CO UNiVtrtSlTY OF JLLINOIS LIBRARY AI URBANA-CHAMPAIGIM GEOLOGY a. 6?~e^(jm FIELDIANA Geology Published by Field Museum of Natural History r APR 2 1 1980 University of rlimo* « UftanaC/iampaign New Series, No. 3 December 5, 1979 Review of the Prothylacyninae, an Extinct Subfamily of South American "Dog-like" Marsupials Larry G. Marshall Visiting Curator of Geology Field Museum of Natural History INTRODUCTION In South America and Australia marsupials evolved to fill the ecological role of terrestrial mammalian carnivores. On northern continents and Africa these same niches were filled by placental carnivores; first by certain creodonts and members of other orders, and later by members of the Carnivora. These faunal differences are clearly related to the long isolation of the South American continent and to the initial presence of marsupials and absence of placental carnivores. The South American group which filled this role is classified into two families— the dog-like Borhyaenidae and the saber-tooth Thy- lacosmilidae (Marshall, 1976a)— in the superfamily Borhyaenoidea. The group is known from beds of Riochican (late Paleocene) through Montehermosan (Pliocene) age. Four subfamilies of Borhyaenidae are recognized: The Hathlyacyninae, which includes small-to- medium-sized omnivores and carnivores which were in part semiar- boreal; the Borhyaeninae and Proborhyaeninae, which included large terrestrial carnivores; and the Prothylacyninae, which in- cluded both large terrestrial carnivores and omnivores. These sub- families are distinguished on the basis of a large number of dental and cranial characters (see Marshall, 1978). Library of Congress Catalog Card No.: 79-51546 ISSN 0096-2651 Publication 1302 1 2 FIELDIANA: GEOLOGY A review of the Borhyaenidae in general and Borhyaeninae in par- ticular is given by Marshall (1978). The purpose of this paper is to review, in detail, the taxonomic history of the Prothylacyninae, to discuss the possible phylogenetic relationships of the included taxa, and to stabilize the group's taxonomy at the generic and specific levels. During the course of this study, I was able to examine, firsthand, all pertinent materials, including type and referred specimens. This work includes discussion and description of some new materials, but is essentially based on a reappraisal of previously known specimens and literature. All diagnoses of the subfamily, genera, and species have been revised and enlarged upon those of previous workers. This study represents an attempt to bring together in one place a modern and expanded treatment of these animals, the relationships of which are now better understood only in hindsight and through the pioneering efforts of a multitude of earlier workers. The fossil localities mentioned below are shown on maps and are discussed in detail in Marshall (1976a, b, c, d; 1977; 1978). The chronology and usage of South American Land Mammal Ages follows Marshall et al. (1977) for the early Tertiary and Marshall et al. (1979) for the later Tertiary. All measurements are in millimeters (mm). Abbreviations Abbreviations used in the text, figure captions, and tables of measurements are as follows: C, canine; ca, approximate measure- ment; I, incisor; L, length; M, molar; P, premolar; W, width. The following abbreviations are used for specimens from institu- tional collections: AMNH, American Museum of Natural History, New York; BM(NH), British Museum (Natural History), London; MACN, Museo Argentino de Ciencias Naturales "Bernardino Rivadavia," Buenos Aires; MLP, Museo de Ciencias Naturales de La Plata, Argentina; MNHN, Museum National d'Histoire Natu- relle, Paris; PU, Princeton University, Princeton, New Jersey; UCMP, University of California Museum of Paleontology, Berkeley. SYSTEMATICS Superfamily Borhyaenoidea (Ameghino, 1894) Simpson, 1930 Diagnosis. -Dental formula 10-4/0-3 Cl/1 P2-3/2-3 M4/4. Extinct South American "dog and cat-like" marsupials of small- to-large MARSHALL: PROTHYLACYNINAE REVIEW 3 size. Lack palatal vacuities. Transverse canal either rudimentary or absent. Strong sagittal and nuchal crests. Lunar small and in con- tact with large magnum. Lacrimal bone extends onto rostrum and usually has large tuberosity developed above lacrimal canal which opens within orbit. Known range.— Riochican through Montehermosan. Family Borhyaenidae Ameghino, 1894 Diagnosis.- Dental formula 13-4/2-3 Cl/1 P3/3 M4/4. Small-to- large size. Skull dolichocephalic to brachycephalic, rostrum robust and well developed. Upper and lower canines usually large, laniary, and with closed roots in adults. Mandibular symphysis typically shallow (may be fused or unfused in adult) and without flange. Man- dibular ramus of subequal depth and breadth below molar series, and without distinct labial bend posteriorly along ventral edge as in thylacosmilids. Masseteric fossa usually shallow. Premolars double rooted. Molars increase gradually or rapidly in size from Ml/1 to M3/4. Protocone large to very reduced. Paracone often reduced. Paracone and metacone approximated on M13. Stylar shelf reduced. Talonids large or reduced, and often imperfectly or not basined. Metaconids often absent; if present always smaller than paraconids. Nasals large and expanded posteriorly. Distinct nasal-lacrimal con- tact. No postorbital bar. Basicranial and basifacial planes parallel. Basisphenoid and basioccipital processes increase in width posteri- orly, neither has a distinct medial keel and both are relatively flat transversely; at suture they form a fairly prominent transverse ridge. Pars petrosa of periotic lacks a tympanic process. Large hypoglossal, postsquamosal and postglenoid foramina present. Known range.— Riochican through Montehermosan. Subfamily Prothylacyninae (Ameghino, 1894) Trouessart, 1898 Diagnosis.— Dental formula 14/3 Cl/1 P3/3 M4/4. Borhyaenids of medium-to-large size. Mandibular symphysis either ligamentous (Lycopsis, Stylocynus), or ankylosed (i.e., immovably united) and rami unfused in adult (Pseudothylacynus) or tightly fused in adult {Prothylacynus). Symphysis extends posteriorly to a point below P3. Large mental foramen below P2. Canine moderately developed, usu- ally not large as in Borhyaeninae or Proborhyaeninae; roots closed in adult (not open as in Proborhyaeninae or Thylacosmilidae). Lower premolars with posterobasal cusp (heel), increasing in size from P, to P3. Pj set obliquely in jaw; P23 are set straight in jaw. P3 usually 4 FIELDIANA: GEOLOGY only moderately well developed, and usually similar in size to M^ Lower molars increase rapidly in size from Ml to M4. Weak antero- basal cingulum on M24; absent on M,. Lower molars, except in Stylocynus, lack metaconid. Well-developed talonids on M13, M4 talonid usually reduced. M1"3 with large protocone and parastyle. Metacrista moderately well developed. Skull dolichocephalic (Lycopsis) or brachycephalic (Prothylacynus). Paroccipital process large. No trace of an ossified auditory bulla. Tympanic process of alisphenoid, tympanic process of pars petrosa and epitympanic sinuses lacking. Shallow floccular fossa on periotic. Foramen lacerum medium rudimentary. Foramen lacerum anterium and posterium large. Foramen ovale large (Prothylacynus) or small (Lycopis longirostrus). Terminal phalanges laterally compressed, sharply pointed, and slightly cleft. Known range.— Colhuehuapian (late Oligocene) through Huay- querian. Pseudothylacynus Ameghino, 1902 Pseudothylacynus Ameghino, 1902, p. 127. Type.— Pseudothylacynus rectus Ameghino, 1902. Distribution.— Colhue-Huapi Formation, Chubut Province, Argentina. Diagnosis.— As for type and only known species. Pseudothylacynus rectus Ameghino, 1902. Figures 1, 2; Table 1. Pseudothylacynus rectus Ameghino, 1902, p. 127. Type.— MACN 52-369, a nearly complete left mandibular ramus with PrM4. Hypodigm.— Type and MNHN Col. 5, a left mandibular ramus with P3-M4. Horizon and Locality.— Both specimens are from the Great Bar- ranca south of Lago Colhue-Huapi, Colhue-Huapi Formation, Chubut Province, Argentina. The type was collected by Carlos Ameghino and the MNHN specimen was collected by Andre Tournouer. Age. —Colhuehuapian. Diagnosis.— Medium-sized borhyaenid. Lower premolars increase in size from Px to P3. PY set diagonally in jaw at 25° angle relative to rest of tooth row. Tooth row tight but not packed, small diastema <0 o < 2 o 1 s o « S *" — II bo H !•« < S 2 « g; s g -a ^ 9 ■3. ^ -e . "o -a a c ^ O • hi . — Fig. 2. Pseudothylacynus rectus Ameghino, 1902. Stereopairs of MACN 52-369 (type), a nearly complete left mandibular ramus with Pj-M^ a, labial; b, occlusal; and c, lingual views. Scale = 3 cm. 6 — J oo ; CD co £ i-j eo J CN <N £ cc cd — — — o £ <N 00 ^ iri kc o cv M J I J 06 oi £ eo ft. J o> o oi d CN © Z z < z 8 FIELDIANA: GEOLOGY between C and Px. Lower molars lack metaconid. Moderately devel- oped talonid on M13, basined lingually but cuspate labially. M4 talonid relatively and absolutely smaller than on M13 and basined throughout its length (no labial cusp). Measurements of MACN 52-369, depth of mandibular ramus below labial side of P3 = 22.5 mm., breadth of same = 10.0 mm.; depth of mandibular ramus below labial side of M3 = 25.0 mm., breadth of same = 12.0 mm. Comments.— Pseudothy lacy nus rectus is possibly involved in the ancestry of Lycopsis torresi, but more probably in the ancestry of Prothylacynus patagonicus. The principal changes involved in such a lineage include increase in body size and slight reduction in size of talonid and protocone. A detailed discussion of the relationships of these taxa is given in the "Comments" section of P. patagonicus. Ameghino (1902, p. 127) was well aware of the similarity of these taxa as he made his comparison of P. rectus with Prothylacynus in the original description of that species. Lycopsis was not recognized as such by Ameghino. Prothylacynus Ameghino, 1891b Prothylacynus Ameghino, 1891b, p. 312. Napodonictis Ameghino, 1894, p. 380. Prothylacocyon Winge, 1923, p. 67. Type.— Prothylacynus patagonicus Ameghino, 1891b. Distribution. —Santa Cruz Formation, Santa Cruz Province, Argentina. Diagnosis.— As for type and only known species. Prothylacynus patagonicus Ameghino, 1891b. Figures 3-8; Tables 2-5. Prothylacynus patagonicus Ameghino, 1891b, p. 312; 1894, p. 380, figs. 47-49; 1898, p. 191, figs. 56, 57a; 1904, p. 21, fig. 9; 1906, fig. 185; Sinclair, 1905, p. 75, pi. II; 1906, p. 372 (with numerous plates and figures); Piveteau, 1961, figs. 21, 22. Agustylus carnifex Mercerat, 1891, p. 54. Prothylacynus carnifex Cabrera, 1927, p. 300, figs. 13, 14. Borhyaena excavata Ameghino, 1894, p. 377 (partim). Prothylacynus brachyrhynchus Ameghino, 1894, p. 380; 1898, p. 189. Napodonictis thylacynoides Ameghino, 1894, p. 381; 1898, p. 189. Type.— Prothylacynus patagonicus: MACN 706-720, a nearly 10 FIELDIANA: GEOLOGY complete left mandibular ramus and attached portion of right sym- physis with left IrM4 and right IrC (706); a left maxillary with M14 (707); and associated postcranial remains (708-720), all of a single individual. Figured by Ameghino (1894, figs. 47-49; 1898, figs. 56, 57a; 1904, fig. 9; 1906, fig. 185). Type.— Prothy lacy nus carnifex: MLP 11-38, a nearly complete mandible with most of dentition present, but broken. Figured by Cabrera (1927, figs. 13, 14). Type.—Prothylacynus brachyrhynchus: MACN 5926, portion of a mandible with left and right rami fused, with left C, alveoli of Pj, P2 complete, alveoli of P3, M14 complete; and right P2 complete, alveoli of P3-Mx, M24 complete. Lectotype.—Borhyaena excavata: MACN 649, a right lower canine. Type.—Napodonictis thy lacy noides: MACN 5931-5937, a com- plete skull with dentition (M4 and P3 erupting; 5931); greater part of a broken right mandibular ramus with dentition (M4 erupting; 5932); and part of an associated and fragmentary skeleton (5933- 5937), all of a single individual. Hypodigm.— The five types and PU 15700, an associated partial skull, mandible, and partial skeleton; MACN 189, a fragment of a right mandibular ramus with roots of M2.3, trigonid and posterior root of M4; MACN 670, isolated right M2 missing parastyle; MACN 11640, posterior half of a left and right mandibular ramus, both with M24 present, but partially broken; MACN 14453, greater part of a skull with most of dentition; AMNH 9561, right maxillary frag- ment with M2'4 (occlusal surfaces of M2 3 are heavily worn and small protocone of M2 and part of that of M3 are missing); BM(NH) M8075, part of a maxilla with teeth; BM(NH) M9178, posterior part of a mandibular ramus with teeth. Horizon and Locality.— All specimens are from the Santa Cruz Formation, Santa Cruz Province, Argentina, and their specific local- ities of collection are as follows: Monte Observation MACN 649, Opposite: Fig. 4. Prothylacynus patagonicus Ameghino, 1891b. Stereopairs of MACN 706 (type), a nearly complete left mandibular ramus and attached portion of right sym- physis with left lx-MA and right IrC: a, labial; b, occlusal; and c, lingual views. Scale = 3 cm. L,.~M '^±,>J.s L* j 1 > 11 Fig. 5. Prothylacynus patagonicus Ameghino, 1891b. MACN 5926, portion of a mandible with left and right rami fused, with left C, alveoli of Px, P2 complete, alve- oli of P3, Mj.4 complete; and right P2 complete, alveoli of P3-Mlt M2.4 complete: a, right lateral; b, occlusal; and c, left lateral views. Scale = 5 cm. 12 Fig. 6. Prothylacynus patagonicus Ameghino, 1891b. MACN 5931, a nearly com- plete skull of a young adult with most of dentition (M4 and P3 are erupting): a, dor- sal; b, right lateral; and c, ventral views. Scale = 5 cm. 13 14 FIELDIANA: GEOLOGY 670 (collected by C. Ameghino 1890-1891); Corriquen-Kaik MACN 706-720, 5926, 5931-5937 (collected by C. Ameghino 1890-1891); Santa Cruz MACN 189, MLP 11-38; Canadon de las Vacas MACN 14453 (collected by A. Bordas 1941-1942); Felton's Estancia BM(NH) M9178 (collected by H. T. Martin in 1903 and sold to BM(NH) in 1905), PU 15700 (collected by J. B. Hatcher in 1896), AMNH 9561 (collected by B. Brown in 1899); MACN 11640 and BM(NH) M8075 (presented by C. Arthur Pearson, June 1902) are without specific locality data. Age.— Santacrucian (early Miocene). Diagnosis.— Medium-sized borhyaenid. Mandibular symphysis ankylosed and both rami tightly fused in adult. P1 set at 30° oblique angle relative to rest of tooth row. P3 large, but not prominent as in species of Cladosictis and Borhyaena. Ph3 relatively shorter and more robust than in other Prothylacyninae. M1'3 with large proto- cone which is cuspate and never basined as in other Prothylacyni- nae. Skull brachycephalic. Foramen ovale large. Description.— Symphysis of jaw extends posteriorly to point below P3-Mj contact. Lower incisors increase slightly in size from Ix to I3. Cheek tooth row relatively short and more packed than in Pseudothylacynus rectus. In length P1(P2/>I>3» in width Pj <^P2 = P3. M14 are proportionately shorter and more robust than in P. rectus. M13 with small to moderately well developed talonid which is flat or slightly cuspate but never basined. Talonid as wide as trigonid on Mj.a, narrower on M3. Small but distinct talonid pre- sent on M4 which is always basined and never cuspate. Weak anterobasal cingula present on M24, but not on Mv Upper incisors increase in size from I1 to I4; increase in size from I1 to I3 gradually, I4 much larger than I3 . In length and breadth P*\ P2<( P3. Premolars are separated from each other and adjacent teeth by small diastems. P1 3 have large posterobasal heel which increases in size from P1 to P3 and is best developed on P2'3. P1 is set obliquely in jaw relative to rest of tooth row. P1 3 are markedly shorter and more robust than in other Prothylacyninae but less so than in Borhyaeninae. In length M1<CM2)>M3, in width M1^M2<M3. M4 protocone reduced but larger than occurs in Borhyaena tuberata. No trace of para- or meta- Opposite Fig. 7. Prothylacynus patagonicus Ameghino, 1891b. MACN 14453, greater part of skull with most of dentition: a, dorsal; b, right lateral; and c, ventral views. Scale = 5 cm. 15 16 FIELDIANA: GEOLOGY conules on M1 4. Paracone smaller than metacone and becoming pro- portionately smaller from M1 to M3. Metacone becomes relatively and absolutely larger from M1 to M3. Paracone and metacone con- nate basally. Paracone dominant cusp on M4. Very large parastyle present on M1"3 which is separated from paracone and not connected to it by a paracrista. A distinct ectocingulum is developed posterior to metacone on M1'3 which encloses pocket between it and meta- crista. Metacrista is well developed and forms major shear surface on M1'3. A small, but distinct, ectoflex is present on M2'3 but not on M1. Skull is brachycephalic. Infraorbital foramen is rather small (diameter = 9.0 mm.; depth = 4.5 mm. on right side of MACN 14453) and opens over point between P^M1 contact. Anterior edge of orbit extends to point above M2. Postorbital process weak and poorly defined. Recess for external auditory meatus is a large shallow U-shaped trough, upper border of which is on line with dorsal edge of occipital condyles; it is bordered posteriorly by a large massive mastoid process, the ventral edge of which reaches a point corresponding to middle of occipital condyle. Ventral edge of mastoid process does not extend forward toward glenoid area of zygomatic arch as in Lycopsis longirostrus. No trace of distinct paroccipital process, nor of a tympanic process of alisphenoid. No evidence of an ossified auditory bulla. A rather ill-defined anterior epitympanic sinus is present; no trace of posterior epitympanic sinus. A large postsquamosal foramen opens directly above dorsal edge of recess for external auditory meatus. A large postglenoid foramen opens onto posterior surface of zygomatic arch just behind and at medialmost edge of postglenoid process near anterodorsal edge of recess for external auditory meatus. A large hypoglossal foramen opens just medial to each occipital condyle, and a very tiny foramen occurs 4.0 mm. anteromedial to latter. A large foramen ovale opens within alisphenoid midway between innermost edge of postglenoid process and outer edge of basioccipital-basisphenoid contact. A very tiny foramen lacerum medium opens 14.5 mm. anteromedial to foramen ovale. Body of petrosal is bulbous and pea-shaped on outer Opposite: Fig. 8. Prothylacynus patagonicus Ameghino, 1891b. Stereopairs of MACN 707 (type), a left maxillary with M1'4: a, labial; b, occlusal; and c, lingual views. Scale = 3 cm. 17 18 FIELDIANA: GEOLOGY Table 2. Measurements of skull of Prothylacynus patagonicus (in mm.). Measurement MACN 5931 MACN 14453 Anterior edge of incisors to posterior edge of occipital condyles 171.0 .... Anterior edge of incisors to posterior edge of secondary palate 83.0 90.0 Interorbital breadth between postorbital processes. . 39.5 42.0 Maximum combined transverse breadth of occipital condyles 42.0 .... Maximum dorso-ventral depth of occipital condyles 19.5 .... Internal breadth between mastoid processes 32.5 .... Maximum width of rostrum immediately anterior to infraorbital foramen • • • • 34.0 Transverse breadth of palate between canines .... 19.0 Transverse breadth of palate between M4's .... 48.0 Ventral edge of occipital condyle to dorsal edge of nuchal crest .... 39.5 Maximum transverse breadth of nuchal crest at level of occipital condyles .... 55.0 surface; a tiny, but distinct, tympanic process is present. Basioccipi- tal and basisphenoid are relatively flat transversely; they lack a distinct medial keel, progressively increase in width posteriorly, and at their contact form a fairly prominent transverse ridge. Posterior carotid foramina and foramen lacerum posterius open into ear region between basioccipital and petrosal. Alisphenoid-squamosal suture lies along innermost edge of glenoid fossa. Measurements of upper canines: left side of MACN 14453, height of crown = 22.0 mm.; anteroposterior diameter of base of crown = 11.0 mm.; transverse breadth of base of crown = 8.0 mm. Measure- ments of lower canines: anteroposterior diameter of base of crown in MACN 706 = 12.0 mm., MACN 5926 = 9.5 mm., MACN 649 = 12.2 mm.; transverse breadth of base of crown in MACN 706 = 8.0 mm., MACN 5926 = 7.8 mm., MACN 649 = 8.8 mm. ^ o -^ CO 00 o >o no<s T W lO lO SJ t- t> OS ■* T ^ . . . ^ ©!©•••• ?J CN CN t- t~ ft. <d CO 'V >o <C 00 <o t- 00 00 00 t» t* oo p ; oq OS OS '. © O O O CO CN tj! in id Tt* eo N O h * N H CN I-H -H r-( ^ .-H » N H 1(3 ^| co t^ t> cd co ■c -J I I * ■a j in i-H p p cn i-j n ri ri n n fi CN -H 1-H O p O O O O A 00 o co 0 ■> o o o O CO p t> t> p CON NH H H as t> CD CO oo iq oq oo as d d 6 oi ffi CO CO 00 00 lO CO -— ■^ ^ ^* ^ i-H o CM o 0) CJi X 00 CO o CO CO I a e CJ .3 e a, O n — CO l> l> p O 1-J CN CN CN CN «-H .-H CN CN CN p cd cd cd id 00 O) OJ H i(J O O i- 1 •- • OS p p id id OS 00 OS a 03 os os io m ^ ^H OS PC z < < < < s Z Z Z Z o o o o SSSSS< a Si S3- ^ ~~ © O CO CD -^ ^ 00 CO CN CN CO CO m O 3) ft H H ' MO lO H H | 0) a Z MLP 1] MACN MACN MACN MACN MACN 19 20 FIELDIANA: GEOLOGY Table 5. Measurements of mandibular ramus of Prothylacynus patagonicus. Depth of ramus Breadth Depth of ramus Breadth Specimen below labial side of Ml of same below labial side of M4 of same MACN 189 28.0 10.2 MACN 706 28.0 12.5 31.0 13.3 MACN 5926 (1) 26.2 12.0 MACN 11640 (1) 28.0 11.3 MACN 11640 (r) 28.6 10.9 Comments.— All of Ameghino's figures of Prothylacynus pata- gonicus are based on the holotype, MACN 706-720. MACN 706, a nearly complete left mandibular ramus, was figured by Ameghino (1894, figs. 47, 48; 1898, fig. 56; 1904, fig. 9; 1906, fig. 185). Although he figured (1894, fig. 48; 1898, fig. 56) this specimen as a complete mandible, he does not mention that the right ramus was ever complete. The right ramus was apparently restored in these figures by making a mirror image of the complete left ramus. The left maxillary fragment of the type, MACN 707, containing M1'4 was figured by Ameghino in 1894 (fig. 49) and 1898 (fig. 57a). Ameghino (1894, p. 380) distinguished "Prothylacynus brachy- rhynchus" (type, MACN 5926) from P. patagonicus in being a little smaller in size and proportionately more massive in structure; in having smaller lower canines which almost contacted, leaving little room for the incisors which "must have been" rudimentary; and in the transverse placement of the Pv The canine of MACN 5926 is cer- tainly smaller than that in the type of P. patagonicus, although these differences are not excessive and they surely represent no more than individual variation. Ameghino's reference to the smaller size of "P. brachyrhynchus" can only be applied to the canine, as in all other features this specimen is as large as or larger than the type of P. patagonicus (table 4). Ameghino's comment that the incisors "ought" to be rudimentary is only an assumption, as neither the incisors nor their roots or alveoli are preserved in this specimen. The specific name "brachyrhynchus" was surely applied in refer- ence to the short massive nature of the jaw and crowding of the anterior cheek teeth. Both of these features are alluded to by Ameghino, but neither is real. The anterior edge of the mandible, in- cluding the incisor alveoli and left canine alveolus, was broken off MACN 5926 and lost. When the left canine was restored onto the specimen, it was placed in the vacuity left by this missing portion of the jaw, and now the canine lies in the area originally occupied by MARSHALL: PROTHYLACYNINAE REVIEW 21 the anterior root of the P1# The canine is thus very close to the P2, giving the jaw the "brachyrhynchus" appearance; an artificial feature and the result of erroneous restoration. "Agustylus carnifex" was erected by Mercerat (1891, p. 54) on the basis of a nearly complete, but poorly preserved mandibular ramus, MLP 11-38. Ameghino (1894, p. 380) recognized this species as a junior synonym of P. patagonicus. Cabrera (1927, pp. 300-301, figs. 13, 14), however, redescribed and first figured MLP 11-38, recogniz- ing it as a valid species of Prothylacynus. He called attention to the fact that "P. carnifex" was smaller than P. patagonicus, it had a more reduced M4 talonid, and the symphysis was longer. He further noted that all of these features agreed with "P. brachyrhynchus" which he seems to have regarded as a synonym of that species, but he is not clear on this matter. The characters used by Cabrera to support his claims are all at- tributable, in large part, to the poor state of preservation of MLP 11-38. The symphysis is long and low because its ventral edge has been removed by erosion, and there is matrix surrounding the sym- physis, giving it a superficially elongated appearance. Similarly, the M4 talonid, although small, is covered by matrix and is not really absent. In size, this specimen agrees well with dimensions of the type of P. patagonicus (table 4), although it does tend, on the whole, to be somewhat smaller. There is little doubt, however, that "A carnifex" is referable to P. patagonicus. The only character used by Ameghino (1894, p. 381) to distinguish "Napodonictis thy lacy noides" from P. patagonicus was that in the former the M4 had "el talon interno mucho mas reducido." He later (1898, p. 189) noted that "A/, thy lacy noides" and P. patagonicus were very similar, but again called attention to differences in mor- phology of the M4. The M4 of the type of P. patagonicus (MACN 707) is slightly longer and wider than the type (MACN 5931-5937) of "A/, thyla- cy noides" (table 3). Morphologically, however, these elements are identical and I can find no trace of the unique features alluded to by Ameghino. Considering the fact that these specimens agree perfect- ly in all other respects, there is little problem in recognizing them as synonymous. The only other specimen referrable to P. patagonicus for which there exists a prior literature reference, is a syntype of ' 'Borhyaena excavata." This "species" was erected by Ameghino (1894, p. 377) on four specimens: MACN 649, a nearly complete left lower canine; 22 FIELDIANA: GEOLOGY MACN 650, part of a left mandibular ramus with two molars; MACN 651, a right lower canine; and MACN 652, part of an M4. All are listed in Ameghino's catalogue in the MACN as belonging to "B. excavata," the first three are labeled "TIPO," and are specifi- cally referred to in his original description of that "species." MACN 649, which is hereby designated as lectotype, is referrable to P. pata- gonicus. MACN 650 and 652 could not be located in the MACN col- lection and their precise identifications are thus not definitely known. MACN 651 is referrable to Borhyaena tuberata (Marshall, 1978, p. 50). Prothylacynus patagonicus is very similar to the Colhuehuapian species Pseudothylacynus rectus. The principal features shared by these species include: relative proportions of jaw, canine, and cheek tooth row; Px notably smaller than P23 and set obliquely in jaw; P^ with a posterobasal heel which increases in relative size from Px to P3; Mh3 with small to moderately well-developed talonid and M4 with much smaller talonid; and with a weak anterobasal cingulum on M2.4, but absent from Ml. The principal differences between these species include: P. pata- gonicus being larger in size; cheek tooth row being more crowded and with P^M,, being proportionately shorter and more robust; Px being relatively smaller than P2.3 and set more obliquely in jaw; talonids on M14 being smaller relative to trigonids and never incipi- ently basined but flat or slightly cuspate; M4 talonid being very reduced but distinctly basined as in P. rectus; and mandibular sym- physis being ankylosed and with both rami solidly fused in the adult. Individually ,these differences are not excessive and there are no characters in P. rectus which would exclude it as a direct ancestral form of P. patagonicus. In fact, most of the features by which they differ are either incipiently developed in P. rectus, or are developed to different degrees in both species. An ancestral-descendant rela- tionship for these species appears probable. Lycopsis Cabrera, 1927 Lycopsis Cabrera, 1927, p. 295. Type.— Lycopsis torresi Cabrera, 1927, p. 295. Distribution.— Santa Cruz Formation, Argentina; and "Monkey Unit" of Honda Group, Colombia. Diagnosis.— Medium to large in size. Protocone large and deeply basined on M1'3, slightly smaller on M4. Paracone becomes smaller MARSHALL: PROTHYLACYNINAE REVIEW 23 from M1 to M3; metacone becomes relatively, and absolutely, larger than paracone from M1 to M3. Upper molars virtually lack stylar shelf; ectocingulum is present labial to paracone; parastyle is large; ligamentous mandibular symphysis, unfused in adult; protoconid trenchant and about twice as high as paraconid. Talonid basin large on M13, small on M4. PI has slight oblique implantation labially at anterior end; P2 and 3 are aligned anteroposteriorly; P3 is well developed but not prominent. Lycopsis torresi Cabrera, 1927. Figures 9-13; Table 6. Lycopsis torresi Cabrera, 1927, p. 295, figs. 11, 12. Anatherium(1) oxyrhynchus Ameghino, 1894, p. 384. Anatherium oxyrhynchus Cabrera, 1927, p. 298. Anatherium oxyrhynchum Roger, 1896, p. 17. Cladosictis oxyrhynchus Sinclair, 1906, p. 447. Cladosictis oxyrhyncha Simpson, 1930, p. 45. Type.—L. torresi: MLP 11-113, a fragment of a left maxilla with P2 and M1'4 nearly complete; a right maxillary fragment with M1'4 complete; part of a left mandibular ramus with P3-M3 complete and talonid of M4; and greater part of right mandibular ramus with C-P3 complete, M23 partially broken, and posterior root of M4; all of a single associated individual. Type.— A. oxyrhynchum: MACN 5930, a left mandibular ramus with alveolus of C, alveoli of Plt posterior half of P2, P3 complete, roots of Mlt talonid of M2, and greater part of M3.4. Hypodigm.— Types only. Horizon and Locality.— Both specimens were collected from the Santa Cruz Formation, Santa Cruz Province, Argentina. MLP 11-113 was collected by C. Berry in 1895 from an unspecified site along the Rio Santa Cruz. MACN 5930 was collected by C. Ame- ghino from Corriquen-Kaik. Age. — Santacrucian. Diagnosis.— Differs from L. longirostrus in being smaller in size, in M4 being wider than M3, in protocone on M1"4 and talonid basin being relatively and absolutely smaller; in P2 and P3 being subequal in size; and in M14 being relatively and absolutely shorter and more robust. Description.— In length Pi<(P2<P3- A distinct posterobasal heel is present on P13. Pl is set at 20° angle relative to rest of tooth row. Talonid is broader than trigonid on M„ subequal on M2, and nar- 24 MARSHALL: PROTHYLACYNINAE REVIEW 25 rower on M3.4. Hypoconid cuspate and larger than entoconid on M^, but noncuspate and subequal to entoconid on M3.4. Distinct hypoconulid visible on Mj.3 in unworn teeth. P2 has well-developed posterobasal heel. Upper molars increase in length and breadth from M1 to M3; M4 is wider than M3. Para- and metaconules distinguishable on M2'4. Metacone very large and con- nate basally with smaller paracone. Parastyle large and connected to protocone by paracingulum along anterior surface of paracone. Small stylar shelf developed on M1 just labial to metacone, but does not connect up with parastyle. No distinct ectocingulum developed posterior to paracone-metacone junction on M1'3. Ectocingulum encloses pocket between it and paracone on M2'3, but not on M1. Large ectoflex present on M3. Paracone on M4 is dominant cusp and is connected to parastyle by large paracrista. Cranial characters are unknown. Comments.— Anatherium(1) oxyrhynchus was erected by Ame- ghino (1894, p. 384) on a partial left mandibular ramus with denti- tion. This specimen, MACN 5930, presently consists of a very fragmentary mandiblular ramus with part of the P2, all of the P3, talonid of the M2, and nearly complete crowns of M3.4. It is accom- panied by numerous tooth fragments and parts of the symphyseal region of the ramus. Judging from Ameghino's original description, this specimen was more complete when he studied it, presenting details no longer preserved. As noted by Ameghino (1894, p. 384), "The space between the two lower canines is very reduced and the incisors are in part atrophied. The canine is strongly inclined anteri- orly, the symphysis is very long and ends almost in a point. The first premolar is situated obliquely or almost transversely; the sec- ond (P2) and third (P3) have the posterior heel atrophied. The talonid of the M4 is well developed. The distance from the anterior part of the mandible to the posterior border of the M4 is 97.0 mm. Pj-M4 measures 76.0 mm. in length. The M4 measures 13.0 mm. in length. The surface of the symphysis measures 48.0 mm. in length and 15.0 mm. in maximum height. The mandible is 19.0 mm. in height below the P2, and this height is conserved more or less to the last molar ..." (translated from Spanish). Opposite Fig. 9. Lycopsis torresi Cabrera, 1927. Stereopairs of complete right C-P3 and alveoli of Ml of MLP 11-113 (type): a, lingual; b, labial; and c, occlusal views. Scale = 3 cm. Fig. 10. Lycopsis torresi Cabrera, 1927. Stereopairs of left P3-M3 of MLP 11-113 (type): a, labial; b, occlusal; and c, lingual views. Scale = 3 cm. 26 mat T| **■■* .. u i ^i -.k »v ' 1 UpJ0 »^n? -B^ * j&^_ ^B ^fl mf Fig. 11. Lycopsis torresi Cabrera, 1927. Stereopairs of right M1'4 of MLP 11-113 (type): a, labial; b, occlusal; and c, lingual views. Scale = 3 cm. 27 o is CO X l- _o "S OJ -<-> CO O E a u ,2, LC Cu II <«-i o JB £3 "3 o CJ CO t>5 j> ~3 CO d CO H-l '> o at "3 _3 co 3 "o CD "o > U O 13 -C -O -LJ -o 'S c CO CO 3 lo E CO % hi a H a -2 3 CO -Q *TT -o CO c s CO E O *j -U <*- l_ _a> CO CO a i- © CO CO OS CO § CO < 5 C4 l> § CS M-. a> O 1-1 35 CO 'c CD _o u CO £> -w CO o S »J CD E o o CO ■4^ *-> O y. C V-. s- ^ CO -J _0) ^ "a E cm' o 1 — 1 CJ 2 1* A 1 X 28 ■C> C5 2 Tf |> «5 © 6 lO CO in o in to CO to -«j. <» 1— 1 l-H 1— 1 i— 1 N in lO oo t^ dd co o6 co in in ie ; © co co t> 1 ci s s 1 I a co e -4 ! tj CO o 2 ^ n co i> © ai i-5 (■ CM i-i co in ^ co co iri i-t ih © © q co in co © q o o 00 t> © CO • m ,_, 00 . i— i co i— i ©' 1—1 CO rH in co * i r o o m ^h js — i .* PH A J3 S P - - 3 s h s *j CO to co •J f f J3 <35 © 3 »n m $ a. Pj -g 22 co co PL, Cm SSS 2 o ■t 00 «0 CO •§eu -3 29 30 FIELDIANA: GEOLOGY In Cabrera's (1927, p. 295, figs. 11, 12) original description of Lycopsis torresi, he specifically compared it with "Anatherium oxy- rhynchus." He noted, by inference, that "A oxyrhynchus" had atro- phied talonids on P23, while in L. torresi these were better developed. The talonids in "A. oxyrhynchus" are indeed smaller than in the type of L. torresi, but these differences are minor and are of no tax- onomic importance. Cabrera further noted that in "A. oxyrhynchus" the depth of the mandibular ramus was relatively constant from P2 to M4, but becomes markedly deeper from P2-M4 in L. torresi. The mandibular ramus of the type of L. torresi has been extensively restored with wax, such that the true structure of this element below the molars is open to question. Similarly, the ramus of the type of 'A. oxyrhyn- chus" is now totally destroyed below the molar series and what its condition was when studied by Ameghino is not surely known. As indicated by tooth wear, the type of ''A. oxyrhynchus" is of a young animal, while that of L. torresi is of a middle-aged adult. As size of the mandibular ramus is subject to age and sex differences, it is of dubious value as a taxonomic character. In all other respects these specimens are virtually identical and they are surely synonymous. Cabrera concluded his discussion on L. torresi by noting that "A oxyrhynchus" possessed none of the characters peculiar to the genus "Anatherium " and that it may warrant being placed in a new genus. His intuition was correct. The type species of Anatherium, A. defossum, is now included within Cladosictis (see Marshall, 1978, p. 21), and the genus Anatherium is no longer valid. Lycopsis torresi differs from L. longirostrus, primarily in its smaller size, in the M4 being wider than the M3, in the metastylar shear being proportionately less well developed, in the lower pre- molars and molars being proportionately shorter and more squat, and in the talonid basin being proportionately and absolutely smaller. These differences are minor and, as I proposed earlier (1977), it is probable that L. torresi represents an ancestral form of L. longirostrus. Opposite Fig. 13. Lycopsis torresi Cabrera, 1927. Stereopairs of MACN 5930, a left man- dibular ramus with alveolus of C, alveoli of Pj, posterior half of P2, P3 complete, roots of Mi, talonid of M2, and greater part of M3.4: a, labial; b, occlusal; and c, lingual views. Scale = 3 cm. 31 32 FIELDIANA: GEOLOGY It is likewise probable that L. torresi evolved from a form not too different from the Colhuehuapian species Pseudothylacynus rectus. Both share the following features: Pj is smaller than P23 and is set at 20° in the jaw relative to the rest of the tooth row; P23 are elon- gated anteroposteriorly and have a distinct posterobasal cusp (heel); talonid basin is distinct and is well developed on M13, smaller on M4; and the hypoconid is cuspate and is larger than the entoconid on Mj.g, subequal and less distinct on M34. In addition, the overall pro- portions of the teeth of these species are similar. They differ in that L. torresi is larger in size and has relatively and absolutely larger talonid basins on M14 (especially M4). The overall morphology of P. rectus does, however, indicate closer affinities with P. patagonicus than with L. torresi. Because of this I propose that P. rectus be regarded as directly ancestral to P. pata- gonicus and that L. torresi evolved from a pre-P. rectus form and not directly from that species. Lycopsis longirostrus Marshall, 1977. Table 6. Lycopsis longirostrus Marshall, 1977, p. 634, figs. 1-4. Type. -UCMP 38061, right half of a skull with C-M4, right den- tary with roots of C, Pi-M4 complete, and greater part of articulated skeleton. Hypodigm.— Type only. Horizon and Locality.— "Monkey Unit" of Honda Group from Upper Magdalena Basin, northeast of village of Villavieja in north- ern part of Department of Huila, Colombia. Type locality is UCMP V-4521. Age.— Friasian (medial Miocene). Diagnosis.— Differs from L. torresi in being larger in size, in M4 being narrower than M3, in protocone on M1"4 being relatively and absolutely larger, and in P2 being larger than P3. In addition, a para- cingulum occurs on M1"3; hypoconid is larger than hypoconulid or entoconid on M12; skull is dolichocephalic; petrosal lacks a subarcu- ate fossa, pars mastoidea, and tympanic process; tympanic process of alisphenoid is lacking; no evidence of an ossified auditory bulla; foramen ovale is small; foramen lacerum medium, carotid canal, and transverse canal are all very small. Description.— A detailed description of UCMP 38061 along with illustrations of the dentition, skull and skeleton are given by Mar- shall (1977). MARSHALL: PROTHYLACYNINAE REVIEW 33 Comments.— Lycopsis longirostrus probably passed through a L. torresi grade at some time in its evolution as all of the characters in the former can be derived with but minor modification from the latter. Considering the overall morphological similarities between these species and their occurrence in consecutive land mammal faunas, an ancestral-descendant relationship is probable. Pseudolycopsis Marshall, 1976d Pseudolycopsis Marshall, 1976d, p. 291. Type. — Pseudolycopsis cabrerai Marshall, 1976d. Distribution.— Arroyo Chasicb Formation, Department of Villa- rino, southwest corner of Buenos Aires Province, Argentina. Diagnosis.— As for type and only known species. Pseudolycopsis cabrerai Marshall, 1976d. Figure 14, Table 6. Pseudolycopsis cabrerai Marshall, 1976d, p. 291, fig. 1. Type.— MLP 57-XI-9-1, a fragment of a palate with left F-M3 and right M23. Hypodigm.— Type only. Horizon and Locality.— Arroyo Chasic6 Formation, Department of Villarino, southwest corner of Buenos Aires Province, Argentina. Age.— Chasicoan (medial to late Miocene). Diagnosis.— Protocone is located toward anterior edge of tooth and lies directly opposite paracone; parastyle is very large on M1 and it has form of spur that projects anteriorly toward P3; main cusp of M1 parastyle is in line with para- and metacone; molars lack ecto- and paracingulum; metacrista is very long and narrow; molars are overall very narrow transversely and are elongated anteropos- teriorly. Description.— Similar in size to Prothylacynus patagonicus. P3 is small, gracile, and lacks a distinct posterobasal heel; its crown is about same height as metacone of M1. Molars increase in length and width from M1 to M3. Protocone is well developed and has a shallow basin on M1"3; protocone decreases slightly in size from M1 to M3. Paracone is small and is fused basally with larger metacone; para- cone becomes slightly smaller from M1 to M3, while metacone becomes relatively and absolutely larger in the same direction. Parastyle on M2 and M3 is shorter than on M1, is cuspate, and is set 34 FIELDIANA: GEOLOGY labiad of a line connecting paracone and metacone. A broad shallow ectoflex is present on M3 opposite metacone. Comments. — In the upper dentition of Lycopsis torresi the molars have a large basined protocone, a large (but not huge) parastyle, and a reduced ectocingulum which appears only on the M1 posterior to the metacone, being absent on M2'3. These same features probably occurred in the ancestor of P. cabrerai and only slight modification of the dentition of L. torresi is required to obtain the dental speciali- zations seen in P. cabrerai. There are, however, two factors which suggest that L. torresi may not in fact be ancestral to P. cabrerai. First, L. torresi is larger than P. cabrerai, and if an ancestral-descendant relationship exists then a diminution in size would have had to occur in this lineage. Second, it appears more likely that L. torresi is ancestral to L. longirostrus (Marshall, 1977) and that within that lineage, and assuming no branchings, there was a tendency for increase in size and for greater development of the protocone and talonid basins. Notwithstanding these factors, L. torresi is basically very similar to P. cabrerai and although the former may not be ancestral to the latter, it certainly approximates what would be expected in an ancestral form of that species. The upper molars of Prothylacynus patagonicus are similar in size to those of Pseudolycopsis cabrerai, although there are characters present in the former which suggest that it is not involved in the ancestry of the latter. These characters include: 1) the protocone, although large and well developed, is cuspate and is not basined; 2) the parastyle is very large and a large ectocingulum occurs on M1"3; and 3) the P3 is large and rather robust, although not to the degree seen in Borhyaena, and it has a distinct posterobasal heel. It appears that unless there occurred a secondary enlargement of the proto- cone and a very marked reduction in size of the parastyle, ectocingu- lum, and P3, then P. patagonicus can probably be eliminated as a potential ancestor for P. cabrerai. Pseudothylacynus rectus may be involved in the ancestry of Pseu- dolycopsis cabrerai, and, for that matter, L. torresi and/or Prothyla- cynus patagonicus as well (see above). The upper dentition of P. Opposite Fig. 14. Pseudolycopsis cabrerai Marshall, 1976d. Stereopairs of left side of MLP 57-XI-9-1, a fragment of a palate with P3-M2 and M3 broken: a, Ungual; b, occlusal; and c, labial views. Scale = 3 cm. 35 36 MARSHALL: PROTHYLACYNINAE REVIEW 37 cabrerai complements the lower dental structure of P. rectus; i.e., they are similar in size and they occlude rather closely. P. cabrerai, however, has a smaller protocone than is expected in the upper den- tition of P. rectus. Nothing definite can be said about the relation- ship of these taxa, except that in their complementary morphology they are similar. Whether or not this similarity is a reflection of direct phylogenetic affinity is presently not known. Stylocynus Mercerat, 1917 Stylocynus Mercerat, 1917, p. 20. Sthylocynus (sic) L. Kraglievich, 1934, p. 62. Type.— Stylocynus paranensis Mercerat, 1917, p. 20. Distribution.— Formation "Entrerriense," near Parana, Entre Rios Province, Argentina. Diagnosis.— As for type and only known species. Stylocynus paranensis Mercerat, 1917. Figures 15-18; Tables 7-8. Achlysictis lelongi Ameghino, 1891, p. 147 {partim, not holotype). Stylocynus paranensis Mercerat, 1917, p. 20n; L. Kraglievich, 1917, p. 278 (as nomen nudum); 1934, p. 62; Cabrera, 1927, p. 278, fig. 3. Type.— MLP 11-94, originally a nearly complete left mandibular ramus with alveoli of incisors and C, P12 present but broken, roots of P3-M,, base of M2, and M34 complete. Since the specimen was described by Mercerat and figured in Cabrera, the ramus has been broken in several places and is now partially restored in plaster. Only M34 remain— Pj, P2, and part of the M2 have been lost. Hypodigm. -The type and MLP 41-XII-13-1112, a fragment of a right mandibular ramus with alveoli of C, P^M! complete; MACN 5893, a fragment of a right maxilla with M2 complete, roots of M1, and posterior alveolus of P3; and MACN 13203, a right maxillary fragment with P2 and M1'2 present but very worn, and alveoli of P3. Horizon and Locality.— All specimens were collected from the "barrancas del rio Parana," "Formacidn Entrerriense," Entre Rios Province, Argentina. Age. — Montehermosan. Opposite Fig. 15. Stylocynus paranensis Mercerat, 1917. Stereopairs of M3.4 of MLP-11-94 (type): a, labial; b, occlusal; and c, lingual views. Scale = 3 cm. 38 FIELDIANA: GEOLOGY Diagnosis.— Largest known species of Prothylacyninae. Mandibu- lar ramus is exceptionally deep and narrow relative to size of teeth. An enormous mental foramen is located below P2. Mandibular sym- physis is broad, ligamentous, and rami are unfused in adult. All lower premolars are large and well developed. Small, but distinct metaconid present on MM. Talonid extremely large and basined on Mi.a, slightly smaller but still large on M4 M1"3, and probably also M4, with very large basined protocones. Description.— Symphysis extends posteriorly to point below anterior root of P3. Two medialmost incisors are larger than labial- most incisor. In length Pl < P2 ) P3. P1 is set at 15° angle relative to rest of tooth row. A distinct posterobasal heel is present on P2.3, but only a hint of this structure occurs on P1. Smal* diastems sepa- rate P13 from each other as well as from adjacent teeth. Lower molars increase in length and width from Mx to M4. Distinct antero- basal cingulum is present on M14. Talonid is wider than trigonid on Mj.g, but is smaller on M4. Hypoconid is markedly larger than ento- conid on Ma (and probably M2), subequal on M3.4. Infraorbital fora- men opens immediately above P3. P23 are aligned in same antero- posterior axis. Metacone is larger than paracone on M1, becoming progressively and absolutely larger on M2. Paracone and metacone are connected basally by low, broad ridge. Small parastyle is pre- sent on M2, and weak labial shelf extends posteriorly from it to point opposite middle of metacone. Metastylar region is well developed, but is not prominent. Comments.— On the type (MLP 11-94), prominent rugosities, probably representing muscle scars, occur along the labial surface of the mandibular ramus and are especially prominent along the upper surface between Ml and M4. These scars probably indicate the pres- ence of a large superficial masseter muscle, and together with the dental specializations (e.g., large protocones and talonids) suggest a predominately omnivorous diet. The name Stylocynus paranensis was applied by Mercerat (1917, p. 20, footnote 1) to a left mandibular ramus with partial dentition, collected from "los depositos sedimentarios terciarios del Parana." The description given by Mercerat is complete, is accompanied by measurements of the specimen, and the species is compared with other Borhyaenidae, especially with Prothylacynus patagonicus. This comparison was well chosen as P. patagonicus was, at that time, the only species of borhyaenid sharing numerous characters with S. paranensis. Nevertheless, L. Kraglievich (1917, p. 278) c — ■S 1 CO 3 E co 3 a ^ I x s " s i 2 £ CD "O C 5 a cq s,J is CO O U. i 39 40 FIELDIANA: GEOLOGY regarded S. paranensis as a nomen nudum, without giving sound justification for such action. Cabrera (1927, p. 278, fig. 3) later redescribed and figured the type, noting (p. 280) that contra L. Kraglievich, the species is perfectly valid. Stylocynus paranensis is one of the most specialized borhyaenids known. The protocone on M1 3 and corresponding talonid basins on M13 (and to a lesser extent on M4) are relatively and absolutely larger than in any other known species. The small, but distinct metaconid on M14, when considered with the large talonid, presents a pair of characters unknown in other members of the family, but resembles the Friasian didelphid Hondadelphys fieldsi (Marshall, 1976c). The enormous size of the anterior mental foramen below the P2 further represents a diagnostic feature found only in this species among Prothylacyninae. In Ameghino's (1891a) original description of Achlysictis lelongi, a species now referred to the saber-tooth family Thylacosmilidae (Marshall, 1976a), he referred a "penultimate" upper molar to that species. This specimen, MACN 5893, is referable to S. paranensis. Although Ameghino (1891a, p. 147) mentioned MACN 5893 first in his description of A. lelongi, he distinctly indicated in his catalogue that the mandibular ramus fragment, MACN 5892, which he illus- trated (1891a, fig. 52), was the "TIPO." Of all known Prothylacyninae, S. paranensis shows special affini- ty only with Lycopsis longirostrus. The upper molars of these species have large protocones, very large metacones which are con- nate basally with much reduced paracones, and distinct parastyles. Differences are that in S. paranensis the protocone is proportionate- ly and absolutely larger, the parastyle is more reduced, and the metacrista is proportionately shorter and more robust. Noteworthy similarities in their lower dentitions include: size (S. paranensis is slightly larger in some dimensions, andL. longirostrus in others), canines are moderately developed, premolars are antero- posteriorly elongated and are set in a relatively straight line in the jaw, P2 is larger than P3, in presence of a distinct posterobasal heel on P23, in M13 having a very large and deeply basined talonid, in M4 Opposite Fig. 17. Stylocynus paranensis Mercerat, 1917. Stereopairs of MLP 41-XII-13- 1112, a fragment of a right mandibular ramus with alveoli of C, PrMj complete: a, labial; b, occlusal; and c, Ungual views. Scale = 3 cm. 41 1 0 a E o u (N s XJ 3 M CO 1 XI be i / k *5 "« < CO O _0> CO 03 Q 0 S OJ C u CD XI -u CA o T3 t"» C 05 CO CO -u CO os _3 t> 53 13 0 CJ hi c QJ S X yi 1 — 1 CO »i K X CD CO C 1 — ' B 0 a ft, CO CO 3 e "S >, 03 ej _3 _c O >, e» > CO 13 X hi o in 6 CD *J fe 00 o a 42 ^ sJ 1 o a. <n 3 C ?■> O «o J3 Jfi qo co % *> '- ■gzz » < < o> X a* cu 43 44 FIELDIANA: GEOLOGY having a distinct talonid basin which is smaller than that on M13, in talonids being markedly wider than trigonids on M12, and in the hypoconid being much larger than the entoconid on M12 but more subequal on M34. Furthermore, the mandibular symphysis is ligamentous in both species and the rami are unfused in adults. The lower jaw and dentition of «S. paranensis differs from L. longi- rostrus in the ramus being proportionately deeper and the ventral border markedly convex, in the presence of a very large anterior mental foramen below the P2, in the proportionately larger size of Pj.3 relative to M14, in the proportionately and absolutely larger talonid basin on M14, in the overall greater robustness of the entire lower dentition and jaw, and, lastly and most importantly, in the presence of a small but distinct metaconid on M14. All in all, these taxa are very similar and of the many points just considered only the presence of a small metaconid in S. paranensis could possibly exclude it as a direct descendent of L. longirostrus which lacks a metaconid. The presence of a metaconid in S. paranen- sis dictates the reappearance of a feature lost (or, better, suppressed) Table 8. Measurements of mandibular ramus of Stylocynus paranensis. Depth of ramus Breadth Depth of ramus Breadth imen below labial of same below labial of same side of P3 side of M4 » 11-94 27.0 11.0 37.0 14.6 Ml-XII-13-1112 29.0 11.0 in L. longirostrus if such a relationship exists. Kurten (1963) has demonstrated the re-establishment of M2, a molar thought to be lost in all Miocene felids, in the dentitions of a part of the population of the modern northern European lynx, Felis lynx. Absence of a meta- conid in L. longirostrus would thus not necessarily bar this species from the ancestry of S. paranensis. Keeping in mind the study of Kurten, the fact that the metaconid in S. paranensis is quite small, and the close agreement of these species in almost every other com- parable character, it seems plausible to regard the metaconid as being secondarily regained. Assuming this to be correct, then an ancestral-descendant relationship for these species is a distinct possibility. MARSHALL: PROTHYLACYNINAE REVIEW 45 SUMMARY OF PHYLOGENETIC RELATIONSHIPS OF PROTHYLACYNINAE Members of the Prothylacyninae are known from beds of Colhue- huapian (late Oligocene) through Montehermosan (Pliocene) in age. In beds of pre-Huayquerian age they are the only medium- to-large terrestrial mammalian omnivore-carnivores on the South American continent. The six species of Prothylacyninae share a suite of characters not found jointly in other borhyaenids. These include: a medium-to- large size; a typically moderately well-developed canine; P13 aligned in relatively straight line with Px being set at slight angle relative to rest of tooth row, less so in later forms; P3 only moderately well developed and not proodont as is typical in other borhyaenids; M1"3 usually with well-developed protocones; talonid moderate to well- developed on M13; metaconid absent except in Stylocynus; P13 typi- cally (except in Prothylacynus) elongated anteroposteriorly and with distinct posterobasal heel which is usually best developed on P2.3; anterobasal cingulum usually present on M14, but weakly developed or absent on Mj in Prothylacynus and Pseudothylacynus. The oldest known species of Prothylacyninae is Pseudothylacynus rectus. P. rectus makes an ideal structural ancestor for Prothyla- cynus patagonicus, and Lycopsis torresi may likewise have evolved from a P. rectus-like form. Evolution of the Prothylacynus patagonicus lineage involved changes toward "brachycephaly" and carnivorous dental specializa- tions. The primary changes from P. rectus to P. patagonicus in- clude: increase in size; crowding of cheek tooth row such that PI comes to lie more obliquely in jaw and premolars and molars become relatively shorter and absolutely more robust; reduction in size of protocone in upper molars and talonid in lower molars; and man- dibular symphysis becoming shortened, ankylosed, and the rami tightly fused in the adult. P. patagonicus is the least specialized of known Prothylacyninae for an omnivorous diet, and shares many features with smaller members of the more carnivorous Borhyaeni- nae, such as Acrocyon sectorius (Marshall, 1978). In contrast, evolution of Lycopsis torresi from a P. rectus-like form involved specializatons toward "dolichocephaly" and om- nivory. The primary changes include: increase in size; elongation of the tooth row as reflected in cheek teeth becoming absolutely and relatively longer anteroposteriorly; increase in size of protocone on ~r. !/; 3 B >. « B fi 3 >> "^ a GQ 0, 03 a 03 03 9 o y o -C U a E a CO o .bf XI CO CO s § O 03 a 5 co 03 e i O 03 c s - - A 8 IN IN Oh Oh V A oT oT s- o a o c ^ S» "o C e "a a p 3 o 03 « -3 CO to 03 c a, 6 "S >> u £ >> J3 ." CO 3 C h-> co ■<~ O H Oh to «4-l ?s So O K) c 03 Oh b£ J3 CO 6«j 2 03 — i ? co a) 03 C. i3 J3 i ^ co a) 03 3. bf" 03 5 -O o CN 6 a -2 "3 "3 A A CM CM Oh Oh V V oT oT Cj 8 U 2 a 03 O O X! CO 3 O c £ CO >A xl 0) s T3 a c 03 0} JS CO CO gj 03 C 5p*3 T3 § -a CO QJ 03 C 5 42 O CN OS Oh CO Oh VI A Oh cm Oh V V oT oT g, CO X) 03 e o a a to 3 03 bo 13 >> _o u J5 O 1 J3 1 3 a 03 u J5 o> CO _o c T3 03 CO 4J c T3 C CO 03 CO T3 C CO -a 03 c 'co g a; a ft, '•V 03 £ u CO *H c 03 03 J2 "CO E CO 3 1c E CO J3 C X> CO CO co 03 u CO 3 £ E 3 CO to 3 B ai >> o O H ,3 to H J >> -B <m o 3 si T3 CQ CJ 2! E 3 '•& 03 s Oh 03 U o CO 03 03 E ^ 03 03 3 03 <C T3 3 «5 .2 9 id hj C 3 CO J2 C 03 CO S2 CO oT N _>> ►, "cO = CO c CO IS _2 03 l^ CO T) 03 s- 03 -*-) 03 c CO ■3 o E 'co CO a CO 3 CJ 0) 3 So 50 CO 6 -C5 03 ft, Oh A 11 a. CM a. V V Oh Oh CO Oh V V C«1 Oh oT V V =7 oT CO 03 'Q a CO s CO J2 >> Ih J3 0) "3 E >, CO — u CB s 03 N CO 3 CO 5 CO s JZ J CN CO S fi 2 £ * S o P c -a .3 < SOh^S „ o Oh 3 o .S 03 -S bn CO r-t W Oh > 46 MARSHALL: PROTHYLACYNINAE REVIEW 47 mybp Epoch S.A. Land Mammal Age Phylogeny 3 - 4 - 5 - AW Oz _|LU CLO MONTEHERMOSAN Stylocynus paranensis 6 - 7 - 8 - 9 - HUAYQUERIAN 10 - 11 - 12 - 13 - ? LU Z LU O o CHASICOAN Pseudolvcopsis cabrerai (HIATUS) 14 - 15 - FRIASIAN Lycopsis longirostrus 16 - 2 I 17 - I 18 - t 19 - 20- SANTACRUCIAN » Lycopsis Prothylacynus torresi patagonicus 21 - \ / 22 - 23 -J \ / 24 - 25 - 26 - 27 - 28 - 111 Z LU o o o _l o COLHUEHUAPIAN \ Pseudothylacynus \ rectus \ \ \ \ (HIATUS) Fig. 19. Proposed phylogeny of the Prothylacyninae. upper molars and talonid becoming larger and basined on lower molars; and in the mandibular symphysis becoming completely liga- mentous and more elongated, and the rami unfused in the adult. Changes in the L. torresi-L. longirostrus lineage involved continu- ation of the trend toward "dolichocephaly" and omnivory— i.e., fur- ther increase in size; greater elongation of cheek tooth row; cheek teeth become proportionately more elongate anteroposteriorly and small diastems develop between premolars and adjacent teeth; in- crease in size of protocone and talonid; and ligamentous mandibular symphysis becomes more elongate. 48 FIELDIANA: GEOLOGY These trends were continued in the evolution of Stylocynus paranensis, and the trend toward size increase of the protocone and talonid reached its greatest development in this species. S. paranen- sis is large and was apparently bear-like in its feeding habits. This species presents some phylogenetic problems, since in size and structure it is remarkably similar to L. longirostrus except that it has a small metaconid on M24, a feature lacking in L. longirostrus. Presence of a metaconid in S. paranensis may represent the reappearance of a "lost" character, in which case that species could have evolved from a form like L. longirostrus. Alternatively, <S. paranensis may represent the last member of an evolutionary line distinct from Lycopsis, the mid-Tertiary record for which is either not known or not recognized. There is no known species of Santacrucian borhyaenid which can confidently be regarded as ancestral to Pseudolycopsis cabrerai, although L. torresi approaches closest in molar morphology to what would be expected in an ancestral form. Prothylacynus patagonicus is clearly too specialized to be ancestral to P. cabrerai. Pseudothyla- cynus rectus from the Colhuehuapian may be ancestral to P. cabrerai, although the latter is known only from upper dentitions and the former only from lowers. Consequently, direct comparison of these taxa is not possible at this time. Diagnostic characters for species of Prothylacyninae are com- pared in Table 9. The probable phylogenetic relationships of these taxa are shown diagramatically in Figure 19. ACKNOWLEDGEMENTS This research was supported, in part, by two grants (1329, 1698) from the National Geographic Society, Washington, D.C. A large part of this study was carried out in the Vertebrate Paleontology Section of the MACN, and I thank Guillermo del Corro, Chief of Sec- tion, and Jose Gallardo, Director, for making collections and work- ing space available to me at that institution. Rosendo Pascual, MLP, allowed study of collections under his care and facilitated research in many ways. For loan of specimens and/or access to rele- vant museum records I extend my sincere thanks to Robert Hoff- stetter, Institut de Paleontologie, MNHN; Roger Hamilton, BM(NH); M. C. McKenna and R. H. Tedford, AMNH; W. A. Clemens, UCMP; and Donald Baird, PU. MARSHALL: PROTHYLACYNINAE REVIEW 49 REFERENCES Ameghino, F. 1891a. Caracteres diagnosticos de cincuenta especies nuevas de mamiferos fbsiles argentinos. Rev. Arg. Hist. Nat., 1, pp. 129-167. 1891b. Nuevos restos de mamiferos fdsiles descubiertos por Carlos Ameghino en el eoceno inferior de la Patagonia austral. Especies nuevas, adiciones y correc- ciones. Rev. Arg. Hist. Nat., 1, pp. 289-328. 1894. Enumeration synoptique des especes de mammiferes fossiles des formations eocenes de Patagonie. Bol. Acad. Cienc, Cbrdoba, 13, pp. 259-452. 1898. Sinopsis geologico-paleontologica. Segundo censo de la Repiiblica Argen- tina. Fol., Buenos Aires, 1, pp. 112-255. 1902. Premiere contribution a la connaissance de la faune mammalogique des couches a Colpodon. Bol. Acad. Cienc, Cordoba, 17, pp. 71-141. 1904. Recherches de morphologie phylogenetique sur les molaires superieures des ongules. An. Mus. Nac. Buenos Aires, 10, ser. 3, vol. 3, pp. 1-541. 1906. Les formations sedimentaires du cretace superieur et du tertiaire de Pata- gonie avec un parallele entre leurs faunes mammalogiques et celles de l'ancien continent. An. Mus. Nac. Buenos Aires, 8, ser. 3, pp. 1-568. Cabrera, A. 1927. Datos para el conocimiento de los dasyuroideos fosiles argentinos. Rev. Mus. La Plata, 30, pp. 271-315. KURTEN, B. 1963. Return of a lost structure in the evolution of the felid dentition. Soc. Sci. Fennica, Commentations Biol., 26, no. 4, pp. 1-10. Kraglievich, L. 1917. Notas paleontol6gicas : Examen critico de un trabajo del Senor Alcides Mercerat. Anal. Soc. Cient. Argent., 83, pp. 262-279. 1934. La antigiiedad pliocena de las faunas de Monte Hermoso y Chapadmalal, deducidas de su comparacion con las que le precedieron y sucedieron. Monte- video : El Siglo Illustrado, 938, pp. 1-136. Marshall, L. G. 1975. The Handel T. Martin paleontological expedition to Patagonia in 1903. Ameghiniana, 12, no. 1, pp. 109-111. 1976a. Evolution of the Thylacosmilidae, extinct saber-tooth marsupials of South America. PaleoBios, 23, pp. 1-30. 1976b. Fossil localities for Santacrucian (Early Miocene) mammals, Santa Cruz Province, southern Patagonia, Argentina. J. Paleontol., 50, no. 6, pp. 1,129-1,142. 1976c. New didelphine marsupials from the La Venta Fauna (Miocene) of Colom- bia, South America. J. Paleontol., 50, no. 3, pp. 402-418. 1976d. A new borhyaenid (Marsupialia, Borhyaeninae) from the Arroyo Chasic6 formation (Lower Pliocene), Buenos Aires Province, Argentina. Ameghiniana, 13, no. 3-4, pp. 289-299. 50 FIELDIANA: GEOLOGY 1977. A new species of Lycopsis (Borhyaenidae : Marsupialia) from the La Venta Fauna (Miocene) of Colombia, South America. J. Paleontol., 51, no. 3, pp. 633-642. 1978. Evolution of the Borhyaenidae, extinct South American predaceous mar- supials. Univ. Calif. Publ. Geol. Sci., 117, pp. 1-89. Marshall, L. G., R. F. Butler, R. E. Drake, G. H. Curtis, and R. H. Tedford 1979. Calibration of the Great American Interchange. Science, 204, pp. 272-279. Marshall, L. G., R. Pascual, G. H. Curtis, and R. E. Drake 1977. South American geochronology : radiometric time scale for middle to late Tertiary mammal-bearing horizons in Patagonia. Science, 195, pp. 1,325-1,328. Mercerat, A. 1891. Caracteres diagnosticos de algunas especies de Creodonta conservadas en el Museo de La Plata. Rev. Mus. La Plata, 2, pp. 51-56. 1917. Notas sobre algunos carnivoros fosiles y actuales de la America del sud. Buenos Aires, pp. 1-21. PlVETEAU, J. 1961. Marsupialia, pp. 585-637. In Traite de Paleontology, 6, no. 1. Roger, O. 1896. Verzeichnes der bisher bekannten fossilen Saugethiere (I. Theil). Ber. Naturw. Ver. Schwaben, Neuberg, 32, pp. 1-272. Simpson, G. G. 1930. Post-Mesozoic Marsupialia, pp. 1-87. In Fossilium Catalogus. I: Animalia. Berlin, W. Junk, 47. Sinclair, W. J. 1905. The marsupial fauna of the Santa Cruz beds. Phil. Soc, 44, pp. 73-81. 1906. Mammalia of the Santa Cruz beds : Marsupialia. Rept. Princeton Univ. Exped. Patagonia, 4, no. 3, pp. 333-460. Trouessart, E. L. 1898. Catalogue mammalium tarn viventium quam fossilium. new ed., Berlin, 5, pp. 665-1,264. Winge, H. 1923. Pattedyr-Slaegter. I. Monotremata, Marsupialia, Insectivora, Chiroptera, Edentata. Copenhagen, 360 pp., 11 pis. (English translation: "The Interrelation- ships of the Mammalian Genera." Copenhagen, 3 vols., 1941-42. Transl. by E. Deichmann and G. M. Allen).