés 15: 169 Achatinida 15:83 Acmaeidae 15:150 Acmaeoidea 15:149 Actinonaias 15:58, 70 Aculifera 15:115 Acutichiton 15:134 Acutichitonidae 15:136 Adenopoda 15:125 Aenigmatectus 15:133 akoratsara, Ampelita 15:83 Alasmidonta 15:60, 70 Archinacelloidea 15:151 Architeuthis 15:181 Arcochiton 15:131 Arcoidea 15:168 Argonauta 15:181 Arhouriella 15:159 Arjamannia 15:3 Autobranchia 15:168 Babinka 15: 171 Bathypolypus 15:181 Bathyteuthidae 15:180 Bathyteuthis 15:181 Batoteuthis 15:191 Belemnoidea 15:179 Belemnotheutis 15:187 Bellerophon 15:148 Bellerophontina 15:155 Choriplax 15:134 Chtenopteryx 15:181 Cirrata 15:191 Cirroteuthidae 15:180 Cirrothauma 15:181 Clavator 15:83 Cocculina 15:149 Cocculinidae 15:151 Cocculinoidea 15:150 Coleoida 15:121, 179 Conchifera 15:115 Coniconchia 15:199 Coniglobus 15:75 Corallidomus 15: 172 Corbicula 15:58, 65 Cranchia 15:181 Cranchiidae 15:190 Aldanella 15:139 Bellerophontoidea 15:143, 147 Crepidula 15:150 Alderia 15:121 bemarahae, Edentulina 15:97 Cryptodonta 15:168 Alluroteuthis 15:181 Bemella 15:142 Ctenodonta 15:168 ambanianae, Ampelita 15:83 Bivalvia 15:47, 57, 65, 115, 139, curiosus, Plasiochiton 15:131 Amblema 15:47, 51, 70 163, 167, 197 Cycloconchoides 15:161 ambongoaboae, Edentulina 15:97 Ambonychiidae 15:173 ambra, Edentulina 15:97 Ampelita 15:83 Amphineura 15:197 Anabarella 15:140 analamerae, Ampelita 15:83 analamerae, Edentulina 15:97 Ancistrocheirus 15:181 anjanaharibei, Ampelita 15:83 ankaranae, Edentulina 15:97 Anomalodesmata 15:168 Antalis 15:122 antankarana, Edentulina 15:97 Antipleura 15: 174 Antipleuridae 15: 168 Aplacophora 15:115 Archaeogastropoda 15:154 Archaeospira 15:140 Archanodon 15: 176 Archimollusca 15:126 bobaombiae, Edentulina 15:97 Bolitaenidae 15:180 Brachioteuthis 15:181 Buchiola 15:168 Bullomorpha 15:121 Caenogastropoda 15:121, 150 Camptochitonidae 15:136 Camya 15:159 Carcassonella 15:151 Cardiola 15:168 Cardiolaria 15: 171 Caudofoveata 15:115 Cephalopoda 15:115, 139, 179, 198 Chaetodermatidae 15:123 Chaetodermomorpha 15:115 Chelodes 15:131, 196 Chelodidae 15:134 Chiroteuthis 15:181 Chiton 15:34, 122, 132 Chitonida 15:34, 120 Choriplacidae 15:134 206 Cyclomya 15:147 Cyclonaias 15:60, 68 Cycloteuthis 15:181 Cymatochiton 15:131 Cymatochitonidae 15:136 Cyrtolites 15:151 Cyrtonella 15:151 Cyrtonellida 15:143 Cyrtosoma 15:125 Davidia 15:159 Decapoda 15:183 Decapodiformes 15:179 Dentaliida 15:120 Diadeloplax 15:133 Diasoma 15:126 Diodora 15:149 Discoteuthis 15:181 Donaldiella 15:3 Dorymenia 15:122 Dreissena 15:47, 51 Dreissenidae 15:47 Dualina 15:173 Ectomaria 15:2 Edentulina 15:97 Elachychiton 15:135 Eledonella 15:181 Elliptio 15: 58, 70 Enoploteuthidae 15:180 Enoploteuthis 15:181 Eotebenna 15:143 Eotomaria 15:13 Epimenia 15:121 Epioblasma 15:70 Euhadra 15:75 Euleptochiton 15:131 Eumegalodus 15: 175 Eunema 15:3 Euomphalinae 15:13 Euomphaloidea 15:144 Euphemites 15:152 Eurekapegma 15:143 Eurystyla 15:83 Ferreiraellidae 15:136 Fissurellidae 15:119 Fissurelloidea 15:148 florensi, Edentulina 15:97 Floripatella 15:149 Fordilla 15:157, 197 Fordillidae 15:159 Frodospira 15:17 Fusconaia 15: 60, 68 Gadilida 15:121 Ganglionata 15:115 Ganglioneura 15:127 gargantua, Helicophanta 15:83 Gastropoda 15:1, 83, 97, 115, 139, 147, 196 Glaphurochiton 15:132 Globonema 15:3 Glyptochiton 15:134 Glyptochitonidae 15:136 Gonatidae 15:180 Gonatopsis. 15:181 Gonatus 15:182 Gotlandochiton 15:131 Gotlandochitonidae 15:136 Graneledone 15:181 griffithsi, Ampelita (Eurystyla) 15:83 griffithsjonesi, Clavator 15:83 Grimpoteuthis 15:181 Groenlandibelus 15:189 Gryphochiton 15:131 Gryphochitonidae 15:136 Guerichia 15:176 Gyronema 15:3 Hapalochlaena 15:181 Helcionella 15:142 Helcionelloida 15:139, 199 Helicophanta 15:83 Helminthochitonidae 15:136 Hepagastralia 15:125 Heraultipegma 15:142 Heterobranchia 15:123 Heteroconchia 15:168 Heterodonta 15: 168, 120 Heteroteuthis 15:181 Histioteuthis 15:181 Hormotoma 15:2 Hubeinella 15:160 Hubeispira 15:140 Hyolitha 15:198 Hyolithes 15:198 Hyolithida 15:198 Hypseloconus 15:144 Idiosepiidae 15:187 Idiosepius 15:181 Incirrata 15:193 Isofilibranchia 15:159, 168 ivohibei, Ampelita 15:83 Januspira 15:200 Japatella 15:181 Jeletzkya 15:187 Josephinae, Ampelita 15:83 Joubiniteuthis 15:181 Katharina 15:181 Kiviasukkaan 15:3 Knightoconus 15:144 Kralovna 15:174 Lamellibranchia 15:159 Lamellodonta 15:157 Lampsilis 15:60, 66 Lasmigona 15:60, 66 Latouchella 15:140 Lekiskochiton 15:132 Lekiskochitonidae 15:136 Lemiox 15:70 Lepetelloidea 15:151 Lepetidae 15:150 Lepidochitona 15:34 Lepidochitonidae 15:33 Lepidopleurida 15:120 Lepidopleuridae 15:136 Lepidopleurina 15:134 Lepidoteuthis 15:181 Leptodea 15:60 Leucotaenius 15:83 Lexingtonia 15:70 Liocranchia 15:181 Lipodonta 15:168 lokii, Tonicella 15:39 Loliginidae 15:180 Loligo 15:182 Longstaffia 15:3 Lophospira 15:3 Lophospindae 15:1 Lophospiroidea 15:1 Lottia 15:149 Lottiidae 15:150 Loxoplocus 15:3 Lycoteuthis 15:181 Lymnaea 15:54 Marocella 15:143 masoalae, Ampelita 15:83 masoalae, Clavator 15:83 Mastigoteuthis 15:181 Mattheva 15:196 207 Matthevia 15:196 Matthevidae 15:134 Medionidus 15:70 Megalodon 15:175 Megalodonta 15:173 Megalodontidae 15:175 Megalomoidea 15:173 Megalonaias 15:60 Mellopegma 15:143 Micropilina 15:119 Modiolopsis 15:159, 172 Modiomorpha 15:168 Modiomorphidae 15:172 Modiomorphoidea 15: 172 Mollusca 15:115, 139, 157, 195 Molluscoidea 15:195 Monoplacophora 15:, 139, 147, 163, 195 Mopalia 15:120 Moroteuthis 15:181 Murchisonia 15:31 Murchisoniina 15:1 Murchisoniinae 15:13 Mytiloidea 15:168 Mytilus 15:55 Naticidae 15:121 Nautilidae 15:179 Nautilus 15:121, 181 Neocoleoidea 15:179 Neocyrtolites 15:150 Neoloricata 15:115, 134 Neomeniomorpha 15:115 Neopilina 15:139, 147 Neopilinidae 15:118, 122, 153 Neotaxodonta 15:168 Neritimorpha 15:125 Nucinella 15:162 Nucinellidae 15:162 Nuculites 15:174 Obliquaria 15:60 Ochmazochitonidae 15:136 Octopoda 15:179 Octopodidae 15:180 Octopodiformes 15:179 Octopoteuthis 15:181 Octopus 15:181 Ocythoidae 15:180 Oegopsidae 15:191 Oelandiella 15:140 Ommastrephes 15:181 Ommastrephidae 15:180 Onychoteuthidae 15:180 Onychoteuthididae 15:192 Onychoteuthis 15:181 Opisthobranchia 15:120, 149 Opisthoteuthidae 15:180 Opisthoteuthididae 15:192 Opisthoteuthis 15:181 Orthotheca 15:198 Orthothecida 15:198 Pagodospira 15:3 Palaeoctopus 15:189 Palaeoheterodonta 15: 168 Palaeoneilo 15:174 Palaeotaxodonta 15:159 Paleochiton 15:131 Paleotaxodonta 15: 168 Patella 15:120, 150 Patellidae 15:150 Patellogastropoda 15:120, 147 Patina 15:150 Paupospira 15:17 Pegias 15:70 Pelagiella 15:139 Pelecypoda 15:157, 171, 197 Permochitonidae 15:136 Phaeohelix 15:75 Pholidoteuthis 15:181 Phyllomenia 15:122 Pilina 15:139 Planctoteuthis 15:190 Plasiochiton 15:131 Platyceratidae 15:152, 198 Pleurobema 15:68 Pleurotomarioidea 15:149 Pojetaia 15:157 Pojetia 15:201 Polyplacophora 15:33, 115, 131, 147, 196 Posidonia 15:176 Potamilis 15:60 Praecardiidae 15:168 Praecardioidea 15:168 Praelamellodonta 15:161 Praenucula 15:168 Preacanthochiton 15:197 Preacanthonidae 15:136 Proplina 15:139 Prosobranchia 15:147 Protobranchia 15:120, 168 Proturritella 15:3 Pseudomyona 15:163 Psychroteuthis 15:181 Pterineidae 15:172 Pterioidea 15:176 Pteriomorpha 15:120, 167 Pteriomorphia 15:163 Pteropoda 15:198 Ptomatis 15:152 Ptychobranchus 15: 60, 68 Ptychozone 15:3 Pulmonata 15:83, 97, 120 Dates of Publication Pyroteuthis 15:181 Quadrula 15:47, 60, 70 ranomafanae, Ampelita 15:83 raxworthyi, Ampelita 15:83 Rhodope 15:121 Rhytida 15:83 Rhytididae 15:83 Rossia 15:181 Rostrochoncha 15:127 Rostroconchia 15:139, 195 Ruedemannia 15:3 rugosa, Edentulina 15:97 Scaphopoda 15:115, 199 Scenella 15:143 Schizolopha 15:23 Scutopus 15:120 Sepia 15:181 Sepiadariidae 15:187 Sepiadarium 15:191 Sepiidae 15:180 Sepioidea 15:187 Sepiolidae 15:180 Sepioloidea 15:181 Sepioteuthis 15:181 Septemchitonidae 15:134 Shanina 15:172 Shaninopsis 15:172 Sigmurethra 15:83 Sinuites 15:151 Sinuitopsis 15:148 Slava 15:173 Solemyoidea 15:162, 168 Solenogastres 15:115 Spirula 15:181 Spirulidae 15:180 Stauroteuthidae 15:180 Stauroteuthis 15:181 Stenotheca 15:143 Stenothecoida 15:160, 197 Stenothecoides 15:197 Sthenoteuthis 15:181 Stoloteuthis 15:181 Streptaxidae 15:97 Streptaxoidea 15:99 Strobilepis 15:133 Strophitus 15:70 Stylommatophora 15:83, 97, 121 Syvestrosphaera 15:151 Tentaculita 15:199 Volume 15(1), October, 1999 Volume 15(2), December, 2000 208 Tentaculites 15:199 Tentaculitidae 15:199 Tergomya 15:139, 147, 199 Testaria 15:115 Teuthoidea 15:190 Thairoplax 15:131 Thermoconus 15:144 Thysanoteuthidae 15:180 Thysanoteuthis 15:181 Tironucula 15:171 Tonicella 15:33 Tonicellidae 15:33 Tonicellinae 15:34 Tonicelloidea 15:34 Toxolasma 15:70 Tritogonia 15:60 Trochonema 15:2 Trochonematidae 15:1 Trochonematoidea 15:13 Trochonemella 15:3 Trochonemlla 15:3 Tryblidia 15:115 Tryblidiida 15:139 Tryblidiidae 15:151 Tryblidioidea 15:147 Tryblidium 15:139, 147 Tuarangia 15:157 Tuarangioida 15:163 Unionidae 15:47, 51, 57, 65 Vampyromorpha 15:187 Vampyroteuthidae 15:180 Vampyroteuthis 15:181 Ventroplicida 15:115 venusta, Tonicella 15:41 Vesconis 15:83 Villosa 15:66 Visceroconcha 15:125 Vitreledonella 15:181 Vlasta 15:172 Vlastidae 15: 172 Watsonella 15:142 Whiteavesia 15:172 Xenoconchia 15:198 Xenoconchidae 15:198 Xianfengoconcha 15:160 Xystera 15:83 Yochelcionella 15:143 CONTRIBUTOR INFORMATION The American Malacological Bulletin serves as an outlet for reporting notable contributions in malacological research. Manuscripts concerning any aspect of original, unpublished research, important short reports, and detailed reviews dealing with molluscs will be considered for publi- cation. Each original manuscript and accompanying illustra- tions must be submitted with two additional copies for review purposes. Text must be typed on one side of 8-1/2 x 11 inch bond paper, double-spaced, and all pages num- bered consecutively with numbers appearing in the upper right hand corner of each page. Leave ample margins on all sides. Form of the manuscript should follow that outlined in the Council of Biology Editors Style Manual (sixth edition, 1994). This can be purchased from the CBE, 11 S. LaSalle Street, Suite 1400, Chicago, IL 60603, U.S.A. Text, when appropriate, should be arranged in sections as follows: 1. Cover page with title, author(s) and address(es), and suggested running title of no more than 50 characters and spaces. Authors should also supply five key words, placed at the base of this page, for indexing purposes. Abstract (less than 5% of manuscript length) 3. Text of manuscript starting with a brief introduction followed by methodology, results, and discussion. Separate sections of text with centered subtitles in cap- ital letters. 4. Acknowledgments 5. Literature cited 6. Figure captions All binomens must include the author and date attrib- uted to that taxon the first time the name appears in the manuscript [e. g. Crassostrea virginica (Gmelin, 1791)]. This includes nonmolluscan taxa. The full generic name along with specific epithet should be written out the first time that taxon is referred to in each paragragh. The gener- ic name can be abbreviated in the remainder of the para- graph as follows: C. virginica. References should be cited within text as follows: Hillis (1989) or (Hillis, 1989). Dual authorship should be cited as follows: Yonge and Thompson (1976) or (Yonge and Thompson, 1976). Multiple authors of a single article should be cited as follows: Beattie et al. (1980) or (Beattie et al., 1980). In the literature cited section of the manuscript refer- ences must also be typed double spaced. All authors must be fully identified, listed alphabetically and journal titles must be unabbreviated. Citations should appear as follows: Beattie, J. H, K. K. Chew, and W. K. Hershberger. 1980. Differential survival of selected strains of Pacific oys- ters (Crassostrea gigas) during summer mortality. Proceedings of the National Shellfisheries Association 70(2):184-189. Hillis, D. M. 1989. Genetic consequences of partial self fertilization on populations of Liguus fasciatus (Mollusca: Pulmonata: Bulimulidae). American Malacological Bulletin 7(1):7-12. Seed, R. 1980. Shell growth and form in the Bivalvia. Jn: Skeletal Growth of Aquatic Organisms, D. C. Rhoads and R. A. Lutz, eds. pp. 23-67. Plenum Press, New York. Yonge, C. M. and T. E. Thompson. 1976. Living Marine Molluscs. William Collins Son and Co., Ltd., London. 288 pp. Illustrations should be clearly detailed and readily reproducible. Fine patterns and screens do not reproduce well. All line drawings should be in black, high quality ink. Photographs must be on glossy, high contrast paper. All diagrams must be numbered in the lower right hand corners and adequately labeled with sufficiently large labels to remain readable with reduction by one half. Magnification bars must appear on the figure, or the cap- tion must read Horizontal field width = x mm or x pum. All measurements must be in metric units. All illustrations submitted for publication must be fully cropped, mounted on a firm white backing ready for reproduction, and have author’s name, paper title, and figure number on the back. All figures in plates must be nearly contiguous. Additional figures submitted for review purposes must be of high quality reproduction. Xerographic reproduction of pho- tomicrographs or detailed photographs will not be accept- able for review. Abbreviations used in figures should occur in the figure caption. Indicate in text margins the appropri- ate location in which figures should appear. Color illustra- tions can be included at extra cost to the author. Original illustrations will be returned to author if requested. Any manuscript not conforming to AMB format will be returned to the author for revision. Final submission of accepted, revised manuscripts should include two typed copies of the text, tables, etc. and an additional copy in electronic form on 3.5” diskette. The electronic version should be readable as non-formatted ASCII files. New Taxa. The Bulletin welcomes complete descriptions of new molluscan taxa. Establishment of new taxa must conform with the International Code of Zoological Nomenclature (1985). Descriptions of new species-level taxa must include the following information in the order as given: higher taxon designation as needed for clarity; fami- ly name with author and date; generic name with author and date; Genus species author sp. nov. followed by numeration of all figures and tables; complete synonymy (if any); listing of type material with holotype and any other type material clearly designated along with complete museum catalogue or accession information; listing of all additional non-type material also with full museum deposi- tion information; type locality; diagnosis and full descrip- tion of material done in telegraphic style including measurements and zoogeographic distribution as neces- sary; discussion; etymology. Descriptions of new supraspe- cific taxa should include type species (for new genus) or type genus (for new family), diagnosis and full description done in telegraphic style, and list of included taxa. Proofs. Page proofs will be sent to the author and must be checked for printer’s errors and returned to the printer within a three day period. Significant changes in text, other than printer errors, will produce publishing charges that will be billed to the author. Mailing. All overseas mailing must be done via airmail. The American Malacological Society will not be responsi- ble for deferred publication of manuscripts delayed in sur- face mail. Charges. There are no mandatory page costs to authors lacking financial support. Authors with institutional, grant, or other research support will be billed for page charges. The current rate is $35.00 per printed page. Acceptance and ultimate publication is in no way based on ability to pay page costs. Reprints. Order forms and reprint cost information will be sent with page proofs. The author receiving the order form is responsible for insuring that orders for any coau- thors are also placed at that time. Submission. Submit all manuscripts to Dr. Ronald B. Toll, Editor-in-Chief, American Malacological Bulletin, College of Natural Sciences and Mathematics, University of Central Arkansas, Lewis Science Center 105, 201 Donaghey Avenue, Conway, AR 72035, U.S.A. (e-mail: rtoll@mail.uca.edu). Subscription Costs. Institutional subscriptions are avail- able at a cost of $48.00 per volume. Membership in the American Malacological Society, which includes personal subscriptions to the Bulletin, is available for $35.00 ($15.00 for students, $45 for affiliated clubs). All prices quoted are in U.S. funds. Outside the U.S. postal zones, add $5.00 seamail and $10.00 airmail per volume within North America, or $10.00 seamail and $20.00 airmail per volume in other locations. For subscription and back-issue information contact Dr. Timothy A. Pearce, Managing Editor, American Malacological Bulletin, Delaware Museum of Natural History, Box 3937, Wilmington, DE 19807-0937, U. S. A. (e-mail: tpearce @delmnh.org). wi 3 9088 iil 3 9088 01457 3547 l 7) wi « < ra o a 2 fe} = =) E Ee 7) = 2 < 2 [o} 7) xr E = C7) a Dea Shah tay Fo ny Oh TRA ED