f * 9 1 c I I - 4 L' ' / v,U Lilftv J. PIERPONT MORGAN PUBLICATION FUND REPORTS OF THE Princeton University Expeditions TO PATAGONIA. 1896-1899 J. B. HATCHER IN CHARGE EDITED BY WILLIAM B. SCOTT BLAIR PROFESSOR OF GEOLOGY AND PALEONTOLOGY, PRINCETON UNIVERSITY VOLUME VI, PALAEONTOLOGY III PRINCETON, N. J. The University STUTTGART Schweizerbart’sche Verlagshandlung (E. Nagele) 1909-28 J. PIERPONT MORGAN PUBLICATION FUND Reports of the Princeton University Expeditions to Patagonia i 896-99 VOLUME VI. MAMMALIA OF THE SANTA CRUZ BEDS I. Typotheria BY WILLIAM J. SINCLAIR PRINCETON UNIVERSITY II. Toxodonta III. Entelonychia IV. Astrapotheria V. Primates BY WILLIAM B. SCOTT PRINCETON UNIVERSITY PRINCETON, N. J. The University STUTTGART Schweizerbart’sche Verlagshandlung (E. Nagele) 1909-28 TABLE OF CONTENTS VOL. VI. Page MAMMALIA OF THE SANTA CRUZ BEDS i PART I. TYPOTHERIA, by William J. Sinclair i INTRODUCTION i CLASSIFICATION OF THE SANTA CRUZ TYPOTHERIA . . 2 Order TOXODONTIA 2 Suborder TYPOTHERIA 2 Relationships of the Santa Cruz Typotheria . . 7 I. With the Toxodonta ..... 7 II. With Typotherium ..... 9 III. With the Rodentia ..... 10 IV. With the Hyracoidea . . . . . 1 1 Conclusions 13 Synonymy 13 Interatheriim: 13 Protypotherium Amegh. ...... 13 Inter atherium (Moreno) Amegh. .... 47 HEGETOTHERIIDyE 67 Hegetotherium Amegh. ...... 67 Pachyrukhos Amegh. . . . . . . 85 TYPOTHERIA INCERTAE SEDIS 98 Protypotherium ....... 99 Inter atherium . . . . . . . . 100 Pachyrukhos . . . . . . . . 107 Hegetotherium Amegh. . . . . . . 109 Bibliography ............ 109 PART II. TOXODONTA, by William B. Scott ...... 111 Classification of the Toxodonta . . . . . . . . 116 Nesodon Owen . . . . . . . . 118 Adinotherium Amegh. ...... 200 Phobereotherium Amegh. ... ... 234 Stenotephanos Amegh. ...... 235 GENERA INCERTAE SEDIS 236 Rhadinotherium Amegh. ...... 236 Palceolithops Amegh. ...... 236 Acrotherium Amegh. ...... 237 Xotoprodon Amegh. ....... 237 v VI CONTENTS. Notohippim; ............ 237 Notohippus Amegh. ....... 237 PART III. ENTELONYCHIA 239 Homalodontotherium Flower ..... 246 Diorotherium Amegh. ...... 286 Order TOXODONTIA 287 BIBLIOGRAPHY 299 PART IV. ASTRAPOTHERIA 301 Astrapotheriiml . . . . . . . . . .301 Astrapotherium Burmeister . . . . . 301 PARTV. PRIMATES 342 Suborder PLATYRRHYNI 343 Cebid.® 343 Homunculus Amegh. ...... 343 Bibliography . 350 INDEX 353 DATES OF PUBLICATION OF THE PARTS OF VOLUME VI. The dates of issue, as printed on the insides of the covers of the various parts, are approximations made in advance, and in each case antedate by a few days the time of actual issue to the sub- scribers. These dates should, therefore, be corrected as follows: Pp. 1-110, Pll. I-XI, published October 19, 1909. Pp. 1 1 1-300, Pll. XII-XXX, published December 16, 1912. Pp. 301-352, Pll. XXXI-XXXVII, published December 15, 1928. LIST OF TEXT ILLUSTRATIONS Page Fig. I. Skulls of Santa Cruz Typotheria, Xf. A. Protypotherium australe, A.M.N.H. B. Inter atherium robustum, A.M.N.H. C. Pachyrukhos moyani. D. Hege- totherium mirabile. .......... 4 Fig. 2. Left hind foot of Protypotherium australe , Xf. A.M.N.H. .... 6 Fig. 3. A. Protypotherium australe , left fore foot, Xf. A.M.N.H. B. Nesodon irnbri- catus, left fore foot, about Xf ....... 7 Fig. 4. A. Hegetotherium mirabile, right hind foot, Xf. B. Nesodon imbricatus, right hind foot, about Xf ......... 8 Fig. 5. A. Protypotherium sp., unworn M-, xfr. A.M.N.H. B. Nesodon imbricatus, M-&-, slightly worn, Xf 8 Fig. 6. A. Protypotherium sp., unworn M3, Xtu A.M.N.H. B. Nesodon imbricatus, two lower molars, Xf. . . . . . . . . . . 9 Fig. 7. Dendrohyrax arboreus, left pes, dorsum, Xf ...... 12 Fig. 8. Protypotherium australe, cranial elements, Xy; . . . . . . 21 Fig. 9. Protypotherium sp., A. Posterior cervicals and anterior dorsals with ribs in place, from left side, Xf. B. the same from below, showing the sternum and calcified costal cartilages, Xf ........ 25 Fig. 10. A. Right manus with opposable pollex and separate centrale, Xy. B. Right pes with opposable hallux, X'i (after Ameghino). Both have been referred er- roneously to Interatherium robustum. ....... 56 Fig. 11. Interatherium robustum, showing curved fronto-nasal suture, Xy 59 Fig. 12. Interatherium excavatum, showing lyrate crest, Xf ... . . . 66 Fig. 13. Hegetotherium mirabile, skull from in front, Xf ...... 67 Fig. 14. Hegetotherium mirabile, occiput showing sutures between cranial elements, Xt- A.M.N.H 71 Fig. 15. Hegetotherium mirabile, showing variation in fronto-nasal suture, Xf . 72 Fig. 16. Hegetotherium mirabile, cranial elements, xf:, A.M.N.H. .... 74 Fig. 17. Nesodon imbricatus, M - much worn, external view, Xy .... 122 Fig. 18. The same, crown view, Xy ......... 122 Fig. 19. Nesodon conspurcatus? Skull, left side, Xf. A.M.N.H. .... 142 Fig. 20. Toxodon burmeisteri: Skull, left side, Xf. La Plata Museum . . . 143 Fig. 21. Toxodon burmeisteri: Left scapula, outer side, Xf. La Plata Museum . 151 Fig. 22. Nesodon imbricatus: Left scapula, outer side, Xf. ..... 151 Fig. 23. Toxodon burmeisteri: Left humerus, dorsum, Xf. La Plata Museum. . . 153 Fig. 24. Nesodon imbricatus: Left humerus, dorsum, Xf . . . . . 153 Fig. 25. Toxodon burmeisteri: Left fore-arm bones, dorsum, Xf. La Plata Museum. 155 vii viii LIST OF TEXT ILLUSTRATIONS. Fig. 26. Nesodon imbricatus: Left fore-arm bones, outer side, X j . . . • 155 Fig. 27. Toxodon burmeisteri: Left manus, dorsum, Xf. La Plata Museum . . 157 Fig. 28. Nesodon imbricatus: Left manus, dorsum, X| ...... 157 Fig. 29. Toxodon burmeisteri: A. left femur, dorsum. B. the same, external view, Also left patella, dorsum. All Xi- La Plata Museum. .... 166 Fig. 30. Nesodon imbricatus: A. left femur, dorsum; B. the same, external view. Also left patella, dorsum. All Xi- ........ 167 Fig. 31. Toxodon burmeisteri: Left tibia and fibula, dorsum, Xj. La Plata Museum . 169 Fig. 32. Nesodon imbricatus: Left tibia and fibula, dorsum, X|. . . . . 169 Fig. 33. A. Toxodon burmeisteri: Left pes, dorsum, Xj. B. the same, external view. La Plata Museum. .......... 176 Fig. 34. Nesodon imbricatus: Left pes, dorsum, X| . . . . . . 176 Fig. 35. Nesodon cornutus, type: Skull, right side, Xi ...... 196 Fig. 36. Nesodon cornuhis, type; Occiput, Xi. . . . . . . . 197 Fig. 37. Nesodon imbricatus: Occiput, Xi . . . . . . . 198 Fig. 38. Nesodon cornutus, type: Forehead, showing horn and fronto-nasal suture, Xf . 199 Fig- 39- Adinotherium ovinum: Left premaxilla with permanent incisors and prelacteal i-, Xi 202 Fig. 40. Adinotherium ovinum: Series showing the variations of the fronto-nasal suture and the frontal horn, Xi . . . . . . . . 204 Fig. 41. Trigodon gaudryi: Skull, left side, about xi-. Ameghino collection, Monte Hermoso Beds ........... 221 Fig. 42. Stenotephanos speciosus: Fragment of right upper maxillary with molars in place. About Xt- Type specimen, La Plata Museum ..... 236 Fig. 43. Homalodontotherium segovice: Second and third upper molars of left side, Xt- Oblique view of grinding surface ........ 253 Fig. 44. Homalodontotherium segovice: Left pes, dorsum, Xi- After Ameghino, except the calcaneum, astragalus and cuboid. ....... 280 Fig. 45. Sus scrofa, Occiput, Xi . . . . . . . . . . 291 Fig. 46. Procavia capensis. Occiput, xf ........ 291 J. PIERPONT MORGAN PUBLICATION FUND Reports of The Princeton University Expeditions to Patagonia, 1896-1899 J. B. HATCHER, in Charge EDITED BY WILLIAM B. SCOTT BLAIR PROFESSOR OF GEOLOGY AND PALAEONTOLOGY, PRINCETON UNIVERSITY VOLUME VI — PALAEONTOLOGY « Part I. Typotheria of the Santa Cruz Beds BY WILLIAM J. SINCLAIR PRINCETON UNIVERSITY (Pp. I-IIO Pis. I-XI) PRINCETON, N. J. The University STUTTGART E. Schweizerbart’sche Verlagshandlung (E. Nagele) I9°9 % 0 8t>1 anterior crescent; pc, posterior crescent; pp. posterior pillar. of the Santa Cruz Typotheria is there any trace of a double deciduous dentition as in Nesodon. From the resemblances in dentition and foot-structure already pointed out it may be inferred that the Toxodonta and Typotheria had a common origin, but the facts at present available do not justify a more extended discussion of relationships. II. With Typotherium. Difficult as it is to ascertain the relationship existing between the Santa Cruz Typotheria and the Toxodonta, it is even more so to determine their degree of kinship with Typotherium , the ancestor of which would appear to be Eutrachytherus , as suggested by Ameghino. From their small size it is quite probable that none are in the direct line of descent culminating in Typotherium , and this seems to be borne out by a consideration of the degree of specialization in dentition and foot-structure which the latter displays. Typotherium is more specialized in dental structure, showing greater complexity of folding in the molar crowns than is attained by any of its Santa Cruz predecessors, but less specialized in foot-structure, pos- sessing a pollex in the manus which has been lost in the most generalized of the Santa Cruz Typotheria ( Protypotherium ) and with digit V of the pes less reduced than in the most specialized of the latter ( Pachyrukhos ). A pollex is figured by Ameghino (1889, pi. 13, fig. 14) in the manus of Pachyrukhos ty pious, but none has been found in any Santa Cruz speci- IO PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. men. The manus in Hegetotherium is unknown and the above state- ment regarding the degree of specialization in foot-structure displayed by Typothevium may, accordingly, require modification in the light of fuller knowledge. III. With the Rodentia. In many features of skull and skeleton the Typotheria resemble the rodents. This is most apparent in the specialized forms like Pachyvukhos. In none of the Typotheria are the following characters peculiar to rodents developed : 1. Persistently growing, chisel-shaped incisors (If of the permanent series, Weber, 1904, p. 480). If of the permanent series is enlarged in some of the Typotheria and may grow persistently, but is modified for cropping and not for gnawing. 2. More or less antero-posterior elongation of the mandibular condyle and corresponding modification of the glenoid fossa to permit backward and forward movement of the mandible. In the Typotheria the condyle is transversely extended, approximately circular in outline, with the glenoid surface flattened, and the movement of the mandible is from side to side. 3. Frequent outward curvature of the crowns of the upper molars and inward curvature of those of the inferior series in hypsodont forms. The reverse is true in the Typotheria. 4. Contact of ascending process of premaxillary with frontal. This process is short and robust in the Typotheria and is widely separated from the frontal by the maxillary. 5. Elongation of the mandibular angle. The angle is evenly convex in the Typotheria. 6. The rodent astragalus is characterized by a broad, short, rather shallow trochlea, with the crests sharp and equally developed, distinct neck and flattened head, convex distally ; trochlea symmetrical to the vertical plane ; fibular and internal malleolar facets vertical; body limited posteriorly; no astragalar foramen. In the Santa Cruz Typotheria, the astragalar trochlea is deeper than in rodents, the crests may or may not be equally developed and the head is globular without antero-posterior flattening. The symmetry of the trochlea with respect to the vertical plane varies in the different families (pp. 2-3). In the other characters they resemble the rodents. 7. The presence of a free centrale in the carpus in all rodents except the Sinclair: typotheria of the santa cruz beds. ii Hystricidse and Ccelogenys , and the general fusion of scaphoid and lunar in all except the Bathyergidae, Ctenodactilidse and the Lagomorpha (Weber, 1904, p. 476). The centrale is always wanting in the carpus of the Typo- theria and the fusion of the scaphoid and lunar does not occur. 8. Presence of a tibial sesamoid in all simplicidentate rodents. This is not present in the tarsus of the Typotheria. The Typotheria resemble rodents in the elongation of the anterior portion of the skull, with reduction in the incisor-canine-premolar series, in the enlargement and often permanent growth of the median incisors (not homologous with enlarged incisors in rodents, see under 1, p. 10), in the development of a mastoid dilatation, which maybe filled with cancellae (Interatheriidae), as in many rodents, and connected with the auditory bulla, in the shape of the proximal articular surfaces between radius and ulna (see Pis. II, fig. 6; IV, fig. 16), in the broad, anteriorly directed transverse processes of the lumbar vertebrae and in several other characters of minor importance. In view of the striking differences in structure indicated in the preceding paragraphs, it seems impossible to interpret these resemblances otherwise than as due to convergence. IV. With the Hyracoidea. Various writers have suggested a more or less intimate relationship between the Typotheria and the Hyracoidea, which the complete material now available has failed to substantiate. In the Hyracoidea the carpus is arranged on the linear plan with separate os centrale, while in the tarsus the astragalus differs from that of any other mammal in possessing a large step-like articulation for the internal tibial malleolus (text fig. 7). On the contrary, in the Typotheria the carpus is strongly interlocking, without centrale, and the internal tibial malleolus is applied to the lateral surface of the trochlea, with no trace of the supporting shelf characteristic of the Hyracoidea. The flat head of the astragalus and the articulation of the fibula with the latter element instead of with the calcaneum also serve to separate the Hyracoidea from the Typotheria. These differences in foot- structure are more than sufficient to offset similarities in the skull, which are confined to a few points, such as the cancellous dilatation of the mastoid, the shape of the posterior border of the palate and the increase in depth posteriorly of the mandible. In all living species of Hyracoidea the malar 12 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. takes part in forming the outer portion of the glenoid cavity and the parietal enters into the postorbital process. Neither of these characters is exhibited by the Typotheria. Again, in the Hyracoidea, the first upper incisor of the permanent series is a persistently growing, downwardly curved tusk, triangular in section, while in the Typotheria this tooth, although growing persistently in some forms, is always antero-pos- teriorly compressed, transversely expanded and functional as a cropping tooth. The molars in Procavia are lopho-selenodont and either brachyodont or short hypsodont, while in the Typo- theria they are extremely hypsodont, developing roots only in the deciduous series. In crown-pattern, they bear less re- semblance to the teeth of the Hyracoidea than do the molars of the early horses and rhinoceroses, which differ from Pro- cavia as fundamentally in foot-structure as do the Typotheria. The so-called Hyracoidea from the Fayum of Egypt ( SctgJia - therium, Megalohyrax ) are as yet known only from fragments of the skull and dentition. It would naturally be supposed that they should bear a closer resemblance to the Miocene Typotheria than to the recent hyraces, if the two groups are related. So far as the available material permits com- parison, this is not found to be the case, the Egyptian forms showing no closer approximation to the Typotheria than do the modern hyraces. Mr. Walter Granger has called the writer’s attention to the apparently constant presence in the Hyracoidea, as well as in these so-called hyracoids from the Fayum, of a superior branch of the alveolar canal, which perforates the base of the coronoid process of the mandible behind the last molar, as in Lepus and the Santa Cruz dipro- todont marsupial Abderites . Although thus shown not to be strictly confined to the Hyracoidea, this perforation is conspicuously absent in the Typotheria and may be interpreted as a further indication of their lack of Hyracoidean affinities. Various pre-Santa Cruz genera (. Archceohyrax , Argyrohyrax ) have been referred to the Hyracoidea. The foot-structure of these is unknown, but the skull and dentition, in the writer’s opinion, are not hyracoidean in character. Too little is known of these forms to warrant a discussion of their relationship to the Santa Cruz Typotheria, but from the available Fig. 7. m Dcndrohy- rax arbor eus, left pes, dor- sum, x | (No. 365, Prince- ton University osteological collection). SINCLAIR : TYPOTHERIA OF THE SANTA CRUZ BEDS. 13 descriptions and figures and from such photographs of the type speci- mens as the writer has examined, they seem to be referable to the same suborder as the forms described in this memoir. CONCLUSIONS. The present investigation demonstrates, in the writer’s opinion, that the Typotheria are to be regarded as a suborder of the Toxodontia, differing from the suborder Toxodonta, to which Nesodon, Toxodon and their allies are referable, in the more primitive structure of the teeth and feet. The origin of the group is entirely unknown, but was probably in South America. No complete phyletic series can be traced from its first appear- ance in the Notostylops beds to its disappearance in the Pampean. The Santa Cruz Typotheria are not ancestral to Typotkerium. No rodent or hyracoid affinities can be claimed for them. SYNONYMY. In working out the synonymy adopted in the systematic part of this memoir the writer has depended largely on a series of photographs of the type specimens taken by Professor W. B. Scott and on the measurements given in Dr. Ameghino’s published descriptions. The specimens in the collection of the Akademie der Wissenschaften at Munich are accom- panied by Dr. Ameghino’s manuscript labels and are practically cotypes. A large number of species which could not be identified with any of the material in the collections studied have been listed, at the end of the present memoir, as Typotheria Incertae Sedis, but, as there explained, it does not follow that these are all invalid. INTERA THERIID Ad. PROTYPOTHERIUM Ameghino. (Plates III; IV; V; VI, Figs. 1-10, 14, 15 ; VII; Text Figs. 1 A, 2, 3 A, 4 A, 5 A, 6A, 8, 9.) Protypotherium Amegh.; Catalogo de la provincia de Buenos Aires en la Exposicion Continentale Sud-amer., March, 1882 ( nomen nudum') ; Observaciones generales sobre el orden de mamiferos estinguidos sud-americanos llamados Toxodontes (Toxodontia), etc., p. 52, 1887 1 Enum. Sistematica, etc., p. 15, 1887. i4 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. Toxodontophanus Moreno ; Patagonia resto de un antiguo continente hoy submerjido, p. 23, 1882 {nomen nudum)', Ameghino, Observaciones generales, etc., p. 64, 1887. Patriarchus Amegh.; Contrib. al conoc. de los Mam. Fos. de la Repub. Argentina, pp. 480-481, 1889. Of all the Santa Cruz Typotheria, Protypotherium is the least specialized in dentition and foot-structure, probably approaching more closely than any of the other genera the ancestral form from which the group as a whole originated. Dentition (Pis. Ill, figs. 1, 3, 5, 6 ; IV, fig. 15; V, figs. 11-i^a, 21, 22). — The dentition is complete and in close series in both jaws. The median upper incisors have broad, antero-posteriorly compressed, strongly curved crowns, covered externally with enamel, producing chisel-like cut- ting edges when worn. These teeth do not grow persistently, but have tapering roots. Externally, the crown is convex transversely, with a faint suggestion of a broad median groove on the anterior face. Internally, it is rendered concave by another broad groove. The second and third incisors are of about the same size. The unworn crowns are pointed, the point lying at the intersection of a prominent enamel ridge on the anterior face with the cutting edge of the tooth. The inner surface of the crown is flat in unworn teeth, but in worn specimens develops a median groove. The canine is incisiform and indistinguishable from the teeth in front. The incisors and the canine are inserted obliquely and im- bricate, the posterior border of each overlapping externally the tooth next succeeding. The canine is lodged entirely in the maxillary. The premolars increase in size posteriorly, but none become completely molariform and all grow from persistent pulps. The first is approxi- mately cylindrical in section, with a prominent ridge on the external face of the crown, which decreases in elevation toward the alveolar border and finally disappears, so that in the worn specimens the first pre- molar is a simple curved 'cylinder. A slight anterior ridge is observ- able in little-worn teeth corresponding to the first of the two external ridges in the remaining premolars. Internally the crown, if well worn, is convex antero-posteriorly, but in less worn teeth a shallow ante- rior groove is discernible. The second, third and fourth premolars may be described together. Each has a deep antero-external groove bounded by two ridges. In all the species represented in the collections Sinclair: typotheria of the santa cruz beds. 15 studied these ridges persist, irrespective of the stage of wear attained. Internally, a groove, less deep than in the molars, divides the lingual aspect of the tooth-crown into a pair of crescentic lobes, of which the anterior is the smaller. The imbrication of the premolars and the succeeding molars is opposed to that of the incisor-canine series. The crown-pattern produced by wear is a shallow basin overshadowed externally by a high cusp marking the termination of the second of the external ridges, preceded by a smaller cusp at the termination of the anterior ridge, the deep notch between being produced by the external groove mentioned above. There are no specimens in the collection with unworn premolars, but in a young individual retaining the milk dentition (No. 9482 American Museum, PI. V, figs. 11-14*7) the first premolar, which is rootless and probably to be interpreted as belonging to the permanent series, is but little worn, showing two internal crescents, of which the anterior is the smaller and less perfect. With the exception of the deep groove inclosed between the anterior horn of the smaller inner crescent and the anterior margin of the ectoloph, and the prominent ridges formed by the same parts, the outer surface of the tooth is broadly concave. The molars decrease in size posteriorly, but are otherwise so much alike that all may be described together. Antero-externally, the ectoloph is rendered slightly undulatory by two ridges corresponding in position to the external ridges already described in the premolars, but less distinct and almost disappearing in worn teeth. Internally, the crown is divided by a reentering fold into a pair of approximately equal lobes. On the margin of the ectoloph a series of cuspules is developed as a result of contact with the teeth of the lower jaw, each ridge on the ectoloph terminating in a cusp. A third cusp marks the junction of the posterior plane surface of the ectoloph with the triturating surface and in the third molar the elongated postero-external corner terminates in a fourth cusp. In No. 9482, American Museum, the third molar had not yet become functional. That of the left side (PI. V, fig. 11 b\ text fig. 5, A ) shows a broadly convex outer wall and two internal crescents joining the ectoloph anteriorly and overlapping posteriorly, giving rise to the deep internal fold. The central portion of the anterior crescent is partly united with the outer wall by a crista-like fold (text fig. 5, A, c). The margin of the ectoloph is serrate, with prominent anterior and less dis- tinct posterior cuspules. The former persists as the cusp at the termina- i6 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. tion of the second external ridge ; the latter is intensified by wear and becomes the third external serration mentioned above. The prominent antero-external serration exhibited in worn teeth is developed from the elongated antero-external angle and is not indicated in unworn molars. A thin layer of cement is quite generally present on the premolars and molars. The crowns of these teeth are characterized by a strong inward curvature, as in all the Santa Cruz Typotheria. The median lower incisors (Pis. IV, fig. 15 ; V, figs. 13, 14, 22) are inclined forward obliquely, continuing the slope of the inner surface of the symphysis. The crowns of the first and second are cylindrical and symmet- rically divided longitudinally by a deep cleft, producing a structure re- sembling the tines of a fork. A similar subdivision into three tines is observable in the lower incisors of Procavia. As the tooth wears, the cleft disappears completely. Apparently the incisors continue to elongate until the adult condition is reached, or even later, as the length of these teeth below the alveolar border is great, but the tooth does not grow persistently, becoming greatly constricted transversely toward the extremity of the root. The third incisor and the canine have broad crowns convex externally, but excavated internally by a broad groove, producing a serrate effect on the cutting edge. The bifurcation of the inferior incisors has been used as a character to separate the genus Patriarchus from Protypotherium, but as the bifurcation is present in all unworn teeth and absent in well worn specimens, with all transitional stages between the two extremes, and as there are no other characters of generic importance to separate individuals with this peculiarity well developed from those without it, Patriarchus can- not be regarded as a valid genus and is here made synonymous with Pro- typotherium. The third incisor overlaps internally on the second, the canine on the third incisor and the first premolar on the canine. The two teeth last mentioned are indistinguishable from each other structurally and in many specimens are of approximately the same size. The second premolar and the teeth succeeding are also inclined to the long axis of the tooth-row, but in a direction opposite to the incisors, canine and first premolar, the anterior margin of each overlapping externally the tooth preceding, instead of overlapping internally, as in the case of the teeth first mentioned. The second, third and fourth premolars increase regularly in size posteriorly, but none is completely molariform. Externally, they are rendered bilobate by a deep groove. Of the two lobes thus produced SINCLAIR I TYPOTHERIA OF THE SANTA CRUZ BEDS. 17 the posterior is the smaller. Internally, there is a deep groove opposite the external groove just mentioned and another broader but shorter furrow anterior to it, which disappears with wear, while the first persists, even in well-worn teeth, and is overshadowed anteriorly by a prominent ridge. The crown-pattern of the unworn premolars is well shown in No. 9559, American Museum, (PI. V, figs. 15, 150) the right ramus of a mandible, in which the second premolar is just erupting. On the external side, the otherwise convex crown is broadly grooved posteriorly, while internally it is rendered concave by the two grooves, already mentioned, of which the anterior in the unworn tooth is the broader, deeper, and more conspicuous. Consequent upon these grooves three serrations are developed on the sharp cutting edge of the crown, of which the second is the most promi- nent. Like the premolars, the molars are also bilobate, but the posterior lobe is larger than the anterior and in the third molar greatly exceeds the latter in size, with a faint trace of a broad external groove and a deeper, more persistent groove internally. The worn surfaces of the anterior and posterior lobes in the first and second molars are triangular in outline with elevated corners, inclosing a basin-shaped depression. This applies also to the anterior half of the third molar, but the posterior half is irregu- larly elliptical in outline and the depression on the triturating surface has its margins interrupted by the shallow grooves described above. Internally, the molars exhibit the same persistent groove, with anterior bounding ridge, as in the premolars. The ectoloph on either side of this groove is slightly concave. Both molars and premolars are hypsodont and cement- covered. In the immature individual figured on Plate V, figures 13, 14, the last molar is just appearing above the alveolar border and the tooth preceding it is but little worn. Each molar is composed of a pair of cres- centic lobes concave internally. The anterior crescent is breached on the inner side by a broad groove corresponding to the shallow anterior groove already mentioned in describing the unworn premolars. The cutting edge of this crescent is sharp, forming a high external cusp. The anterior horn is recurved, terminating in a small blunt tubercle. The posterior horn is produced beyond the anterior extremity of the posterior crescent, giving rise to the persistent internal ridge and groove observable even in worn molars. The posterior lobe incloses a depression bounded by sharp edges, formed by the postero-external crescent and a long straight ridge (text fig. 6, A, posterior pillar) trending posteriorly from the anterior horn of the i8 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. former. The depression inclosed by these elements rapidly disappears, leaving no trace of enamel lakes. The posterior lobe of the third molar differs from that of the second only in its greater length antero-posteriorly. Milk Dentition (PI. V, figs, n-14.) — The milk-premolars may be readily recognized by the presence of roots. So far as can be ascertained, the order of replacement seems to be the normal one. In the dental series figured on Plate V, figures 1 1 and 12, the first and second permanent molars are in place and already somewhat worn, the first more so than the second. The third is just appearing below the alveolar border. An- terior to these are three double-rooted deciduous teeth. Of the two roots the anterior is the larger. DpA is almost completely molariform, showing the same arrangement of external and internal grooves as in the permanent molars. Dp- and dp- are deeply grooved externally, as in their permanent successors, but the groove is continued but a short distance above the alveolar border up the outer side of the anterior root. In- ternally, the crowns are slightly grooved, producing the same separation into anterior and posterior lobes as in the permanent premolars. Px seems to have had no deciduous predecessor. The crown-pattern of the decidu- ous premolars (PI. V, fig. 1 1 a) is the same as that of their permanent suc- cessors, so far as can be determined from the partly worn teeth available for comparison. As already explained, this consists of a pair of internal crescents and a broadly concave ectoloph. The depressions inclosed be- tween the crescents and the ectoloph are shallow and soon disappear. Only in little worn teeth are isolated lakelets, due to the partial oblitera- tion of these depressions, observable on the triturating surface of the crown (PI. V, fig. 11). Sufficient material is not available for determining the order of replacement of the incisors and canine. In the lower jaw, the eruption of the true molars follows the same order as in the superior series. Dp^ is identical in pattern with the molars, while Dp-3- and Dpa resembles their permanent successors. As in the superior series, PT appears to have been without a deciduous predecessor, as the little-worn tooth is hypsodont. The order of replacement of the milk den- tition is believed to have been from behind forward, judging from the fact that P^oes not appear until some time after the eruption of the third and fourth premolars (PI. V, fig. 15). Skn/l (Plates III; V, fig 21). — There is little difference, apart from size, between the skulls of the various species, the specific characters de- Sinclair: typotheria of the santa cruz beds. i9 pending more on dimensions than on structural peculiarities. In side view (PI. Ill, figs. 1, 5, 6), the upper profile of the skull slopes forward gradu- ally from the elevated parietal region, with a more abrupt slope back- ward, especially in P. attenuatum , in which the parietal tract is consider- ably elevated and globose. The orbits are central and widely open posteriorly, of moderate size and less prominent than in Inter atherium, Hegetotherium or Pachyrukhos. The rostrum is laterally compressed and excavated longitudinally. The facial portion of the premaxillary is broad, with short ascending process and attenuated anterior extremity, terminat- ing in a prominent anterior nasal spine. On the palatal surface, the incis- ive foramina are confined almost entirely to the premaxillae, touching the maxillae posteriorly along the line of the premaxillo-maxillary suture. The canine is implanted entirely in the maxillary, the suture between the pre- maxillary and the maxillary passing between the third incisor and the canine. The maxillary forms the entire lower surface of the zygomatic arch, supporting anteriorly a small descending process. It extends as far back as the posterior border of the glenoid cavity and forms with the malar a prominent buttress preventing the lateral displacement of the lower jaw. The maxillary enters also into the anterior and inferior bound- ary of the orbit. Its ascending process is acute and greatly prolonged, extending almost as far as the middle of the orbit, where it is received in a deep notch in the frontal. The facial expanse of the maxillary is broadly excavated horizontally. At the posterior margin of this excava- tion, the large infraorbital foramen perforates the base of the elevated orbi- tal rim. Neither maxilla nor premaxilla is firmly united with the nasals, which, owing to a lack of support, are frequently crushed down into the narial chamber. The lachrymal is scale-like, with but little facial expan- sion. It supports a prominent tubercle, beneath which, on the orbital surface, is the lachrymal duct. A small “ lachrymal ” tubercle is present also on the maxillary (PI. Ill, fig. 5). The malar is a long, narrow ele- ment confined entirely to the zygomatic arch, where it is supported beneath by the maxillary and is overlapped above by the temporal process of the squamosal. Anteriorly, it touches the orbital rim, while posteriorly it ex- tends slightly beyond the maxillary process. The root of the temporal process of the squamosal and a large part of the cranial expanse of the same element are dilated and filled with cancellae (PI. Ill, fig. 7), the chambers communicating with each other as well as with the tympanic 20 PATAGONIAN EXPEDITIONS : PALEONTOLOGY. cavity. The dilatation is not confined to the squamosal but extends also to the mastoid, with which the former is closely fused, resulting in the in- clusion of the external auditory meatus in a puffy mass of bone extending from the postglenoid process to the base of the paroccipital. A view of the upper surface of the skull (PI. Ill, fig. 2) shows that the nasals are long, broad behind, where they are in contact with the frontals, but tapering anteriorly to round points, which project slightly beyond the terminal narial opening. The nasals are convex in cross section anteri- orly, but are excavated longitudinally farther back, producing a sigmoid cross-section. As previously explained, the nasals are but slightly sup- ported on either side by the maxillary and premaxillary. A more or less elongated frontal process extends between the nasal and maxillary, while a second process, in contact with the lachrymal, is applied to the outer surface of the ascending process of the maxillary, excluding the latter from the orbit. The interorbital tract is approximately plane (in some specimens slightly concave, but this may be due to crushing). It is slightly elevated in the region of the ascending maxillary processes and sometimes also along the line of the interfrontal suture, which is more or less persistent. The temporal ridges are low, but sharply defined, and converge rapidly to form a low but strong sagittal crest. The postorbital processes are long and pointed, the posterior border coinciding with the temporal ridges. Just back of these processes the brain case is greatly constricted, but rapidly expands posteriorly, and in P. attenuatum superi- orly. The sagittal and lambdoidal crests join each other at a right angle. The latter bifurcates, sending one branch forward as the superior bound- ary of the temporal process of the squamosal and the other downward and forward over the inion to the base of the paroccipital process, bound- ding the lateral expanse of the distended squamosal and mastoid (PI. Ill, fig. 2). The temporal fossa is continued backward to the lambdoidal crest as a deep trough floored by the parietal and squamosal. Many foramina pierce the latter element, unlike Hegetotherimn, in which most of the foramina in this region perforate the parietal. On the back of the skull (PI. Ill, fig. 4) there is a large expanse of cancellous squamosal and mastoid, but so completely are these elements fused that the suture between them is not apparent. The supra- and exoccipitals are also fused. Dorsally, the supraoccipital is expanded, overlapping the mastoid and supporting prominent tubercles for the recti Sinclair: typotheria of the santa cruz beds. 21 capitis muscles. About midway, the occipital plate is strongly constricted transversely, lodging the mastoid foramen on the suture between the mas- toid and the occipital at the point of greatest constriction of the latter. Inferiorly, the occipital expands, supporting long fang-like paroccipital pro- cesses, the points of which curve forward. Above the foramen magnum the occipital surface is convex, sometimes broadly so, sometimes with a slight median keel. The condyles are oblique and semi-cylindrical in shape. The palate (PI. Ill, fig. 3) is concave in all dimensions, but less deeply so antero-posteriorly than transversely. The palatine extends as far Protypotherium australe, zygomatic arch removed to show the arrangement of the cranial ele- ments in the temporal fossa, x p am, auditory meatus ; as, alisphenoid ; c, condyle ; f, frontal ; m, mastoid ; o , optic foramen ; os, orbitosphenoid ; ov, foramen ovale and lacerum medium ; p, parietal ; pi, palatine ; pt, pterygoid ; s, squamosal. forward as the anterior lobe of M-. Opposite the posterior margin of the last molar, it is constricted by deep notches, beyond which it expands again, terminating in a pair of triangular rugose processes formed in part by the palatine and in part by the alisphenoid. Each process is supported by a strong, transverse, plate-like buttress of the alisphenoid, the squamosal taking no part in its formation, unlike the structure of this region in Hegetotherium (cf. text figs. 8 and 13). The pterygoid is a small, thin plate, terminating in upwardly curved hamular processes. Between the pterygoid and alisphenoid plate is a deep fossa. In all the specimens examined the narial border has been injured and it is not possible to ascertain whether a posterior narial spine was present or not. Two pos- Fig. 8. / P S 22 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. terior palatine foramina are present, the larger perforating the maxillary opposite the posterior lobe of the fourth premolar about half way between the base of that tooth and the median suture, while a second smaller fora- men perforates the maxillo-palatine suture. The basioccipital supports a median keel, which does not extend beyond the suture between this ele- ment and the basisphenoid. The pear-shaped bullae are hollow and are more or less completely fused with the basioccipital, producing a different arrangement of the cranial foramina in this region from that characterizing Hegetotkerium. Anteriorly, where most prolonged, the bullae are greatly flattened in the horizontal plane. The glenoid surfaces are almost flat, presenting downward and backward. They are separated from the post- glenoid process by a wide fossa, accommodating the non-articular portion of the madibular condyle when the mouth is widely open. Owing to the fusion of the basioccipital with the tympanic bulla, the carotid foramen is shifted posteriorly, fusing with the foramen lacerum posterius at the base of the paroccipital process. Traces of a septum dividing the two foramina are observable. Large condyloid foramina are present, in addition to which one or more small foramina pierce the basi- occipital on either side of the median keel in line with the carotid-lacerum posterius foramina. A single large foramen (probably vascular) is lodged in the groove between the tubular auditory meatus and the dilated mas- toid, while in Hegetotkerium three or four foramina are present in this region. Two foramina, lying side by side, perforate the suture between the meatus and the postglenoid process (postglenoid foramina). The foramen ovale and foramen lacerum medium are confluent, with vestiges of a dividing septum. Within the orbit, a deep groove leads forward to the infraorbital foramen. Posteriorly the course of this groove is back- ward and downward, terminating in the posterior palatine notches. Va- cuities within this groove communicate with the olfactory chamber. The foramen rotundum and spheno-orbital foramen are confluent. The pos- terior narial canal is connected with the spheno-maxillary fossa as in the hyraces. Several smaller foramina within the orbit are probably vascular. Complete fusion of the opposite halves of the mandible has taken place, leaving no trace of suture (PI. IV, fig. 15). The horizontal ramus increases suddenly in depth beneath the last molar and as suddenly decreases in depth posteriorly, producing a prominent convexity in this region, which is carried to a much greater extreme in Inter atherium. The Sinclair: typotheria of the santa cruz beds. 23 angular portion of the mandible does not project below this convexity. Posteriorly, the angle extends far beyond the condyle. The free border is strongly curved inward, inclosing a deep submaxillary fossa. The masseteric fossa is shallow, but its boundary is well defined and almost circular in outline interiorly. The coronoid process is high, thin and sharply pointed, projecting far above the condyle. Its anterior border is S-shaped, and strongly inclined forward interiorly. The condyles are broadly oval in outline, wider externally than internally and almost flat both antero-posteriorly and transversely. The articular surface presents upward and forward. Posteriorly, the capitulum supports a non-articular projection, which fits into the postglenoid fossa, preventing the backward dislocation of the mandible when the mouth is widely opened. A large mental foramen is present beneath the fourth premolar or the anterior lobe of the first molar, and one or two smaller foramina are in the symphysial region beneath the canine. On the inner surface of the ramus, the inferior dental canal is circular in outline, with a more or less well defined groove leading into it from above, a structure far more strikingly developed in Hegetotherium . Vertebral Column, Ribs and Sternum. — -The atlas (PI. V, figs. 18-20) is characterized by broad transverse processes, with irregularly lobate free border, but little basal constriction and no canal for the vertebral artery. The neuro-arterial foramina are large and completely inclosed anteriorly by strong bony bars. The neural arch is wide, with a large median tubercle at its anterior margin. The narrow inferior arch also supports a tubercle at its anterior margin, but a much smaller one than that on the arch above. The axis (PI. V, figs. 16, 17) may be readily recognized by the small size of the arterial canal perforating the base of the transverse process. The neural spine is strong and hatchet-shaped, projecting beyond the centrum posteriorly, where it terminates in a point. This has been broken off in the specimen figured. The odontoid is flattened dorso-ventrally, its superior surface being flush with the floor of the neural canal, unlike Hegetotherium. Interiorly, the centrum is keeled and supports a pair of tubercles at its posterior margin. The centra of the third and fourth cervicals are also keeled, with similar inferior tubercles, but both keels and tubercles are absent from the posterior members of the series. The neural spines and transverse processes have 24 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. not been preserved in the anterior cervicals available for study. In the sixth, the transverse process supports a large hatchet-shaped inferior lamina and a prominent diapophysis. The former is wanting in the seventh cervical and the diapophysis is greatly elongated. No complete dorso-lumbar series is known, but, from analogy with the closely related Inter atherium, it may be assumed with a high degree of probability that there were fifteen dorsals and seven lumbars. The anter- ior dorsals (PI. VI, fig. 8) have high narrow spines strongly inclined backward, with the tips terminating in knob-like expansions. The spines of the posterior dorsals (PI. VII) are wide antero-posteriorly, with obliquely truncated, dorsally flattened summits. The change from back- ward to forward inclination of the neural spines occurs at about the thir- teenth dorsal. The spines of the anterior lumbars resemble those of the posterior dorsals, but increase in length and are more slender posteriorly. The transverse processes of the lumbars are broad flat blades, curving forward and downward. The width of these processes increases in the posterior lumbars. Prominent anapophyses are developed in the posterior dorsals and anterior lumbars, but decrease in size toward The end of the lumbar series. Large metapophyses and strongly interlocking zygapoph- yses are also characteristic of the posterior dorsals and lumbars (PI. VI, figs, io, 14, 15). Three sacrals are in contact with the ilium and two caudals are fused with them, making five vertebrae in the sacral complex (PI. VI, fig. 10). A dorsal intervertebral fenestra is present between the first and second sacrals. The remaining sacrals and the first and second caudals have the neural arch firmly fused, without trace of fenestrae. The neural spine of the first sacral is imperfectly preserved, but appears to have been similar to that of the posterior lumbars. The spines of the remaining members of the caudo-sacral series are solidly fused, as in Interatherium , forming a narrow plate, with the dorsal margin transversely expanded and flattened. The tail is known to have been long, eighteen free caudals being pre- served with one specimen (No. 15,161). The proximal caudals have broad transverse processes and short, transversely compressed spines. Pos- teriorly, the spines rapidly decrease in height and the transverse processes elongate antero-posteriorly, but decrease in transverse diameter. Begin- ning with the seventh free caudal, the neural arch is greatly reduced and the transverse process limited to a small tuberosity at the anterior and SINCLAIR : TYPOTHERIA OF THE SANTA CRUZ BEDS. 25 posterior margins of the centrum. The distal caudals are mere cylinders, without neural arch or transverse process. The ribs (text fig. 9) are slender, with little tendency toward flattening, even in the anterior members of the series. Distally, near the junction with the costal cartilage, the rib shaft expands slightly. The costal carti- lages were calcified and are preserved in No. 15,352 (text fig. 9, B). Fig. 9. A, Protypotherium sp. Posterior cervicals and anterior dorsals with ribs in place, viewed from the left side, x f (No. 15,352). B, the same from below, showing the sternum and calcified costal cartilages, x f . The sternum consists of six pieces, of which four are mesosternal. The first sternal segment (figured separately in PI. VI, fig. 7) is strongly keeled interiorly in P. australe , but this keel is wanting in No. 15,352. The mesosternal segments are hour-glass shaped, with the inferior sur- face flat. The xiphisternal segment is longer than any of those preced- ing. Its inferior surface is also flat. Posteriorly, the transverse diameter increases slightly. Appendicular Skeleton. — Several more or less complete scapulae of P. australe are in the collections, of which the best preserved is figured on Plate IV, figure 10. The coracoid border above the suprascapular notch is moderately convex throughout, the convexity continuing un- broken over the vertebral border as far as the inferior angle. The axillary 26 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. border is also convex centrally, but immediately above the glenoid fossa and below the inferior angle it is concave. A high narrow spine is present, terminating in a slender acromion process and supporting a large met- acromion, the distal portion of which has not been preserved. The body of the bone is very thin and has been slightly deformed by crushing, pro- ducing an undulatory surface throughout the floor of the supraspinous fossa. The infraspinous fossa is triangular in outline and deeply concave transversely. The glenoid surface (PI. IV, fig. 11) is continued forward to the extremity of the large bicipital tubercle. Apart from this prolong- ation, the outline of the glenoid fossa is approximately circular. The margins are but little elevated. The fossa is slightly concave in all diameters. The coracoid process is short, with its free margins curved inward. The proximal extremity of the humerus (Pis. IV, figs. 6, 7 ; VI, fig. 1) is the stoutest and heaviest portion of that bone. The head, which is hemi- spherical in shape, projects a considerable distance beyond the posterior margin of the shaft. The tuberosities are low, the greater tuberosity rising but little above the level of the head. It is separated from the lesser tuberosity by a broad bicipital groove. Antero-externally, the proximal surface of the shaft is flattened, supporting a broad triangular area, the apex of which is continued distally as the deltoid ridge. Some distance above the distal extremity of the deltoid ridge, on the inner surface of the shaft, is a prominent rugosity for muscular attachment, corresponding in position to the area of insertion of the coracobrachial muscle in the rabbit. The shaft is slightly curved antero-posteriorly and somewhat flattened laterally. The distal end is expanded transversely. The supinator ridge is very slightly developed, but the inner epicondyle is prominent and a large internal epicondylar foramen is present. The distal articular surface is characterized by the great prominence of the inner lip of the trochlea and the absence of any sharp demarcation be- tween the latter and the capitellum. Posteriorly, both lips of the trochlea are sharp (PI. IV, fig. 7), the inner exceeding the outer in elevation, but farther forward the outer lip terminates suddenly at the margin of the capitellar expansion (PI. IV, fig. 6). No perforation unites the anconeal and coronoid fossae. The radius (Pis. IV, figs. 12-14; VI, fig. 2) is strongly curved antero- posteriorly. The shaft is elliptical in cross-section proximally, but dis- SINCLAIR: TYPOTHERIA OF THE SANTA CRUZ BEDS. 27 tally it becomes wedge-shaped, the point of the wedge corresponding to the sharp interosseous margin, while the head of the wedge forms the convex anterior surface of the shaft. The head is oval, narrower inter- nally than externally and slightly cupped for reception of the humeral capitellum. The inner narrower portion of the oval is occupied by a surface inclined downward and inward, slightly convex transversely and articulating with the inner portion of the humeral trochlea. A slight projection on the anterior margin of the radial head separates the margins of these two surfaces. The ulnar surface is slightly convex transversely and flattened vertically. The carpal articular surface is small in propor- tion to the width of the shaft at this extremity. It is irregularly oval in outline, concave in all dimensions and shows faint traces of division into two surfaces for the scaphoid and lunar respectively (PI. IV, fig. 13.) The styloid process is quite short and inconspicuous. The ulnar shaft (Pis. IV, figs. 16, 17; VI, fig. 3) is strongly curved laterally. The interosseous border is almost straight and strongly rugose, and the posterior border smooth and broadly sigmoid in outline. On the outer surface the shaft is excavated longitudinally for about half its length below the sigmoid cavity. The olecranon is heavy, posteriorly flattened and strongly rugose distally and projects as far forward as the coronoid process. The latter is quite sharp, the humeral and radial articular sur- faces meeting at an acute angle, as in many of the rodents. Distally, the anterior margin of the shaft curves outward, leaving a flattened area above the radial facet (PI. IV, fig. 16). The latter is convex from side to side and plane proximo-distally. In contrast with Hegetotherium , the styloid process is long, narrow and more sharply pointed. The manus is tetradactyl, with interlocking carpus and metacarpus (PI. V, figs. 3, 4, 9 ; text fig. 3, A). The second and. third digits are of approxi- mately the same size, the third slightly exceeding the second in length, with the axis of the foot passing between them. The fourth digit is much shorter and feebler and the fifth greatly reduced. No trace of the first remains. Proximally, the scaphoid presents a broad, antero-posteriorly convex surface for articulation with the radius. Distally, it is in contact with the magnum by a long, narrow process (scaphoid + centrale) support- ing a small, plane, quadrangular facet, and with the trapezoid by a large crescentic facet, plane in dorso-palmar section and slightly concave trans- versely. Externally, it articulates with the lunar by a small, oval, proximo- 28 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. distally concave facet on the outer side of the process in contact with the magnum. Internally, the palmar surface supports a heavy rugose tuber- cle. The lunar is a large element, wedge-shaped in dorsal view, with a broad, strongly convex proximal surface for the radius. Distally, it is in contact with the magnum and unciform. The surface for the former is divided into two portions, a small dorsal convex area, wedge-shaped in outline, and a large, irregularly quadrangular palmar tract, strongly concave in dorso-palmar section. The unciform facet is triangular, slightly con- cave, and but little differentiated from the surface for the cuneiform. In- ternally and distally, there is a small triangular concave facet for the scaph- oid. Externally, contact with the cuneiform is secured by a broad sur- face of irregular outline, concave in dorso-palmar section and convex proximo-distally. The cuneiform is broadly grooved proximally for the styloid process of the ulna. On the palmar surface is a large oval flat facet for the pisiform, distally an oval concave facet for the unciform and internally a small, crescentic, convex surface for the lunar. Externally, the cuneiform supports a large protuberance almost half as large as the bone itself. The pisiform is a large T-shaped element, the cross bar of the T articulating with the cuneiform. The shaft is quite heavy, broadly convex in cross-section dorsally and more sharply angulate on the palmar surface. Distally, the shaft is expanded and is either abruptly truncate (P. australe ) or terminates in a broad convexity [P. prczrutilum). The trapezium is convex internally and slightly rugose, with no trace of an articular surface for the pollex. Externally, it articulates with the second metacarpal, which is deeply excavated for its reception, while proximally it is in contact, by a small triangular facet, with the trapezoid. The latter is a triangular ele- ment, broad dorsally, but tapering toward the palmar margin, where it supports a heavy tubercle. Proximally, there is a triangular facet for the scaphoid, concave in dorso-palmar section and slightly convex transversely, and, distally, a similar facet convex in both dimensions for the second metacarpal. Internally, there is slight lateral contact with the trapezium, while externally and distally the surface for the magnum is but little dif- ferentiated from the metacarpal facet. The magnum is narrow dorsally, but increases in proximo-distal diameter toward the palmar margin. It is in contact with the second metacarpal, the trapezoid, the scaphoid, the lunar and the unciform, and distally rests upon the third metacarpal. The surface for the second metacarpal is narrow dorsally, but increases in width Sinclair: typotheria of the santa cruz beds. 29 toward the palmar margin. It is irregularly triangular in outline, slightly convex in dorso-palmar section and strongly concave transversely. The trapezoidal surface is narrow and broadly convex in dorso-palmar section, while the facet for the scaphoid is concave from side to side and also dorso-palmarly. The lunar surface is sigmoid in the latter direction, slightly convex palmarly, and concave dorsally. It is confluent dorsally and externally with the irregularly triangular, plane facet for the unciform. The metacarpal facet is quadrangular in outline and deeply concave in dorso-palmar section, while transversely it is plane. The unciform is a^ large element, irregularly tetrahedral in form, supporting proximally a large, transversely sigmoid surface for the cuneiform and proximo-inter- nally a concave semicircular facet for the lunar. Internally, there is a large surface, slightly concave proximo-distally, for the third metacarpal and the magnum. Distally, there are two surfaces for the fourth and fifth metacar- pals respectively. That for the former is triangular and concave in all dimensions, while the surface for metacarpal V is oval, concave in dorso- palmar section and slightly convex transversely. Proximally, the meta- carpus is strongly interlocking. The articular surfaces are sufficiently well shown in the figure (PI. V, fig. 9) to make detailed description unneces- sary. The shafts are transversely flattened and bear well marked plantar keels, distally. The distal articular surfaces of the proximal phalanges are confined to the distal and palmar surfaces, while the proximal surfaces of the second row present upward and forward, indicating a moderate amount of angularity in the position of these elements (see restoration of the skeleton, PI. VII). The ungual phalanges are laterally compressed hoofs, with slight, more or less well defined terminal clefts. All the digits sup- ported terminal phalanges, but in none of the specimens studied has the terminal phalanx of digit V been preserved. The ilia (PI. VI, figs. 9, 10) are broadly expanded, with deeply concave gluteal fossae and prominent anterior superior and posterior inferior spines. Between the anterior and posterior superior spines the crest of the ilium is inclined forward obliquely. Inferiorly, the iliac margin is broad, flat- tened and channeled longitudinally by a shallow groove. The neck is stout and the tubercle for the rectus femoris rather prominent and rugose. Ilio-pectineal eminences are practically absent. The ischium is broadly expanded posteriorly, with prominent spine and tuberosity. The ischial ramus and the pubis are slender. 30 PATAGONIAN EXPEDITIONS ! PALEONTOLOGY. The straight femoral shaft is slightly flattened transversely (Pis. IV, figs. 1-3 ; VI, fig. 4). The head is large and hemispherical, with a deep pit for the round ligament (not shown in the figures). The major tro- chanter slightly exceeds the head in elevation, from which it is separated by a shorter notch than in Hegetotherinm. The trochanter minor is very large and sharply pointed. A prominent third trochanter is present, situ- ated on the opposite side of the shaft and slightly farther down than the trochanter minor. The condyles are large, the inner exceeding the outer A in width and posterior extension. Both are slightly convex transversely. The patella (PI. IV, figs. 8, 9) is quite narrow, tapering distally. An- teriorly, it is strongly convex in all directions and rugose for tendinous attachment. Posteriorly, it is divided by a median keel into two equal concave articular surfaces, occupying the entire posterior aspect of the bone except at the extreme distal end. The tibia and fibula (Pis. IV, figs. 4, 5; VI, fig. 5) are usually unfused both proximally and distally, but in some specimens partial fusion has taken place distally. The tibial shaft is slightly arched inward and the fibular shaft outward, inclosing a large lens-shaped interosseous fossa. Of the condylar surfaces, the outer is the larger and more nearly circular. The spine is comparatively inconspicuous. The proximal fibular facet presents outward and downward, and is completely overhung by the outer condyle. It is a dumb-bell-shaped surface, sigmoid in antero-posterior section and plane transversely. The tibial shaft is triangular in cross- section proximally, becoming circular in cross-section toward the middle portion of the shaft. Externally, it is deeply excavated longitudinally, while the internal surface is quite smooth and slightly convex transversely. The cnemial crest is prominent, terminating in P. australe in a rugose tubercle. Distally, the tibial shaft is slightly expanded transversely, sup- porting externally a large rugose surface for ligamentous union with the fibula. The distal articular surface (PI. IV, fig. 5), which is placed some- what obliquely with respect to the median plane of the shaft, is divided by a prominent keel into two approximately equal grooves for the crests of the astragalar trochlea. The internal malleolus is quite long, support- ing a plane surface for lateral articulation with the internal astragalar crest. The fibula (Pis. IV, fig. 18; VI, fig. 5) is quite slender, slightly arched externally and greatly compressed laterally, with strongly marked interos- Sinclair: typotheria of the santa cruz beds. 3i seous ridge. Proximally, the shaft is expanded both antero-posteriorly and transversely, with the oval surface for the tibia terminal. Distally, the fibula is in lateral contact with the tibia by a broad rugose surface. In- ternally, the distal end articulates with the outer astragalar crest by a cres- centic, plane facet, while distally a broad, antero-posteriorly concave surface is in contact with the calcaneum. The peroneal groove is quite distinct. The pes (PI. V, figs. 1, 2, 5-8, 10) is tetradactyl and paraxonic, the median digits are of approximately the same size and the lateral digits reduced. The hallux is wanting. The astragalar trochlea is long, moderately deep, with the crests sharp and equally developed. The neck is long and the head globular. The ectal facet is dumb-bell shaped, wider proximally than distally and strongly concave in proximo-distal section. The sustentacular facet is a flattened oval in outline, convex in all diam- eters and is supported largely by the neck of the astragalus. The fibular and internal malleolar facets, on the lateral surfaces of the body of the astragalus, are vertical, and the entire body is approximately symmetrical to the vertical plane, differing in this respect from Hegetotherium (cf. Pis. 11, fig. 1 9; V, fig. 1). The most striking feature of the calcaneum is the large fibular facet, which is almost as large as the ectal surface. Both are strongly convex proximo-distally and slightly convex transversely. The sustentaculum is heavy and deep in the dorso-plantar diameter, with its free margin grooved. The sustentacular facet is oval in outline and concave. The cuboidal facet is much less deeply concave than in Hege- totherium. Unlike that genus, there is no articulation between the cal- caneum and the navicular. The tuber is moderately elongated and rather heavy, supporting a large rugose area distally. The navicular is deeply cupped proximally for the head of the astragalus, while distally it sup- ports three facets for the cuneiform bones (PI. V, fig. 8). It may be readily distinguished from that of Hegetotherium (PI. II, fig. 16) by the small size of the mesocuneiform facet and the absence of an articular surface for the calcaneum. The external cuneiform facet is subcircular in outline and almost plane, that for the mesocuneiform oval and slightly convex in all diameters, while the facet for the internal cuneiform is cres- centic and convex in all diameters. Externally, the navicular is in con- tact with the cuboid by a large reniform facet, slightly convex in dorso- plantar section and concave proximo-distally. The internal plantar tuberosity is quite long, perhaps representing a coossified sesamoid, such 32 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. as occurs in a free state in many of the hystricomorph rodents, with which it coincides in position. The proximal surface of the cuboid is strongly convex in dorso-plantar section and broadly concave transversely. Internally, it is in contact with the navicular by a large oval facet concave in dorso-plantar section and convex proximo-distally, below which are two oval facets, with their major axes at right angles, for contact with the outer cuneiform. The facet for the fourth and fifth metatarsals is a large, irregularly oval area, concave in all dimensions and not differentiated into separate articular surfaces for these two elements. The peroneal groove is deep and the overhanging tubercle very large. Of the cunei- form series, the outer is the largest. It is irregularly quadrilateral in out- line dorsally, decreasing in transverse diameter toward the plantar surface, where it supports a long process terminating in a blunt tubercle. The proximal surface is semicircular in outline and almost plane. Of the two oval facets for the cuboid, the dorsal is convex in proximo-distal section and almost a plane surface at right angles to this, while the plantar sur- face is plane. Internally, there is a narrow, almost plane surface for the mesocuneiform proximally and two semicircular, approximately plane facets distally for the second metatarsal. The surface for the third meta- tarsal is reniform, wider dorsally than at the plantar margin, convex from side to side and concave in dorso-plantar section. The mesocuneiform is quite small. Its proximal surface is semicircular and almost plane. Externally and proximally, there is a narrow, plane surface for the outer cuneiform. Distally, the facet for the second metatarsal is sigmoid in dorso-plantar section and convex transversely. Internally and proximally, there is a single facet, variable in form, for the internal cuneiform. The latter is scale-like, terminating distally in a rugose tubercle without any trace of an articulation for the hallux. The third and fourth metatarsals are of approximately the same length and weight, while the second and fifth are much shorter and more slender, the fifth slightly exceeding the second in length. Proximally, the metatarsals interlock with each other and with the tarsus. The proximal surfaces are sufficiently well shown in the figure (PI. V, fig. io) to dispense with further description. The shafts are slightly flattened. Well developed keels are present distally. Except for their superior size, the phalanges are indistinguishable from those of the manus. The terminal phalanges are laterally compressed hoofs, with- out clefts. Sinclair: typotheria of the santa cruz beds. 33 Restoration of the Skeleton (PL VII). — Compared with Inter atherium (PI. IX), the restored skeleton of Protypotherium shows the proportionately greater length of the limbs. The hind feet have been given a digitigrade position because of the length, depth and sharpness of definition of the astragalar trochlea. The depth of the posterior portion of the thorax is conjectural, as are also the lengths of the spines of the median dorsals and cervicals. The tail has been drawn to scale from a much smaller speci- men, probably P. prcerutilum , in which eighteen free caudals are present. It may have been longer than is represented in the restoration. Species. — Three species are represented in the collections at Princeton University and the American Museum of Natural History (P. australe , prcerutilum , attenuatum ). Within the limits of each there is considerable range in size, all transitions occurring between the minimum for the largest species ( P '. australe ) and the maximum for the next smaller form ( P . prcerutilum ), making their separation a more or less arbitrary matter. P. attenuatum shows less range in size and is characterized by fairly well marked cranial peculiarities, by which it may be readily recognized. In the absence of detailed stratigraphic work, with systematic collecting from definite levels, it is impossible to avoid confusing individual variations with true mutations of specific value, and this may account for the diffi- culty in separating P. australe from P. prcerutilum. Protypotherium australe (Moreno) Ameghino. (Pis. Ill, Figs. 1-4; IV ; V, Figs. 1-3, 6-10, 16-20 ; VI, Figs. 7-10, 14, 15 ; VII ; Text Figs. lA, 2, 3A, 4A, 8.) Toxodontophanus australis Moreno ; Patagonia, resto de un antiguo con- tinente hoy submerjido, p. 23, 1882 ( nomen nudum). Protypotherium ( Toxodontophanus) australe Ameghino ; Observaciones generales sobre el orden de mamiferos estinguidos sud-americanos llamados Toxodontes, etc., p. 64, 1887. Patriarchies palmidens Amegh.; Contrib. al Conoc., etc., p. 481, PI. 15, figs. 2-3, 1889. Patriarchus furculosus Amegh.; Revista Argent, de Hist. Nat., I, p. 292, 1891. Patriarchus distortus Amegh.; ibid., p. 293, 1891. Protypotherium distortum Amegh.; Enum. Syn., etc., p. 13, 1894. 34 PATAGONIAN EXPEDITIONS : PALEONTOLOGY. Patriarchus altus Amegh.; Revista Argent, de Hist. Nat., I, p. 293, 1891. Protypotherium altum Amegh.; Enum. Syn., p. 13, 1894. ? Proty pot herium lineare Amegh.; Enum. Syn., etc., pp. 13-14, 1894. This is the largest and also the most abundant species of Protypotherium in the Santa Cruz beds. It may be readily recognized by its large size (skull length .099-. 1 12). There is considerable individual variation in the size of the skull and teeth, but with a sufficiently large suite of specimens it is possible to detect all transitions between the extremes in dental and cranial measurements. Specimens in the Munich collection, determined by Ameghino as Patri- archus palmidens and P. furculosus , are old individuals of P. australe. The latter differs from P. prcerutilum, a smaller and more slenderly constructed form, in the greater width of the molar crowns, which, although they may in some of the smaller individuals approach those of P. prcerutilum in antero-posterior diameter, are always much wider. From P. attenuatum the species under consideration differs in its much greater size and in the less pronounced dilatation of the parietal region of the skull. The specimens selected for measurement and illustration represent fairly well the extremes of size variation within the limits of the species. Nos. 15,828, 15,340, 15,551, 15,598 of the Princeton collection, and Nos. 9565, 9149 and 9566 of the American Museum collection, are figured on the accompanying plates. The following localities are represented, the figures referring to the number of individuals from each: Rio Gallegos (3), Canon de las Vacas (2), Felton’s estancia, Rio Gallegos (7), Halliday’s estancia, Rio Gallegos (4), Killik Aike (7), West of Killik Aike Ranch, Rio Gallegos (1), two miles west of Killik Aike, Rio Gallegos (1), five miles south of Coy Inlet (1), ten miles south of Coy Inlet (1), seventeen miles north of Cape Fair- weather (1), Mount of Observation (2). Measurements. No. No. No. No. 13,828. 9565- 1 5,39s ■ 914.9. Skull, maximum length .... .112 .108 n length, premaxillae to condyle inclusive .1085 •1055 .099 u greatest width across arches . .0645 .061 .062 u interorbital width .032 .031 .0273 SINCLAIR : TYPOTHERIA OF THE SANTA CRUZ BEDS. 35 No. No. No. No. 15,828. 9565- 15, 598- 915.9 Skull, least width of brain case “ height of occiput .... “ width of occiput .... .0175 .021 .0425 •0155 .020 .044 .017 Palate, length ..... .070 .067 * .0615 “ width at Mi .... Mandible, length including Iy .0245 .100 .021 * . IOI .0235 “ depth below Py .0176 .016 .0166 “ “ “ posterior margin of .0293 .027 .0255 Upper dentition, length Ii-Mi .063 .060 •0575 .058 “ “ “ Pi-Mi on alveolar border ...... .045 .042 .0405 •0395 Upper dentition, length Mi-Mi on alveolar border ...... .0245 .023 .0225 .021 P-, width ...... .0056 .0067 .007 “ greatest transverse diameter .0034 .0026 .0023 Ii, width ...... .005 .0055 .005 .005 “ greatest transverse diameter .0027 .0024 .002 .002 Ii, width ...... .0055 .0055 •0053 .0052 “ greatest transverse diameter .0027 .0026 .0023 .002 C, width ...... .005 .005 .005 .005 “ greatest transverse diameter .003 .0022 .0022 .002 Pi, antero-posterior diameter just above triturating surface .... .005 .005 .005 .004 “ transverse diameter in same plane .003 .0027 .0028 .0025 Pi, antero-posterior diameter just above trit- urating surface ..... .0058 .0055 .0053 .0045 “ greatest transverse diameter in same plane. .004 ■0037 •0035 .0035 Pi, antero-posterior diameter just above trit- urating surface ..... .0065 .006 .006 .0055 “ transverse diameter in same plane .0045 .0045 .004 .004 Pi, antero-posterior diameter on triturating surface externally .... .006 .006 .0055 .0048 “ transverse diameter at widest part .005 .0045 .0045 .0045 Mi, antero-posterior diameter on triturating surface externally .... .0088 .008 .008 .0075 “ transverse diameter at widest part .0058 .005 .0054 .0052 Mi, antero-posterior diameter on triturating surface externally .... .0085 .0073 .0073 .0065 “ transverse diameter at widest part .005 .005 .0045 .0048 Mi, antero-posterior diameter on triturating surface externally .... .008 .0075 .0066 .0065 “ transverse diameter at widest part .0045 .0045 .004 .004 * Approximate. 36 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. No. No. No. No. 15,828. 9565- '5,59s- 9149. Lower dentition, length Iy-My .059 .0565 “ “ “ Py-My on alveolar border ...... .044 .0415 •039 Lower dentition, length My-My on alveolar border . . ... .026 .0235 .0228 Iy, width of crown ..... .002 .002 “ greatest transverse diameter .0023 .0019 Iy, width of crown ..... .0027 .0025 .0026 “ greatest transverse diameter .0025 .002 .002 Iy, width of crown ..... .004 .004 “ greatest transverse diameter .0025 .002 C, width of crown ..... .005 .0045 .0045 “ greatest transverse diameter PT, antero-posterior diameter on triturating .0027 .0026 .0023 surface ...... .0045 .0055 .005 “ greatest transverse diameter Py, antero-posterior diameter on triturating .003 .0022 .0025 surface ...... l n O q .0045 b 0 “ greatest transverse diameter Py, antero-posterior diameter on triturating .003 .003 .0025 surface ...... .0055 .0053 .005 “ greatest transverse diameter Py, antero-posterior diameter on triturating .0034 .003 .003 surface ...... .0067 .0065 .006 “ greatest transverse diameter My, antero-posterior diameter on triturating .004 .0035 .0035 surface ...... .009 .008 .007* “ greatest transverse diameter M^, antero-posterior diameter on triturating .0045 .004 .0035 surface ...... .0085 .0075 .0073 “ greatest transverse diameter My, antero-posterior diameter on triturating .0042 .004 .0035 surface ...... .010 .009 .0083 “ greatest transverse diameter .0037 No. .0032 .0032 I565I- Atlas, approximate width across transverse processes ...... .032 Atlas, width across anterior articular surfaces .020 “ “ “ posterior “ “ .020 “ “ of neural arch .009 Axis, length including odontoid, but exclusive of distal epiphysis .... .0205 * Approximate. Sinclair: typotheria of the santa cruz BEDS. 37 No. 15,551. Axis, width across condyles ...... .019 Axis, height of neural spine above floor of neural canal .016 No. 15,828 First dorsal, length of centrum ...... .0105 Second “ “ “ “•..... .0105 “ “ “ “ spine measured along anterior margin .026 Third “ “ “ centrum ...... .0105 “ “ “ “ spine measured along anteror margin . .029 Fourth dorsal “ “ centrum ...... .0105 “ “ “ “ spine measured along anterior margin .0265 Fifth dorsal, length of centrum .0115 Sixth “ “ .01 1 5 First lumbar, length of centrum ..... .015 “ “ width across transverse processes . .014 “ “ “ of neural spine at tip .01 1 Second lumbar, length of centrum, minus epiphyses . .014 “ “ width across transverse processes .021 Third lumbar, length of centrum, lacking anterior epiphysis .0165 “ “ width across transverse processes . • .028 Fourth lumbar, length of centrum, lacking anterior epiphysis .015 No. 9566. “ “ width across transverse processes .032 .031 “ “ “ of neural spine at tip ... .010 .0075 Fifth lumbar, length of centrum ..... .017 “ width across transverse processes . .032 .031 “ “ “ of neural spine at tip . .0076 .006 Sixth lumbar, length of centrum ..... .0175 “ “ width across transverse processes .036 .0325 “ “ “ of neural spine at tip . .0083 .006 Seventh lumbar, length of centrum minus epiphyses . .015 “ “ of neural spine at tip .0075 .0045 Scapula, length, coracoid process to inferior angle .0835 “ greatest transverse width ..... •043 No. 914.9. “ 'antero-posterior diameter of glenoid fossa .0195 “ transverse diameter of glenoid fossa .012 .0095 Humerus, length ........ .104 .094 .086 greatest antero -posterior diameter at proximal end .0265 .023 .022 “ “ transverse diameter at proximal end .022 .0187 “ “ “ “ distal end . .027 .023 Radius, length ........ .082 .075 “ antero- posterior diameter of head .... .0075 .0066 “ transverse diameter of head ..... .0105 .0104 “ width of distal end ...... •0155 •013 38 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. Ulna, length ...... “ width at upper margin of sigmoid cavity “ “ “ coronoid process . “ antero-posterior diameter of distal end . Carpus, width, trapezium — cuneiform . Metacarpal II, length ..... “ “ greatest width distally . “ III, length .... “ “ greatest width distally . “ IV, length .... “ “ greatest width distally . “ V, length ..... “ “ greatest width distally . Manus, digit II, first phalanx, length “ “ “ “ “ width proximally “ “ “ “ “ “ distally . “ “ “ second “ length “ “ “ “ “ width proximally “ “ “ “ “ “ distally . “ “ “ ungual “ length “ “ “ “ “ width proximally “ “ III, first phalanx, length “ “ “ “ “ width proximally “ “ “ “ “ “ distally . “ “ “ second “ length “ “ “ “ “ width proximally “ “ “ “ “ “ distally . “ ungual “ length “ “ “ width proximally IV, first phalanx, length “ “ “ width proximally “ “ “ “ distally . “ second “ length “ “ “ width proximally “ “ “ “ distally . “ ungual “ length “ “ “ width proximally V, first phalanx, length “ “ “ width proximally “ “ “ “ distally . Pelvis, length ..... “ transverse diameter at acetabulum “ width of ilium at greatest expansion “ width at neck of ilium No. No. 15,828. 914.9. .1055 .0955 .0146 .014 .016 .015 .009 .0088 .0205 .030 .0074 .032 .007 .025 .0062 .0145 .005 .015 .007 .005 .0105 .006 .0045 .0105 .0045 .0145 .007 .005 .0105 .0055 .004 .010 .0045 .0128 .006 .0042 .009 .005 .0035 .009 .0038 .010 .0048 .003 .120 .025 •0135 No. 9566. .100 .050 .0195 .013 SINCLAIR! TYPOTHERIA OF THE SANTA CRUZ BEDS. 39 Pelvis, antero-posterior diameter of acetabulum “ transverse “ “ “ Femur, length ..... transverse diameter proximally . “ “ “ at middle of shaft “ “ “ distally Patella, length ..... “ width Tibia, length ..... “ transverse diameter proximally “ “ “ distally . Fibula, length Calcaneum, length Astragalus, length “ length of trochlea “ width “ “ “ “ head Metatarsal II, greatest width distally “ III, length . “ greatest width distally “ IV, length . “ “ greatest width distally “ V, length . “ “ greatest width distally Pes, digit II, ungual phalanx, length tt tt n “ III, proximal tt n tt tt tt tt “ “ second (( a n u u a “ “ ungual it it it “ IV, proximal (( it tt a tt tt it it width proximally length width proximally “ distally length width proximally “ distally length width proximally length width proximally “ distally length No. No. No. 15,828. 9H9- 9566. .015 .012 .0155 No. 15540 .1055 .0995 .0925 .027 .0248 .013 .0105 .0095 .025 .022 .0178 •0135 .0095 .010 .120 .1125 . 108 .026 .0235 .014 .Oil . 104 .102 No. 15,828 .0365 •035 .020 .019 .014 .012 .0085 .0085 .008 .007 .007 .040 .009 .042 .0085 •0315 .0065 No. 9T49- .010 .004 .019 .0086 .0063 •0135 .0072 .0054 .0136 .0055 .019 .008 .006 .013 second 40 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. Pes, digit IV, second phalanx, width proximally No. 9H9 .0066 44 a a 44 44 “ distally O O “ “ “ ungual “ length 013 a u 44 a 44 width proximally 005 “ “ V, proximal “ length 015 n a a 44 a width proximally 0065 n a 44 44 “ distally 0045 “ “ “ second “ length .0092 a a u u width proximally ■Ln O q 44 44 44 44 44 “ distally .004 “ “ “ ungual “ length 009 44 44 44 44 44 width proximally .004 Protypotherium prcerutilum Ameghino. (PL V, Figs. 21, 22.) Protypotherium prcerutilum Amegh.; Enumeracion Sistematica, etc., p. 15, 1887. Protypotherium compressidens Amegh.; Revista Argentina de Hist. Nat., I, p. 292, 1891. ? Protypotherium convexidens Amegh.; Revista Argentina de Hist. Nat., I, p. 292, 1891. Patriarchus leptocephalus Amegh.; Revista Argentina de Hist. Nat., I, p. 293, 1 89 1 ; Enum, Syn., etc., p. 14, 1894 (listed); Segundo Censo, etc., p. 150, 1898 (listed). Considerable hesitancy is felt in admitting this as a valid species. The only positive character of specific value seems to be the narrowness of the superior molars in proportion to their length. Although the length antero-posteriorly may be the same as in some of the smaller individuals of P. australe , the width is always less. The characters, apart from size, enumerated in the original descriptions (Ameghino, 1887, 15; 1889, 478) do not seem sufficient to warrant the separation of P. prcerutilum from P. australe. In the absence of exact stratigraphic data, the species as it stands now is a rather ill-defined assemblage of individuals intermediate in size between P. attenuatum and P. australe. Specimens in the Munich collection labelled by Ameghino P. leptocephalum are about the size of P. prcerutilum and might be grouped with the latter without overstepping the bounds of probability. Some do not differ greatly from medium sized individuals of P. australe. Sinclair: typotheria of the santa cruz beds. 4i Specimens referred for the present to P. prcerntilmn have been obtained from the following localities: Ten miles south of Coy Inlet (4), Felton’s estancia, Rio Gallegos (1), Ilalliday’s estancia, Rio Gallegos (1), Killik Aike, Rio Gallegos (2). Measurements. No. No. No. /SjS6. 15,161- 9486. Skull, maximum length ....... .0925 “ length, premaxillre to condyle inclusive .088 “ greatest width across arches ..... .061 •05-45 “ interorbital width ....... .030 “ least width of brain case ...... .020 “ height of occiput ....... •015s “ width of occiput ........ .046 Palate, length ......... .054 “ width at posterior margin of M-2- ..... .0206 .020 Mandible, length including Iy ..... .082 “ depth below Py ...... .015 “ “ posterior margin of My .023 Upper dentition, length Ii-M-4 ...... •0495 “ “ “ Pi-M-4 on alveolar border . •034 .0352 “ “ “ ML-M-2- “ “ “ .020 •0185 Ii, width .......... .0032 “ greatest transverse diameter ...... .0016 Ii, width .......... .0035 “ greatest transverse diameter ...... .002 C, width .......... •0035 “ greatest transverse diameter ...... .002 PL, antero-posterior diameter just above triturating surface .0022 .0036 “ transverse diameter in same plane ..... .002 .0025 Pi, antero-posterior diameter just above triturating surface .0035 .0038 “ transverse diameter in same plane ..... .0025 .0025 Pi, antero-posterior diameter just above triturating surface .0045 .0046 “ transverse diameter in same plane ..... .0032 •0035 Pi, antero-posterior diameter on triturating surface externally . .0042 .0046 “ transverse diameter at widest part ..... .0035 •0037 Mi, antero-posterior diameter on triturating surface externally . .0062 .0066 “ transverse diameter at widest part ..... .0045 .0045 Mi, antero-posterior diameter on triturating surface externally . .0062 . .006 “ transverse diameter at widest part ..... .004 .004 Mi, antero-posterior diameter on triturating surface externally . .0065 .0061 “ transverse diameter at widest part ..... .0034 •0035 Lower dentition, length Iy-My ...... .046 “ “ “ Pj-Mg- on alveolar border . •0343 |oo l df .0194 42 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. Iy, width of crown ...... “ greatest transverse diameter .... I2, width of crown ...... “ greatest transverse diameter .... y, width of crown ...... “ greatest transverse diameter .... C, width of crown ...... “ greatest transverse diameter .... PT, antero-posterior diameter on triturating surface . “ greatest transverse diameter .... P2, antero-posterior diameter on triturating surface “ greatest transverse diameter .... Py, antero-posterior diameter on triturating surface . “ greatest transverse diameter .... Py, antero-posterior diameter on triturating surface . “ greatest transverse diameter .... My, antero-posterior diameter on triturating surface “ greatest transverse diameter .... M2, antero-posterior diameter on triturating surface “ greatest transverse diameter .... My, antero-posterior diameter on triturating surface “ greatest transverse diameter .... Atlas, width across transverse processes “ “ “ anterior articular surfaces . “ “ “ posterior “ “ “ “ of neural arch .... “ “ “ inferior “ . . . . Axis, length including odontoid .... “ width across condyles ..... Cervical series, length ...... Sixth-fifteenth dorsal, length ..... Humerus, length ....... “ greatest transverse diameter at proximal end “ “ “ “ “ distal Radius, antero-posterior diameter of head “ transverse “ “ “ . “ width of distal end ..... Pelvis, width of ilium ...... “ “ “ neck of ilium ..... “ antero-posterior diameter of acetabulum “ transverse “ “ “ Femur, length ....... “ transverse diameter at proximal end “ “ “ “ middle of shaft . No. No. N,386- 15,161 .001 5 >j~> 00 0 0 q q .0016 .003 .002 .0035 .002 .003 .002 .0032 .002 .0043 .0024 .005 .0026 .0057 .003 .0063 .0026 .0075 .0025 .031 .018 .0185 .007 .005 •079 .016 .018 .0058 .008 .0115 .019 -0195 .011 .012 .0117 .012 .011 .084 .086 .022 .0095 No. 9486. .0305 .0205 .019 .004 .0205 .018 .068 .100 SINCLAIR : TYPOTHERIA OF THE SANTA CRUZ BEDS. No. No. No. 15,386. 15,161. 94.86. Femur, transverse diameter at distal end . .020 Tibia, length ...... .097 .100 “ transverse diameter proximally .020 “ “ “ distally .01 1 Fibula length ...... .0945 Calcaneum, length . . . ' . .029 •033 Astragalus, length ..... .018 “ “of trochlea . .Oil “ width of head. .006 43 No. '5,71-2. Tarsus, width ..... Metatarsal II, length “ “ greatest width distally “ III, length “ “ greatest width distally “ IV, length “ “ greatest width distally “ V, length “ “ greatest width distally The following measurements of the manus are from a specimen numbered 15,364, in size between P. australe and P. attenuatum and therefore refer Carpus, width, trapezium — unciform Metacarpal II, length “ “ greatest width distally “ III, length “ “ greatest width distally “ IV, length “ “ greatest width distally “ V, length “ “ greatest width distally Digit II, proximal phalanx, length . “ “ “ width proximal ly “ second ii ii ii ii “ ungual a a III, proximal ii ii ii ii “ second ii ii “ ungual “ distally length . width proximally “ distally length . width proximally length . width proximally “ distally length . width proximally length . width proximally ed to P. proemtiliim. .0165 .0245 .005 .0362 .0073 .0035 .0064 .026 .0046 intermediate .018 .029 .0065 .030 .0063 .026 .0055 .015 .0042 .015 .0065 .005 .01 1 .0055 .004 .01 1 .004 .015 .0065 .0046 .0115 .0054 .011 .004 44 PATAGONIAN EXPEDITIONS : PALEONTOLOGY. Protypotherium attenuatum Ameghino. (Pis. Ill, Figs. 5-7 ; V, Figs. 4, 5 ; VI, Figs. 1-6.) Protypotherium attenuatum Amegh.; Enumeracion Sistematica, etc., p. 15, 1887. Protypotherium globo sum Amegh.; Revista Argentina, etc., p. 291, 1891. Protypotherium icochiloides Amegh.; Enumeration Synoptique, etc., pp. 14-15, 1894. Patriarchus rectus Amegh.; Revista Argentina de Hist. Nat., I, p. 293, 1891. Protypotherium attenuatum may be readily distinguished from the other species by its smaller size (length of skull .08-09 as contrasted with .099-112 in P. australe), and the marked convexity of the brain-case just posterior to the fronto-parietal suture, a character used by Ameghino in defining P. globo sum, a species here regarded as synonymous with P. attenuatum. A careful comparison of the skeleton, so far as known, with that of P. australe fails to show characters of specific value, apart from size. P. icochiloides was originally defined as transitional to the genus Icochilus [Inter atherium). The molars and premolars are said to be like those of Protypotherium , but the mandible is stout, with the horizontal ramus short, low in front and very high behind. The measurements accompanying the description agree closely with the dimensions of the mandible in No. 9187 Am. Museum (PI. Ill, fig. 6) which has been referred to P. attenuatum , and the two species are regarded as syn- onymous. Specimens in the Munich collection determined by Ame- ghino as P. icochiloides and Patriarchus rectus are certainly the same as P. attenuatum. The latter is represented in the Princeton and American Museum col- lections by a considerable amount of material illustrating the structure of the skull and limbs. The following localities have afforded specimens : Killik Aike (4), two miles west of Killik Aike, Rio Gallegos (1), south of Santa Cruz (1), ten miles south of Coy Inlet (1), five miles south of Coy Inlet (1), Coy Inlet (1). Nos. 15,665, 15,341, and No. 9187 American Museum, are illustrated in the accompanying plates. Sinclair: typotheria of the santa cruz beds. 45 No. No. \ 9187. 15665- Skull, maximum length ........ b CO Ln .0885 “ greatest width across arches ...... .0515 “ interorbital width ........ .027 .027 “ least width of brain case ...... .016 .0155 “ height of occiput ........ .0165 .020 “ width of occiput ........ •03 55 .037 * Palate, length from alveolus of Ii to palato-narial border, along median line ......... .049 * .0525 Palate, width at Ml ........ .0195 .0195 Mandible, depth below P^ ...... •0135 “ “ “ posterior margin of .0216 Upper dentition, length 11— Ml ...... .0466 .048 “ “ “ Pi-Mi on alveolar border . •0323 •033 “ << << MI-MI “ “ ' “ .0185 .018 U, width of alveolus ........ .0035 .0039 11, width of crown ......... .0038 “ greatest transverse diameter ...... .001 5 11, width of crown ......... .0036 .004 “ greatest transverse diameter ...... .0016 .0022 C, width of crown . . . * .0036 .004 “ greatest transverse diameter ...... .0016 .002 Pi, antero-posterior diameter just above triturating surface .003 .0027 “ transverse diameter in same plane ..... .0023 .002 PI, antero-posterior diameter just above triturating surface •0035 •0035 “ transverse diameter in same plane ..... .0026 .0027 PI, antero-posterior diameter just above triturating surface .004 .0045 “ transverse diameter in same plane ..... .003 .0035 PI, antero-posterior diameter on triturating surface externally . .004 .0044 “ transverse diameter at widest part ..... .0032 .004 Mi, antero-posterior diameter on triturating surface externally . .006 .0065 “ transverse diameter at widest part ..... .004 .0046 Ml, antero-posterior diameter on triturating surface externally . .0058 .006 “ transverse diameter at widest part ..... .0035 .004 Ml, antero-posterior diameter on triturating surface externally . .0057 .0055 “ transverse diameter at widest part ..... .003 • 0033 Lower dentition, length Iy-Mg ...... •043 “ “ Py— M^ on alveolar border . .0326 “ “ M^— M-3 on alveolar border .0183 Iy, width of crown ....... .0014 “ greatest transverse diameter ...... .0012 1^, width of crown ... .... .0019 “ greatest transverse diameter ...... .001 5 * Approximate. No. I5,3N- 46 PATAGONIAN EXPEDITIONS I PALEONTOLOGY. No. No. No. 9i87- 15,665. I5,34I- Ij, width of crown ........ .003 “ greatest transverse diameter ..... .0017 Py, antero-posterior diameter on triturating surface •0035 “ greatest transverse diameter ..... .0025 Py, antero-posterior diameter on triturating surface . y .0046 “ greatest transverse diameter . . . .0025 My, antero-posterior diameter on triturating surface . .0055 “ greatest transverse diameter ..... .0029 M^, antero-posterior diameter on triturating surface . .0056 “ greatest transverse diameter ..... .003 M-g, antero-posterior diameter on triturating surface . .0075 “ greatest transverse diameter ..... .0025 Humerus, length ........ .074 “ greatest antero-posterior diameter proximally . .0175 “ greatest transverse diameter distally . .0167 .0165 Radius, length ........ .058 .056 “ antero-posterior diameter of head •005 .0047 “ transverse diameter of head .... .0072 .0066 “ width of distal end ...... .01 1 .010 Ulna, antero-posterior diameter at upper margin of sigmoid cavity .010 .0103 U U U U U << ii i i .009 .0096 Pelvis, width of ilium at greatest expansion .0155 “ “ “ “ “ neck .0095 “ antero-posterior diameter of acetabulum .0109 “ transverse diameter of acetabulum .010 Femur, length ........ .0775 •0 79 “ greatest width at proximal end .... .0195 .019 “ “ “ “ distal end .... .0185 .017 Patella, length ........ .01 1 “ width ........ .007 Tibia, length ......... .091 .0905 Tibia, transverse diameter at proximal end .0186 .0165 “ “ “ “ distal end .... .01 1 .0095 Fibula, length ........ .086 “ antero-posterior diameter at proximal end .007 “ “ “ “ “ distal end . .0067 Metacarpal II, length ....... .024 “ III, “ .027 .026 “ IV, “...... .021 .021 “ V, “ .0122 Calcaneum, length ....... .0285 .0265 Astragalus, length ....... .0165 .0145 “ width of body ...... .007 .0062 “ “ “ head ...... .0065 .0055 Sinclair: typotheria of the santa cruz beds. 47 Metatarsal II, length “ III, “ “ IV, “ « v, “ Digit, III, first phalanx, length “ “ second “ “ “ “ ungual “ “ Digit IV, first “ “ “ “ second “ “ “ “ ungual “ “ No. No. No. 9187. 15,665. 1 5,341 . .026 .0246 •035 •037 .036 .0265 .016 .0113 .01 1 .0162 .01 1 .0105 INTERATHERIUM (Moreno) Ameghino. (Plates VI, Figs. 11, 12, 13, 16-21 ; VIII, IX; Text Figs. 1 B, 11, 12.) Interatherium Moreno ; Patagonia resto de un antiguo continente hoy submerjido, p. 23, 1882 [no men mid uni). Ameghino; Observ. gen. sobre los Toxodontes, etc., p. 63, 1887. Icochilus Ameghino; Contrib. al Conoc., etc., pp. 469-474, 1889. Tembothevium Moreno; fide Ameghino, Observ. gen. sobre los Toxo- dontes; etc., p. 65, 1887. A large amount of material from localities near the east coast of Pata- gonia is referable to this genus. All parts of the skeleton are well represented. Dentition (PI. VIII, figs. 16, 18, 20-27). — The dental formula varies slightly, owing to the loss in some individuals of the lateral incisor and occasionally the canine, and the presence in others of supernumerary teeth. The median incisors are enlarged and functional as cropping teeth. The crown continues to grow more or less during life, but in old individuals the width of the crown rapidly decreases above the alveolar border. The enamel layer, in worn specimens, is confined to the anterior surface. If any is present on the inner side of the crown, it must extend but a short distance above the cutting edge, as none is here observable in little-worn teeth. The crown of the second incisor is smaller and lower than the first with the enamel layer also external. The margins of the cutting edges of the first and second incisors are closely applied without imbrica- tion, forming a perfectly symmetrical crescent. The third incisor is still smaller. In certain individuals it is well developed in proportion to its 48 PATAGONIAN EXPEDITIONS : PAL/EONTOLOGY. size, while in others it is quite small, present on one side only or entirely absent. This character, which does not seem to be due to difference in age, has been applied to the separation of two of the species (/. robustum and /. extensuni). The second and third incisors are usually separated by a slight diastema. Both I- and the canine have laterally compressed pointed crowns. A supernumerary canine is sometimes present (PI. VIII, fig. 26) or the canine may be entirely wanting (PI. VIII, fig. 16. Age character). Between I- and the canine a diastema is always present, and usually also between the latter and P-. The first upper premolar is a simple-crowned, cylindrical, laterally compressed tooth, strongly curving inward and backward. The remaining premolars and molars may be readily distinguished from those of Protypotherium by the great elongation of the antero-external angle and the strong development of the outer ridges and the intervening groove. The second premolar is incompletely molariform, while the third and fourth resemble the molars. This is a distinct advance over Protypotherium , in which none of the pre- molars have attained the molariform condition. In the second premolar of Interatherium the crown is deeply grooved internally, producing a pair of lobes, of which the anterior is the smaller. Antero-externally a deep groove is lodged between two ridges, as in Protypotherium , but the groove is proportionately much deeper and the ridges higher. The same arrange- ment of inner lobes and outer ridges, is observable in the third and fourth premolars and in the molars, but in the latter the inner lobes are equal in size. The antero-external portion of the crown is greatly elongated. As the crown wears down, the external groove with its bounding ridges dis- appears and there remains only the elongated antero-external angle (PI. VIII, fig. 18). The triturating surfaces of the inner lobes are deeply cupped. Plate VIII, figure 20 shows the pattern of the unworn second and third premolars. The molar-pattern appears to be very similar to these, but no specimen in the collection shows an unworn tooth of this series. The crown pattern consists essentially of two internal crescents separated by a deep groove, and an ectoloph concave anteriorly and con- vex posteriorly. The antero-external groove lies between the anterior horn of the first inner crescent and the ectoloph. The crescentic lobes are cuspidate internally, but the cusps soon wear down. A cement layer 1 is present as in Protypotherium. In the inferior series, the incisors are pronate with the crowns of the Sinclair: typotheria of the santa cruz beds. 49 first and second notched by a deep internal groove. These teeth are proportionately broader than in Protypotherium and the groove does not bifurcate the summit of the crown as in the median lower incisors of that genus (Pis. V, fig. 22 ; VIII, fig. 25) but notches it as in I3 of Pvotypo- therium. The third incisor is cylindrical, with several shallow internal grooves. The canine and first premolar are simple cylindrical teeth, the former showing, in unworn specimens, a slight internal groove. The worn crowns of the second, third and fourth premolars and the first and second molars consist of a pair of triangular to oval lobes joined by a narrow isthmus like a figure 8 (PI. VIII, fig. 25). The third molar is rendered trilobate by the elongation of the posterior lobe and its constric- tion by a broad groove on the outer side. In slightly worn teeth it is seen that this lobate structure is developed from a pair of crescents having parts homologous with the elements of the crown pattern in the unworn lower molars of Protypotherium (PI. V, fig. 14 or, text fig. 6, A). A thin layer of cement is present on the lower premolars and molars. Milk Dentition. — The only difference between the deciduous premolars and their permanent successors is the presence of roots in the former (PI. VIII, figs. 22, 24). On the presence or absence of this character two genera have been established, Inter atherium with rooted premolars and Icochilus with these teeth hypsodont. These are merely the immature and adult stages of one and the same form. Interatherium has priority and must be retained as the proper designation for the genus. The order of tooth replacement is the normal one, but the milk teeth are retained in position until after the eruption of the third molar. It cannot at present be ascertained whether the first premolar has a deciduous predecessor, as it is a single-rooted tooth and has the same shape in specimens with rooted deciduous premolars as in those with hypsodont premolars. Sknll (PI. VIII, figs. 16-19). — The skull is short and broad, with heavy arches and prominent crests. The facial region is decidedly short, the orbit lying farther forward than in Pvotypotherinni. The premaxillae are heavy, with the ascending process short or wanting (PI. VIII, fig. 16). A prominent anterior nasal spine is present. On the palatal surface, the premaxillae are strongly arched antero-posteriorly and deeply excavated by the anterior palatine foramina, which extend posteriorly beyond the line of the premaxillo-maxillary suture. Along the line of this suture the rostrum is constricted vertically, producing a strong upward arching 50 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. of the alveolar border between the second incisor and the second pre- molar (PI. VIII, fig. 1 6). The facial expanse of the maxillary is broadly concave both antero-posteriorly and transversely. It is continued upward as a robust, V-shaped bar inclosed in a deep notch in the frontal, widely separating the nasal from the lachrymal. The maxillary is extensively involved in the anterior and inferior margin of the orbit and in the tem- poral arch, where it extends as far back as the posterior margin of the glenoid cavity, forming with the malar a strong buttress preventing lat- eral dislocation of the mandible. A large descending maxillary process is developed beneath the orbit and the whole lower surface of the temporal process of this bone is deeply pitted for muscular attachment. The orbits are circular, quite prominent and widely open posteriorly. They are bounded anteriorly and interiorly by the maxillary, the orbital portion of which rises above the facial tract as a vertical plate (PI. VIII, figs. 16, 17). The lachrymal is entirely orbital, a “ lachrymal ” tubercle being developed from the maxillary. Almost directly beneath this tubercle, the circular infraorbital foramen perforates the maxillary at the junction of its orbital and facial portions. Posteriorly, the orbit is bounded by the anterior extremities of the malar and squamosal. These elements have the same mutual relations as in Protypotherium , but the malar is proportionately heavier than in the latter genus. The squamosal is distended and filled with cancellae, the distention involving also the mastoid. In superior view (PI. VIII, fig. 17; text figs. 11, 12) the nasals are seen to be very broad, with square-cut tips and either straight (/. extension) or curved fronto-nasal suture (/. robustmn). They are not firmly united with the premaxillae and maxillae and are frequently crushed down into the narial chamber. A broad, V-shaped maxillary process widely separates the nasal from the lachrymal. The interorbital tract is flat, with persistent median suture. The postorbital processes are short and robust, with blunt points. Their posterior borders give rise to the temporal ridges, which are heavy and converge rapidly at an acute angle to form a prominent sagittal crest in I. robustum and /. extension , while in I. excavation they are lyrate, lower and converge much farther back on the parietal to form a short low crest. Back of the postorbital processes, the brain case attains its maximum constriction, but expands rapidly posteriorly. The lamb- doidal and sagittal crests join at a right angle, inclosing, with the superior border of the temporal process of the squamosal, a deep temporal fossa, Sinclair: typotheria of the santa cruz beds. 5i the floor of which is perforated by numerous foramina situated in both the parietal and the squamosal. The latter is dilated, as in Protypotherium , and firmly coossified with the mastoid and the auditory meatus. The back of the skull (PI. VIII, fig. 19) is circular in outline superiorly, with strong rugosities on the occipital surface for muscular attachment. The distended squamosal and mastoid are broadly exposed, but the suture between them can no longer be distinguished. The occipital elements are likewise completely fused, as in Protypotherium. As in that genus, the occipital is constricted transversely, the mastoid foramen perforating the suture between the mastoid and the occipital at the point of greatest constriction of the latter. The foramen magnum is circular in outline and the condyles obliquely directed and semicylindrical in shape. The palate is deeply concave anteriorly, but posteriorly the concavity is about the same in degree as in Protypotherium. The sweep of the tooth rows is lyrate, approaching horizontality back of the first molar. The anterior palatine foramina are large, extending posteriorly beyond the premaxillo-maxillary suture. The posterior palatine foramina emerge opposite the fourth premolar on the maxillo-palatine suture, or a short distance anterior to it. The palatines extend well back of the last molar and terminate in a pair of very large, heavy, triangular processes propor- tionately much larger than in Protypotherium and similarly supported ex- ternally. The pterygoid appears to have been small and scale-like, though it is not preserved in any of the specimens available, nor can the shape of the posterior narial border be determined. The bullae are pear-shaped and considerably flattened anteriorly, where most prolonged, as in Proty- potherium. The auditory meatus is long, tubular and directed almost horizontally. The basi-occipital is keeled inferiorly, as in Protypotherium. The arrangement of the cranial foramina is the same as in the latter genus. The mandible is short, heavy and very deep (PI. VIII, figs. 16, 25), with the rami firmly coossified, without trace of suture. The symphysis is pronate, tapering anteriorly, as in Protypotherium. Back of the symphy- sis, the depth of the mandible rapidly increases until a maximum is reached at a point vertically below the third molar. Beyond this point, the angle rapidly decreases in depth. As in Ptoty pother ium, it extends well beyond the condyle and has the inferior and posterior border inverted, inclosing a large submaxillary fossa. The coropoid process is sharp- pointed and slender, with strong posterior inclination. The coronoid 52 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. margin of the ascending ramus is sigmoid in outline, as in Protypotherium. The sigmoid notch is broad and the condyle irregularly elliptical in out- line, broader externally than internally, with its margin overhanging the neck. The groove terminating in the inferior dental canal is much longer than in Protypotherium. Two mental foramina are present, one beneath the canine and the other beneath the last premolar, or the anterior part of the first molar. Sometimes the latter foramen is doubled. The masse- teric fossa is almost circular, as in Protypotherium , but less sharply defined anteriorly and with proportionately stronger transverse ridges for muscu- lar attachment. Vertebral Column; Ribs and Sternum. — The atlas (PI. VIII, figs. 13- 15, 28) may be readily recognized by the prominence of the neural spine, the robustness of the transverse processes and the position of the canal for the vertebral artery. The latter perforates the base of the transverse process at the margin of the posterior cotylar surface, emerging on the lower surface of the process near its anterior margin. Between the point of emergence and the neuro-arterial canal, the artery lay in a groove (sometimes enclosed as a foramen, PI. VIII, fig. 13) between the trans- verse process and the anterior cotylus. The neuro-arterial canals are large and are inclosed anteriorly by robust bony bars. The transverse processes vary somewhat in shape and degree of expansion. In I. robus- tum and /. extension , there is but little basal constriction, while the free border is broadly expanded, especially antero-externally. In I. excava- tion, there is considerable basal constriction and the free border is less expanded (PI. VIII, fig. 28). The anterior margin of the neural arch supports a large and very prominent spine-like tubercle. A smaller sharp-pointed spine projects backward from the posterior margin of the inferior arch. The axis (PI. VIII, figs. 11, 12) is characterized by a large, hatchet- shaped neural spine, which varies slightly in shape in specimens referred to one and the same genus. In some, it is strongly convex dorsally, with round anterior and posterior extremities. In others, the dorsal margin is horizontal or nearly so. The transverse processes are slender, sharp- pointed and directed posteriorly. Their bases are perforated by the canal for the vertebral artery. The odontoid is short and robust, its dorsal surface lying in the same plane as the floor of the neural canal. The centrum is strongly keeled interiorly, the keel bifurcating posteriorly. The Sinclair: typotheria of the santa cruz beds. 53 deep concavities on either side of the keel are bounded externally by the inferior edges of the transverse processes. The slender spines of the remaining cervicals increase in length pos- teriorly. Anteriorly, they are directed vertically, but posteriorly they incline backward like the anterior dorsal spines. They are seldom com- pletely preserved. Hatchet-shaped transverse processes begin on the third cervical and the differentiation into diapophysis and inferior lamina is already apparent in the fourth. The diapophysis alone is present in the seventh. Interiorly, the centra of the anterior members of the series are strongly keeled, but the keel soon bifurcates, terminating posteriorly in a pair of tubercles. These tubercles are absent in the fifth, sixth and seventh cervicals and the centra are broadly keeled. The dorso-lumbar vertebral formula is definitely known to be twenty- two, of which fifteen are dorsals. The anterior dorsals have long, slender, posteriorly directed spines, which, at about the sixth dorsal, begin to decrease in length, to increase in antero-posterior diameter and to expand at the tip into a dorsally flattened, triangular area. From the ninth dor- sal onward, the spines are very broad antero-posteriorly, with the tips flat- tened and oblong in outline. The backward inclination of the neural spines changes between the tenth and eleventh dorsals. The lumbar spines are similar to those of the posterior dorsals, but in a single speci- men referred to /. extension (No. 15,041) the tips of the spines are broadly expanded dorsally, irregular in outline and strongly rugose. In another specimen (No. 9557 American Museum collection) which it has not been possible to determine specifically, the neural spines of the lumbars are bifid posteriorly and the dorsal flattened area is quite narrow. The transverse processes are broad, flat blades, curving forward. Prominent metapophyses and anapophyses are present on the posterior dorsals and lumbars, the anapophyses decreasing in length in the posterior members of the lumbar series. Strongly interlocking zygapophyses are a feature of the lumbar and posterior dorsal vertebrae. Five vertebrae form the sacrum, three of which are in contact with the ilium and two belong to the caudal series (PI. VI, fig. 11). The second and third sacrals and the caudals are firmly coossified by their neural arches, zygapophyses, transverse processes and centra, while the first sacral is free. The coossified neural spines of the sacral complex form an elongated plate, flattened dorsally. 54 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. The tail was probably long and heavy, but its exact length cannot be determined from the material available. The most complete specimen has sixteen free caudals in series (PI. IX). The proximal caudals (PI. VI, figs. 1 6, 17) have robust transverse processes, which decrease in length and increase in antero-posterior diameter posteriorly. These soon bifur- cate, forming a process at either extremity of the elongated, hour-glass- shaped centrum (PI. VI, figs. 18-20). The neural arch disappears at about the tenth caudal, a pair of processes at the anterior and posterior extremities of the centrum alone remaining. Chevrons are present (PI. IX). In proportion to the size of the animal, the ribs (PI. IX) are quite robust, the first rib exceeding in width that of Protypotherium. It is broadly expanded antero-posteriorly at the proximal and distal ends. Farther back in the series the ribs are cylindrical. The sternum is imperfectly known, as but three segments are present in the most complete specimen (No. 15,401). The first segment (PI. VI, fig. 21) is dagger-shaped, with its anterior half, corresponding to the blade of the dagger, sharply keeled inferiorly. The two mesosternal segments preserved are hour-glass-shaped and doubly keeled inferiorly, the keels diverging anteriorly and posteriorly. Appendicular Skeleton. — The scapula (Pis. VI, fig. 13; VIII, fig. 8) varies somewhat in shape with the species, but it is quite probable that a part of the difference is due to crushing. In all the specimens the cora- coid border is strongly convex, the convexity continuing unbroken over the vertebral border as far as the inferior angle, which is quite prominent. The axillary border is convex in /. robustum and concave with elevated margin in a specimen referred to I. extensum (No. 15,041), but this may be due to crushing. Some slight differences in the shape of the coracoid border exist, but these can be more readily appreciated after an examina- tion of the accompanying figures (Pis. VI, fig. 13 ; VIII, fig. 8) than from a description, however detailed. The surface of the supraspinous fossa is slightly undulatory in /. robustum and strongly convex in I. ex- tensum, but just how much of the convexity is due to crushing is hard to determine. The infraspinous fossa is concave transversely in both species, but more so in the one last mentioned. The spine is high, with narrow, flattened crest and short metacromion. The neck is short and thick. The glenoid cavity is elliptical in outline, concave in all diameters, and Sinclair: typotheria of the santa cruz beds. 55 continued forward to the extremity of the large bicipital tubercle, as in Protypotherium. The coracoid process is scarcely indicated. No trace of a clavicle has been observed. The humerus (PI. VIII, fig. 7) is of about the same length as in Pachyrukhos , but may be readily identified, among other characters, by its larger head, stouter shaft, the greater transverse expansion of the distal end and the entire absence of an internal epicondylar foramen. The shaft is strongly curved antero-posteriorly, laterally compressed proximally and transversely expanded distally. The head is large, overhanging the shaft posteriorly. The tuberosities are not prominent, not exceeding the head in elevation. The deltoid ridge is broad and strong. Distally, the supinator ridge is not very well defined, resembling in this respect Pro- typotherium. The inner epicondyle is very heavy and the inner epicon- dylar foramen entirely wanting. The distal articular surface is similar to that of Protypotherium , with the exception that the inner lip of the trochlea is shorter, not so sharp and less completely separated from the inner epicondyle. A supratrochlear foramen is wanting. The radius and ulna are short and very heavy. The radial shaft (PI. VIII, fig. 10) is strongly arched antero-posteriorly and expanded trans- versely both proximally and distally. The head is oval in outline, its proximal surface concave externally for the humeral capitellum and flattened internally for contact with the inner lip of the trochlea. The articular surface for the ulna is convex transversely, plane proximo-distally, and evidently permitted considerable freedom of motion. The neck is constricted antero-posteriorly and expanded transversely. Distally, the radius is irregularly triangular in cross section, with a large oval concave facet for the scaphoid and lunar. The ulnar articulation is concealed in anterior view by a prominent flange on the outer margin of the distal end, proportionately larger than the similar structure in Protypotherium. The styloid process is short and the groove for the extensor tendons especially conspicuous and deep. The ulna (PI. VIII, fig. 9) exhibits the same decided lateral curvature as in Protypotherium , with strongly sigmoid posterior border. The olecranon is especially heavy, with the subcutaneous portion broad and almost flat and the area of insertion of the triceps quite rugose. The olecranon and coronoid processes have the same degree of anterior exten- sion. Distally, the shaft decreases greatly in weight and is transversely 56 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. flattened. Its anterior margin terminates in a sharp flange similar to that described in Protypotherium. At the level of the radial tubercle the shaft is antero-posteriorly expanded, but rapidly decreases in width distally. The styloid process is long, with a globular head. The radial articular surface is oval in outline and plane. The manus (PI. VIII, fig. i) is tetradactyl. A pentadactyl manus, here reproduced as text figure io, A, with large opposable thumb and sepa- rate os centrale is referred by Ameghino (1891$, p. 394, fig. 96) to Inter- atherium (fcochilus) robustum. No manus of Interatherium in the Prince- ton or American Museum collections has more than four digits and B Fig. io. A, Right manus with opposable pollex and separate centrale, x f. B, Right pes with oppos- able hallux, x I (after Ameghino). Both have been referred erroneously to Interatherium rob?istum. neither in Interatherium nor in any of the Santa Cruz Typotheria does the centrale occur as a separate element. In shape and arrangement, there is the closest similarity between the carpal elements of Interatherium and Protypotherium. Some slight differences in the shape of the various articular surfaces are noticeable. The proximal surface of the scaphoid is deeply excavated antero-posteriorly in contrast with its even convexity in Protypotherium. On the lunar the facets for the unciform and cunei- form are wedge-shaped in outline and differentiated from each other, unlike Protypotherium. The cuneiform does not differ greatly from the corresponding element in ProtypotJierium. The ulnar surface is continued externally over the neck of the prominent external tubercle. The trape- zium is a small element, laterally compressed proximally, with surfaces for SINCLAIR! TYPOTHERIA OF THE SANTA CRUZ BEDS. 57 the trapezoid and the second metacarpal. Distally, it is spherical and quite rugose, without the slightest trace of a surface for the pollex. The trapezoid, magnum and unciform do not differ sufficiently from those of Protypotherium to call for separate description. The metacarpals inter- lock proximally to about the same extent and in the same manner as in the latter genus. Distally, keels are present on the palmar surfaces. The phalanges resemble those of Protypotherium. Slight terminal clefts may or may not be present in the unguals, both conditions occurring in the same specimen. The ilia (PI. VI, figs, n, 12) are long and narrow, with deeply concave gluteal fossae, prominent superior and inferior borders and inconspicuous spines. The crest of the ilium, as in Protypotherium , is inclined forward obliquely. Inferiorly, the ilium is flattened, not excavated longitudinally as in the latter genus. The neck is robust, with prominent tubercle for the origin of the rectus femoris. Prominent ilio-pectineal eminences are developed. The ischia are broadly expanded and fan-like, with scarcely perceptible ischial spine. The descending ramus and the pubis are slen- der, inclosing a large, oval obturator foramen. The acetabulum is circular and deeply cupped, with narrow cotyloid notch. The straight femoral shaft (PI. VIII, figs. 3, 4) is more compressed antero- posteriorly than in Protypotherium. The head is large and globular, and impressed by a deep pit for the round ligament. The greater trochanter is low, not exceeding the elevation of the head. The lesser trochanter is proportionately larger than in Protypotheriimi and the third trochanter quite small and inconspicuous. The distal end is flattened transversely, with prominent condyles, of which the inner projects slightly beyond the outer. The patellar trochlea is proportionately wider and shallower than in Protypotherium. The patella (PI. VIII, figs. 29, 30) seems disproportionately large. It is almond-shaped in outline, strongly convex in all dimensions anteriorly and quite rugose. Posteriorly, the surface for contact with the femoral trochlea is slightly differentiated into two broadly concave facets. The tibia and fibula (PI. VIII, figs. 5, 6) are quite firmly fused proxi- mally, but less completely so distally, where the suture remains distinct. The tibial shaft is slightly curved inwardly, the straight fibula spanning the arc. Proximally, the tibial articular surfaces are circular and slightly concave transversely, with inconspicuous spine. The shaft is triangular 58 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. in cross-section proximally, but becomes oval in section toward the distal end, where it is expanded transversely. Internally and proximally, it is broad and almost a plane surface, while externally it is concave, the prominent cnemial crest forming the dividing line between the two sur- faces. Distally, the articular surface closely resembles that of Protypo- therium. The trochlea is evenly divided by a broad median keel. The internal malleolus is robust, with a large, slightly concave surface for the internal astragalar crest. Distally, the internal malleolus terminates in a hook-shaped process directed outward and downward (PI. VIII, fig. 5). The fibula is a slender element, firmly coossified with the tibia proxi- mally and articulating with it distally by suture. The shaft is irregularly oval in cross-section, except at the ends, where it is transversely com- pressed proximally and expanded, with triangular outline, distally. The latter extremity supports a crescentic, plane surface for the outer astraga- lar crest and a second crescentic surface antero-posteriorly sigmoid in sec- tion for the calcaneum. The peroneal groove on the posterior surface of the distal end is quite conspicuous. The pes is tetradactyl, with no trace of the hallux. Ameghino (1891 b, p. 393, fig. 95, reproduced as text fig. 10, B ) refers to Interatherium (. Icochilus ) robustum a large hind foot, with strongly opposable hallux. An examination of the numerous complete hind feet in the Princeton and American Museum collections shows that this foot, like the fore foot men- tioned on page 56 does not pertain to Interatherium or to any of the Santa Cruz Typotheria. The astragalus of Interatherium closely resem- bles that of Protypotherium , differing in the slightly shallower trochlea, proportionately shorter neck and the lesser degree of median constriction in the calcaneal facet. The calcaneum also is strikingly like that of Pro- typotherium, differing in the reniform shape of the astragalar facet and its sharper differentiation from the fibular facet than in the latter genus. With the exception of minor differences in the shape and proportions of the facets, the tarsus is closely similar to that of Protypotherium and need not be described in detail. The third and fourth metatarsals are equal in length, as are also the second and fifth, but shorter than the other two. All the metatarsals interlock proximally with each other and with the tar- sus. Well developed keels are present on the distal plantar margin. The phalanges can be distinguished from those of the manus only by their greater size. They are slightly expanded transversely both proximally and SINCLAIR : TYPOTHERIA OF THE SANTA CRUZ BEDS. 59 distally. Those of the first row have the distal articulations confined to the distal and palmar surfaces, while in the second row the distal surface is in part dorsal, indicating some angularity in the position of the digits (PL IX). The terminal phalanges are laterally compressed hoofs, some- times with slight terminal clefts. Restoration. — The restored skeleton (PI. IX) shows to advantage the large head, long back, heavy tail and short limbs. A digitigrade gait is inferred for the same reasons as those already given (p. 33) in discussing the gait of Protypotherium. Interatherium robustum (Ameghino). (Pis. VI, Figs. 11, 12, 16-20; VIII, Figs. 2, 5-8, 10-16, 19, 26, 29, 30 ; IX ; Text Figs. 1 B, 1 1.) Icochilus robustus Amegh.; Revista Argentina de Hist. Nat., I, p. 393, footnote, fig. 97, but not figs. 95 and 96, 1891 ; Segundo Censo, etc., p. 150, fig. 17, 1898 (listed). Characterized by the presence of a large persistent I-, and curved fronto-nasal suture. The temporal ridges are acutely convergent and not lyrate as in I. excavatum. The following localities have afforded specimens of this species : Ten miles south of Coy Inlet (5), five miles south of Coy Inlet (1), eight miles south of Coy River (1), Coy Inlet (1), Killik Aike (1), seven miles south of Mt. Leone (1), Canon de las Vacas, thirty-five miles south of Santa Cruz (nodule layer) (1), Canon de las Vacas (2), Felton’s estancia, Rio Gallegos (1), Halliday’s estancia, Rio Galle- gos (1), Rio Gallegos (1), Monte’s Casa, seven miles south of Coy River ( 1). As in the preceding lists, the figures in parentheses indi- cate the number of specimens from each locality. Nos. 15,401, 15,100, 15,348 15,293, and Nos. 9129 and 9263 of the American Museum collection are illustrated in the accompanying plates. Measurements. No. No. No. No. 9263. 15,34.8. 15,100. I5>40i. Skull, length premaxillse to inion .082 •0735 .080 “ greatest width across arches . b Cn 00 .0505 .0585 “ interorbital width .... . .031* * Approximate. Fig. 11. Interatherium robustum , show- ing the curved fronto-nasal su- ture, x y (No. 15,300). 6o PATAGONIAN EXPEDITIONS : PALEONTOLOGY. No. No. No. No. 9263. N, 348- 15,100. I5A0I Skull, least width of brain case .... .016 .0145 .0145 “ width of occiput ...... .0445 •0455 “ height “ “ above foramen magnum . .017 .014 .018 “ width of palate at Pi .014 .014 Mandible, length including Iy .... .071 .0635 .0745 .072 “ depth below Py .017 •0135 .017 .016 “ “ “ Mf .023 .023 .024 .024 Upper dentition, length Ii-M-i .... •043 .0405 .0455 .0444 “ “ “ Pl-M- on alveolar border . .029 .0285 .028 H 1 g 1“ .015 .0144 .0144 I-, width of crown at cutting edge .... .0054 .0044 .0045 .005 J2. (< <( t< (( a u .003 .002 .002 .003 J.3 <( << <( .0018 .0015 .0016 C, antero-posterior diameter ..... P-, antero-posterior diameter just above triturating .0017 surface ........ .0035 .0035 P-, greatest transverse diameter in same plane . P-, antero-posterior diameter on triturating surface .0025 .0025 externally ........ .0036 .004 P-i, greatest transverse diameter on triturating surface PL, antero-posterior diameter on triturating surface .0032 .0038 externally ........ .0042 .0046 PL, greatest transverse diameter on triturating surface ML, antero-posterior diameter on triturating surface .0035 .004 externally ........ .0046 .005 Ml, greatest transverse diameter on triturating surface ML, antero-posterior diameter on triturating surface .0035 .004 externally ........ .0043 .0046 ML, greatest transverse diameter on triturating surface Mi, antero-posterior diameter on triturating surface .0035 .0036 externally ........ b 0 01 .0045 Mi, greatest transverse diameter on triturating surface .003 .0034 Lower dentition, Iy— My ..... •043 “ “ length Py-Mg on alveolar border .02 7 “ “ “ My-Mg “ .OI45 Iy, width ........ .0022 .002 “ greatest transverse diameter .... .0016 Iy, width ........ .002 .0025 “ greatest transverse diameter .... .0016 Iy, antero-posterior “ .... .002 “ greatest transverse “ .... .0015 C, antero-posterior “ .... .0015 “ greatest transverse “ .... .0012 Sinclair: typotheria of the santa cruz beds. 6i No. No. No. No. 9263. 15,348. 15,100. 15,401. Py, antero-posterior diameter on triturating surface .0018 “ greatest transverse diameter .... .0012 Py, antero-posterior diameter on triturating surface .003 “ greatest transverse diameter .... .002 Py, antero-posterior diameter on triturating surface .0035 “ greatest transverse diameter .... .0025 Py, antero-posterior diameter on triturating surface .0038 “ greatest transverse diameter .... .003 My, antero-posterior diameter on triturating surface . * .0045 “ greatest transverse diameter .... .003 My, antero-posterior diameter on triturating surface . .004 “ greatest transverse diameter .... .003 My, antero-posterior diameter on triturating surface . .0056 “ greatest transverse diameter .... .003 Atlas, width across transverse processes . .0305 .030 .029 “ “ “ anterior articular surfaces . .0205 .019 .018 “ “ “ posterior “ “ . . .0175 .017 “ “ of neural arch ..... .009 .008 .008 “ “ “ inferior arch ..... b 0 .0048 Axis, length including odontoid .... .017 .016 “ height of spine above floor of neural canal .016 “ length of spine ...... .018 .020 .018 “ width across condyles ..... .018 .017 Vertebral column, length, atlas-sacrum • 301 “ “ “ first-fifteenth dorsal . .1385* “ “ first sacral-second caudal . .047 “ “ length, third-eighteenth caudal . .2215 Scapula, length, coracoid process to inferior angle .048 .052 “ greatest transverse width .... .0265 .028 “ antero-posterior diameter of glenoid fossa .01 1 • “ transverse diameter of glenoid fossa .007 Humerus, length ....... .0635 .055 .059 .0605 “ antero-posterior diameter of head .017 .05 “ transverse “ “ “ .0155 “ width of distal end ..... .0185 .Ol6 .016 Radius, length ....... .0415 •043 .0415 “ antero-posterior diameter of head .0048 “ transverse “ “ “ . .0067 “ width of distal end ..... .0085 .OO95 .0087 Ulna, length ........ .0525 •055 .055 “ width at upper margin of sigmoid cavity .OIO .0095 “ “ coronoid process .... .OIO .0095 * Approximate. 62 PATAGONIAN EXPEDITIONS ! PALAEONTOLOGY. No. No. No. No. 9263. '5,348. 15,100. 15, ■ l-oi Ulna, antero-posterior diameter of distal end .0055 .0065 .006 Metacarpal II, length ...... .013 .014 “ “ greatest width distally .0046 .0046 “ III, length ...... .015 •0155 “ “ greatest width distally .0045 .0048 “ IV, length ...... .013 .013 “ “ greatest width distally .0045 .0046 “ V, length ...... .0095 .009 “ “ greatest width distally .004 .0035 Manus, digit II, first phalanx, length .0086 “ “ “ “ “ width proximally .0045 “ “ “ “ “ “ distally .0034 “ “ “ second “ length .0055 “ “ “ “ “ width proximally .0038 “ “ “ “ “ “ distally .0025 “ “ III, first phalanx, length .009 .009 “ “ “ “ “ width proximally .0045 .0048 “ “ “ “ “ “ distally . .0035 .0035 “ “ “ second “ length .0065 .0065 “ “ “ “ “ width proximally .004 .004 “ “ “ “ “ “ distally . .003 .003 “ “ “ terminal phalanx, length .008 “ “ “ “ “ width proximally. .0032 “ “ IV, first phalanx, length .009 “ “ “ “ “ width proximally .0045 “ “ “ “ “ “ distally .0035 “ “ “ second phalanx, length . .0065 “ “ “ “ “ width proximally . .004 “ “ “ “ “ “ distally .003 “ “ “ terminal “ length . .008 “ “ “ “ “ width proximally . .0033 “ “ V, first “ length . .007 “ “ “ “ “ width proximally .004 “ “ “ “ “ “ distally .0025 “ “ “ second “ length .0048 “ “ “ “ “ width proximally .003 “ “ “ “ “ “ distally . .0025 Pelvis, length ....... .070 .074* “ width at anterior margin of acetabulum .028 •035 “ “ of ilium at greatest expansion .Oil .01 1 “ “ at neck of ilium ..... .0085 “ acetabulum, antero-posterior diameter .0097 * Approximate. SINCLAIR : TYPOTHERIA OF THE SANTA CRUZ BEDS. 63 Pelvis, acetabulum, transverse diameter . Femur, length ..... “ transverse diameter proximally . “ “ “ at middle of shaft “ “ “ distally . Patella, length ..... “ width ..... Tibia, length ...... Tibia + fibula, transverse diameter proximally Tibia, transverse diameter distally . Fibula, length ..... Calcaneum, length ..... Astragalus, length ..... “ length of trochlea . “ width of trochlea . “ “ “ body “ “ “ head Metatarsal II, length .... “ “ greatest width distally “ III, length .... “ “ greatest width distally “ IV, length .... “ “ greatest width distally “ V, length .... “ “ greatest width distally Pes, digit II, first phalanx, length “ “ “ “ “ width proximally “ “ “ “ “ “ distally . III, “ “ length “ “ “ width proximally “ “ “ “ distally . “ second phalanx, length . “ “ “ width proximally . “ “ “ “ distally. “ terminal phalanx, length “ “ “ width proximally IV, first phalanx, length . “ “ “ width proximally “ “ “ “ distally . “ second phalanx, length. “ “ “ width proximally. “ “ “ “ distally “ terminal phalanx, length No. No. 9263. 15,338. .058 .085 .017 .011 .008 .005 .005 No. No. 15,100. 15,301. .0095 .059 .0155 .0075 .016 .014 .007 .007 .0675 .0645 .018 .0105 .009 .061 .0195 .01 1 5 .0085 .005 .007 .005 .0165 .0047 .021 .0056 .022 .0055 .0175 .0045 .0093 .0047 .0035 .0106 .006 .0042 .0074 .0045 •0035 .0075 •003 .01 1 .0054 .004 .007 .0045 •0033 .0075 64 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. Pes, digit IV, terminal phalanx, width proximally “ “ V, first phalanx, length “ “ “ “ “ width proximally “ “ “ “ “ “ distally . “ “ “ second phalanx, length . “ “ “ “ “ width proximally “ “ “ “ “ “ distally “ “ “ terminal phalanx, length “ “ “ “ “ width proximally No. .003 .0099 .005 .0035 .006 .004 .003 .007 .003 Interatherium extensum (Ameghino). (Pis. VI, Fig. 13 ; VIII, Figs. 3, 4, 1 7, 18.) Icochilus extensus Amegh. ; Contrib. al Conoc., etc., pp. 471-472, PI. 15, figs. 4-9, 1889; Enum. Synoptique, etc., p. 15, 1894 (listed); Segundo Censo, etc., p. 150, 1898 (listed). Icochilus hegetothevoides Amegh.; Enum. Synoptique, etc., p. 17, 1894; Segundo Censo, etc., p. 15 1, 1898 (listed). Separated from /. robustum by characters of somewhat doubtful im- portance. The third upper incisor is frequently entirely wanting, or may be present on one side and absent on the other, and is much smaller than the corresponding tooth in I. robustum , which persists even in aged indi- viduals. In the best preserved specimen referred to /. extensum the fronto-nasal suture is straight, while in I. robustum it is curved. The absence of lyrate temporal ridges separates it from I. excavatum. Represented by two specimens in the Princeton collection No. 15-041 from a locality ten miles south of Coy Inlet and No. 15,666 from Coy Inlet. Measurements. No. No. 15,04.1. 15,666. Skull, length premaxillse to inion 0815 “ greatest width across arches b Cn OC “ interorbital width 034 “ least width of brain case 0145 “ width of palate at P- . 0135 .014 “ width of palate at M-2 . .019 .0185 Upper dentition, length I-L-M-2- 0435 .0446 “ “ “ Pi-Mi .029 .0295 “ “ “ Ml-Ml 015 •0155 Sinclair: typotheria of the santa cruz beds. 65 No. No. 15,04-1. 15,666. P- width of crown at cutting edge ....... .005 .005 J2. i> “ << “ l i .003 .003 C, antero-posterior diameter ........ .0017 “ transverse “ ........ .001 Pi, antero-posterior “ just above triturating surface .0024 .0021 “ transverse “ in same plane ..... .0015 .0015 F-, antero-posterior “ jusf above triturating surface •0035 .003 “ transverse “ in same plane ..... .0025 .0025 Pi, antero-posterior “ on triturating surface externally .004 .004 “ transverse “ at widest part ..... .0035 .0032 Pi, antero-posterior “ on triturating surface externally .0043 .0045 “ transverse “ at widest part ..... .004 .004 Mi, antero-posterior “ on triturating surface externally .0048 .0052 “ transverse “ at widest part ..... .0045 .0045 Mi, antero-posterior “ on triturating surface externally .0045 .0048 “ transverse “ at widest part ..... .004 .0042 Mi, antero-posterior “ on triturating surface externally .005 .0055 “ transverse “ at widest part ..... •0035 •0035 Atlas, greatest width across transverse processes. .... .028 “ width across anterior articular surfaces ..... .017 “ “ of neural arch ........ .0073 “ “ “ inferior arch ........ .0055 Axis, length including odontoid ....... .017 Pelvis, length .081 Femur, length .063 Tibia, length .068 Astragalus, length .0115 Calcaneum, length .0195 Metacarpal II, greatest length ....... •oi35 “ III, length .0158 IV, “ 6 .013 Metatarsal II, “ dorsally ....... .017 “ III, “ .0208 “ iv, “ “ .0202 “ V, “ “ .016 Interatherium excavatum (Ameghino). (Pis. VIII, Fig. 28, Text Fig. 12.) Icochilus excavatus Amegh. ; Contrib. al Conoc., etc., pp. 472-473, PI. 15, figs. 10-13, 1889. This species may be readily recognized by the lyrate temporal ridges (text fig. 12). A photograph of a specimen in the collection of Dr. 66 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. Fig. 12. Ameghino, determined by him as I. excavatum , shows this feature plainly. The dental characters given in the original description (Ameghino, 1889, pp. 472-473) have not been found to be of specific value. The third upper incisor is about as large as in /. extension. The posterior border of the nasals is curved, as in I. robustum. Characters of specific importance are exhibited by the atlas (PI. VIII, fig. 28) in which the transverse processes are more sharply constricted proximally and less widely ex- panded distally. Represented by a crushed skull and mandible asso- ciated with the atlas and axis, the posterior portion of the vertebral column, the pelvis and portions of both hind limbs, No. 15,043, from a locality ten miles south of Coy Inlet. Several immature skulls also showing the lyrate crest are probably to be referred to this species. These are No. 9750 Am. Museum, No. 15,291 and No. 15,146 from the same locality as No. I5.043- The adult individual (No. 15,043) is considerably crushed, so that only approximate dimensions can be given. Interatherium excava- tion, showing the lyrate crest, xf (No. 15,043). Slightly diagrammatic. Measurements. Length of skull ..... Interorbital width ..... Length of mandible including incisors Atlas, width across transverse processes “ width across anterior articular surface “ “ of neural arch “ “ “ inferior arch Axis, “ across condyles “ length of odontoid .... Fourteenth dorsal, length of centrum Fifteenth “ “ “ “ . . First lumbar Second “ “ “ . Third “ Fourth “ “ “ “ . . Fifth “ “ “ “ . . No. 15,043. .080 .0305 .0665 .025 .0185 .007 .0045 .017 .004 .0105 .0105 .01 1 .0105 .0115 .0115 .012 Sinclair: typotheria of the santa cruz beds. 67 Sixth lumbar, length of centrum ..... Seventh “ “ “ “ Pelvis, length ......... “ greatest (width of ilium ...... “ width of neck of ilium . ... “ antero-posterior diameter of acetabulum . “ transverse “ “ . Femur, transverse diameter proximally .... “ “ “ across minor and third trochanters Patella, length ........ “ width No. i5,°43 ■ .01 1 5 .01 1 .069 .0093 .008 .0094 .0095 .015 •0135 .012 .0065 Fig. 13. HEGE TO THERIIDAi. HEGETOTHERIUM Ameghino. (Plates I, II ; Text Figs. I, D ; 4, A ; 13-16.) Hegetotherium Amegh. ; Enumeracion Sistematica, etc., p. 14, 1887 Selatherium Amegh.; Enumeration Synoptique, etc., p. 19, 1894. This genus is represented in the collec- tions by a single species, Hegetotherium mirabile. With the exception of the ver- tebral column and ribs, pelvis, scapula and fore foot, the skeleton is fairly well known. Dentition (PI. I, figs. 1, 3, 5-5 b). — In Hegetotherium , the median upper incisors are enlarged, cropping teeth, growing from persistent pulps, with the enamel layer confined, as in the Rodentia, to the outer side of the crown. The enamel is marked by a large number of minute, longitudinal, parallel flutings, occasionally crossed by growth lines. These teeth are implanted obliquely and converge anteriorly (text fig. 13). The remaining incisors and the canine are cylindrical, single-rooted teeth, separated from each other and from the median incisor by short diastemata. The canine may be close to the first premolar, or slightly spaced from it, and is implanted entirely in the maxillary. The true shape Hegetotherium mirabile , skull from in front, xf (No. 1 5,542). 68 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. of the crowns of these teeth cannot be ascertained, as the enamel-covered portions have been worn away, leaving merely cylinders of dentine. In all the specimens examined the teeth are well worn, but the patterns de- veloped by attrition in the molars and premolars and the outlines of the crowns are characteristic, and cannot be confused with those of any other genus. The first premolar is more or less cylindrical in shape, depending on the amount of wear to which it has been subjected. The crown is slightly curved, with the convexity of the curve directed anteriorly; the buccal surface is convex ; the lingual surface is sometimes slightly grooved. The worn area on the occlusal surface varies in different indi- viduals from a deep notch to a slight, postero-internally directed concavity. The second premolar is triangular, with the base of the triangle directed postero-internally. Externally, the anterior portion of the crown is chan- neled by a groove between two ridges (PI. I, fig. i). The third and fourth premolars are molariform and may be described together with the molars. Externally, the crowns are convex, the convexity flattening out anteriorly and posteriorly. An antero-external groove is present in each, but its depth and distinctness decrease with the age of the individual and the amount of wear to which any particular tooth has been subjected. Internally, the crowns are convex except in the third molar, which is slightly grooved on the inner side posteriorly, the depth of the groove decreasing as the tooth wears down. The antero-external angle of the molars and molariform premolars is slightly elongated. Posteriorly, these teeth are rectangular in outline except in the third molar, where the pos- tero-external angle is considerably elongated. The degree of elongation of this portion of the third molar varies somewhat in different individuals. The triturating surface of the crown of each tooth (P--M-) is traversed by two ridges of dentine trending inward from two serrate cusps in the enamel of the buccal face. These serrations interlock with the external crescents of the lower molars. The molars and molariform premolars are invested externally with a thin layer of cement. All are hypsodont. The first molar is the largest, the remaining members of the series decreasing anteriorly and posteriorly. The primary crown pattern must have been quite superficial, as the surfaces developed by attrition show no traces of enamel lakes or other indications of a deeply indented crown. From P- to M- all the teeth are implanted obliquely, the antero-external portion of each projecting beyond the postero-external angle of the tooth next Sinclair: typotheria of the santa cruz beds. 69 preceding. As in all the Santa Cruz Typotheria, the crowns of the upper molars and premolars curve inward. In the inferior series (PI. I, fig. 5), the median and second pair of in- cisors are greatly enlarged, with the enamel confined to the outer surface. These teeth are inclined forward obliquely, both shearing against the tips of the median upper incisors, and like the latter growing from persistent pulps. The third incisor and the canine are cylindrical in shape and are directed forward. The first premolar is also cylindrical, is slightly larger than the teeth preceding it and differs from them in its erect position. As in the case of the vestigial members of the superior series, the enamel- covered portions of the crowns have been entirely worn away, leaving cylinders of dentine. The third incisor, canine and first premolar are separated from each other and from the second incisor and second pre- molar by short diastemata. In some specimens in which the vestigial teeth are less worn, the third incisor is close to the second, and owing to its oblique position, overhangs the posterior border of the latter. The absence of the lower canine has been used by Ameghino to separate the genus Selatherium from Hegetotheriwn. All the specimens on which this determination is based are more or less broken in the region of the lower canine, and even admitting the correctness of this diagnosis, it is a ques- tion whether the loss of such a minute vestigial tooth would justify a generic distinction. In the opinion of the writer, Selathermm is invalid and should be united with Hegetotheriwn. As in the upper series, the premolars become increasingly molariform posteriorly. The shape of the second premolar varies with the amount of wear to which it has been sub- jected. In less worn teeth, it is triangular with the apex of the triangle directed forward. Externally, the crown is impressed with a narrow groove. Internally, it is broadly grooved. As the tooth wears down, the outer groove disappears and the angularity of the crown decreases, until finally it assumes a cordate outline in cross-section (PI. I, figs. $a, 5^). The third and fourth premolars are fully molariform and may be described with the first and second molars. Externally, the crown is divided by a deep V-shaped notch into two crescents, while internally the tooth wall is smooth and broadly convex, differing markedly in this respect from Pro- typotherium. In the fourth premolar and first and second molars the postero-internal corner is elongated. The third molar is trilobate exter- nally, the second groove being much shallower than the first. Internally, 70 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. the crown is marked by a broad shallow concavity opposite the second external groove. The crowns of the lower molars and molariform pre- molars curve outward, and, as in the superior series,, are invested with a thin layer of cement. The milk dentition is unknown. Skull (PI. I, figs. 1-4, 6; text fig. 13). — The skulls in the collection at Princeton University and the American Museum of Natural PI istory belong to individuals of approximately the same age and show but a small amount of variation in size, so that the measurements given in the description of H. mirabile may be regarded as a fair average. In side view (PI. I, fig. 1) the upper profile of the skull is broadly con- vex, the surface sloping gradually forward and rather rapidly backward from the parietal eminences. The orbit is approximately central and almost entirely inclosed posteriorly by the strong postorbital frontal process and the abruptly truncated anterior extremity of the zygomatic process of the squamosal. Unlike Protypotheriwn and Inter atherium, the maxillary is entirely excluded from taking any part in the orbital border by the malar, the free border of which is produced anteriorly as a broad plate overhanging the preorbital expanse of the maxillary (text fig. 13), bound- ing externally a deep fossa and, in side view, almost concealing the infra- orbital foramen, which lies above the anterior margin of the first molar. Anterior to the lachrymal, a small portion of the maxillary enters into the lateral surface of this plate. The lachrymal is about equally divided into orbital and extraorbital moieties, with the tear duct double, one canal open- ing within the orbit behind the rather large lachrymal tubercle and the other on the orbital margin below the tubercle. The maxillary is in contact with the frontal by a small process extending between the nasal and lach- rymal (PI. I, fig. 2). 0 This process has less posterior extension than in either Protypotherium or Inter atherium, in which it is produced well beyond the anterior margin of the orbit, while in Hegetotherium it does not extend much beyond the fronto-nasal suture. A small process of the frontal is received between it and the nasal. Anteriorly, the nasal and maxillary are in broad contact, preventing the premaxillary from touching the frontal. Posteriorly, the maxillary gives rise to a long narrow process applied to the inferior inner surface of the zygoma and terminating a short distance anterior to the glenoid cavity. The posterior border of this zygomatic process of the maxilla originates opposite the posterior margin of M-. In SINCLAIR : TYPOTHERIA OF THE SANTA CRUZ BEDS. 71 Protypotherium it arises farther forward opposite the posterior border of M-, while in Pachyrukhos its origin lies a considerable distance posterior to the last molar. The premaxillse are heavy, increasing in width superiorly, with the ascending process proportionately shorter than in Protypotherium. On the palatal surface the premaxillae almost entirely inclose the incisive foramina. The arches are heavy, their lateral surface being formed entirely by the squamosal and malar in contrast with the structure in Protypo- therium and Interatherium , in which the maxillary enters largely into the lateral surface of the arch. In Hegetotherium the maxillary enters into the inferior inner surface of the arch, sending a narrow process applied to the malar almost as far posteriorly as the glenoid cavity. The deep, narrow zygomatic process of the squamosal terminates abruptly at the orbital rim and, with the postorbital process of the frontal, al- most closes the orbit posteriorly. The malar ter- minates a short distance anterior to the glenoid cavity, while in Protypotherium it bounds the glen- oid cavity laterally. The latter presents posteriorly as well as inferiorly. It is much wider than long, concave in section from side to side and convex antero-posteriorly. The postglenoid process, which is closely fused with the spout-like, posteriorly di- rected auditory meatus, is separated from the glen- oid cavity by a deep fossa for accommodation of the non-articular portion of the mandibular condyle when the mouth is widely opened. The mastoid is dilated, inclosing a large oval chamber (PI. I, fig. 6) connected by a canal with the tympanic cavity. The sutures between the mastoid and the squamosal have been completely obliterated by fusion. This is usually also the case with the suture between the squamosal and the auditory meatus, but in No. 9223 American Museum (text figs. 14, 16) the latter is clearly apparent, as is also the mastoideo- posttympanic suture. Superiorly (PI. I, fig. 2), the nasals are seen to be broad and slightly arched, increasing in width posteriorly, and extending as far back as the anterior margin of the orbit. The relation existing between frontals and Fig. 14. m s Hegetotheriu m mirabile, occiput showing the sutures between the cranial ele- ments, x f (No. 9223, American Museum). am, auditory meatus ; c, condyle ; mp, mastoideo- posttympanic suture ; ms, mastoideo -supraoccipital suture ; pp, posttympanic- paroccipital suture. Ter- minology after Roth (1903) 72 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. nasals has been made a basis for specific distinction, Hegetotherium cunea- tum being founded on an individual in which a tongue of the frontals is produced between the nasals (text fig. 15, A), but as transitions maybe observed between this extreme and the normal in Hegetotherium mirabile (text fig. 15, B, C), it seems probable that these are merely individual Fig. 15. Hegetotherium mirabile , showing variation in the shape of the fronto-nasal suture, x f . A, No. 9156 American Museum collection; B, No. 15,341 ; C, No. 15,542. variations. The interorbital tract is quite broad, plane in some individ- uals, slightly concave in others, and is perforated by one or more supra- orbital foramina. The interfrontal suture persists. The postorbital pro- cesses are large and bilobate. Their posterior borders are confluent with the slight temporal ridges, which converge far back on the parietal, giving rise to a low sagittal crest. The brain case is proportionately wider than in Protypotherium or Inter at her ium , but not as wide proportionately as in PacJiyrukhos. Numerous foramina perforate the parietal tract. The outer margin of the temporal fossa is sharply defined by the elevated border of the zygomatic process. This crest terminates a short distance anterior to the inion, from which it is separated by the mammilated sur- face of the squamosal, inclosing a deep groove continuous anteriorly with the temporal fossa. In Protypotherium and Inter at tier ium it is produced to the inion, joining the lambdoidal crest. In some individuals the inter- parietal suture remains distinct ; in others it is entirely obliterated. The suture between parietal and supraoccipital lies a short distance anterior to the lambdoidal crest, exposing a narrow tract of the supraoccipital in superior view (PI. I, fig. 2). On the back of the skull (PI. I, fig. 4; text fig. 14) the mastoid is SINCLAIR: TYPOTHERIA OF THE SANTA CRUZ BEDS. 73 broadly exposed. The occipital elements are fused to such an extent that it is impossible to differentiate between supra- and exoccipitals. Dorsally, the occipital is broad, a wing-like expansion overlapping the mastoid dila- tation. About midway between the lambdoidal crest and the foramen magnum the width of the occipital decreases sharply. This constriction lodges the mastoid foramen, which perforates the occipito-mastoid suture. When not obliterated by fusion, the mastoideo-posttympanic suture trends outward and downward from this point (text fig. 14) for about two-thirds of its course, thence rising toward the border of the occiput at a point opposite the lower margin of the auditory meatus. Opposite the con- dyles the occipital attains its maximum expansion, but does no reach the border of the occiput, where the post-tympanic area is exposed. In- teriorly, the exoccipital portion is greatly elongated, forming the fang- shaped paroccipital process. The occiput is divided by a slight ridge extending from the lambdoidal crest to the foramen magnum. The latter is elliptical in outline, with the condyles obliquely placed. The condylar surfaces are divided into posterior and inferior moieties by a low but per- fectly distinct ridge, coinciding with the inferior border of the condylar surface as seen in posterior view (PI. I, figs. 3, 4). These surfaces are convex in all dimensions, but are comparatively flat when contrasted with the semi-cylindrical condyles of Protypothevium. The palate (PI. I, fig. 3) is moderately concave. The tooth-rows are straight posteriorly, converging rather sharply anterior to the second pre- molar and again becoming straight or nearly so. The palatine extends as far forward as the middle of the first molar. Opposite the middle of the second molar the maxillo-palatine suture is perforated by the posterior palatine foramina. The narial border is emarginated by a strong spine. Posteriorly, the palate resembles that of Protypothevium , terminating in a pair of processes with knob-like extremities, which are partly developed from the palatines and partly from the alisphenoids. Each process is sup- ported externally by a plate-like buttress (text fig. 16, as) developed from the alisphenoid and squamosal, directed outward and backward and coinciding with the line of suture between the alisphenoid and the squa- mosal, while in Protypothevium it is directed mainly outward. The ptery- goid is not preserved in any of the skulls examined, but, with the alisphenoid plate, incloses a deep fossa as in Protypothermm. The squa- mosal is peculiar in that it sends a long process upward and forward to 74 PATAGONIAN EXPEDITIONS : PALEONTOLOGY. contact with the frontal (text fig. 16), from which it is excluded in Proty- potherium by the alisphenoid and parietal, and a second process forward and downward, forming the postero-external margin of the alisphenoid plate. The bullae are large, heart-shaped, hollow structures, convex in all dimensions and tapering to a point anteriorly. The auditory meatus is long and tubular, with its orifice opening posteriorly. It is frequently fused with the squa- mosal. The basioccipital is sharply keeled for a short distance anterior to the foramen magnum. The distribution of the cranial foramina differs in several essential particulars from their arrangement in Protypotherium and Inter athe- rimn (cf. Pis. I, fig. 3 ; III, fig. 3). The carotid foramen perforates the suture between the basioccipital and the tympanic opposite the middle of the bulla and is widely separated from the foramen lacerum posterius, with which it is confluent in Protypotherium and Inter atherium. The latter foramen is situated between the basioccipital and the tympanic at the inner anterior margin of the base of the paroccipital process. On the external side of this process is a deep pit for the tip of the stylohyal. Be- tween the auditory meatus and the mastoid are three or four foramina apparently opening into lateral vascular sinuses (text fig. 16). A large postglenoid foramen perforates the suture between the squamosal and the auditory meatus at the margin of the postglenoid fossa. The foramen ovale and the foramen lacerum medium are confluent. The Eustachian canal opens on the suture between the tympanic and the basisphenoid at the elongated anterior extremity of the bulla. Within the orbit a large, elon- gated orbito-nasal vacuity, communicating with the infraorbital foramen, opens into the olfactory chamber. Several smaller foramina (text fig. 16) are probably vascular. The foramen rotundum and spheno-orbital fora- men are confluent. A large spheno-palatine foramen connects the spheno- maxillary fossa with the posterior narial cavity. Fig. 16. Hegetotherium mirabile, arch removed to show the arrangement of the cranial elements in the temporal fossa, Xy (No. 9223, American Mu- seum). am, auditory meatus , as, alisphenoid ; b, bulla ; c, condyle ; f, frontal ; in, mastoid ; o, optic foramen ; os, orbitosphenoid ; p, parietal ; pi, palatine ; jt, squamosal. The pterygoid has been broken off. Sinclair: typotheria of the santa cruz beds. 75 The mandibular rami (PI. I, figs, i, 5) are firmly coossified at the sym- physis, producing a long, spout-like depression, the slope of which is con- tinued forward by the pronate incisors. The horizontal ramus is some- what thicker than in Protypotherium and increases more gradually in depth posteriorly. The masseteric fossa extends to the angular margin, limited anteriorly and inferiorly by a strong everted flange, with hook- shaped anterior extremity. The angular portion of the mandible is deep, projecting posteriorly beyond the condyle and inferiorly below the lower border of the horizontal ramus, with sharply convexAutline. The coronoid process is imperfectly preserved in all the specimens examined. It is thin and narrow, projecting well above the condyle, from which it is sepa- rated by a narrow sigmoid notch. Anteriorly, the coronoid border is slightly convex, rising almost vertically from the alveolar margin, in strik- ing contrast with its strong forward inclination in Protypotherium. The condylar surface is a flattened oval in outline, wider externally than internally, convex in all dimensions and presenting forward. The pos- terior portion of the capitulum is not occupied by an articular surface, but by a roughened area of irregular outline fitting into the postglenoid fossa, as in Protypotherium. The neck is far less sharply differentiated than in the last named genus. Externally, on either side, there is always a large mental foramen (sometimes double) in the symphysial region, and several small foramina varying in number and position on the external surface of the horizontal ramus. Internally, the opening of the inferior dental canal is large and is rendered doubly conspicuous by a long, deep groove, lead- ing into it from above. Vertebral column (PI. II, figs. 22-27). — The vertebral column is very incompletely known. The atlas (PI. II, figs. 25-27) differs from that of . Protypotherium in the shape of the transverse processes and the arrange- ment of the vertebral foramina. The transverse processes are fan-shaped expansions, with lobate free border, and are rather sharply constricted at the base. The posterior third of the base of the process is thickened, the arterial canal entering at the margin of the posterior constriction and emerging on the lower surface of the process. Anterior to the point of emergence of the canal, the thickness of the basal portion of the process greatly decreases. In the atlas of Protypotherium this canal is not present. The neuro-arterial canal sometimes lacks the anterior bony bridge, appearing as a deep groove. Both conditions occur in the same 76 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. specimen. In Pachyrukhos , the anterior bar is apparently regularly want- ing. The inferior arch supports a strong median tubercle. The axis (PI. II, figs. 23, 24) is not sufficiently well preserved in any of the specimens examined to show the shape of the spine. Unlike Pro- typotherium , the odontoid is strongly curved upward and the transverse processes are perforated by a large arterial foramen, which is quite small in Protypotkerium . The centrum is keeled interiorly and supports a pair of large tubercles at its posterior margin. Little can be said regarding the remaining cervicals, as in all the speci- mens available the transverse processes have been broken off, and in the majority of cases the neural spines also. In the third cervical (PL II, fig. 22) the neural spine is low, broad and directed vertically. The centra of the third to the fifth cervicals inclusive are keeled interiorly, the keel becoming bifid posteriorly and terminating in a pair of tubercles. The transverse processes are perforated by a large arterial canal. A single lumbar vertebra, from which the transverse processes and the neural spine are missing (PI. I, fig. 7), is interesting in that it shows the strongly interlocking character of the zygapophysial articulations. Rather prominent metapophyses are present. Anapophyses are but slightly developed. The centrum is hour-glass-shaped, with prominent median keel. Appendicular Skeleton. — The scapula has not been preserved with sufficient completeness to warrant an attempt at extended description. Enough remains in one specimen to show the presence of a high spine and prominent metacromion. The humerus (PI. II, fig. 13) is of about the same size as the corre- sponding element in the larger individuals of Protypotkerium australe , but differs structurally in several important details. The shaft is strongly compressed laterally. The head is of about the same shape as in Pro- typotkerium , the greater tuberosity rising farther above the level of the head than in the latter genus. The lesser tuberosity is comparatively insignificant and is separated from the greater tuberosity by a wide bicip- ital groove. The area of insertion of the deltoid is broad and flat, the margins of this area converging distally to form a sharp ridge. Distally, the shaft expands transversely. The supinator ridge is somewhat better defined than in Protypotkerium and the inner epicondyle heavier. Capi- tellum and trochlea are differentiated to about the same extent as in Pro- Sinclair: typotheria of the santa cruz beds. 77 typotherium , with the inner lip exceedingly prominent. In fact the whole distal articular surface of the humerus is hardly to be distinguished from that of Protypotherium. The olecranon and coronoid fossae are separated by a very thin lamina of bone, which is sometimes perforated by a supra- trochlear foramen. An inner epicondylar foramen is always present. The head of the radius (PI. II, figs. 3, 4) is irregular in outline, much wider in transverse diameter than antero-posteriorly. The ulnar surface is almost plane transversely. The anterior margin supports a strong me- dian convexity, unlike the smooth, evenly convex, anterior margin of the radial head in Protypotherium. The outer two thirds of the head is cupped for articulation with the humeral capitellum. The inner third is concave antero-posteriorly, slightly convex transversely and slopes sharply inward. It articulates with the inner lip of the trochlea. The neck is transversely flattened, especially on the anterior side, where it is overhung by the head of the radius. The shaft is almost straight, while in Protypotherium it is strongly arched forward (cf. PI. IV, fig. 14). It is circular in cross-section proximally, but becomes triangular distally and at the extreme distal end irregularly quadrangular, owing to the development of grooves to accom- modate the extensor tendons. The carpal articular surface covers almost the entire distal end. The facets for the scaphoid and lunar are undiffer- entiated. The entire surface is deeply concave antero-posteriorly and more broadly concave transversely. The facet for the ulna is a small crescentic, antero-posteriorly concave surface on the ulnar side of the dis- tal end. The ulnar shaft (PI. II, figs. 6, 7) is much straighter than in Protypo- therium, exhibiting none of the lateral curvature characteristic of that genus. The olecranon process is proportionately about as long and wide as in Protypotherium — but is much narrower transversely. Anteriorly, it projects slightly beyond the coronoid process. The greater sigmoid cav- ity is narrower transversely than in Protypotherium. Its coronoid por- tion meets the lesser sigmoid cavity at a right angle, resembling closely the arrangement of these parts in many rodents. Both are approximately plane surfaces. The shaft is flattened laterally, lacking the strongly devel- oped interosseous ridge and the deep concavity on the inner face which are present in Protypotherium. Toward the distal end the shaft is ex- panded antero-externally, giving rise to a narrow, sharp ridge trending downward and outward. The inferior extremity of the ulna is heavier 78 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. than in Protypotherium. The radial articulation is an almost plane sur- face. The styloid process is proportionately shorter and the articular surface for the cuneiform flatter than in Protypotherium. The manus is unknown. The pelvis is known from several fragments, the best preserved speci- men (No. 15,093, PI. II, fig. 14) being an incomplete right half, retaining portions of the ilium and ischium, with the pubis broken off at the inferior margin of the acetabular cavity. Compared with the pelvis of one of the larger individuals of Protypotherium australe , the gluteal surface of the ilium is seen to be less deeply excavated and the posterior inferior spine less prominent. The inferior iliac margin is broad and flat, separated from the gluteal fossa by a sharp ridge trending forward from the origin of the rectus femoris muscle. In Protypotherium the area thus inclosed is rendered concave by a median groove. The ischial spine is large and the ilio-pectineal eminence quite prominent, while in Protypotherium it is scarcely indicated. The acetabulum is larger than in the latter genus and the cotyloid notch somewhat wider. Internally, the surface for articula- tion with the auricular process of the sacrum seems to extend farther for- ward than in Protypotherium. The femur (PI. II, figs. 8-10) differs from that of Protypotherium in proportion rather than in detail. The head is globular, with a well-marked pit for the ligamentum teres (not shown in the figures), which varies in distinctness and depth in different individuals. The great trochanter rises above the head about as far as in Protypotherium, but is separated from it by a somewhat wider notch. The digital fossa is much wider than in the last named genus. The trochanter minor and third trochanter occupy the the same relative positions as in Protypotherium , but the third trochanter is proportionately larger. Bistally, there is scarcely any difference between the two genera in the details of the condylar surfaces. The patella (PI. II, figs. 11, 12) is an elongated element of oval shape, tapering gradually to a rounded apex. The anterior surface, which is strongly convex in all diameters, is quite rugose. The posterior surface is divided by a median ridge into two approximately equal, concave articular surfaces, extending to the apical extremity, but separated from the upper edge by a narrow, roughened area. The tibia and fibula (PI. II, figs. 1, 2) are firmly coossified both proxi- mally and distally. The two condylic surfaces for articulation with the SINCLAIR I TYPOTHERIA OF THE SANTA CRUZ BEDS. 79 femur are approximately equal. Both are circular in outline. The inner is strongly concave from side to side and very slightly convex antero- posteriorly ; the outer is strongly convex from before backwards and plane transversely. The spine is inconspicuous. The shaft is irregularly tri- angular proximally, but in its central portion becomes laterally flattened, in contrast with Protypotherimn , in which this portion of the shaft is cir- cular in cross-section. A strong cnemial crest terminating in a prominent tubercle extends about one-third the length of the shaft. The shaft is strongly arched inward, the straight, slender fibula forming the chord of the arc. The shaft of this element is semi-circular in section, with the inner surface flat. Distally, both bones are firmly welded, the laterally expanded distal end of the tibia forming with the fibula a broad flat area on the anterior side of the symphysis and a deeply excavated depression on the posterior side. The internal malleolus is very long and deeply grooved postero-internally for the lodgement of tendons. The fibula is also produced distally, but does not extend so far as the internal malleolus. The surface for articulation with the astragalar trochlea (PI. II, fig. 2) is deeply sunk between them. The outer lip of the trochlea is lodged in a deep, V-shaped notch along the line of suture between tibia and fibula. The groove for the inner trochlear lip is U-shaped in transverse section and extends down the outer surface of the internal malleolus. The fibula supports three facets distally, a large, flat facet for articulation with the lateral surface of the body of the astragalus, an almost flat, trapezoidal facet in contact with the lateral process of the astragalar body, when the foot is in the position of extreme flexion, and a small, lemniscus-shaped facet for contact with the calcaneum. The groove for the peroneal tendons is well defined and the external malleolus prominent. No complete specimen of the pes is preserved, but by putting together parts of two specimens (Nos. 15,392 and 15,298) it has been possible to reconstruct an almost complete tarsus and the entire metatarsus (PI. II, fig. 19). The astragalar trochlea is wider than in Protypotherimn , but the margins are unequally developed, the external being high and sharp and the internal low and round. The former fits into the deep groove between the distal ends of the tibia and fibula. The trochlear surface is short and not produced as far backward as in Protypotherimn. The astragalar body supports two processes, as in Pachyrukhos. The external process rises from the distal outer corner of the body of the astragalus. Its dorsal sur- 8o PATAGONIAN EXPEDITIONS ! PALAEONTOLOGY. face supports a facet, confluent with the crescent-shaped facet for articula- tion with the inner side of the fibula, which is in contact with the small, trapezoidal fibular facet, already described, when the foot is in extreme flexion. The inner process rises from the plantar margin of the astraga- lar body. Internally, the inner margin of the trochlea articulates with the tibial malleolus. The neck is long and directed rather more obliquely than in Protypotherium . The head is spherical, fitting into a cup-shaped de- pression in the navicular. On the plantar surface (PI. II, fig. 20) the sus- tentacular facet is reniform in outline, but is wider distally than proximally. Transversely, it is almost a plane surface, but in proximo-distal section is broadly S-shaped. The astragalar facet is dumbbell-shaped, constricted at the middle and wider at the ends. It is concave in proximo-distal section and plane transversely. The calcaneum (PI. II, figs. 15, 19) may be readily distinguished from that of Protypotherium by the small size of the fibular facet, which is hardly differentiated from the ectal facet, and by its extensive articulation with the navicular. The ectal facet is much longer and wider than in Protypo- therium. It is convex in proximo-distal section, approximately plane transversely and presents forward and inward. The sustentaculum is wedge-shaped, decreasing in width toward the free border, while in Protypo- therium it is slightly wider at the free border than at the base. The sus- tentacular facet is circular and concave in all dimensions, but especially proximo-distally. The surface for the cuboid (PI. II, fig. 15) is concave in dorso-plantar section, more so than in Protypotherium. Internally, it is confluent with a small, almost plane facet for the navicular, which is entirely wanting in Protypotherium. The tuber has much the same shape in both genera, but the rugosities at the free end are proportionately smaller in Hegetotherium . The navicular is deeply cupped proximally for articulation with the round head of the astragalus, while distally (PI. II, fig. 16) it supports two heart-shaped facets for the ecto- and mesocuneiforms. The facet for the ectocuneiform is plane in both dimensions ; that for the mesocunei- form is concave transversely and convex in the direction at right angles to this. On the inner side a small, semicircular, almost plane facet sup- ports the extremity of the internal cuneiform. An oval, plane facet on the outer side articulates with the cuboid. Another, for contact with the calcaneum, lies between the cuboidal facet and that for the head of the Sinclair: typotheria of the santa cruz beds. 8i astragalus on the dorsal surface of the navicular. At the proximal inter- nal extremity a strong tubercle is developed, as in all the Santa Cruz Typotheria. The most striking difference by which the navicular of Hegetotherium may be differentiated from the corresponding element in Protypotherium is in the large size and characteristic shape of the meso- cuneiform facet. The cuboid is much longer than in Protypotherium , but the calcaneal facet is smaller. Distally, it supports a large concavity common to the fourth and fifth metatarsals. Proximally, there is a single oval facet for the calcaneum, convex in dorso-plantar section and slightly concave trans- versely. Internally, there are a large, oval, almost plane surface for the navicular, two small, oval, approximately plane surfaces for contact with the external cuneiform, and a small crescentic, plane facet for the third metatarsal. The plantar surface supports a heavy tubercle overhanging the peroneal groove. The inner cuneiform is an elongated, scale-like element, articulating proximally with the navicular and applied to the inner side of the meso- cuneiform and the second metatarsal. The mesocuneiform is unknown. The outer cuneiform differs greatly from that of Protypotherium in the extent of its articulation with the second metatarsal. Proximally, it sup- ports an almost plane, oval surface for the navicular ; distally, a dumb- bell-shaped facet concave in both diameters for the head of the third metatarsal ; on the outer side, two small, oval facets for the cuboid and internally, three facets, a large oval facet, deeply concave proximo-distally, for articulation dorsally with the head of the second metatarsal, a circular plantar facet, concave in proximo-distal section and plane transversely, for the plantar process of the same metatarsal, and a narrow, B-shaped surface for the mesocuneiform. The plantar surface supports a prominent, pedun- culate knob. The metatarsus is tetradactyl, the axis of the foot passing through the third digit, which is much the longest. The first digit is entirely wanting ; the second and third are large, the third exceeding the second in length ; the fourth is much smaller and the fifth greatly reduced, but supporting phalanges. Proximally, the metatarsals overlap, the amount of imbrica- tion increasing toward the inner side of the foot, until the second attains the extreme degree of overlap on the outer cuneiform just described. Proximally, the articulations between the third and fourth metatarsals 82 PATAGONIAN EXPEDITIONS : PALEONTOLOGY. interlock, as in Protypotherium (Pis. II, fig. 1 8 ; V, fig. 2). Distally, strongly developed keels are present on the plantar surfaces. No associated series of phalanges are preserved. Such representatives of the first and second rows as are known are scarcely distinguishable, except in point of size, from those of Protypotherium , except that the phalanges of digit V in Hegetotherium are extremely slender. The ter- minal phalanges (PI. II, fig. 21) maybe readily recognized by their broad, hoof-like character and terminal clefts. Hegetotherium mirabile Ameghino. (Plates I, II; Text Figs. I , D \ 4, A ; 13-16.) Hegetotherium mirabile Amegh.; Enum. Sistematica, etc., p. 14, 1887. Hegetotherium strigatum Amegh.; ibid., p. 14, 1887. Lydekker, Anales del Museo de La Plata, Palaeontologia Argentina II, article 3, p. 8, PI. I, figs. 3, 4, 1893. Hegetotherium cuneatum Amegh.; Revista Argentina de Hist. Nat., I, p. 291, 1891. Hegetotherium costatum Amegh.; ibid., p. 291, 1891. Selatherium pachymorphum Amegh.; Enum. Synoptique, etc., p. 20, 1894. Selatherium remissum Amegh.; ibid., p. 20, 1894. The preceding account of the osteology of Hegetotherium is based entirely on remains of this species, which is represented in the collections at Princeton University and the American Museum of Natural History by a large suite of specimens from localities as follows : Killik Aike (5), two miles west of Killik Aike (1), Canon de Palo (1), ten miles south of Coy Inlet (7), five miles south of Coy Inlet (3), Coy Inlet (1), seventeen miles north of Cape Fairweather (1), ten miles north of Cape Fairweather (1), Cape Fairweather (2), Felton’s estancia, Rio Gallegos (2), Halliday’s estancia, Rio Gallegos (1), seven miles south of Monte Leone (1), fifteen miles south of Monte Leone (1), forty miles south of Santa Cruz (1), ten miles north of Coy Inlet (1). In each case the figures in parentheses refer to the number of individuals * from the locality mentioned. As but one species is represented, the generic and specific characters cannot be separated. Nos. 15,542, 15,432, 15,505, 15,093, 15,298, 15,43b 15,176, 15,392, and Nos. 9223 and 9156, American Museum, are figured on Plates I and II. Sinclair: typotheria of the santa cruz beds. 83 Measurements. Skull, maximum length ...... “ greatest width across arches just anterior to glenoid fossa ....... Skull, interorbital width ...... “ least width of brain case .... “ height of occiput ...... “ width of occiput ...... Palate, length from alveolus of U to palato-narial border, along median line ..... Palate, width at Ml ...... Mandible, approximate length including Iy “ depth at Py . “ “ “ posterior margin of My “ “ below condyle ..... Upper dentition, length U-Ml .... “ “ “ Pi-Mi on alveolar border . “ “ “ M-— M-^- “ “ “ Ii, width at alveolar border ..... II, antero-posterior diameter ..... “ transverse “ . I-, antero-posterior “ “ transverse “ C, antero-posterior “ “ transverse “ P-, antero-posterior “ “ transverse “ P-, antero-posterior “ “ transverse “ P-, antero-posterior “ on triturating surface “ transverse “ “ “ “ PI, antero-pOsterior “ “ “ “ “ transverse “ “ “ “ Ml, antero-posterior “ “ “ “ transverse “ “ “ “ Ml, antero-posterior “ “ “ “ “ transverse “ “ “ “ Ml, antero-posterior “ “ “ “ “ transverse “ “ “ “ Lower dentition, length Iy-M3 . . . . “ “ “ Py-My on alveolar border . “ “ “ My-My “ Iy, width of crown at alveolar border 7 (( a u ( < t( (( ±2, * No. I5i542- .118 .070 •037 .0275 .021 .045 .0675 .0245 .104 .0165 .0255 .055 .062 .041 .0225 .008 .002 .0019 .0018 .0016 .0021 .0022 .003 .0025 .0041 .0034 .006 .004 .0065 .0042 .008 .005 .007 .0049 .0079 .0044 •0595 .0386 .023 .005 .0044 84 PATAGONIAN EXPEDITIONS *. PALAEONTOLOGY. No. 15,54-2- I3, antero-posterior diameter . .0015 “ transverse “ .0017 C, antero-posterior “ .002 “ transverse “ .002 Py, antero-posterior “ .0025 “ transverse “ .0022 Py, antero-posterior diameter on triturating surface .0035 “ greatest transverse “ a a ( ( .002 Py, antero-posterior “ t if it .006 “ greatest transverse “ a t t tt .0035 Py, antero-posterior “ u n n .0066 “ greatest transverse “ a a it .0035 My, antero-posterior “ < a t i .0072 “ greatest transverse “ a a a .0035 My, antero-posterior “ it it tt .007 “ greatest transverse “ it a a .0035 My, antero-posterior “ a a a .009 “ greatest transverse “ a a a .003 Atlas, transverse breadth .042 “ breadth between anterior articular surfaces .228 “ width of neural arch .007 “ “ “ inferior “ .006 Axis, length including odontoid .024 “ width across anterior articular surfaces .022 “ length of odontoid .0072 Humerus, length .... “ width of distal end Radius, length .... “ width of proximal end “ “ “ distal “ Ulna, length ..... “ greatest width of olecranon “ “at coronoid process Femur, length “ width of proximal end . “ “ distal “ Patella, length . • . “ width Tibia, length .... .025 .0765 .01 18 .014 .099 .0156 .0147 No. 1 5 A3 1- .112 .029 .027 No. 15,298. .0225 .012 .128 No. I5A76- .103 .026 .115 Sinclair: typotheria of the santa cruz beds. 85 No. No. No. I5A3I ■ 15,298. I5392 Tibia + fibula, greatest transverse diameter proximally. .029 .029 it U (( tl distally .0245 .026 Calcaneum, length .... .038 .038 Astragalus, length .... .0195 .020 “ greatest width of body . .0165 .016 “ width of head .008 .0085 Metatarsal II, length •0395 “ width of proximal end .0077 “ “ distal “ . .009 Metatarsal III, length .0426 “ width of proximal end .0075 “ “ distal “ .008 Metatarsal IV, length .038 width of proximal end .0065 “ “ distal “ .0065 Metatarsal V, length .0273 “ width of distal end .003 PACHYRUKHOS Ameghino. (Plates X, XI ; Text Fig. I, C.) Pachyrukhos Amegh.; Nuevos restos de mam. fos., etc., Bol. Acad. Nac. Cien. Cordoba, VIII, entr. 1, pp. 160-162, footnote, 1885. Pachyrucos Amegh.; Contrib. al Conoc., etc., pp. 422-436, 918, 1889. Pcedotherium Burmeister; Anales del Mus. Nac. de Buenos Aires, T. Ill, p. 179, 1888 [fide Ameghino). This genus includes the smallest representatives of the Typotheria in the Santa Cruz beds. Like Hegetotherium , to which it is closely related, but one recognizable species appears to be present. Three almost com- plete individuals are preserved in the American Museum collection, illus- trating all parts of the skeleton except the anterior dorsal region, and on these the following description is based, supplemented wherever possible by the less complete material in the Princeton collection. Dentition (PI. X, figs. 1, 3, 13). — As in Hegetotherium , the median upper incisors are greatly enlarged, cropping teeth, growing persistently, and supported entirely by the premaxillae. In the moderately worn speci- mens studied the enamel is confined entirely to the anterior surface of the crown. The remaining incisors, the canine and the first premolar are entirely wanting, producing a long diastema between the median incisor 86 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. and the second premolar. The remaining upper teeth are in close series. All are rootless, with long, inwardly curving crowns and in all the crown pattern has been modified by wear, producing a structure closely similar to that developed in Hegetotherium. The second premolar is triangular in cross-section, decreasing in width anteriorly. Antero-externally, there is a shallow groove bounded by ridges and on the lingual side of the crown a broad groove situated somewhat farther back than the external one. The margin of the ectoloph is serrate, two or three cusps being developed, of which the first and second mark the position of the external ridges just mentioned, while the third is not located with respect to any ridge and corresponds in position with the posterior serration on the margin of the ectoloph already described in Protypotherium. The third and fourth premolars are molariform and may be described with the molars. In these teeth the ectoloph is almost plane externally, except for the slight antero-external groove with its bounding ridges. In the molars, another slight ridge is present, terminating at the apex of the notch be- tween the two most prominent points on the margin of the ectoloph. Internally, the crown is evenly convex. The posterior margin is almost at right angles to the ectoloph except in M-, while the anterior margin meets it at an acute angle. Two cross-crests are developed as a result of wear, terminating externally in sharp serrations at the margin of the tooth-crown. The first molar is the largest, the series decreasing in size posteriorly. M- is narrower posteriorly than anteriorly, with the postero- external angle produced as in Hegetotherium , but the degree to which this is developed varies somewhat with the state of wear. The premolars and molars are arranged in an imbricating series, each overlapping externally the tooth just preceding. A layer of cement is usually present. In the inferior series (PI. X, fig. 13) the first and second incisors are pronate, growing persistently, like those of Hegetotherium , which they further resemble in having the enamel limited to the anterior surface. The third incisor, canine and first premolar are absent, producing a long diastema similar to that in the superior series. The remaining teeth are unspaced. P2 is incompletely molariform, P3 and Pt completely so. The first mentioned tooth is triangular in cross-section, tapering anteriorly. The inner wall is convex posteriorly and broadly grooved anteriorly, while externally the crown is divided by a deep groove into two crescentic lobes, of which the posterior is the larger. The third and fourth premolars Sinclair: typotheria of the santa cruz beds. 87 and the first and second molars are convex internally, while externally they are divided by a deep groove into two crescentic lobes. M3 differs from the teeth preceding in the greater antero-posterior diameter of the crown, which is externally trilobate and broadly concave postero-internally. A layer of cement is present as in the superior series. The milk dentition is unknown. Skull (PL X, figs. 1-4; text fig. 1, C). — The skull is most rabbit-like in appearance, due to the convexity of the dorsal profile, the slender ros- trum, long diastema, large prominent orbits, slender zygomatic arches and great posterior depth of the mandible. Dorsally, the upper profile slopes forward sharply from the region of the mastoid dilatation, where it meets the plane of the occiput at an acute angle. The slender rostrum is grooved longitudinally from the region of the infraorbital foramen to the margin of the premaxillary above the alveolus of the median incisor. The premaxilla is long and heavy, without ascending process. Laterally, it is broadly grooved and perforated by numerous foramina. Dorsally, it is in contact with the nasal, but not firmly united therewith. Inferior ly, the long incisive foramina extend beyond the premaxillo-maxillary suture, unlike Hegetotherium ( cf. Pis. I, fig. 3 ; X, fig. 3). The facial portion of the maxillary is also excavated longitudinally. Its thin walls are perfor- ated by a net-work of foramina, recalling this region in the rabbit. The maxillary is excluded from the orbital margin by the lachrymal and malar, as in Hegetotherium. It forms the larger part of the orbital floor and appears as a narrow zone in front of the lachrymal at the margin of the plate-like elevation formed from the maxillary, lachrymal and malar, which bounds the orbit anteriorly. A thin vertical lamina developed from the maxillary incloses a groove, at the lower extremity of which the infra- orbital foramen lies. The zygomatic process of the maxillary extends almost to the glenoid cavity, its posterior border originating some distance back of the last molar. The facial expansion of the lachrymal is large, but, owing to the number of foramina perforating it, this portion of the bone is usually more or less broken. A prominent tubercle is developed on its orbital margin, concealing in external view the lachrymal duct. Contact between frontal and lachrymal is reduced by a narrow bar of the maxillary. The orbits are almost circular, very large and prominent and but slightly constricted posteriorly by the slender styliform postorbital processes. The zygomatic arches are slight, resembling those of Hegetothevium in the 88 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY . arrangement of the several elements. The delicate zygomatic process of the squamosal is not dilated, as in Hegetotherium , but the inflation of the mastoid is proportionately much greater than in any of the other genera, lodging a large spherical cavity communicating by a canal with the tympanic chamber (PL X, fig. 4). In superior view (PI. X, fig. 2) the long, blunt-pointed nasals are seen to increase greatly in width posteriorly, receiving a broad tongue of the frontals between them. They are separated from contact with the lachrymal by a long, narrow bar of the maxillary. The nasals are convex in cross-section anteriorly, but flatten out posteriorly. The interorbital tract is approximately plane transversely, the interfrontal suture persist- ing. Unlike Hegetotherium , the postorbital processes are slender and spine-like. No marked constriction of the brain-case succeeds them. The temporal ridges are exceedingly slight, appearing as faint lyrate crests, originating just back of the postorbital processes and converging posteriorly, but not uniting. The most prominent feature of the back of the skull is the greatly dis- tended mastoid bulla, which projects posteriorly beyond the condyles. These bullae show but slight contact with the squamosal anteriorly and are separated from the parietal by a groove opening into the cerebral chamber. Internally, they are hollow and, as in Hegetotherium , are con- nected by a canal with the tympanic cavity (PI. X, fig. 4). The sutures between the various occipital elements can no longer be distinguished. Dorsally, the supraoccipital appears on the upper surface of the skull between the mastoid bullae. Posteriorly, there is a strong median con- vexity for lodgment of the vermis of the cerebellum, bounded on either side by a broad longitudinal concavity. The degree of transverse con- striction of the occipital in the region of the mastoid foramen is propor- tionately much less than in Hegetotherium. The foramen magnum and occipital condyles are similar to those of the last named genus. The paroccipital processes are quite long, increasing greatly in antero-posterior diameter a short distance below their bases, where they form a thin plate, from the posterior margin of which a narrow, transversely compressed process is continued downward, gradually curving forward toward the tip (PI. X, figs. 3, 4). The palatal surface of the skull recalls that of Hegetotherium. The tooth-rows are slightly arched, inclosing an area deeply concave both SINCLAIR: TYPOTHERIA OF THE SANTA CRUZ BEDS. 89 antero-posteriorly and transversely, but decreasing in concavity anteriorly, so that the palatal floor of the rostrum is practically flat. The anterior palatine foramina, as already described, are proportionately longer than in Hegetotherium , extending posteriorly beyond the premaxillo-maxillary suture. The posterior palatine foramina are in advance of the maxillo- palatine suture, perforating the maxillary opposite the anterior part of M-. The posterior narial border is divided by a strong median spine, as in Hegetotherium. In structure, the pterygoid region is practically the same in both genera, the only difference of importance appearing in the absence of the squamosal process, which in Hegetotherium is applied externally to the external alisphenoid plate (text fig. 1 6, page 74). The bullae are heart shaped, pointed anteriorly, but considerably flatter inferiorly than in Hegetotherium. The meatus is long, tubular and directed posteriorly, opening on a level with the postglenoid fossa. The distribution of the cranial foramina is similar to that in Hegetothe- rium. The infraorbital foramen opens into a deep fissure within the orbit connected by various vacuities with the olfactory chamber. Owing to the rupture of the thin interorbital septum in all the specimens in which this region is exposed, it is impossible to locate the position of the foramina opening at this point, but it is probable that a confluent foramen rotundum and spheno-orbital foramen and a large spheno-palatine foramen would be found to be present, as in Hegetotherium. An irregular opening at the anterior margin of the bulla marks the position of the foramen ovale and lacerum medius. The internal carotid enters the skull between the basi- occipital and the bulla near the suture between the basioccipital and the basisphenoid. The condylar foramina are rather large and are situated close to the condyle, while farther forward several additional foramina on either side perforate the basioccipital. The foramen lacerum posterius is inconspicuous, lying back of the bulla near the base of the paroccipital process. The presence of a postglenoid foramen cannot be definitely ascertained, owing to fracture in all specimens in which this region is exposed. ' Above the fossa for the tip of the stylohyal a pair of foramina, probably vascular, emerge at the edge of the auditory meatus. On the occipital surface the mastoid foramen occupies its usual position at the point of greatest transverse constriction of the occipital plate. The mandible (PI. X, figs. 1, 13) differs from that of any other genus of Santa Cruz Typotheria in transmitting an external branch of the inferior 90 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. dental canal, which opens into the masseteric fossa close to the anterior border of the latter (PL X, fig. i). The symphysis is firmly coossified and spout-like. The horizontal ramus increases rapidly in depth pos- teriorly, the coronoid border is strongly inclined backward, and the angle broadly rounded. The coronoid process is small, sharp-pointed and quite delicate and is accordingly broken in the majority of specimens. The condyle is approximately circular in outline and almost flat, present- ing upward and forward. The inferior margin of the angle has an out- wardly directed, sharp ridge as in Hegetotherium. Vertebral Column and Ribs. — The atlas (PI. X, figs. 24-26) is peculiar in having the neuro-arterial foramina reduced to notches and in the irregu- larly lobate outline of the free borders of the transverse processes. Each arch supports a single, prominent, median tubercle. Owing to the broken condition of the superior arch and the absence of good contacts, it is not possible to be certain whether the tubercle is at the anterior or posterior border, but, judging from the structure of this region in Protypotherium and Hegetotherium , it was probably anterior. The posterior third of the base of the transverse process is perforated by the vertebral artery, as in the last named genus. The axis is incompletely preserved in one specimen (No. 15,744). As in Hegetotherium , the odontoid curves upward sharply. The neural spine is almost complete, differing considerably in shape from that of Protypo- therium (cf. Pis. V, fig. 1 6 ; X, fig. 23). Anteriorly and posteriorly, it has about the same degree of extension and is evenly convex above. Its pos- terior margin is grooved below to receive the tip of the neural spine of the third cervical, when the neck is flexed upward. This spine is short, broad and directed vertically. The spines of the fourth, fifth and sixth cervicals are delicate, sharp-pointed, and incline forward. The transverse processes are heavy, with prominent anterior and posterior tubercles. The differentiation of the diapophysis from the inferior lamella is first observable in the fifth cervical. The centra are keeled inferiorly. The number of dorsal vertebrae is not positively known, but probably is not more than fifteen, as in Inter at herium. Eight lumbars are present (seven in Interatherium ) and five vertebrae are coossified in the sacral complex, three true sacrals and two caudals. The anterior dorsals are either crushed or concealed by other bones which cannot be removed without injury to the specimens. The anticlinal vertebra appears to be the Sinclair: typotheria of the santa cruz beds. 9i thirteenth dorsal, and has a broad, vertically directed spine. It is followed by two dorsals, in which the spines are similar in shape to those of the anterior lumbars and are directed forward. The spines of the posterior lumbars are very heavy, strongly inclined forward and deeply grooved posteriorly, with a prominent tubercle developed on either side of the groove about half way between the posterior zygapophyses and the tip of the spine (PI. X, fig. 31). The transverse processes are broad, flat blades, with truncated extremities, curving outward and forward. Prominent anapophyses are developed on the posterior dorsals and anterior lumbars, decreasing in size posteriorly. The zygapophyses in the posterior dorsals and in the lumbar series interlock as strongly as in Hegetotherium. Five vertebrae are coossified in the sacrum, of which three are in con- tact with the ilium and two are referable to the caudal series. In all the specimens (Nos. 9242, 9283, 9481 of the American Museum collection) the neural spines of the anterior sacrals have been broken off. Those of the posterior sacral and the coossified caudals are very heavy, with flat summits (PI. X, fig. 18). Little can be said regarding the tail, as but four free caudals close to the sacrum are preserved in No. 9242. These are elongated, with little indication of transverse processes, and suggest a short tail. Appendicular Skeleton. — The scapula is so imperfectly preserved in all the specimens available that its shape cannot be determined. In the restored skeleton of Pachyrukhos (PI. XI) it has been given approximately the same shape as in Protypothevimn. The spine is high and the meta- cromion robust. The clavicle is a slender rib-like element compressed laterally and gently arched longitudinally (PI. X, fig. 5). At the distal end it expands, but has been so damaged by fracture that its exact shape cannot be ascertained. The humerus (PI. X, fig. 1 1 ) is almost a miniature of that of Hegeto- therium. The head is large, overhanging the shaft posteriorly. The greater tuberosity is prominent, rising above the level of the head, as in Hegetotherium , while the lesser tuberosity is inconspicuous. Distally, the correspondence with Hegetotherium is very close. There is the same well- defined separation of the articular surface into trochlea and capitellum, the same elongation of the inner lip of the trochlea, strong development of the inner epicondyle and short, but sharply defined, supinator ridge. An internal epicondylar foramen is present and in all the specimens 92 PATAGONIAN EXPEDITIONS : PALEONTOLOGY. available the septum separating the olecranon and coronoid fossae is perforated. So far as it is possible to make comparison, the radius and ulna (PI. X, figs. 9, io, 12) closely resemble the corresponding elements in Megeto- therium, and do not call for separate description. The only difference of importance appears in the shape of the styloid process of the ulna, which is slightly inclined internally, while in Megetotherium it is straighten The mutual relationships of the carpal and metacarpal elements is much the same as in Protypotherium (cf. Pis. V, fig. 3 ; X, fig. 14) except that metacarpal II apparently articulates with the scaphoid. But two specimens of the carpus are available for study (Nos.- 9551 and 9481, American Mu- seum) and in both of these the proximal end of the second metacarpal projects some distance above the magnum, excluding the latter from con- tact with the trapezoid. Owing to the partial relief employed in mount- ing the only specimen with this portion of the carpus complete (No. 9481 of the American Museum collection), it is not possible to verify this obser- vation by an examination of the articular surfaces of the carpal elements. Ameghino (1889, PI. 13, fig. 14) figures a slender pollex in Pactiyruk- hos typicus from the Monte Hermoso beds. No trace of a pollex is pre- served in the otherwise almost complete fore foot of No. 9481 of the American Museum collection, but this cannot be regarded as conclusive, as the trapezium is also wanting. The metacarpals decrease in strength externally, as in Protypotherium. The third is the longest. Well devel- oped keels are present on the distal palmar surfaces. The proximal phalanges are slightly arched and taper towards the distal end. The proximal articular surface is almost at a right angle to the shaft, is cres- centic in outline and slightly concave. Distally, the articulation is c to the distal and palmar surface, while the distal articulation of the second phalanx is about equally developed on both dorsal and plantar surfaces. Distally, the terminal phalanges are flattened transversely and hoof-like, without trace of terminal clefts, differing in this respect from Pachyrukhos typicus. The pelvis (PI. X, figs. 17, 18) is well preserved in one of the specimens in the American Museum collection (No. 9242). The gluteal surface of the ilium is slightly concave in all dimensions, in contrast with its deep concavity in Protypotherium and Inter at herium. Anteriorly, it is obliquely truncated from above forward, with prominent anterior and posterior Sinclair: typotheria of the santa cruz beds. 93 superior spines. The posterior inferior spine, and the ischial spine and tuberosity are also prominent. The symphysial portion of the pubis and the ischial ramus are broad, and the superior ramus of the pubis slender. The patella (PI. X, figs. 19, 20) is an almond-shaped element, convex in all dimensions anteriorly, while posteriorly it supports two equal concave facets for the femoral condyles. The femur (PI. X, fig. 8) is stout, with the shaft slightly arched anteri- orly, prominent minor, major and third trochanters, and heavy condyles. Except for the straighter shaft in Hegetotherium and the greater size, there is little or no difference between the femora of the two genera, the descrip- tion of that of Hegetotherium already given applying equally well to Pachyrukhos. Compared with the length of the femur, the tibia (PI. X, fig. 7) is pro- portionately much longer in Pachyrukhos than in Hegetotherium. This fact, correlated with the length and strength of the inner digits of the pes, the comparatively short fore limb and other structural peculiarities pres- ently to be mentioned, demonstrates in the writer’s opinion that Pachy- rukhos was a jumping animal. Tibia and fibula are firmly fused, both proximally and distally, the fusion at the distal end extending up the shaft for a distance somewhat greater than one third its total length. The tibial shaft is laterally compressed, with short, but sharp, cnemial crest. Inter- nally, the proximal portion of the shaft is convex and externally concave, and is slightly curved inward longitudinally, the nearly straight fibula forming the chord of the arc. The distal third of the shaft, formed by the fusion of tibia and fibula, is elliptical in cross-section. The proximal articular surfaces resemble closely the corresponding region in Hegeto- therium. They are approximately circular in outline, the inner trans- versely concave and antero-posteriorly slightly convex, and the outer convex in both directions. The spine is inconspicuous. The distal end closely parallels in structure the fused tibia and fibula of the rabbit. Both external and internal malleoli are long, the latter exceeding the former in this respect. Both are grooved posteriorly for transmission of tendons, the peroneal grooves being especially deep. The surface for the astragalar trochlea is similar in every respect to that in Hegetotherium and does not call for for separate description (see p. 79, and Pis. II, fig. 2 ; X, fig. 16). The American Museum is especially fortunate in possessing two skele- tons with the hind limbs complete and the elements of the pes in their 94 PATAGONIAN EXPEDITIONS ! PALAEONTOLOGY. natural positions. Here also, correspondence with Hegetotherium is ex- ceedingly close. The astragalar trochlea (PI. X, fig. 15) is wide, shallow and bilaterally symmetrical, with the outer crest sharper and higher than the inner. The trochlear surface is confined to the dorsal aspect of the bone, but is not produced as far backward as in Protypotherium. The shallowness of the trochlea and its small posterior extension suggest a plantigrade position for the foot as indicated in the restoration (PI. XI). The proximal and distal corners of the astragalar body on the inner and outer sides respectively are prolonged as in Hegetotherium. The fibular facet is a vertical crescentic plane surface ; that for the internal malleolus a convex surface sloping inward. The neck is long and obliquely directed toward the inner side of the foot, and the head hemispherical. The cal- caneum is almost a replica on a smaller scale of that of Hegetotherium , the shape and arrangement of the facets being practically the same in each (see description p. 80 and Pis. II, fig. 15, X, fig. 27). Slight differences appear in the top-shaped rather than circular sustentacular facet, the more oblique position of the navicular facet and the slightly shorter and more convex astragalar facet in Pachyrukhos as compared with Hegetotherium. The tarsals have not been disturbed from their original position in No. 9481 of the American Museum collection (PI. X, fig. 15) and, so far as it is possible from the type of mounting employed, to compare them, they are exactly the same in shape and position as in the last named genus (pages 80, 81 and Pis. II, fig. 19, X, fig. 15). No trace of a hallux remains. The inner cuneiform, although not present, seems, from its articular sur- faces, to have been scale-like as in Hegetotherium. The metatarsals inter- lock strongly proximally, the second articulating with the middle of the outer cuneiform and the third with the cuboid. The second and third are the most robust, the latter exceeding the former in length. The fourth is more slender than the second and slightly shorter, and the fifth quite slender and shorter than the fourth. Well developed keels are present on the plantar surfaces. The phalanges closely resemble those of the fore foot, differing only in size and greater robustness. Their natural position is probably more strongly angulate than is represented in the restoration (PI. XI). The terminal phalanges (PI. X, fig. 15) are trans- versely expanded distally as minute hoofs without trace of median cleft. Restoration (PI. XI). — The restored skeleton shows effectively the much greater length and strength of the hind limb as compared with the SINCLAIR i TYPOTHERIA OF THE SANTA CRUZ BEDS. 95 fore limb. If the pes were placed in a digitigrade position the spinal column would incline obliquely upward. The approximate correctness of the position of the hind feet in the restoration receives additional con- firmation from specimen No. 9481 of the American Museum collection, which shows the death pose of an animal smothered by a fall of the vol- canic ash of which the Santa Cruz formation is largely composed. The back is strongly arched, more so than is shown in the restoration, and the long hind feet project forward between the fore feet, and are strongly flexed, more so than seems probable in a digitigrade form. The evidence furnished by the astragalar trochlea in favor of a plantigrade position for the hind foot has already been mentioned (p. 94). The great length and strength of the hind limb, the shortness of the fore limb and the planti- grade pes are regarded as indicative of a saltatorial gait. The fore foot may have been as digitigrade as in the rabbit. The shape of the scapula is largely hypothetical. It has been given the same shape as in Proty- potherium. The tail may have been longer, but the small size of the proximal caudals does not seem to justify such an inference. Pachyrukhos moyani Ameghino. (Plates X, XI, Text Fig. 1, C.) Pachyrukhos moyani Amegh.; Nuevos restos de mamif. fos. oligocenos, recogidos por el Prof. Pedro Scalibrini, etc., p. 160, 1885. Pachyvucos moyani Amegh.; Contrib. al Conoc., etc., pp. 430-431, PI. 13, figs. 28, 29, 34, 35, 1889. Pachyvucos ncevius Amegh.; Contrib. al Conoc., etc., p. 430, PI. 13, fig. 30, 1889. Pachyvucos absis Amegh.; Contrib. al Conoc., etc., pp. 429-430, PI. 13, figs. 32-33, 1889. But one species of Pachyrukhos is represented in the collections studied. It may be distinguished from Pachyrukhos typicus of the Monte Hermoso horizon by the different shape of the postorbital process which is exceed- ingly slender in the Santa Cruz form and large, triangular and perforated by a foramen in the later species, by the shallower lower jaw, less heavy rostrum and the possession of pointed, uncleft, terminal phalanges. Doubtless other characters of specific importance would appear if it were possible to make comparison directly with the type. A specimen in the 96 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. Munich collection labelled by Ameghino P. ncevius has the mandible less thick transversely and the teeth somewhat smaller than in the typical P. moycini. Possibly P. nceuius is to be regarded as a distinct species but for the present it may be included with P. moycini. The following Santa Cruz localities have afforded specimens of the latter species, the figures referring in each case to the number of indi- viduals : Halliday’s estancia, Rio Gallegos (2), Felton’s estancia, Rio Gallegos (10), Rio Gallegos (1), Killik Aike, Rio Gallegos (6), four miles east of Killik Aike (1). Nos. 15,743, 15,744 and Nos. 9283, 9481, 9242 and 9219 of the American Museum collection are figured on the accompanying plates. Measurements. No. No. No. *5,743- 9283. 9481. Skull, maximum length ..... .0825 .0765 “ greatest width across arches .0485 •0455 “ interorbital width ..... .021 .0195 “ height of occiput ..... •0135 “ width “ .030 .0265 Palate, length along median line from alveolus of 11 to palato-narial border ..... .049 Palate, width at Ml ...... .0165 Mandible, length including median incisors .0665 .061 .0666 “ depth at Po- . .0105 .0105 .010 “ “ “ Mg .015 •0135 •0135 Upper dentition, length U-M 3- .... .0425 •039 “ “ “ P-l-M-1 on alveolar border . .0244 .0227 “ “ “ M-l-M-1 “ “ “ .014 .0126 P-, width at alveolar border .... .0065 P-1, antero -posterior diameter .... .0036 “ transverse “ .... .0022 PI, antero-posterior “ .... .004 “ transverse “ .... .0025 PI, antero-posterior “ .... .0042 “ transverse “ .... .0028 Ml, antero-posterior “ .... M i-O O q “ transverse “ .... .003 Ml, antero-posterior “ .... .0045 “ transverse “ .... .003 Ml, antero-posterior “ .... .0045 “ transverse “ .... .0025 SINCLAIR : TYPOTHERIA OF THE SANTA CRUZ BEDS. 97 No. No. No. 15,743- 9283. 9481. Lower dentition, length Iy-Mg •0395 “ “ “ P j— Mg- on alveolar border . •0245 “ “ “ mt-m¥ “ U .0145 Ij, width of crown at alveolar border .0044 T u a a (( u a A2 .0023 Pg-, antero-posterior diameter .003 “ transverse “ . . .0015 P3-, antero-posterior “ .0035 “ transverse “ . .0022 P^, antero-posterior “ .0038 “ transverse “ . .0022 My, antero-posterior “ . .0045 “ transverse “ . .0025 M^-, antero-posterior “ .0045 “ ti'ansverse “ . .0025 M-j, antero-posterior “ .0055 “ transverse “ . . .002 Humerus, length ..... •055 .056 “ antero-posterior diameter of proximal end. .01 18 “ transverse diameter of distal end .012 Radius, length ..... •053 .050 width of proximal end .005 “ “ “ distal “ .0075 .0075 Ulna, length ..... .059 “ greatest width of olecranon . .0066 “ “ “ at coronoid process .0065 Manus, metacarpal II, length .020 “ “ “ width proximally . .004 “ “ III, length .0225 “ “ “ width proximally . .0032 “ “ “ “ distally .0035 “ “ IV, length .0185 “ “ “ width proximally . .0034 “ “ V, length .0145 “ “ “ width proximally . .002 Proximal phalanx, digit II, length .0104 i< << << jjj n .010 Second “ “ “ “ .006 Proximal phalanx, digit IV, length • .0095 Second “ “ “ “ •0055 Terminal “ “ “ “ •0055 Proximal “ “ V, “ .0085 Second “ “ “ “ .0042 98 PATAGONIAN EXPEDITIONS : PAL/EONTOLOGY. No. No. No. No. 1 5,743- 9283 ■ 94Sl- 9242 Pelvis, length ...... .0665 “ width at anterior margin of acetabulum .032 “ of ilium at greatest expansion .013 “ “ at neck of ilium .0062 .007 Femur, length ...... .070 “ transverse diameter proximally . .015 “ “ “ through minor and third trochanter ...... .014 Femur transverse diameter distally .0145 Patella, length ...... .01 1 “ width .0055 Tibia -(- fibula, length ..... CO O .090 transverse diameter proximally .016 •0155 “ “ “ distally .012 .0118 Calcaneum, length ..... .021 .022 Astragalus, length ..... .0105 .0105 “ width of body .... .008 “ “ “ head .... .0045 Metatarsal II, length ..... .030 .0265 .029 “ “ width proximally .0042 .004 “ “ “ distally .... .005 .0048 Metatarsal III, length ..... •033 .030 .031 “ “ width proximally . .0035 .003 .004 “ “ “ distally .0045 .0045 “ IV, length ..... .0266 .029 “ “ width proximally . .003 b 0 Ul “ V, length ..... .0214 .023 “ “ width proximally .004 Proximal phalanx, digit II, length •0135 .0142 Second “ “ “ “ . . .009 Proximal “ “ III, “ •0133 Second “ “ “ “ . . .0095 Terminal “ “ “ “ . . .0072 Second “ “ “ “ . . .0075 Terminal “ “ “ “ . . .0055 Proximal Second V, .0105 .005 TYPOTHERIA INCERTPE SEDIS. A large number of Typotheria have been described by Dr. Ameghino from the Santa Cruz Beds, which it has not been possible to recognize in the collections studied. It does not follow that these are necessarily in- SINCLAIR ! TYPOTHERIA OF THE SANTA CRUZ BEDS. •99 valid species, although some may be individual variants of species de- scribed in the present memoir. The collections studied, although exten- sive, are not sufficient to settle this point. It therefore seems best to list these forms separately, quoting for each the full bibliography and a trans- lation of the most complete description, with some of the characteristic measurements given by the author. PROT YPOT H ERIUM. Protypotherium diversidens Ameghino. Protypotherium diversidens Amegh. ; Revista de Hist. Nat., I, p. 292, 1891 ; Enum. Syn., etc., p. 13, 1894 (listed) ; Segundo Censo, etc., p. 150, 1898 (listed). Of relatively small size. Is easily distinguished by the upper premo- lars which have an external vertical groove, broad on the crown but nar- rowing and suddenly disappearing toward the base, instead of running the whole length of the tooth as in the other species. Length P--M- .024 C91, 292). Protypotherium diastematum (Ameghino). Patriarchus diastemcitus Amegh. ; Revista Argentina de Hist. Nat., I, p. 293, 1898, Enum. Syn., etc., p. 14, 1894 (listed) ; Segundo Censo, etc., p. 150, 1898 (listed). Of the same size as Patriarchus rectus (synon. Protypotherium prceru- tilum ) or slightly larger, from which it is distinguished by the first upper premolar which is small, cylindrical and separated from the second pre- molar by a short diastema. The upper dental series is a little more arched. Length P--M- .029 ; breadth of palate at M- .018 ; at Mx .0205 ; at PA013 (’91, 293). Protypotherium claudum Ameghino. Protypotherium claudum Amegh.; Contribucion al Conocimiento de los Mamiferos Fosiles de la Republica Argentina, p. 480, 1889; Enum. Syn., p. 13, 1894 (listed); Segundo Censo, etc., p. 150, 1898 (listed). This species, of quite small size, intermediate between P. prcerutilum and P. attenuatum , is removed from all the others by some characters of importance. I (i. e., Ameghino) know of it a fragment of the right ramus of the mandible with the two last premolars and the first two true molars. IOO PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. P3 is entirely different from the shape seen in all the other species. The inner side is almost plane, presenting a notable analogy to that of P. obstruction* \ however it displays three faintly marked vertical crests which bound two broad but entirely superficial depressions. The outer side is distinguished by a broad and shallow perpendicular depression which does not form a reentering fold, a depression which replaces the deep and narrow groove that the other species have in the hinder part of the outer side, forming a sharp entering fold. The perpendicular pos- terior face, instead of being narrow and rounded as in the other species, is of the same width as the rest of the tooth and grooved perpendicularly in the middle. The grinding surface is of regularly elongate or elliptical shape, somewhat flattened on the inner side but not divided into unequal lobes as in the other species. It measures .0035 by .0025 PT has an internal vertical groove and another external, opposite, as in almost all of the other species, but the two lobes are of nearly equal size, the anterior slightly the larger. It measures .004 by .0026. The first and second molars are bilobate by opposite vertical grooves as in the other species, with the anterior lobe considerably smaller than the posterior. On the inner side the former is convex, ending behind in a broad and rounded column, but the posterior lobe is depressed and somewhat excavated in its anterior part. Each molar measures .0065 by .003. Length P3-M2 inclusive .021. Two mental foramina are present, a larger one under P¥ and a smaller one beneath PT. Depth of hori- zontal ramus at MT .016 (’89, 480). INTERATHERIUM. Interatherium trilineatum (Ameghino). Icochilus trilineatus Amegh. ; Enum. Synoptique, etc., p. 16, 1894 ; Se- gundo Censo, etc., p. 15 1, 1898 (listed). This species is of the same size as /. externum, from which it is easily distinguished, as well as from the other species, by the form of the upper molars and premolars. Each of these teeth has on the external face a broad and deep groove which divides it into two lobes and on each lobe there is a narrow and deep groove which divides it into two vertical columns. On the external face there are thus four columns separated by * A Monte Hermoso species. Sinclair: typotheria of the santa cruz beds. ioi three grooves of which the median one is much broader and the other two very narrow (’94, 16). Interatherium anguliferum Ameghino. Interatkerium anguliferum Amegh. ; Enum. Synoptique, etc., p. 18, 1894 ; Segundo Censo, etc., p. 151, 1898 (listed). Size intermediate between /. rodens and /. supermini from which it is distinguished by the absence of the first upper premolar. The upper canine is well developed and isolated by diastemata in front and behind. P--A have the two perpendicular columns of the antero-external angle very strong. Length from the anterior border of the canine to the pos- terior border of the first upper true molar .017 (’94, 18). Interatherium anomalum (Ameghino). Icochilus anomalus Amegh.; Enum. Synoptique, etc., p. 16, 1894; Se- gundo Censo, etc., p. 15 1, 1898 (listed). Of the same size as Icochilus extensus. This species is easily dis- tinguished by the atrophy and disappearance of many teeth. The ex- ternal lower incisor is very small. The lower canine is extremely small and isolated before and behind by diastemata of considerable length, while in almost all the other species of this genus this tooth is, on the contrary, well developed and recumbent on the incisors of which it has the form. The first lower premolar is well developed, with two opposite per- pendicular grooves, one internal and the other external (’94, 16). Interatherium truncum (Ameghino). Icochilus truncus Amegh. ; Enum. Synoptique, etc., p. 16, 1894; Segundo Censo, etc., p. 151, 1898 (listed). Almost of the same size as Icochilus extensus. It is distinguished by the presence of a very small first lower premolar, and by the absence of the lower canine. In the place of the canine there is quite a long dia- stema which separates the external incisor from the first premolar (’94, 16). Interatherium undulatum (Ameghino). Icochilus undulatus Amegh.; Contrib. al Conoc., etc., p. 473, PI. 15, fig. 14, 1889; Enum. Synoptique, etc., p. 15, 1894 (listed); Segundo Censo, etc., p. 150, 1898 (listed). 102 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. Intermediate in size between I. extensum and /. excavatum from which it is distinguished by the quite different conformation of the molars and premolars, especially on the outer side. The premolars have the two antero-external crests a little broader but lower, without projecting beyond the plane of the outer face, separated by a shallow groove and followed by a groove identical with the preceding one, whence it results that the external side of these teeth has three low vertical elevations separated by two shallow grooves which produce a gently undulating appearance very different from that seen in I. extensum and /. excavatum . The true molars are even more different, since the two antero-external crests are completely fused into a single convex column which is quite broad. Each molar has thus externally two low columns corresponding to the two in- ternal lobes, separated by a shallow depression opposite to the deep fur- row of the inner side. The posterior upper molars, and especially M-, have much compressed crowns. The dental series is almost straight. Diameters of P- .0035 X .0025 ; PA .0037 X .003 ; M1 .004 X .0027 ; M- .004 x .0024 ; M- .004 X -002 ; length of space occupied by the five last upper molars .0185 (’89, 473). Interatherium lamellosum (Ameghino). Icochilus lamellosus Amegh.; Enum. Synoptique, etc., p. 15, 1894 ; Se- gundo Censo, etc., p. 150, 1898. Species of small size. The superior canine is well developed, of the same size and almost of the same form as the external incisor. They are separated the one from the other by a small diastema. These two teeth are laterally compressed and have the form of trenchant blades. The external incisor, or third, is separated from the second by a considerable diastema. Another diastema, a little longer, separates the first premolar from the canine. P- is placed opposite P- and carries a vertical groove with its angle antero-external. Length from anterior border of I1 to pos- terior border of M- .037 (’94, 15-16). Interatherium multident atum (Ameghino). Icochilus multidentatus Amegh.; Enum. Synoptique, etc., p. 17, 1894; Segundo Censo, etc., p. 15 1, 1898 (listed). Quite small. It is distinguished in having eight molars above and below on each side, of which the five anterior teeth are premolars. In Sinclair: typotheria of the santa cruz beds. 103 the upper jaw, the canine and the first two premolars are very small and in continuous series with the other teeth. In the lower jaw, the first two premolars have the form of small canines and the second is separated from the third by a small diastema. The first two lower premolars and the canine are unspaced. Length of space occupied by the eight upper molars .028 (’94, 17). Interatherium interruptum Ameghino. Interatherium interruptum Amegh. ; Enurn. Synoptique, etc., p. 18, 1894 ; Segundo Censo, etc., p. 15 1, 1898 (listed). Size of I. supermini. Distinguished easily by the presence of the first upper premolar which is placed against the anterior part of P-, and by the complete absence of the canine. There is a long diastema which sepa- rates the first premolar from the external incisor. The margin of this diastema carries a deep longitudinal groove which has the same direction as the dental series. Length of the seven upper molars .025 (’94, 18). Interatherium curtum (Ameghino). Icochilus curtus Amegh. ; Enum. Synoptique, etc., p. 17, 1894; Segundo Censo, etc., p. 15 1, 1898 (listed). Of the same size as I. extensum. This species is easily distinguished by the great shortening of the last lower molar which is hardly longer than and by its posterior lobe which is convex on the external side without vestige of the perpendicular groove seen on the same tooth in the other species. is nearly .005 long. The last two molars occupy a space of .009 (’94, 17). Interatherium crassirame (Ameghino). Icochilus crassiramis Amegh. ; Enum. Synoptique, etc., pp. 16-17, 1894 ; Segundo Censo, etc., p. 15 1, 1898 (listed). Of the same size as /. extensus. Distinguished very well by the first lower premolar which has the form of a well-developed canine, being iso- lated in front and behind by fairly large diastemata. The lower canine has the form of an incisor, resting anteriorly on the external incisor. The second lower premolar is elliptical, without perpendicular internal groove. Of the outer groove only traces are visible (’94, 16-17). io4 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. Interatherium dentatum Ameghino. Interatherium dentatum Amegh. ; Enum. Synoptique, etc., p. 1 8, 1894; Segundo Censo, etc., p. 15 1, 1898 (listed). Approaching I. supermini in size. Distinguished by possessing an extra premolar in the upper jaw, that is to say five instead of four. The first two premolars are conic, the first or anterior being separated from the second. Complete upper dental series measures .041 (’94, 18). Interatherium senile (Ameghino). Icochilus senilus Amegh.; Enum. Synoptique, etc., p. 15, 1894; Segundo Censo, etc., p. 150, 1898 (listed). This species is of the size of /. extensus , but a little more robust. It is easily distinguished by the second lower premolar which is not bilobate but elliptical in contour and consequently without vertical groove either on the internal or the external face. The third and fourth lower premolars as well as the true molars are larger than in the other species. The sec- ond upper premolar is also of elliptical contour and without groove. There is a diastema of considerable length between the lower canine and the first premolar. Length of the seven lower molars .032 (’94, 15). Interatherium rotundatum (Ameghino). Icochilus rotundatus Amegh.; Contrib. al Conoc., etc., pp. 473-474, PI. 15, figs. 1 5- 1 6a, 1889; Enum. Synoptique, etc., p. 15, 1984; Segundo Censo, etc., p. 150, 1898 (listed). This species is separated by very marked characters. The skull is more prolonged anteriorly. I1 very large, I- much smaller and I- still smaller. The upper incisors are closely crowded. P1 is placed against the anterior part of P-. There is no upper canine, and there is a long diastema between the external upper incisor and the first premolar. Length, anterior part of I- to posterior part of M- .046 ; length of diastema between I- and P- .008 (’94, 15). Interatherium brevifrons Ameghino. Interatherium brevifrons Amegh.; Enum. Synoptique, etc., p. 18, 1894; Segundo Censo, etc., p. 15 1, 1898 (listed). Species much smaller in size than I rodens and with all the dentition in continuous series without diastemata. The anterior part of the skull SINCLAIR : TYPOTHERIA OF THE SANTA CRUZ BEDS. 105 is very short. Ix is very large and I- and I- very small. The canine is well developed. Length from the anterior part of I1 to the posterior part of M- .027 (’94, p. 18). Interatherium rodens (Moreno) Ameghino. Interatherium rodens Moreno ; Patagonia, resto de un antiguo continente hoy submerjido, p. 23, 1882 [nomen nudum). Ameghino; Observa- ciones generales sobre el orden de mamiferos estinguidos sud-ameri- canos llamados Toxodontes, etc., p. 63, 1887; Enumeracion Sis- tematica, etc., p. 15, 1887; Contrib. al Conoc., etc., pp. 467-468, PI. 15, figs. 20-26, 1889; Enum. Synoptique, etc., p. 18, 1894 (listed); Segundo Censo, etc., p. 151, 1898 (listed). Size of a. rabbit. Represented by numerous fragments of mandibles and the anterior part of a skull with almost all the dentition. The pre- maxilla is short and thick. Ix is three millimeters wide at the base of the crown. The base of I- is two millimeters and that of I- less than one millimeter. The base of the crown of the canine is also less than one millimeter. P1 is very small, almost rudimentary and of an acutely con- ical shape ; it is closely appressed to P-. P- has a more elongate crown. It has a small internal groove which dies away on the base of the crown before reaching the root, dividing the crown into two internal lobes, the anterior much smaller than the posterior. On the anterior part of the outer side is a deep and narrow groove which hardly extends to the base and divides the crown into two external lobes, the anterior very narrow and ridge-like, the posterior much broader. P- and P-, successively larger, have the same general form but with the inner lobes more equal. The three premolars with very long crowns have the bases divided into separate but very short roots. The true molars have an internal groove like the premolars, which extends to the base, dividing the crown into two equal lobes. The antero-external groove also extends to the base, but near the middle of the anterior face there is a second broader groove forming three well-marked external columns. The three true molars are completely open at the base. The seven molars form a much more pro- nounced arc of a circle than in the following species. The symphysial part of the mandible, much compressed laterally, has the incisors and canine in continuous series and inclined forward. Only IT is placed in the anterior narrowest part of the symphysis, with its crown transverse. io6 PATAGONIAN EXPEDITIONS ! PALAEONTOLOGY. The others are placed on the sides with the principal diameter in the direc- tion of the dental series. The anterior part of the crown is a little broader than the root, with the apex bilobate by an internal groove which gradually disappears on wear, but the canine, which has not the groove, has an acutely conical crown. PT, which is separated from the canine by a short diastema, is similarly shaped but implanted vertically. Px is low and bilobate by two opposite grooves, each lobe elliptico-triangular, and the two of almost equal size, the base divided into two separate and quite long roots. P¥ and PT and Mx and Mx are composed of two elliptico- triangular, almost equal prisms separated by two opposite grooves which extend to the completely open base. Each prism is a little narrower on the inner than on the outer side, without reentering folds or salient ridges. M3 is composed of three unequal subprismatic parts, the median twice as large as the anterior and the last one half smaller but on the inner side fused with the second. Length of space occupied by the seven upper molars .021 ; length of space occupied by the seven lower molars .023 (’89, 467-468). Founded on an immature individual with the milk denti- tion. A photograph by Professor Scott of the anterior part of the skull referred to above shows very clearly by the texture of the bone that the individual is far from adult. Interatherium supernum Ameghino. Interatherium supermini Amegh. ; Enumeracion Sistematica, etc., p. 15, 1887 ; Contrib. al Conoc., etc., pp. 468-469, PI. 15, figs. 17-19$, 1889 ; Enurn. Synoptique, etc., p. 18, 1894 (listed) ; Segundo Censo, etc., p. 1 5 1 , 1898 (listed). This species is one-third larger than I. rodens , from which it is easily distinguished by the molars which are shorter and broader, the interme- diate molars being of almost equal length and width, while in /. rodens the antero-posterior diameter of the crown always considerably exceeds the breadth. Further, in the premolars, the internal reentering fold is deeper than in I. rodens but is not extended into a vertical groove, whence the inner side is not bilobate but of vertically convex surface and notably narrower than the external surface. P- differs from P- and P- in being more compressed, with crown much longer than wide, resembling the cor- responding tooth in /. rodens , but without the internal groove of the lat- ter. All the upper premolars and molars have the same antero-external Sinclair: typotheria of the santa cruz beds. 107 ridge as in /. rodens , but less accentuated in the molars and with the second depression broader, already well-marked in P- and P-. All the upper teeth except P- have the anterior part considerably broader than the posterior. P- measures .0045 X .002. P- and PA are .004 long on the outer side but only .0025 on the inner ; maximum breadth .0035. The true molars are somewhat broader on the inner side and diminish slightly in size from M- to M-. Length, P1 to M- .028 following the curve of the series and only .026 in a straight line. A small fragment of the anterior part of the skull with the premaxilla shows that behind the first incisor, in the palate, were two large incisive foramina as in Pachyrukhos, instead of the incisive notch of Typotherium. The lower dentition presents nothing peculiar except for the slightly larger size, but the teeth are in a more continuous series, the short diastema between the canine and PT having disappeared. Length PT to .027 ; depth of jaw at Mt .015. The symphysis measured on the outer side is .014 long. There are three mental foramina, the anterior and largest below the canine, the second and smallest below P2 and the third beneath PT. I have of this species the two rami somewhat broken with three molars of the first dentition still but little worn and the first true molar. The de- ciduous molars, bilobate by two opposite grooves, have separate roots. These teeth, deciduous molars and true molars, have the peculiarity of a small notch or cavity on the inner side of each lobe, very superficial and soon disappearing on wear, but which evidently correspond to the two inner folds of Toxodon , Protoxodon , Colpodon, etc., as also of many pachy- derms, such as Homalodontotherium , Macrauchenia , etc. (’89, 468-469). PACHYRUKHOS. Pachyrukhos teres Ameghino. Pachyrucos teres Amegh.; Contrib. al Conoc., etc., p. 429, PI. 13, figs. 25-27, 1889; Enum. Synoptique, etc., p. 19, 1894 (listed); Segundo Censo, etc., p. 15 1, 1898 (listed). One half smaller than P. typicus * but is very similar in general con- formation. P- is not so disproportionately smaller than P- as in the latter, but is only slightly so. Its anterior part is also less compressed than in the other species and is of perfectly elliptical section. P-, is slightly larger, is of the same shape as P- and both are inserted obliquely. * A Monte Hermoso species. io8 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. M- is somewhat smaller than M-3-, narrower behind and without the third posterior cusp which distinguishes this tooth in P. typicus. The palate is more concave in its posterior part, opposite PA, M1 and M-, ascending rapidly between M-. Pg is considerably smaller than P3, but divided equally into two lobes by a perpendicular external groove as is not the case in P. typicus. The anterior lobe is notably narrower than the posterior and the inner side slightly grooved in its anterior part. The symphysis is very little shorter than in P. typicus , its posterior part extending beneath the anterior lobe of PT, but is much more horizontal. The mental foramina are two, both quite small. The first is a little forward of the anterior part of P^, and the second below the space between Pg and P^, considerably more forward than in P. typicus. Length of space occu- pied by the six upper molars .019 ; width of palate at P- .010 ; width of palate at M- .0135 ; length of space occupied by the six lower molars .019; width of IT .0037 I width of .002 ; length of symphysis inferiorly .0125; length of inferior diastema .006; depth of ramus below MT .011 (’89, 429). Pachyrukhos trivius Ameghino. Pachyrucos trivius Amegh. ; Contrib. al Conoc., etc., p. 429, PI. 13, fig. 31, 1889; Enum. Synoptique, etc., p. 19, 1894 (listed); Segundo Censo, etc., p. 151, 1898 (listed). A little smaller than the preceding species, although the dental series is of the same length, but the bones are more delicate, the mandible is lower and thinner and the molars, especially P^ of very different shape. This tooth is triangular and might be described as formed by a single lobe with anterior appendix, which ends forward in a thin vertical crest. This appendix is well demarcated from the lobe by a small external groove, but on the inner side is united with it into a single smooth plane. The mandible is distinguished by the almost uniform depth of the hori- zontal ramus below the teeth. On the most complete fragment of the mandible at my disposal there is no sign of the first mental foramen since the fragment is broken at this point. The second foramen, quite small, is below the anterior part of PT, somewhat farther back than in the pre- ceding species. The six lower molars occupy a space .018 in length; depth of jaw at MT .0095 (’89, 429). SINCLAIR : TYPOTHERIA OF THE SANTA CRUZ BEDS. IO9 HEGETOTHERIUM. Hegetotherium convexum Ameghino. Hegetotherium convexum Amegh. ; Revista Argentina de Historia Natural, I, pp. 1 33-i 34. fig- 30, 1891 ; Enum. Synoptique, etc., p. 19, 1894 (listed) ; Segundo Censo, etc., p. 15 1, 1898 (listed). Size comparable to Hegetotherium mirabile or slightly smaller. M- con- siderably smaller than M-, with postero-external angle not pronounced and with the inner face very convex. Antero-posterior diameter of M- .007 ; of M- .005 ; width of palate between the third molars .020 (’91, 134)- EIegetotherium minum Ameghino. Hegetotherium minum Amegh. ; Enum. Synoptique, etc., p. 19, 1894 ; Se- gundo Censo, etc., p. 15 1, 1898 (listed). This species is distinguished by its size, a little smaller than H. striga- tum (synon. H mirabile ) and by the horizontal ramus of the mandible being very low and shortened in front. Length, anterior border of IT to posterior border of .041 ; depth of mandible below MT .014 (’94, 19). Hegetotherium anceps Ameghino. Hegetotherium anceps Ameghino ; Revista Argentina de Hist. Nat., I, p. 242, 1891, Enum. Syn., etc., p. 19, 1894 (listed) ; Segundo Censo, etc., p. 15 1, 1898 (listed). Size of Hegetotherium convexum ; M- much smaller than M- but with inner face depressed and more or less grooved perpendicularly. Antero- posterior diameter of M- .007 ; the same for M- .004 ; length of space occupied by the second and third molars .012 (’91, 242). BIBLIOGRAPHY. Ameghino, Florentine. 1882 Catalogo de la provincia de Buenos Aires en la Exposicion Continentale Sud-amer., March, 1882. 1885 Nuevos restos de mamiferos fosiles oligocenos, recogidos por el Prof. Pedro Scalabrini y pertenecientes al Museo provincial del Parana. Bol. Acad. Nac. de Cienc., T. VIII, p. 5, 1885. Also published separately in an octavo of 205 pages. 1887a Observaciones generales sobre el orden de mamiferos estinguidos sud-americanos llamados Toxodontes (Toxodontia) y sinopsis de los generos y especies hasta ahora conocidos, pp. 1-66, 1887. 1887b Enumeracion sistematica de les especies de mamiferos fosiles coleccionados por Carlos Ameghino en los terrenos eocenos de la Patagonia austral. La Plata, 1887. I IO PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. 1889 Contribucion al conocimiento de los mamiferos fosiles de la Republica Argentina (Actas de la Academia Nacional de Ciencias en Cordoba, T. V.). 1891a Nuevos restos de mamiferos fosiles descubiertos por Carlos Ameghino en el eoceno inferior de la Patagonia austral. — Especies nuevas, adiciones y correcciones, pp. 1-44, Aug., 1891. Ibid. Revista Argentina de Historia Natural, Tomo I, Entrega 5a, pp. 289-328, Oct., 1891. 1891b Los monos fosiles del eoceno de la Republica Argentina. Revista Argentina de His- toria Natural, Tomo I, Entrega 6a, pp. 383-397, 1891. Figures and describes I. robustum , also a fore and hind foot incorrectly referred to this genus and species. 1894a Enumeration synoptique des especes de mammiferes fossiles des formationes eocenes de Patagonie (Boletin de la Academie Nacional de Ciencias en Cordoba, T. XIII, p. 259). In the citations from this paper given in the text the pagination is that of the separate edition. 1894b Sur les ongules fossiles de l’Argentine. Revista del Jardin Zoologico de Buenos Aires, Tomo II, Entrega 7, pp. 193-303, 1894. 1898 Sinopsis geologico-palaeontologica. Segundo Censo de la Republica Argentina, T. I, p. 1 13, Buenos Aires, 1898. Cope, E. D. 1897 Toxodontia. American Naturalist, Vol. XXXI, pp. 485-492, June, 1897. Lydekker, R. 1893 A study of the extinct ungulates of Argentina. Anales del Museo de la Plata, Palaeon- tologia Argentina, II, article 3, 1893. Moreno, Francisco P. 1882 Patagonia, resto de un antiguo continente hoy submerjido. Conferences de la Sociedad Cientifica Argentina, Conferencia del 15 de Julio. Buenos Aires, 1882. Roth, Santiago. 1903 Los ungulados sudamericanos. Anales del Museo de La Plata, Seccion Palseontologica V, pp. 1-36, Pis. I-IV, La Plata, 1903. Scott, W. B. 1904 The Miocene Ungulata of Patagonia. Rept. British Assn. Adv. Sci., pp. 589-590, 1904. Sinclair, Wm. J. 1908 The Santa Cruz Typotheria. Proceedings of the American Philosophical Society, Vol. XLVII, pp. 64-78, figs. 1-10, April, 1908. Weber, Max. 1904 Die Saugetiere, pp. 589, 702, 705. Zittel, K. A. von. 1893 Handbuch der Palseontologie, pp. 490-500. PATAGONIAN EXPEDITIONS I* PALAEONTOLOGY. EXPLANATION OF PLATE I. PAGE Fiof. i. Hegetotiif.rium mirabile Fig. 2. “ “ Fig. 3. Fig. 4. “ *• Fig. 5. Fig. 5 a. “ “ Fig. 5^. “ “ Fig. 6. “ “ Fig. 7. Skull and mandible, left side (No. 1 5,542). 70 “ from above (No. 15,542). . . 72 “ “ below “ “ . . 73 “ “ behind “ “ . . 73 Mandible crown view (No. 15,542). . 69, 75 Second lower premolar, Xf (No. 15,542). 69 “ “ “ “ (No. 15,432). showing outline of less worn tooth. . 69 Mastoid cavity from behind (No. 15,505). The canal from the mastoid cavity opens into the tympanic chamber at x. m, mastoid ; am, auditory meatus ; e, exoccipital ; pp, paroccipital pro- cess ; t , tympanic. . . . . 71 Lumbar vertebra from in front showing interlocking zygapophyses (No. 15-093) 76 All the figures except 5 a and 5$ are natural size. (vol. vi) Patagonian Expeditions Vol.vi Plate i Bruce Horsfall del Werner & Winter, Frankfort °M , lith Hegetotherium PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. Fig. i. Fig. 2. Fig. 3. Fig. 4. Fig- 5- Fig. 6. Fig. 7. Fig. 8. Fig. 9. Fig. 10. Fig. 11. Fig. 12. Fig. 13- Fig. 14. Fig. 15- Fig. 16. Fig. 17. Fig. 18. Fig. 19. Fig. 20. EXPLANATION OF PLATE II. PAGE FIegetotiierium mirabile : Right tibia and fibula, front view (No. 15.298) 79 “ “ Right tibia and fibula, distal end (No. i5.298) 79 “ “ Right radius, inner side (No. 15,542). 77 “ “ “ “ proximal end “ “ 77 “ “ “ “ distal “ “ “ 77 “ “ “ ulna, front view “ “ 77 “ “ “ “ inner side “ “ 77 “ “ “ femur, front view (No. 15,431). 78 “ “ “ “ from behind “ “ 78 “ “ “ “ distal end “ “ 78 “ “ Patella, front view (No. 15,298). . 78 “ “ “ back “ “ “ . 78 “ “ Right humerus from in front (No. I5N76) 76 “ “ Right half of pelvis (No. 15,093). . 78 “ “ Inner side of right calcaneum (No. 15.298) showing navicular facet. . 80 “ “ Right navicular, distal aspect (No. 15.298). 80 “ “ Proximal articular surfaces of the meta- tarsals of the right pes (No. 15,392). 81 “ “ Third and fourth metatarsals (No. 15,392) showing the peg and socket articulation at the proximal end. Slightly restored from another specimen. . . . . . 81 “ “ Right pes (composite). Calcaneum, cuboid, navicular and ectocuneiform are from specimen No. 15,298. The astragalus, entocuneiform and meta- tarsals are from No. 15,392. The calcaneum has been turned out- ward slightly to show the facet for this element on the navicular. . 80, 81 “ “ Right astragalus, plantar aspect (No. *5.392) 79 Patagonian Expeditions Vol.vl Plate ii. Bruce Horsfall del Werner & Winter. Frankfort . Hegetotherium I r Sinclair: typotheria of the santa cruz beds. Fig. 21. Hegetotherium mirabile: Terminal phalanx (No. 15,298). 82 Fig. 2 2. i 4 “ Third cervical from the left side (No. 15-542) 76 Fig. 23. i t “ Axis from below (No. 15,542). 76 Fig. 24. i i “ “ “ the left side (No. 15,542). 76 Fig- 25- i i “ Atlas “ above “ “ 75 Fig. 26. u “ “ “ behind “ “ 75 Fig. 27. ( ( “ “ “ below “ “ 75 All the figures are natural size. o (VOL. VI.) PATAGONIAN EXPEDITIONS : PALEONTOLOGY. EXPLANATION OF PLATE III. PAGE Fig. i. Protypotherium australe: Skull and mandible, left side (No. 9565 American Museum). . . . 19 Fig. 2. “ “ Skull from above (No. 9565 American Museum). ..... 20 Fig. 3. “ “ Skull from below (No. 15,598). . . 21 Fig. 4. . “ “ “ “ behind (No. 9565 American Museum). . . . . . 21 Fig. 5. “ attenuatum : Skull, leftside (No. 15,665). . . 44 Fig. 6. “ “• “ and mandible, left side (No. 9187 American Museum). The lachrymal is broken off in this specimen. ..... 44 Fig. 7. “ “ Internal structure of the right mas- toid, showing cancellae (No. 15.665) 20 All the figures are natural size. (vol. VI.) Patagonian Expeditions Vol.vi Plate hi Bruce Horsfall del Werner& Winter, Frankfort °M. PR o typo the rium t ■ PATAGONIAN EXPEDITIONS : PAL/EONTOLOGY. EXPLANATION OF PLATE IV. Fig. i. Protypotherium AUSTRALE Right femur from in front (No. 15,340). PAGE Fig. 2. 4 4 4 4 “ “ “ behind “ “ 30 Fig. 3. 4 4 4 4 “ “ distal end “ “ 30 Fig. 4. 4 ( 4 4 “ tibia, front view “ “ 30 Fig- 5- 4 ( 4 4 “ “ distal end “ “ 30 Fig. 6. 4 i 4 4 Left humerus from in front (No. 15.828) 26 Fig. 7. 4 4 4 4 Left humerus from behind (No. 1 5828). 26 Fig. 8. 4 4 4 4 Patella, front view (No. 9149 Ameri- can Museum). .... 30 Fig. 9. 4 4 4 4 Patella, back view (No. 9149 Ameri- Museum). ..... 30 Fig. 10. 4 4 4 4 Scapula, side view (No. 15,828). 26 Fig. 11. 4 4 4 4 “ distal end “ “ Slightly restored from another specimen. 26 Fig. 12. 4 4 4 4 Left radius, proximal end (No. 15,828). 27 Fig. 13. 4 4 4 4 “ “ distal end (No. 15,828). . 2 7 Fig. 14. 4 4 4 4 “ “ side view “ “ 27 Fig. 15- 4 4 4 4 Mandible, crown “ “ “ 16, 23 Fig. 16. 44 4 4 Left ulna, inner side “ “ 2 7 Fig. 17. 4 4 4 4 “ “ front view “ “ 27 Fig. 18. 4 4 4 4 “ fibula, inner side (No. 9149- American Museum). 31 All the figi ares are natural size. (VOL. VI.) Patagonian Expeditions Vol.vi Plate iv. Bruce Horsfall del Werner* Winter, Frankfort °M PR o typo the rium PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. EXPLANATION OF PLATE V. Fig. 1. Protypotherium australe : Left pes, dorsum (No. 9149 American Museum). . . . . . Fig. 2. ( i “ Third and fourth metatarsals showing the peg and socket articulation at the proximal end (No. 9149 Ameri- can Museum). . . . . Fig. 3- u “ Left manus, dorsum (No. 9149 Amer- ican Museum). . . . . Fig. 4- u attenuatum : Right manus, dorsum (No. 15,341). Fig. 5- ( i “ Left pes, dorsum (No. 1 5,341). The astragalus is turned slightly toward the inner side of the foot. . Fig. 6. (( australe : Right calcaneum, inner side (No. 15.828) Fig. 7- ( ( “ Right astragalus, plantar aspect (No. 15.828) Fig. 8. a “ Right navicular, distal aspect (No. 15.828) Fig. 9- u “ Proximal articular surfaces of the metacarpals of the left manus (No. 9149 American Museum). Fig. 10. ( i “ Proximal articular surfaces of the metatarsals of the left pes (No. 9149 American Museum). Metatarsal II is supplied from another specimen. Fig. 1 1. i < sp. : Upper deciduous and permanent dentition, crown view (No. 9482 American Museum). Fig. 1 1 a. u “ Px and Dp-, crown view, Xf (No. 9482 American Museum). . . . . Fig. 1 ib. u “ M-, crown view, Xf (No. 9482 American Museum). ...... Fig. 1 2. (( “ Upper deciduous and permanent dentition, side view (No. 9482 American Museum). Fig- i3- ( i “ Lower deciduous and permanent dentition, side view (No. 9482 American Museum). Fig. 14. ( ( “ Lower deciduous and permanent dentition, crown view (No. 9482 American Museum). PAGE 3i 33 27 27 3i 3i 31 32 29 33 18 18 18 18 18 18 Patagonian Expeditions Vol.vi. Plate v. Bruce Horsfall del Werner & Winter. Frankfort Pr o typo the rium SINCLAIR: TYPOTHERIA OF THE SANTA CRUZ BEDS. Fig. 14 a. Protypotherium sp.: M-g, crown view, X-f- (No. 9482 American Fig- T5- << Museum). ...... “ Right ramus of mandible with P2 not yet Fig. 15 a. i fully erupted (No. 9559 American Museum). ...... “ P-2 showing the unworn crown (No. 9559 Fig. 16. i American Museum). . . . . australe : Axis from the left side (No. 15,551). . Fig. 17. i i “ “ “ below “ “ Fig. 18. < < “ Atlas “ above “ “ Fig. 19. i i Slightly restored. . . . . “ Atlas from behind (No. 1 5, 5 5 1 )• Fig. 20. i ( Slightly restored. . . . . “ Atlas from below (No. 15,551). Fig. 21. a Slightly restored. . . . . prtlrutilum : Skull, palatal view (No. 15,386). . Fig. 22. i ( “ Lower dentition, crown view (No. All the 15-386) figures except 11 a, 11b, 14# and 15# are natural size. 18 i7 i7 23 23 23 23 23 40 40 (VOL. Vi) PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. Fig. I. Fig. 2. Fig- 3- Fig. 4. Fig- 5- Fig. 6. Fig. 7. Fig. 8. Fig. 9. Fig. 10. Fig. 1 1 . Fig. 12. Fig. 13- Fig. 14. Fig- 15- Fig. 16. Fig. 17. Fig. 18. Fig. 19. Fig. 20. Fig. 21. EXPLANATION OF PLATE VI. Protypotherium attenuatum : Right humerus, front view (No. 15,341). The perforation in the coronoid fossa is artificial. “ “ Right radius, front view (No. 15.341) . • “ “ Left ulna, outer side (No. 15,341). “ “ “ femur from in front (No. 15.341) “ “ Left tibia and fibula from in front (No. 15,341) “ “ Left half of pelvis (No. 15,341). “ australe: First sternal segment from below (No. 15.828). “ “ Anterior dorsal vertebrae from the right side (No. 15,828). . “ “ Pelvis from the left side (No. 9566 American Museum). Slightly re- stored from other specimens. “ “ Pelvis, sacrum and posterior lumbars from above (No. 9566 American Museum). . Interatherium robustum : Pelvis and sacrum from above (No. 15,401). Partly restored from other specimens. . “ “ Pelvis from the left side (No. 15,401). Partly restored from other specimens. “ extensum : Left scapula, outer side (No. 15,041). Protypotherium australe: Sixth lumbar from in front (No. 15,828). “ “ First lumbar from the left side (No. 15.828) Interatherium robustum : Caudal vertebra from above (No. 15,401). (( (( U U (( (( (< (( a u (( u u a u (( u a <( << u a a u a u a (( u u (( (( “ sp.: First sternal segment from below (No. 9129 American Museum). . . . . . All the figures are natural size. PAGE 26 27 2 7 30 30 30 25 24 30 24. 30 53. 57 57 54 24 24 54 54 54 54 54 (VOL. VI.) HI Werner & Winter, Frankfort Bruce Horsfall del. Patagonian Expeditions Vol.vi. Plate vi. Protypgtherium, Interatherium PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. EXPLANATION OF PLATE VII. Protypotherium australe : Restoration of the skeleton about one half natural size. Based on specimens in the collections at Princeton Uni- versity and the American Museum of Natural History. PAGE 33 (VOL. Vi) Patagonian Expeditions Vol.vi. Plate vii. Bruce Horsfall del Werner & Winter, Frankfort °- A . hth PR o typo the rium PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. Fig. I. Fig. 2. Fig- 3- Fig. 4. Fig. 5- Fig. 6. Fig. 7. Fig. 8. Fig. 9. Fig. 10. Fig. 1 1. Fig. 12. Fig. 13. Fig. 14. Fig- 15- Fig. 16. Fig. 17. Fig. 18. Fig. 1 9. EXPLANATION OF PLATE VIII. PAGE Interatherium : Composite fore foot, based primarily on the fore foot of I. robustum (No. 15,348). The trape- zium is drawn from No. 15,041 (/. externum). The phalanges of digits II, IV, V and the terminal phalanx of digit III are from No. 15.401. 56 robustum : Right pes, dorsum (No. 15,401). Slightly restored from other speci- mens. The cuboid is broken above metatarsal V. .... 58 extensum: Left femur from in front (No. 15,041). 57 “ “ “ “ behind “ “ 57 robustum : Right tibia and fibula from in front (No. 15.401) . 57 “ Right tibia and fibula distal end (No. 15.401) . The elements' are slightly dislocated. . . . . . 57 “ Left humerus from in front (No. 15, 100). Partly restored from other specimens. 55 “ Right scapula, side view (No. [5,348). 54 sp. Left ulna, outer side (No. 9557 American Museum). . . . . . . 55 “ Right radius, from in front (No. 15,401). . 55 robustum: Axis, side view (No. 15,401). The transverse processes are restored from other specimens. . . . 52 Axis from below (No. 15,401). The transverse processes are restored from other specimens. . . . 52 “ Atlas from above (No. 15,401). . . 52 “ “ “ below “ “ . . 52 “ “ “ behind “ “ . . 52 “ Skull from the left side (No. 9263 American Museum). . . . 49-52 extensum: Skull from above (No. 15,041). . . 49-52 “ “ “ below “ “ . . 49—52 robustum: Skull from behind (No. 9263 American .Museum). ..... 49-52 Patagonian Expeditions Vol.vi Plate viii Bruce Horsfall del. Werners Winter. FrankfortfM. I N T E RAT HERIUM Sinclair: typotheria of the santa cruz beds. Fig. 20. Interatiierium sp. : Second and third upper premolars of the left side, showing pattern of unworn crowns, Xf (No. 15,187) 48 Fig. 21. < t “ Lower milk molars, crown view (No. 15,021). 49 Fig. 2 2. <4 “ “ “ “ side “ “ 49 Fig. 23. 4 4 “ Upper “ “ crown “ (No. 15,541). 49 Fig. 24. 4 4 “ “ “ “ side 49 Fig. 25. ( 4 “ Mandible, crown view (No. 15,554). 49- 5 1 Fig. 26. 4 4 robustum : Part of the upper dentition of an old individual showing supernumerary canine (No. 15,293). 48 Fig. 27. 4 4 sp. : Part of the upper dentition of a young indi- vidual with alveoli for supernumerary canine and first premolar, probably due to the pres- ence of both milk and permanent teeth (No. 15-201) 48 Fig. 28. 4 4 excavatum : Atlas from above (No. 15,043). 52 Fig. 29. 4 4 robustum: Patella, front view (No. 15,348). Slightly restored from other specimens. 57 Fig. 30. 4 4 “ Patella, rear view (No. 15,348). Slightly restored from other specimens. All the figures except 20 are natural size. 57 (VOL. Vi) PATAGONIAN EXPEDITIONS ! PALAEONTOLOGY. EXPLANATION OF PLATE IX. Interatherium robustum : Restoration of the skeleton based on two speci- mens of this species, Nos. 15,401 and 15,348, supplemented by a third specimen referred to I. extensum (No. 15,041). About two thirds the natural size. ........... PAGE 59 (VOL. Vi) INTERATHERIUM PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. Fig. I. Fig. 2. Fig- 3- Fig. 4. Fig. 5- Fig. 6. Fig. 7. Fig. 8. Fig. 9. Fig. 10. Fie. 11. Fig. 12. Fig. 13- Fig. 14. Fig. 15- Fig. 16. Fig. 17- EXPLANATION OF PLATE X. PAGE Pachyrukhos moyani: Skull and mandible from the left side (No. 9283 American Museum). Slightly re- stored from other specimens. . . 87-90 “ “ Skull from above (No. 15,743). Partly re- stored from other specimens. . . 87-90 “ “ Skull from below (No. 15,743). Partly re- stored from other specimens. . . 87-90 “ “ Skull from behind (No. 1 5,743). The canal connecting the mastoid chamber with the tympanic cavity is seen on the right side. 87-90 “ “ Left clavicle, outer side (No. 9481 American Museum). . . . . . . 91 “ “ Left hyoid apparatus, inner side (No. 15.743) “ “ Right tibia and fibula (No. 9242 American Museum). ...... 93 Right femur (No. 9242 American Museum). 93 Left ulna from in front (No. 9219 American Museum). ...... 92 Left ulna, inner side (No, 9219 American Museum). ...... 92 Left humerus (No. 9481 American Mu- seum). Partly restored from other specimens. . . . . . . 91 Left radius, outer side (No. 9481 American Museum). . . . . . . 92 Right half of mandible, crown view (No. 15,743). Partly restored from other specimens. . . . . . . 86, 89 Left manus, dorsum (No. 9481 American Museum). ...... 92 Right pes, dorsum (No. 9481 American Museum). ...... 94 Distal end of left tibia and fibula (No. 1 5»743) 93 Pelvis, left half (No. 9242 American Mu- seum). Partly restored from other spec- 92-93 imens. . Bruce Horsfall del Werner & Winter, Frankfort Pachyrukhos Sinclair: typotheria of the santa cruz beds. Fig. 1 8. Fig. 19. Fig. 20. Fig. 21. Fig. 22. Fig- 2 3- Fig. 24. Fig- 25. Fig. 26. Fig. 27. Fig. 28. Fig. 29. Fig. 30. Fig. 31- Pachyrukhos moyani : Pelvis and sacrum from above (No. 9242 American Museum). . . . . “ “ Patella from in front (No. 9242 American Museum). ...... “ “ Patella from behind (No. 9242 American Museum). ...... “ “ Glenoid fossa of right scapula (No. 15,743). The outer margin faces the lower edge of the plate. ...... “ “ Right astragalus, plantar aspect (No. 9242 American Museum). . . . . “ “ Axis and third to sixth cervicals from the left side (No. 15,744). . . . . “ “ Atlas from below (No. 15,744). u << “ “ above “ “ “ “ “ “ behind “ “ “ “ Right calcaneum, dorsal aspect (No. 9242 American Museum). . . . . “ “ Anterior dorsal vertebra from the left side (No. 15,743) “ “ Anterior dorsal vertebra from the left side (No. 15,743). “ “ Posterior dorsal vertebra from the left side (No. 15743) “ “ Posterior lumbar vertebra from above (No. 15,743). The transverse processes are restored in outline from other specimens. All the figures are natural size. 92-93 93 93 9i 94 90 90 90 90 94 90 90 90 9T (VOL. Vi) PATAGONIAN EXPEDITIONS! PALAEONTOLOGY. EXPLANATION OF PLATE XI. Pachyrukhos moyani : Restoration of the skeleton based on three specimens in the collection of the American Museum of Natural History (Nos. 9242, 9481, and 9283). Two thirds the natural size. .... PAGE 94. 95 (VOL. Vi) Patagonian Expeditions Vol.vi. Plate, xl W.J.S.& F.v.l del. Werner & Winter. Frankfort Iith. Pachyrukhos J. PIERPONT MORGAN PUBLICATION FUND Reports of Princeton University Expeditions to Patagonia, 1896-1899 J. B. HATCHER, in Charge , . ■ ; . EDITED BY WILLIAM B. SCOTT BLAIR PROFESSOR OF GEOLOGY AND PALAEONTOLOGY, PRINCETON UNIVERSITY . genian \n$tir0. VOLUME VI — PALEONTOLOGY Part Part II. Toxodonta of the Santa Cruz Beds III. Entelonychia of the Santa Cruz Beds BY WILLIAM B. SCOTT PRINCETON UNIVERSITY PRINCETON, N. J. The University STUTTGART E. Schweizerbart’sche Verlagshandlung (Nagele & Dr. Sproesser) 1912 Issued December 10, 1912 Press of The New era Printing Company Lancaster. Pa. MAMMALIA OF THE SANTA CRUZ BEDS PART II. T0X0D0NTA. BY WILLIAM B. SCOTT, Princeton University. THE known representatives of this suborder range from rather small to very large and massive animals, always much exceeding in bulk and stature the contemporary Typotheria, but, in Santa Cruz times, equalled or even surpassed by the Entelonychia, not to mention the Astrapotheria, some species of which are the largest mammals known from the Santa Cruz beds. In this group the dental formula is very generally unreduced : If Cf, Pf, Mf, though there is some individual variability in the number of small and functionally unimportant teeth between the incisor tusks (i- and 3) and the cheek-teeth. Only in Phobereotherium is there a significant difference from the usual formula in the absence of the upper median incisors. A highly characteristic feature of the suborder is the arrangement of the incisors and the remarkable changes in appearance and relative size which these teeth undergo during the lifetime of the individual animal. The second upper and third lower incisors (i- and ?) are much enlarged and grow throughout life, or, at least, to very advanced age, and form moder- ately large tusks, such as are found in no other subdivision of the order. Indeed, this arrangement occurs in no other known group of Santa Cruz ungulates except in the family Proterotheriidae of the Litopterna (see Vol. VII, p. 8) in which, however, these tusks are relatively small. The very large tusks, of the Astrapotheria are canines. The teeth, especially the tusks and the true molars, display a strong tendency to hypsodontism, the former growing from persistent pulps and the latter forming their roots at an advanced stage of wear. The crowns are incompletely covered I 12 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. with enamel, which forms broad vertical bands. In the post-Santa Cruzian genera complete hypsodontism is attained. Except in Phobereotherium , the median upper incisor (iA) is broad, antero-posteriorly compressed and somewhat chisel-shaped ; it early de- velops a root, toward which the crown contracts and is thus steadily reduced in size by abrasion. The second upper incisor (i-) is the tusk and is placed in the same transverse line as i-; it is of D-shaped cross- section, terminating in a sharp point, and somewhat resembles the bayonet- like canine tusk of the peccaries, but is abraded on the posterior face. The third upper incisor (i-) is very small, hardly more than vestigial, and can have had little functional importance. The same is true of the canines in both jaws, which may even be absent, though probably only as an indi- vidual abnormality. The two median lower incisors (iy and ?) are broad, chisel-shaped and rooted and diminish in size with age and wear. The inferior tusk (iy) is laterally compressed and inclined forward, more or less procumbent, and is obliquely truncated by the abrasion of the upper tusk. In the Pampean Toxodon all the teeth are hypsodont and the lower incisors are all strongly procumbent, so that the appearance of the anterior teeth is quite different from that seen in the Santa Cruz toxodonts. The premolars are all of simpler pattern than the molars and, though having very high crowns, they early form roots in the Santa Cruz repre- sentatives of the suborder, which distinguishes them from those of the contemporary Typotheria. In toxodonts from the Pampean beds the premolars are completely hypsodont. The upper molars are all strongly curved, with the convexity outward, those of the opposite sides almost meeting in the middle line of the palate ; they also are partially hypso- dont in the Santa Cruz genera, completely so in the later forms. As Sinclair has pointed out (p. 8), the molars are constructed on the same plan as those of the Typotheria and have some resemblance, though not a close one, to the rhinocerotic pattern, with broad external wall and two oblique transverse crests. The grinding surface is much complicated by spurs and crotchets given off from the outer wall and transverse crests, though in the genera of the post-Santa-Cruzian formations the molar- pattern is greatly simplified. The lower molars are composed of two crescents, each with an internal pillar. The posterior pillar appears to be common to all of the groups of the Santa Cruz ungulates, though in some of the genera of the Litopterna SCOTT: TOXODONTA OF THE SANTA CRUZ BEDS. I 1 3 it is much reduced or even absent. This pillar divides the valley of the posterior crescent into two parts and does not close it internally, as it does in the Santa Cruz Typotheria. Like the upper molars, the lower ones are incompletely hypsodont and form roots, at a comparatively late stage. The milk-dentition differs in several respects from the permanent one. The incisors, which are rooted before eruption, are all more or less chisel- shaped, not forming tusks, and thus the appearance of the young skull with milk-teeth is very different from that of the adult and has led to great confusion in the nomenclature. The grinding teeth are quite similar to their permanent successors, but are more brachyodont and dpq is molariform. Ameghino has reported the very extraordinary fact that in Nesodon and Adinotherium there is a complete and functional pre-lacteal dentition, which is not known to be the case in any other mammal, although the same author has shown that pre-lacteal grinding teeth are present in the tapir. The skull, though of very much larger size and quite different appear- ance and proportions, is yet constructed upon essentially the same plan as that of the Typotheria, the toxodont skull having much less resemblance to that of the Rodentia. The cranium is relatively narrow and the facial region broad, without such a long and slender rostrum as characterizes the Typotheria. On the other hand, the skull and mandible are deeper dorso-ventrally, which gives a peculiar appearance to the head. The occiput, which differs considerably in the various genera and species of the group, is proportionately narrower and higher than in the Typotheria. The auditory apparatus is of the same exceptional structure as is found throughout the order and in the Toxodonta has few special peculiarities. The post-tympanic portion of the squamosal (or perhaps, as Roth main- tains, the mastoid) contains a large, oval cavity, which communicates with the cavity of the tympanic bulla, but it is not so inflated as to form an external protuberance on the surface of the skull, as it is in such Typotheria as Pachyrukhos and Hegetotherium. On the other hand, this part of the squamosal forms a larger proportion of the occipital surface than in the Typotheria or Entelonychia and the external audi- tory meatus has a more elevated position than in the other suborders, owing to the extensive union of the post-tympanic and postglenoid pro- cesses with the intervening mastoid portion of the periotic (the protuber- I 14 PATAGONIAN EXPEDITIONS'. PALAEONTOLOGY. ancia petrosa of Roth). This position of the mastoid between the two processes of the squamosal is an altogether exceptional one, the normal place, when it appears on the surface of the skull, being between the ex- occipital and the squamosal. Inward, the process is extended to a contact with the tympanic bulla, with which it may coossify. The hyoid apparatus possesses the peculiarity (unique among mammals, so far as is now known) of being attached to the anterior , instead of the posterior, end of the tympanic bulla and in the adult skull the stylohyal is firmly ankylosed in that position. In the Santa Cruz Typotheria the hyoid is loosely connected with the skull and is very rarely found in posi- tion, but its point of attachment is different from that of the Toxodonta, though likewise very exceptional. In this group the junction of the par- occipital process with the tympanic leaves no room for the hyoid in its usual position and it is displaced to the external side of the bulla, near the posterior end, where a cylindrical fossa may be seen in some of the genera. The neck is short and thick and the trunk is very long, with long and deep thorax. The development of the neural spines is quite different in the various genera. In the comparatively small Adinotherium the spines of the back and loins have their tips in nearly the same horizontal plane, making the profile an almost straight line, while in Nesodon the spines of the anterior thoracic vertebrae are greatly elongated and form a decided hump at the shoulders, an arrangement which is exaggerated in Toxodon . The thoracic vertebrae have the pedicles of the neural arch perforated by large foramina for the passage of the spinal nerves. The sacrum is long, broad and depressed. Caudal vertebrae are not yet known in the Santa Cruz representatives of the suborder, but from the character of the sacrum it may be inferred that the tail, as in Toxodon , was heavy and of moderate length, in contrast to the long and slender tail of most of the contempo- rary Typotheria. The limbs are in all cases relatively short and heavy and the feet sur- prisingly small in proportion to the size of the animal ; the limb-bones differ greatly in size and massiveness in the various genera of the sub- order, but agree quite closely in structure among all the forms in which these elements are known. One striking difference between the earlier and the later genera is in the shape of the scapula, which in the Santa Cruz animals is relatively much broader than in the Pampean types and scott: toxodonta OF THE SANTA CRUZ beds. I 15 bears two very prominent metacromia, given off from the spine, of which only the proximal one is retained and in greatly reduced form in Toxodon. This double metacromion does not occur in theTypotheria ; there is merely a small, well-defined process near the distal end of the spine. The clavicle, which is retained in some at least of the Typotheria (see p. 91), has not been found in any member of the present suborder, yet it is not at all unlikely that it may have persisted as late as Santa Cruz times, as is suggested by the prominent acromion. The unexpected discovery of a vestigial clavicle in an artiodactyl of the upper Oligocene (Scott, ’94, 136) shows that only a rare and fortunate accident could bring to light so small and loosely attached an element. The pelvis is quite different from that of the Typotheria, in which the ilium is elongate, slender and more or less trihedral. In the Toxodonta the ilium has lost the trihedral form and its anterior plate is greatly ex- panded and everted, in a degree which varies with the size of the animal and reaches its maximum in the large Pampean species, in which the pelvis has quite an elephantine appearance. In the Santa Cruz genera the fore and hind limbs are of approximately equal length, but in the Pampean Toxodon the hind limb is much longer, which results in the elevation of the rump and depression of the shoulders, neck and head and gives a very curious appearance to the skeleton. The humerus is short and quite stout, becoming very massive in the later and larger representatives of the suborder. The fore-arm bones are separate in all of the known genera and the ulna remains large and heavy throughout the series, while the radius is relatively slender. The femur is the longest of the limb-bones and in the Santa Cruz genera retains a prominent third trochanter, which is lost in the Pampean forms, in which also the femoral shaft is much compressed and flattened antero-pos- teriorly, which gives the bone a decided resemblance to that of the ele- phants. In all of the known genera of the suborder the leg-bones are ankylosed at the proximal, but not at the distal end, a very peculiar arrangement ; the tibia has the shaft strongly compressed laterally, but has a very broad proximal end, so that the interosseous space is very wide. The fibula is heavy and unreduced and retains throughout the series a large articulation with the calcaneum. As already noted, the feet are curiously small in proportion to the size of the skeleton. Though assuredly derived from five-toed ancestors, all Il6 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. of the known genera are tridactyl, but always retain a vestige of the fifth metacarpal, and are of mesaxonic symmetry, which is thus more constant and definite than in the Santa Cruz Typotheria, which have tetradactyl feet with either paraxonic or mesaxonic symmetry. The carpus is of the com- pletely interlocking type, the scaphoid articulating with the magnum and the lunar with the unciform, thus closely resembling the structure of the perissodactyls, notably the early rhinoceroses. On the other hand, the tarsus retains a highly primitive character ; the astragalus has a trochlea with very shallow groove, a short neck and convex head, which rests exclu- sively upon the navicular, being widely removed from the short cuboid, while the calcaneum bears a very large and prominent facet for the fibula ; the ento- and mesocuneiforms are coossified. The ungual phalanges, espe- cially that of the median digit, are broad and heavy and quite like those of such perissodactyls as Palceosyops in the Santa Cruz genera, and in the Pampean are more rhinocerotic in form. Classification of the Toxodonta. Ameghino has divided the Toxodonta into several family groups, the Toxodontidae, Xotodontidae, Haplodontheriidae, Nesodontidse, but, as he has predicted, this example will not be followed here. He complains that South America fossils are treated differently from those of other continents, saying: “Comme les faunes eteintes sudamericaines et celles des autres continents sont toujours jugees avec un criterium distinct, on pretend qu’il n’y a pas de raison pour separer les Nesodontidce des Toxodontidcz ; on en dira certainement autant des Haplodontheriidce, et on en fera probable- men t trois sous-familles de celle de Toxodontidcz ” (’07, 89). This complaint necessitates some remarks upon the significance of the family as applied to extinct organisms. Broadly speaking, there are among palaeontologists two contrasted methods of using the family group. One method, that of Cope, for example, is to apply the Linnaean concep- tion to the fossil forms and to treat the family purely as a matter of defi- nition, the fauna of each horizon being classified without reference to preceding or succeeding groups. The other method, first suggested, I believe, by Schlosser (’86) is to regard the family as a phylogenetic series and to include in it not merely the main stem of the series, but also such side-branches as are not themselves so distinct and so widely ramified as to constitute other families. Just how large and important a side-branch SCOTT : TOXODONTA OF THE SANTA CRUZ BEDS. I 1 7 shall be to demand recognition as a distinct family is, of course, largely a matter of individual judgment, and perfect agreement on this point is hardly to be expected. If classification is to be the concrete expression of genetic relationship, the second method is clearly the better and more logical one. For example, all of the horses, from the lower Eocene to the present, are by this method included in the Equidae, both the genera which can be placed in the main line of descent and such side-branches as Anchitherium, Hippidium, etc., as are given off at different levels from the principal stem. On the other hand, if, as is probably true, the Palae- otheriidae are derivatives of the same stock, they diverge and ramify so much as to form a separate family. An even more instructive example is that of the rhinoceroses, of which Osborn (’ i o, 557-8) recognizes two families, the Hyracodontidae and Rhinocerotidae, and divides the latter into four subfamilies. Personally, I prefer the classification which includes all of the rhinoceroses in one family, dividing this into three subfamilies, for the true rhinoceroses, the cursorial Hyracodonts and the presumably aquatic Amynodonts respec- tively, but Osborn’s arrangement will suffice for the comparison. The family of the Rhinocerotidae includes much more diversified forms than are known among the Toxodonta. Whether we consider the dentition, the skull, or the feet, we find far greater differences between such genera as the Oligocene Trigomas, the Miocene Teleoceras and the Pleistocene Elasmotherium than can be found among the known Toxodonta. To this single family are referred animals with and without horns, with single and with paired nasal horns, with frontal horns present or absent, with and without incisors and canines, with grinding teeth brachyodont and comparatively simple, or hypsodont, cement-covered and highly complex, with tetradactyl or tridactyl feet, which are either long or slender, or short and extremely heavy. Comparing the rhinoceroses with the toxodonts, I can see no valid reason for making more than one family for the Santa Cruzian and later members of the suborder. From the foregoing considerations the Notohippidae have been excluded, because I have had but little opportunity to examine the animals of this group and they are still incompletely known, but they are doubtless enti- tled to separate family rank. The same is true of the Leontiniidae of the Pyrotherium beds, which are so aberrant that they are usually included in the Entelonychia. I am confident, however, that this reference is erroneous. PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. I 18 NESODON Owen. (Plates XII-XXVI.) Nesodon Owen; Rep’t Brit. Ass. Adv. Science, 1846, p. 67. Toxodon Moreno (non Owen) ; Patagonia, resto de un continente hoy submerjido, Buenos Aires, 1882. Colpodon Burmeister (in part) ; Anales del Mus. Nac. de Buenos Aires, T. Ill, Entr., XIV, 1885, p. 161. A strapotherium Burmeister (in part) ; Descr. Phys. Repub. Argent., T. Ill, 1879, p. 517. Protoxodon Ameghino ; Observ. gener. sobre el orden de mamiferos esting. sud-amer. llamados Toxodontes, p. 62. La Plata, 1887. Atryptherium Ameghino; Enumeracion sistematica, etc., 1887, p. 18. Scopotherium Ameghino ; Ibid. Adelphotherium Ameghino; Ibid., p. 16. Gronotherium Ameghino ; Ibid., p. 17. Acrotherium Ameghino (in part) ; Ibid. Nesothevium Mercerat; Rev. del Museo de La Plata, T. I, 1891, p. 386. Among the Santa Cruz ungulates by far the commonest is the genus Nesodon , the remains of which occur in remarkable abundance. Almost all parts of the skeleton are fully represented in the collections, sacral and caudal vertebrae alone excepted, and even the successive stages of develop- ment, from earliest youth to extreme age, may be readily followed. As Ameghino has well shown (’91, 357; ’94^, 230), the animal undergoes remarkable changes of appearance during the course of development, and to these changes are largely attributable the many names which have been applied to it. Though not the largest of Santa Cruz mammals, none of which are gigantic, the species of Nesodon are among the larger and heavier forms and there is no great range of variation in size. Dentition (Pis. XIII, XVI-XIX). — Attention has repeatedly been called, especially by Ameghino and Lydekker, to the extraordinary changes in the character and appearance of the dentition, which in this genus take place in the course of individual development. Without a continuous series of skulls, representing all the steps of the transformation, one would hardly venture to suggest that the earlier and later stages were of the same animal. Normally, the number of teeth is still unreduced and the formula is that common to all the early groups of placentals, viz., If, Cf, SCOTT: TOXODONTA OF THE SANTA CRUZ BEDS. I 1 9 Pi, Ml. Supernumerary teeth occasionally appear among the premolars, and, in certain cases, some of the smaller teeth may be suppressed, but these appear to be abnormalities. A. Upper Jaw (Pis. XIII, XV, XVI, XVII, figs. 2, 3 ; XX, fig. 5).— The median incisor (i~) is a large tooth, which greatly changes in shape and appearance with age. When freshly erupted and unworn (PI. XVI, fig. 1) it is very broad and massive and the exposed portion is of uniform width; the masticating surface has an enamel-pit, or “mark” which, how- ever, is quite shallow, and its floor is irregularly pitted with numerous de- pressions, while the anterior and posterior borders are trenchant edges. On the anterior face the medial border is raised into a more or less prom- inent ridge and, external to this, is a broad, shallow concavity, while the outer moiety of the face is convex. Wear speedily obliterates the pit on the masticating surface and the concavity and convexity of the anterior face also soon disappear, for they have no great vertical extension, leaving that face gently convex, or nearly plane (PL XVII, fig. 2). For some time the base of this incisor remains open and the tooth continues to grow, the crown, in front view, having much the appearance of the scalpriform incisor of the rodents. Thick enamel covers the anterior face and is re- flected around the medial and external borders, but leaves the posterior face uncovered. With the formation of the root, which takes place when the animal is in early adult life, i1 assumes quite a different appearance (PI. XVI, figs. 3-6), narrowing to the base and having a more or less tri- angular anterior surface. As growth has now ceased, the tooth becomes smaller with advancing age, as the crown is worn down by use. When the first incisor is erupted and for some time after it has come into use, there is still no visible trace of i-, di- being still in place, and for a considerable period after its eruption i- is much smaller and less conspicuous than i- (PI. XVI, fig. 3) but it grows from a persistent pulp and throughout the whole life of the animal, or, at least, to extreme old age, eventually far surpassing i1 in size and especially in length. (Cf. PI. XVI, figs. 4-6.) It is these changes in the relative size of the median and second incisors which are the chief factor in the surprisingly altered appearance of the dentition in the successive stages of the animal’s devel- opment. I-, which thus becomes a formidable tusk, with considerable resemblance in shape to the lower canine of the boar, is recurved and sharply pointed and of trihedral cross-section, with the apex formed by 120 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. the rounded anterior border. Enamel is not present on the posterior face, which is bevelled by the abrasion of h, but covers the front and sides in a thick layer. Apparently, there is a sexual difference to be noted in the development of the tusk-like i-, which is heavier and more robust in some individuals than in others of similar age, and there is nothing in the skull-structure to indicate that this difference is specific rather than sexual. The first and second incisors are implanted in the same transverse line and form the front of the broad muzzle, but i- is inserted in the same fore- and-aft line as the premolars and is quite invisible from the front. This tooth is very much smaller than i1 or i- and of simple, irregularly style- like form, but, in the unworn condition, with enamel-pit on the cutting surface; it cannot have had much functional importance. To a varying degree, the enamel is absent from more or less of the inner face. A short diastema in front of and behind it gives to this small tooth an isolated position. The canine is slightly larger than i-, which it resembles in form and in the presence of an enamel pit on the masticating surface, while still un- worn ; it stands quite near p1- and, at first sight, appears to be one of the premolar series. This tooth also must have been functionally insignificant. The seven cheek-teeth (Pis. XIII, XV, XVII, figs 1-5) form a contin- uous series, progressively increasing in size from pL to m-, though all the premolars are of simpler pattern than the molars. These teeth are all strongly curved, with the convexity turned outward, as in Toxodon , and almost meet those of the opposite side in the median line of the palate. When first erupted, they have open bases and no roots, which are devel- oped later, first in the premolars and then successively in the molars, m- remaining rootless till a very advanced period. In freshly erupted and unabraded condition, the premolars all have a very deep central pit, or valley, completely enclosed by the external wall and a continuous ridge, which joins the outer wall in front and behind and bounds the valley anteriorly, internally and posteriorly. This valley is much complicated by spurs and pillars given off from the enclosing wall, their number and prominence differing in the successive teeth. In the first premolar, p1, which is the smallest and simplest of the series and is implanted by a single root, there is no spur, but the valley is in two portions, an anterior and much deeper part and a posterior, shallower part. When the tooth SCOTT: TOXODONTA OF THE SANTA CRUZ BEDS. I 2 I has been moderately abraded, the two portions of the valley are converted into two separate enamel lakes, of which the anterior one is the larger and deeper, and the posterior one speedily disappears altogether. In the fully adult animal the anterior lake is also worn away, leaving a smooth surface of dentine partially enclosed in a wall of enamel. The second premolar is considerably larger than p1 and is inserted by two roots ; it has a prominent spur from the external wall, which divides the valley into two parts, and a second and much smaller spur projects into the valley from the posterior part of the enclosing ridge. Behind this ridge is a second valley, which is smaller and much shallower than the principal one and has several small, circular enamel-pits in its floor, and at the antero-internal angle of the crown is an enamel-lined pocket. The third premolar resembles the second, but is somewhat larger, and the spur which projects forward from the posterior ridge is better developed. After a short period of wear, in both p- and p®, the principal valley becomes a narrow, antero-posteriorly directed, enamel-lined slit, and the posterior, shallower valley is obliterated, but the pits remain for a time as four or five very small enamel-lakes, and the antero-internal pocket persists as a lake. The minute lakes of the posterior border soon disappear, but the central valley and the anterior lake remain until the animal may be called old. The fourth premolar, which is the last of the permanent teeth to come into use, is of a pattern similar to that of p- and p®, but slightly more complicated ; the crista or spur from the external wall is more prominent, and that from the posterior crest is divided into three or more ridges, while the posterior valley is divided into two well-defined parts, each with two or more pits in its floor; the antero-internal pocket is smaller than in p®. In quite an advanced state of wear, p- differs from the preceding premolars and has a certain resemblance to a worn molar in the retention of the crista and the consequent Y-shaped valley. There is, however, no real difference of structure and the dissimilar appearance is largely due to the very late eruption of p-, so that, at a given stage, it has suffered less abrasion than the other premolars. The first of the premolar series has an irregularly oval grinding sur- face, while in the others this surface is approximately square, but, as the crowns contract toward the base, these teeth become smaller through abra- sion, as age advances. The external wall of p--p- shows, more or less 122 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. obscurely, a division into two nearly equal lobes by a shallow median groove. The upper molars are entirely different from the premolars both in shape and pattern, being, in the fresh or little worn state, elongate antero-pos- teriorly and narrow transversely and having an incompletely lozenge- shaped masticating surface. The angle between the anterior and external walls is quite acute and overlaps and projects prominently external to the tooth in front, giving to the molars the appearance of being set en echelon , while the outer faces of the premolars all lie in nearly the same vertical plane. The relative proportions of the three molars and the shape of their crowns change much with age and wear. When first brought into use, m- appears to be much smaller than m-, which, in turn, is consider- ably larger than m1, but, as m 1 greatly enlarges in antero-posterior diam- eter toward the base, this tooth becomes both relatively and absolutely Fig. 17. Nesodon imbricatus : Last upper molar (m-) of the right side, much worn, external view, Xi- E mel band white, dentine dark. Fig. 18. The same, crown view, Xy. Enamel bands with double contour. larger with advancing age, until it is much the largest of all the grinding teeth, though in very old animals, when m1 begins to develop roots, the crown grows smaller toward the base and diminishes in size with abrasion. The external wall in the molars shows no division into lobes and is quite smooth, but there is reason to think that the posterior lobe is considerably larger than the anterior. The enamel does not cover the entire crown, but is restricted to vertical bands, the extent of which varies in the different teeth and also changes scott: toxodonta OF THE SANTA CRUZ beds. 123 with the degree of abrasion. In nearly all of the grinding teeth the enamel covers the external wall and is reflected over upon the anterior and pos- terior faces, but covers them only partially. On the inner side the band of enamel is narrower. The much abraded m-, shown in Text-figs. 17, 18, has the posterior half of the crown bare of enamel and in m^ and m^ there is none on the inner side. In the unworn condition, the essential pattern of the molars may be readily made out, but the homologies of some of the elements are far from satisfactory determination. From the external wall run two principal transverse crests, of which the anterior one forms the front wall of the crown and the posterior one arises about midway in the fore-and-aft length of the outer wall. These two crests are separated by a broad valley, which narrows upward along the inner face of the tooth, until the two crests meet and coalesce, closing the entrance to the valley. From the external wall, nearly half-way between the two transverse crests, arises a very prominent spur or crista, which divides the valley into two parts and is much thicker and more prominent than in the premolars. Another spur projects from the outer wall behind the posterior crest, which it joins, enclosing a deep pocket, partially open on the inner side. It might, how- ever, be more accurate to regard this second spur as the posterior crest and to say that the valley is divided into three parts by two parallel spurs. There is a difference between various teeth in this respect ; in some, the third of the transverse ridges appears to be the posterior crest, or metaloph, in others the fourth. Still a third spur arises much higher up on the inner face of the external wall and encloses a second posterior pocket. With the progress of abrasion, the appearance of the molars becomes greatly changed ; the cross-crests, which in early stages are nearly trans- verse, grow more and more oblique, the principal valley is closed on the inner side by the junction of the transverse crests, but long retains its Y-shape, owing to the prominence of the spur which divides it, while the two pockets are converted into lakes, of which the posterior, being larger and deeper, persists longer, but is eventually obliterated ; in very old teeth it even disappears in m-. In the fully adult animal, with moder- ately worn teeth, there is a certain resemblance to the molar pattern of the perissodactyls, especially of the rhinoceroses, but the unworn teeth show that this resemblance is entirely superficial and without taxonomic significance. 124 PATAGONIAN EXPEDITIONS PALEONTOLOGY. B. Lower Jaw (Pis. XIII, XVIII, fig. 4). — As in the case of the upper incisors, those of the mandible greatly change their shape and appearance with age, though the changes are hardly so striking. All of the lower incisors are more or less strongly procumbent, but there is a considerable variation in the degree of procumbency among the individuals of the same species, though it never approximates the condition seen in Toxodon. When first erupted, iT is a broad, much depressed, scalpriform tooth, con- tracting gradually to the base and of trihedral cross-section, with thickened medial border, which forms the base of a narrow triangle, and thinning toward the external border, the apex of the triangle. When the root has been formed and growth ceases, abrasion causes the tooth to become nar- rower and thicker and assume a more definitely trihedral form. The second incisor is similar, but somewhat broader, and under- goes much the same changes of form. The third, on the other hand, grows throughout life from a persistent pulp and in the adult is a large and characteristic tusk, which bites behind the upper tusk and abrades its posterior surface, while i-3 itself is obliquely truncated by the wear. In this stage of development the tooth points obliquely upward and forward and its principal diameter is antero-posterior ; in shape, it is trihedral, with the base of the triangle formed by the antero-inferior border and the apex by the postero-superior. I3 is not erupted until after iy and i ^ have been in use for some time, and at first appears as a thin, depressed and flattened tooth, with rounded cutting edge and smaller in every dimension than h ; its principal diameter is transverse. It is the per- sistent growth of the tusk-like incisors (i- and i-3) with the diminution ol the others by abrasion, that so remarkably changes the appearance of the animal in passing from the youthful to the fully adult condition. In the stage in which i-3 is just erupted, all the incisors are closely crowded together and are arranged in imbricating fashion, each tooth being extensively overlapped by the one external to it. In older stages this overlapping is much reduced, partly by the growth of the jaw and widen- ing of the symphysis, partly by the narrowing of iy and 2, as they are worn down. The canine is a small tooth, of little or no functional importance, which is inserted somewhat internal to h and is usually in contact with the first premolar ; it has a narrow, compressed crown, with irregularly rounded, trenchant edge, convex on both internal and external sides, except that there is a shallow depression on the anterior portion of the inner surface. SCOTT: TOXODONTA OF THE SANTA CRUZ BEDS. 125 The first premolar is a small, single-rooted tooth, with compressed- conical crown, but is complicated by a narrow valley on the posterior side, enclosed between two ridges which run down from the apex of the crown. The remaining lower premolars (p2, 3, t) have essentially the pattern of the molars, with some differences, but develop roots at a much earlier stage and therefore have less decidedly hypsodont crowns. On the ex- ternal side is a vertical groove, running down to the base of the crown and dividing it into two lobes, which are of crescentic form, with valleys on the inner side, giving the bicrescentic pattern so very frequent among primitive ungulates of many different groups. These internal valleys, however, are shallow and have but small vertical extension and are soon obliterated by wear. The second premolar, the simplest of the last three, has imperfectly formed crescents, especially the anterior one, and there is no spur, or pillar, in the posterior crescent. On the external face, the groove is broad and open and the two lobes are of nearly equal size. In the third premolar the external valley is a narrow groove, and the posterior lobe is distinctly larger than the anterior ; the posterior horn of the anterior crescent projects obliquely backward as a spur and encloses a deep fossa with a prominent spur in the valley of the posterior crescent, and there is also a deep pit in the posterior valley. At a very early stage of wear, even before the eruption of pT, the inner opening of both these pits is obliterated and the pits are converted into small enamel lakes. The fourth premolar is like the third, but the posterior lobe exceeds the anterior in size to a greater degree, though much less than in the molars. The lower molars increase progressively in size from mi to nn, the pre- dominance of the latter growing more marked with age, as in the case of the corresponding upper tooth, but even more strikingly. These teeth are all decidedly hypsodont and do not form roots until the animal is beginning to grow old. They have the bicrescentic pattern already described in the premolars, but the difference in size between the two lobes is much greater than in the premolars, the posterior crescent being far larger than the anterior, and the disproportion attains its maximum in mi. The valleys on the inner side of the crescents have a much greater vertical extent than in the premolars and the posterior valley much more than the anterior. The spur in the posterior valley is sepa- rated by a deep cleft from the hinder horn of the anterior crescent and is itself marked by a deep pit and there is a similar, but larger and deeper, 126 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. pit in the floor of the posterior valley. At a certain stage in the abrasion of each molar, but not more than one at the same time, there are three enamel-lakes in the masticating surface ; of these, the anterior one is formed by the isolation of the slit between the posterior horn of the an- terior crescent and the spur of the posterior valley, the second by the fossa in that spur itself, and the third by the pit in the floor of the pos- terior valley. The small differences of structure between the molars and premolars thus become clear, only when several stages of growth and abrasion are available for comparison. The last molar (mj) is remarkable for the great antero-posterior extension and transverse narrowness of the hinder lobe. As the posterior border of the tooth slopes strongly back- ward and downward, the antero-posterior length of the crown increases as the tooth is worn down and begins to decrease only in old individuals. Milk Dentition (Pis. XVII, fig. i ; XVIII, fig. 6 ; XIX, figs. 1-3, 7-8 ; XXI, figs. 1-2). — This temporary series of teeth is complete, all of the antemolars having predecessors in it, and there are some striking differ- ences between the corresponding teeth of the two series, especially among the incisors, where the absence of tusks from the milk-teeth makes a great difference in the appearance of the skull and jaws, a difference which has led to the creation of many synonyms. A. Upper Jaw. — The first incisor (di1) is much the largest of the series and, while still quite unworn, considerably resembles the permanent i1, but is, of course, actually smaller and proportionally thinner antero-posteriorly. When the root has been formed and the crown considerably worn, the appearance of the tooth is much changed, as it then has a low and very broad crown, contracting abruptly into the long and slender root. The second incisor (di-) is not placed in the same transverse line with di— , as are i- and i-, but behind it and in line with the cheek-teeth. It is at first much smaller than di1 and of quite a different shape, being trihedral with rounded angles, except the postero-external one, which is a sharp ridge. In the more advanced and rooted stage the crown has become much broader, resembling that of di1, but smaller. The third incisor (di— ) is the smallest of the series and has a simple, laterally compressed crown, the principal diameter of which is antero-posterior. The canine resembles di- in form, but is somewhat larger ; the temporary incisors and canine all have a shallow enamel-pit, or mark, on the cutting surface, but this is soon obliterated by wear. scott: toxodonta OF THE SANTA CRUZ beds. 127 The milk-premolars are, generally speaking, molariform, with some dif- ferences, especially in the anterior ones. The first (dp1) is so early shed and replaced, that it is difficult to find examples of it and I have seen no unworn specimens. As ordinarily found, it has a small subquadrate crown, with two external lobes, and an enamel-lake on the grinding surface, the remnant of the internal valley. The second (dp—) is much larger in both diameters and has two equal external lobes, each with a prominent median rib. The valley is T-shaped, directed longitudinally, with a narrow, slit- like opening on the inner side ; three small spurs invade the valley, two from the posterior and one from the external wall. Posterior to the main valley is a fossette, with three small, circular pits in its floor, which, on abrasion, give rise to three minute lakes, a character of the permanent premolars. The special peculiarity of this tooth, in addition to the presence of the posterior fossette and its pits, is the large size and anterior prolongation of the postero-internal cusp (tetartocone), which extends for- ward almost to a contact with the antero-internal (deuterocone). The third milk-premolar (dp—) is still larger than dp-, especially in the antero-posterior diameter, and is almost exactly like a molar, except for the presence of a prominent median rib on the antero-external lobe. The fourth (dp-) is the largest of the temporary series and, like dp-, has a median rib on the antero-external lobe, but it is less prominent' ; otherwise it is like a molar. Dp- and - are set en Echelon, like the true molars, the antero-external angle of each overlapping the preceding tooth. B. Lower Jaw. — The incisors are broad, scalpriform teeth, which in- crease in size from diT to dig and, when freshly erupted, with a median ridge on the postero-superior face ; when worn and rooted, diT has con- siderable resemblance to a human incisor. These teeth are thin and an- tero-posteriorly compressed and dix and x are much less distinctly tri- hedral than their permanent successors : dig does not become a tusk, but develops a root and ceases growth very early, and is abraded at the tip like the other incisors, not obliquely truncated as is ix. All the milk-in- cisors are somewhat procumbent, but less decidedly so than the perma- nent ones, and each one at first extensively overlaps the tooth in front of and internal to it, but in older stages this overlapping is much reduced, as the crowns are narrowed by wear. The canine is in close juxtaposi- tion to di¥, which it somewhat resembles in form, but is very much smaller. 128 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. Owing to the very early replacement of dpT by px, I have seen no ex- ample of the former. The second milk-premolar resembles px, but in N. imbricatus a small separate pillar stands at the entrance of the anterior valley and a spur from the anterior crescent almost closes the posterior valley, which, when well-worn, becomes a lake in the grinding surface of the posterior crescent. The third and fourth of the series (dpx and T) resemble molars except in their lower crowns and the earlier formation of roots. Pre-lacteal Dentition. — Ameghino has announced (’02, 80; ’04, 10) the exceedingly surprising discovery that in Nesodon and Adinotherium a third functional dentition, the pre-lacteal, is present. Traces of such a dentition had already been observed in the embryos of certain mam- mals, but no case was known in which these teeth were fully developed and functional, and to find them in so advanced and specialized a group as the toxodonts and at so late a geological date as the Santa Cruz epoch, is indeed most extraordinary. According to Ameghino’s latest account: “ L’avant-premiere serie des Nesodontides est constitute par trois incisives, une canine et trois molaires de chaque c6te, qui sont remplacees par les memes dents de la premiere serie” (’04, 11). “Chez Nesodon et Adino- therium la denture definitive comprend le nombre complet de sept molaires, dont les quatres anterieures sont des rempla^antes et les trois dernieres des persistantes de la premiere serie. La premitre strie est constitute par les quatre caduques anttrieures et les trois persistantes posttriures, les sept molaires, restant en fonction en merne temps durant une certaine ptriode de la vie de ces animaux. A un age moins avanct, avant d’entrer en fonction la premitre persistante, la strie n’ttait constitute que par les quatre caduques. Les caduques, leur tour, ont ttt les remplacantes d’une strie anttrieure, Tavant-premitre strie qui ne comprenait que trois molaires correspondantes a la premitre, deuxitme et troisieme molaires. Done, les caduques 1 a 3 ttaient prtctdtes par les avant-caduques cor- respondantes mais la quatrieme caduque n’ttait pas prtctdte d’une avant- caduque ” (1. c., pp. 14-15). To this description I can add but little, as the material in the Princeton and New York collections exhibits hardly any remnant of the pre-lacteal series, but in one specimen each of Nesodon and Adinotherium (see text- fig. 39, p. 202) the pre-lacteal upper tusk has persisted alongside of the permanent i- and is much smaller than di~ and of entirely different shape, SCOTT: TOXODONTA OF THE SANTA CRUZ BEDS. 129 being a cylindrical, quill-like tooth. During my visit to La Plata, how- ever, I was convinced by the study of the materials in Dr. Ameghino’s collection that, so far at least as the mandibular incisors and canines are concerned, there were actually functional teeth of the pre-lacteal series in Nesodon. One mandible (PI. XIX, fig. 6), in particular, was especially instructive, as it showed the milk-incisors and canines in process of erup- tion and entirely unworn, with the roots of the pre-lacteal series still in place. Unfortunately, the photographs which I made of this and other specimens in a corresponding stage of development are not satisfactory for reproduction, but apparently the same individual is figured by Ameghino (op. cit., fig. 1, p. 12). This remarkable phenomenon loses something of its entirely exceptional character through Ameghino’s recently published observation of a pre- lacteal premolar in the tapir. The succession of the teeth in Nesodon may be followed out in much detail. The first of the permanent teeth to appear is mi, in both jaws, which becomes functional before any of the milk-teeth are shed. This is followed by pi, dpi being shed so early that it is rare to find an individual retaining it and, in some cases at least, the replacement takes place before mi comes into use. Next follows m2 and then dii is shed and replaced by ii, while di2 still continues in function. The second and third incisors, the canines, second and third premolars are replaced in rapid succession and before the last molar is erupted ; m3 then comes into use, and, last of all, dp4 is replaced. This account, which is drawn chiefly from the material in the Princeton collection, does not exactly agree with the description given by Ameghino (’94^, 233) according to which pi is not erupted until after m3 has ap- peared. No doubt, there is some individual variability in the order of succession. Ameghino (’91, 357 ff.; ’94^, 231-234) has summed up the individual changes in the dentition of Nesodon , dividing them into twelve stages, which are here reproduced in abbreviated form, with the addition of a stage for the prelacteal dentition. Stage A. — Prelacteal dentition of three incisors, a canine and three premolars, i- small, simple and styliform. 130 PATAGONIAN EXPEDITIONS: PALAEONTOLOGY. Stage i. — Milk-dentition of dif ; dc-f ; dpt. The dp are the first teeth erupted and the last are the external incisors and canines ; all the teeth in continuous series. Stage 2. — Milk-teeth in function and no permanent tooth yet visible. In the mandible the four internal incisors are of equal size and the two external ones a little larger, and all the teeth have open roots. The two rami of the mandible are in contact, but have not yet begun to ankylose. Stage j. — Mf in function, but mf not visible. The anterior part of the skull is elongated, producing short diastemata on each side of the upper canine, and the first upper incisors, diminished by wear, are separated by a space. Lower teeth still in continuous series ; roots are closed in diT-dpT and roots begin to appear in dp^- and ^ : the germs of most of the per- manent ante-molars are clearly visible, the two rami of the mandible are ankylosed, and the sagittal crest begins to appear. Stage 4. — Mi considerably worn and mf coming into use; milk-teeth much abraded and dpi and i with four separate roots ; germs of the per- manent incisors well developed and visible in the sides of the jaws. Germs of pf quite large and placed above dpx and below dpy. The masticating surface of all the teeth in use has the appearance of that belonging to an adult animal. The mandible described and figured by Owen as N. imbricatus is in this stage of dental development and was mistakenly believed by him to be an adult individual. Stage 5. — Mf well abraded; none of the milk-teeth yet replaced. Germ of m- small and placed in the upper part of the maxillary against the base of m- ; that of m3, also small, is placed in the base of the ascend- ing ramus ; germs of pf very small and hardly visible. This stage cor- responds to Nesodon as defined by Mercerat. Stage 6. — Germs of mf quite large, but not yet erupted; replacement of incisors begins : germs of pf very small. Stage 7. — Mf erupted, but not yet worn; germs of pi larger, dc, dpi-! shed and replaced. Stage 8. — Mf enters into use; dpi replaced by pf. Stage 9. — All the permanent cheek-teeth in use and have open bases. I1 much broader and thicker than i~, which is thin and pointed and has its tip nearly on a level with that of i-. Cheek-teeth gradually increasing in size ; m- with relatively small grinding surface, and mi but slightly larger than mi; upper molars with cross-section a parallelogram. This stage corresponds to Adelphotherium Mercerat. SCOTT: TOXODONTA OF THE SANTA CRUZ BEDS. 1 3 1 Stage 10. — Animal has attained its full development. I1 has decreased in size by abrasion, while i~, growing from a persistent pulp, is longer and much heavier than i1. Mi is a little larger than m^, a much more marked difference than in the preceding stage. Upper molars decreased in antero-posterior diameter through abrasion and of more square shape ; roots begin to appear on the molars and those of the premolars to close. Sagittal crest high and thin. This stage corresponds to the genus Neso- t her him of Mercerat. Stage 11. — This stage is well distinguished by the relatively enormous size of m- in proportion to the other teeth ; all the molars have the antero- posterior diameter reduced by abrasion and the upper molars have a cross- section of trapezium-shape ; all the molars, except m-, have well-formed roots ; mi has one long root, but open and undivided. Sagittal crest very high and compressed. This stage corresponds to the genus Protoxodon as defined by Mercerat. Stage 12. — Mi is proportionally very large and the other teeth much smaller ; at a still more advanced age m- forms several roots. The Pampean genus Toxodon has a dentition which may be described as that of Nesodon somewhat reduced in number and considerably simpli- fied in pattern, but thoroughly hypsodont, all the teeth growing from per- sistent pulps. Lydekker’s brief statement may be here quoted with advantage: “This genus ... is characterized by a somewhat Rodent like dentition, in which all the teeth grow from persistent pulps, the normal dental formula in the adult being if, cf, pf,* mf. The first upper premolar may, however, frequently be wanting in the adult, and occasion- ally the corresponding lower tooth is present. The upper molars are characterized by their height and decidedly triangular form, the middle lobe being but slightly developed and almost disappearing in an advanced state of wear, while the main cleft on the inner side continues to the base of the crown even in the most advanced state of wear and there is no posterior valley. The lower molars are very narrow, with three internal folds, the third tooth of this series being much larger than either of the others. In both jaws the premolars are much simpler than the molars. Both pairs of upper incisors are scalpriform, and while the first and second * Lydekker has z/i, an obvious misprint, as is shown by the statement immediately following. 132 PATAGONIAN EXPEDITIONS : PALEONTOLOGY. lower incisors are flattened and proclivous, the third is triangular and raised above the level of the other two. All three are however worn to nearly the same transverse line by the upper teeth and their alveoli are nearly in the plane of the inferior border of the jaw” (’93, 13, 14). While in general character the dentition of Toxodon is thus of the same type as that of Nesodon , there are several significant differences which should be emphasized. (1) The teeth are all conspicuously hypsodont and grow throughout life. (2) I1 is thin antero-posteriorly, broad and scalpriform, while i- is more trihedral and tusk-like. There is a marked difference in the relative size of these teeth in the different species ; in T. ftlatensis, for example, i- is much longer and more tusk-like than i but in T. burmeisteri these proportions are reversed. (3) The third upper incisor and the canine have disappeared, and there is a long diastema between the incisors and the cheek-teeth. (4) The lower incisors are much more strongly procumbent than in Nesodon , pointing almost directly forward and with their inferior faces in nearly the same plane as the ven- tral side of the symphysis. It and 2 are much broader and more flattened than in Nesodon , but i-3 retains something of its trihedral and tusk-like form, though worn to the same transverse line as the others. (5) The premolars are reduced in number and in relative size and importance, and are even less molariform than in Nesodon. (6) The upper molars are more triangular and have a much simpler pattern than those of Nesodon; the posterior valley has disappeared and only the principal spur, which projects into the main valley from the external wall, is retained. In the genus Xotodon even this spur has been lost. The Skull of Nesodon (Pis. XIII-XV; XVIII, fig. 1) is remarkable in many ways and strikingly different from the typical ungulate skull, but closely similar to that of Toxodon ; its most characteristic peculiarity is in the remarkable structure of the auditory region, which Roth (’03) was the first to point out and has so strongly emphasized. The upper profile of the skull may be nearly straight from the occiput to the end of the nasals, or may rise quite steeply from the forehead to the occiput, a difference which does not appear to be of specific value. The orbits are well forward, the anterior rim being in advance of the middle of the skull, making the cranial region considerably longer than the facial, but the brain-case proper is short, narrow and of small capacity. The great development and spread of the zygomatic arches give to the cranium a width which scott: toxodonta OF THE SANTA CRUZ beds. 133 is deceptive as to its actual capacity. The sagittal crest is thin and sharp and is often divided by a deep, narrow groove into two closely approximated parallel ridges. Though much longer proportionately than in Toxodon , the crest is yet quite short and divides anteriorly into the supraciliary ridges, which die away upon the forehead without reaching the orbital border. The orbit has a more inferior, as well as a more anterior, position than in Toxodon , the roof not forming a protuberance above the level of the forehead. The facial region is short, but very deep dorso-ventrally, especially when the mandible is in position ; in fact, the whole skull is curiously short, deep and massive in proportion to the size of the animal. The anterior nares are rather small and terminal in position, presenting almost directly forward, though with a slight inclination up- ward and backward, one of the most marked differences from the skull of Toxodon. The posterior surface of the skull, which can hardly, with propriety, be called the occiput, as it is so largely made up of elements which are not occipital, is low, very broad and almost D-shaped, with lateral and supe- rior borders describing a continuous and nearly semicircular curve. The occiput proper is of quite a different shape ; it is very broad at the base, forming here nearly the whole width of the posterior surface of the skull, but above the foramen magnum it is sharply constricted, expanding again, though but moderately, to the summit of the inion. The large area on each side which is thus left open, is filled by the element which Roth calls (’03) the pars mastoidea of the petrosal, but for reasons which will be explained in a subsequent section, is here regarded as being more probably the post-tympanic portion of the squamosal ; the relations of the parts are apparently much as in the pig. The upper view of the skull (PI. XIV) is very peculiar. From this point of view the outline is that of a somewhat irregular, truncated tri- angle, with the occiput forming the base, though the greatest width is over the zygomatic arches. The downward and backward slope of the occipital surface makes much of that surface visible and even the sessile condyles are prominently conspicuous. A highly characteristic feature is the great prominence of the auditory meatus and post-tympanic process at the supero-external angles of the occiput, and the elevated position of the zygomatic arches where their dorsal borders pass into the lambdoidal crest. 134 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. The brain-case is very narrow, making the temporal openings very large, giving the skull an almost reptilian appearance and allowing the pterygoid processes of the alisphenoids to be plainly visible. The forehead broadens greatly and abruptly in front of the postorbital constriction, and the face narrows very gradually forward from the orbits, though widening again very slightly at the muzzle. Another very characteristic feature of this skull is the manner in which the facial region narrows dorsally, the narrow and convex nasals forming the whole upper surface in front of the orbits. The solid and massive premaxillaries project for a considerable distance in front of the nasals, but much less than in Toxodon , in which the nasals are greatly reduced in length. The base-view (PI. XV) of course corresponds closely with the upper view in the form of the outline, but shows somewhat more distinctly that the widening of the muzzle, which is very slight, and, indeed, hardly per- ceptible, begins at the anterior end of the maxillaries, at the insertion of the canines. This widening is very much greater and more abrupt in Toxo- don:, and even in the Santa Cruz genus Adinotherium it is much more decided than in Nesodon . This view also shows clearly the peculiar form of the posterior nares, the prominent lateral projection of the pterygoid processes of the alisphenoids and the great breadth of the auditory region, on each side, external to the paroccipital process. The mammillate tym- panic bullae appear unduly small, as their principal diameter is dorso- ventral. The broken stylo-hyals are also visible, attached to the anterior side of the bullae, a highly exceptional arrangement. The basi-occipital is quite long and stout, though narrowing between the tympanic bullae and expanding again behind them, and bears a moder- ately prominent median keel ; it takes no part in the formation of the condyles, which are confined to the exoccipitals. The latter are very low and wide, but form relatively little of the occipital surface, except at the base, and having no crest. The foramen magnum is low and very broad, of transversly oval shape. The condyles also are low and broad, quite closely approximated ventrally, very widely separated dorsally. Though they are almost sessile, they project quite strongly backward, owing to the slope of the exoccipitals, and are fully visible when the skull is seen in side-view. Each exoccipital bears a very long paroccipital process which is heavy and trihedral proximally, becoming much more slender and rounded distally. The supra-occipital is very large and forms the greater scott: toxodonta OF THE SANTA CRUZ beds. 135 part of the occiput proper ; its dorsal border is a heavy and prominent crest, which is relatively short from side to side, as the supra-occipital does not expand greatly above the constriction of the exoccipitals. In the young skull the exposed surface of the supra-occipital is nearly plane, but in the adult animal it becomes quite deeply concave, in which stage the bone is very thick and massive, with cancellous interior, but without any distinct sinus. The basisphenoid is also long, passing forward into the posterior nares and tapering anteriorly ; near the middle of its course, on each side, is a massive projection for the attachment of the alisphenoid. Very little of the presphenoid is visible, as it is mostly covered and concealed by the pterygoids. The alisphenoid is very narrow antero-posteriorly, but has a very considerable dorso-ventral extent and sends down a long, narrow descending process, or pterygoid plate, which ends distally in a prominent knob. The limits of the orbitosphenoid are not distinctly shown in any skull that I have examined. The parietals are relatively small and especially narrow, forming but little of the side-walls of the cranium ; posteriorly, at the junction with the supraoccipital, they are very thick and massive, thinning forward, and have a diploetic structure, but I have seen no sinus in them. For most of their length, they carry the sagittal crest, but anteriorly the ends diverge, to receive between them the posterior extensions of the frontals. The cranial cavity, as seen in a longitudinally bisected skull, is enclosed almost entirely by the occipitals, sphenoids, parietals and squamosals, the frontals being but little involved. The cerebral fossa is small and ap- parently does not extend back over the cerebellar fossa ; the olfactory fossae are surprisingly small in view of the otherwise primitive character of the brain. Nothing can be learned from the material at my disposal regarding the number and course of the cerebral convolutions. The squamosals are very large, forming much the greater portion of the side-walls of the cranium and even part of the floor, where the bone is incurved internal to the glenoid cavity. This cavity is a saddle-shaped surface, narrow and convex antero-posteriorly, broad and slightly con- cave transversely; its outer end forms a prominent protuberance, much as in the horse, but larger and more conspicuous. Separated from the cavity by a notch, which is very broad externally and narrows inward, is the large and swollen-looking postglenoid process, which is filled with 1 36 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. spongy bone, and, so far as I have observed, does not contain a sinus. This process is closely applied to a high thin plate of bone, which is the element usually called the mastoid portion of the periotic (the protuber- ancia petrosa of Roth) and is interposed between the postglenoid and the post-tympanic processes, an altogether exceptional position. The post-tym- panic portion, or pars Serrialis is a very remarkable and anomalous struc- ture ; it is extensively exposed on the posterior surface of the skull, of which it makes up a large part, larger than in the Typotheria or the pig. Its dorsal portion is inflated and contains a large, oval cavity, lying above the long, tubular auditory meatus ; at the antero-internal corner of this cavity is a small, tubular passage, which communicates with the cavity of the tympanic bulla through the inner end of the auditory meatus. The surface exposure of the post-tympanic varies considerably with the age of the animal ; in the very young skull it is relatively much less extensive and forms a smaller proportion of the posterior surface of the skull ; along the base, or ventral side, of the occiput it is completely over- lapped by the exoccipital and the paroccipital processes form the infero- external angles of the posterior surface, since, at this stage, there are no mastoid processes. In the adult skull, on the other hand, the post-tym- panic is much increased in relative size and, along the base of the occiput, projects well externally to the exoccipital ; the paroccipital processes are no longer at the infero-external angles of the occiput, the development of the mastoid processes displacing them from this position. The periotic, or petrosal, is relatively large and of the usual dense, shining texture. The internal surface has a very small and shallow fossa for the flocculus of the cerebellum and from the anterior end is given off a blunt, inwardly projecting and recurved hook. The internal auditory meatus is large and conspicuous and deep within it may be seen the divi- sion into two canals for the seventh and eighth cranial nerves respectively. On the external face the periotic is concave and is closely applied, but not ankylosed, to the large, hollow capsule of the post-tympanic, which it par- tially encloses. The mastoid portion (as it is here called, without dis- cussion of the general homologies of this element, the protuberancia petrosa of Roth) is, in the very young animal, very small and inconspicuous, for the postglenoid and post-tympanic processes are in contact distally and but very little of the mastoid is visible between them. As the skull grows, however, the mastoid portion increases rapidly, becoming a large plate in SCOTT: TOXODONTA OF THE SANTA CRUZ BEDS. 137 the adult, completely separating the postglenoid and post-tympanic pro- cesses and projecting freely below them, ending in a conspicuous, bluntly pointed mastoid process, of which there is no trace in the very young stage. Toward the mesial side it reaches and is closely applied to the tympanic bulla with which it may become fused. The tympanic, in the very young skull, is a large, oval, thin-walled and moderately inflated bulla, with long axis directed antero-posteriorly, in striking contrast to the small, scale-like tympanic of the Litopterna. In the adult skull the bulla has a different form, being more mamillate in shape, with the principal diameter dorso-ventral, somewhat as in the pig, but not in such an exaggerated degree, and throughout life it remains hollow and free from cancellous bone. The external auditory meatus is a very long tube, only the outer end of which is visible in the intact skull, and which passes upward, outward and backward to a very elevated ter- mination, which has nearly the same relative position as in Sus. This auditory region of the skull is, in Nesodon , quite different in its details from that seen in the small Typotheria of the Santa Cruz. Disre- garding the extraordinary specialization of this region in Pachyrukhos , the other and less extremely modified genera, such as Hegetotherium and Protypotherium , have a less extensively developed post-tympanic process, which occupies much less of the occipital surface and its cavity is filled with coarsely cancellous bone. The postglenoid and post-tympanic pro- cesses are more widely separated than in Nesodon and the space between them is filled by the tubular auditory meatus ; the mastoid is not visible externally and there is, of course, no mastoid process. The zygomatic process of the squamosal in Nesodon is moderately elongate, very deep dorso-ventrally, plate-like, thin and compressed later- ally ; its dorsal border is continuous with the lambdoidal crest and rises steeply upward and backward. The anterior end of the process is abruptly truncated and slightly convex, and forms little or none of the orbital boundary. The jugal is very large, especially in dorso-ventral diameter, but thin and laterally compressed ; the posterior border is excavated to receive the rounded anterior end of the zygomatic process and it extends far back beneath the latter, but not reaching to the glenoid cavity, as it does in the Santa Cruz Typotheria. Anteriorly, it rests upon the zygomatic process of the maxillary, extending upward to a con- tact with the lachrymal and excluding the maxillary from the rim of the 1 38 PATAGONIAN EXPEDITIONS: PALAEONTOLOGY. orbit, of which the jugal forms the whole inferior and most of the anterior border ; it has no postorbital process, or even angulation. As a whole, the zygomatic arch is long and heavy, though laterally compressed, somewhat as in the rhinoceros, and curves out boldly from the side ol the skull, enclosing a large temporal opening and forming a very high external border for the temporal fossa. A characteristic feature of the arch is the steep rise upward and backward to the junction with the occipital crest. The lachrymal, though not very large, is a very conspicuous bone, 01 elongate oval shape, ending above and below in a point. It forms the dorsal moiety of the anterior border of the orbit and bears a heavy, prominent spine ; the foramen is within the rim of the orbit. The frontals are short antero-posteriorly, but broad; their hinder ends are sharply contracted into narrow processes, which pass between the divergent anterior ends of the parietals and bear low, inconspicuous temporal ridges, which die away before reaching the orbital border. The forehead is broad and is usually plane longitudinally, but may have a short, abrupt descent over the orbits (see Lydekker, ’93, PL XIV, fig. 1). The postorbital processes are prominent and the frontals extend well out- ward to form a roof over the deeply set orbits. Anteriorly, the frontals usually have short, triangular nasal processes between the hinder ends of the nasal bones, but these processes may be absent (JV. marmoratus) . At the postorbital constriction a cross-section shows that the frontals are very thick and filled with a dense mass of cancellous bone, but anteriorly sinuses appear, extending over the orbits, though these sinuses are com- paratively small and cause no protuberance of the forehead. The nasals are long, quite broad and strongly convex from side to side, curving downward laterally, so that the suture with the maxillary and premaxillary is well below the upper contour. In the most abundant species, N. imbricatus, the nasals diverge posteriorly and their hinder ends are separated by the triangular processes of the frontals, each nasal termi- nating in a sharp point. In the much less common N marmoratus , ad- mitting that to be a distinct species, the nasals are in contact throughout their length and their hinder ends together form a convex curve, which is received into a corresponding emargination of the frontals. Anteriorly, the nasals have broad, regularly rounded tips and project for a very short distance in advance of the premaxillaries. scott: toxodonta OF THE SANTA CRUZ beds. 139 The premaxillaries are relatively very large ; the ascending ramus is a broad, thin plate, which forms an appreciable portion of the face and has a suture with the nasal of 4-6 cm. in length. Thus, the nasals and pre- maxillaries completely enclose the anterior nares and exclude the max- illaries. The narial opening is large, especially in dorso-ventral diam- eter, of cordate shape and presenting almost directly forward, though having a slight obliquity, due to the inclination of the ascending rami. The two premaxillae are in contact in quite a long median symphysis, but, so far as I have been able to observe, even in aged skulls, they are never coossified. At the antero-superior angle of this symphysis each bone has a blunt, rugose prominence of no great length or height, though together they form a conspicuous keel, which is much larger in Toxodon . The alveolar portion is broad and deep, to carry the great incisors, and the palatine processes are also large, especially in length, narrowing behind to triangular plates, which separate the palatine pro- cesses of the maxillaries. The incisive foramina, which are relatively very small, are completely separated from each other by the heavy spines. The maxillaries are very large and make up the greater part of the facial region ; the pre-orbital portion has great dorso-ventral depth and slopes upward and inward toward the median line, thus causing the dorsal narrowing of the face, which is so striking and characteristic a feature of this skull. In undistorted specimens this facial portion, above the alveolar process, is often quite deeply concave, especially in the antero-posterior direction. The alveolar portion is also very large, in correlation with the great size of the teeth, and forms a large mass beneath the orbit; posteriorly, this mass ends in a rounded, prominent border, which projects freely and is not continued backward by the pala- tines. The infraorbital canal is very short and its anterior opening is placed near the orbit over m-. The palatine processes of the maxillaries are long and rather narrow, though broadening posteriorly, and deeply concave, following the remarkable curvature of the grinding teeth. The palatines are short and take but little share in forming the bony palate, hardly extending to the line of m-; they are composed of two quite distinct portions, a broad anterior part, which forms the hindmost region of the bony palate, and the part which encloses the posterior nares. The two portions are demarcated by a very sharp constriction a little behind the last molars. On the sides of the posterior nares the palatines 140 PATAGONIAN EXPEDITIONS : PALEONTOLOGY. have unusually broad ventral surfaces and each ends behind in a promi- nent, everted and rugose knob, part of which is formed by the pterygoid process of the alisphenoid. The posterior border of the palatines is also very broad and is deeply impressed by the large pterygoid fossa. The pterygoids are narrow slips of bone, which proximally almost conceal the presphenoid and end distally in recurved and very prominent hamular processes. The posterior nares, which are entirely behind the teeth, form a small opening, remarkably so in proportion to the size of the skull. The whole posterior palatine region is highly characteristic in Nesodon and differs in several respects from that seen in the Santa Cruz Typo- theria, as figured by Sinclair. (Cf. this volume, Pis. I, fig. 3; III, fig. 3; V, fig. 21 ; VIII, fig. 18; X, fig. 3.) In the latter the palatines take a much larger share in the formation of the bony palate, to a degree which varies in the different genera, and the posterior nares are displaced farther backward by the production of the median suture between the two pala- tines, so that the opening presents more backward and less ventrally. The hard palate is thus one of the most curious features in the skull of Nesodon; it is broadest behind, narrowing forward to p1-, thence widening slightly to the space behind the anterior incisors (i1, -), and is remarkably concave transversely, which is chiefly due to the great prominence of the alveolar processes of the maxillaries and is much more strongly marked in the adult than in the young animal with brachyodont milk-teeth. The cranial foramina are, in several respects, like those of the suillines and also those of the Litopterna. (See Vol. VII, p. 3.) There is a large and conspicuous sphenorbital foramen, a small optic foramen and a large foramen lacerum anterius, the two latter in close juxtaposition, but no alisphenoid canal, foramen ovale or rotundum. There is thus no open- ing visible between the foramen lacerum anterius and the foramen lacerum medium, the latter a very large opening, which no doubt transmitted both the second and third branches of the trigeminal nerve, as well as the eustachian canal and the internal carotid canal ; there is no other carotid canal on the inner side of the tympanic bulla, such as occurs in the Santa Cruz Typotheria. The foramen lacerum posterius is likewise a large and irregular opening and the stylo-mastoid foramen is of unusual size. The glenoid foramen is a conspicuous opening between the postglenoid process and the mastoid portion of the periotic. The condylar foramen, which is rather small, is placed near the foramen lacerum posterius. scott: toxodonta of the santa cruz beds. i 4 i The mandible is large and heavy. The ascending ramus is high and rather broad, with regularly curved and rounded angle, the free border oi which is not thickened, as it is in the rhinoceroses, nor yet especially thin, as it is in the tapirs, and is quite flat on both sides. The condyle is ses- sile, broad transversely, narrow antero-posteriorly and convex in both directions ; on the posterior face of the internal half is a well-defined ar- ticular surface for the postglenoid process and from the inner end depends a broad, hook-like plate of bone, a very exceptional feature, the purpose of which is not obvious. The sigmoid notch is shallow and the coronoid process quite low, but yet rises above the level of the condyle ; the mas- seteric fossa is very obscurely marked. On the anterior border of the ascending ramus the linea obliqua externa is not well defined, but the in- ternal one is very prominent distally and encloses a deep fossa behind m-g. The horizontal ramus is long, deep dorso-ventrally and stout, though laterally compressed. In the very young animal the two rami are sepa- rate, but they early become indistinguishably fused in a long, deeply con- cave symphysis, which extends back nearly to the middle of px. The chin is narrow, rounded and inclined, rising steeply forward from the posterior end of the symphysis and thus very different from the flattened, procumbent chin of Toxodon. There are two mental foramina, one be- neath pT and the other under mx, while the inferior dental foramen has quite a low position. Little of the hyoid apparatus is preserved in connection with any of the skulls, but enough to show some very unusual features. The hyoid is attached to the antero-internal corner of the tympanic bulla. Flow very exceptional this arrangement is, may be made clear by the statement of Flower (’85, 144), who says, in speaking of the tympano-hyal : “it is always in relation to the hinder edge of the tympanic bone, generally more or less surrounded by it.” In Nesodon , on the contrary, a style-like bone is attached to the anterior edge and, in the adult, is ankylosed with it and seems to represent the coalesced tympano- and stylo-hyals. In one very young skull, with the milk-dentition still in use, a very small, cylin- drical tympano-hyal is firmly attached to the antero-internal corner of the bulla, but is so extremely short that it cannot possibly represent the long and prominent bone, which is unfortunately broken in all of the adult skulls. (See PI. XV.) In the young skull mentioned there is also a separate stylo-hyal, which is long, slender, laterally compressed 142 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. and moderately curved, with convexity behind ; the proximal end is simple, without expansion or process, but on the posterior side, near the distal end, is a low, roughened ridge. No other hyoid element has yet been found. It is not practicable for me to make any detailed comparison between the skulls of Nesodon and Toxodon , for of the latter only a cast is at my Fig. 19. Nesodon conspurcatus ?: Skull, left side, X j. American Museum of Natural History. disposal, and in this cast but few of the sutures are shown. Nevertheless* instructive results may be obtained from a general comparison. The skull of Toxodon is manifestly of the same type as that of Nesodon , but has undergone considerable modification and exhibits many differences. The pre-orbital or facial region of the skull is somewhat longer in proportion to the cranial region than in the Santa Cruz genus. The occiput is much higher in relation to its width, and is more deeply and uniformly con- cave, with the lambdoid crest projecting prominently backward ; the fora- men magnum is less depressed and the occipital condyles are much more widely separated, especially ventrally. The auditory region is very differ- ent in being less specialized and more as in Pvotypotherimn than in Ne- sodon and more like the young skull than the adult of the latter genus, but in this respect there is considerable variation in the Pampean genus. scott: toxodonta OF THE SANTA CRUZ beds. 143 The tubular auditory meatus, the aperture of which has the same elevated position as in Nesodon , is much more extensively exposed between the postglenoid and post-tympanic processes than in the latter. No part of the mastoid portion of the periotic would appear to be visible and, of Fig. 20. course, there is no mastoid process, the paroccipital processes forming the infero-external angles of the occipital surface, as in Protypotherium and the immature Nesodon. The sagittal crest is much shorter and widens very gradually into the broad forehead, which has a sudden, though short, descent to the nasals; The orbits have a more elevated position, their roof rising a little above the level of the forehead. The postorbital processes of the frontals are 144 PATAGONIAN EXPEDITIONS ! PALAEONTOLOGY. not so prominent, but there is a distinct postorbital angulation of the jugals. The zygomatic arches are of similar plate-like and laterally com- pressed character, but have a relatively greater dorso-ventral depth, especially the jugal portion ; posteriorly, the arches do not rise so high upon the lambdoid crest. The nasals are much shorter, giving an entirely different shape to the anterior nares, which are no longer terminal, but very oblique, presenting dorsally as much as forward. The two premaxillae are firmly coossified and the dorsal portion of their symphysis is raised into a prominent, mas- sive and rugose crest, which is much longer and more conspicuous than in Nesodon. The muzzle broadens anteriorly very much more strongly than in the latter, but the face, especially the edentulous part, is propor- tionately shallower dorso-ventrally. The hard palate is very much alike in the two genera, but in Toxodon it is narrower at the facial constriction and widens more posteriorly, and anteriorly much more. The posterior nares are placed farther back, as the palatines remain in contact in the median line for a greater distance. The mandible has a very different appearance in the two genera. In Toxodon the symphysis is much longer, extending back as far as and, though deeply concave, is shallower anteriorly and has no anterior wall, owing to the complete procumbency of the incisors. The jaw narrows forward to p4, and thence expands strongly, becoming very broad at the incisive alveolus, and there is no rise at the chin, the whole anterior region of the mandible being broad, depressed and with nearly flat ven- tral surface. The ascending ramus is very high and rather narrow and the curve of the angle rises more than in Nesodon , without the down- ward curvature seen in the latter. The sigmoid notch is almost obsolete and the coronoid low and weak, not rising to the level of the condyle. The latter projects internally more than in Nesodon , but has no such dependent, plate-like process from the inner end. Vertebral Column , Ribs and Sternum (PI. XII, fig. 2). — The vertebral formula is not definitely known, since no individual has yet been ob- tained with completely preserved back-bone. A careful examination of the available material leads me to the conclusion that the formula was nearly the same as in Toxodon , viz., C. 7, Th. 16-17, L. 4-5, though there is some reason to think that the lumbar region was relatively longer than in the Pampean genus, at least in certain individuals. scott: toxodonta OF THE SANTA CRUZ beds. 145 The neck is short and heavy, considerably shorter than the skull. The atlas (PI. XXIV, figs. 5, 6) is very short antero-posteriorly, but extremely wide transversely, owing to the great development of the transverse processes, and quite depressed dorso-ventrally, and even aside from the processes, the width is the greatest of the three diameters. The anterior cotyles are very broad, deeply concave, quite widely separated ventrally and even more so dorsally. The neural arch is stout and shows only a moderate transverse convexity and, on each side, it bears a large and deep depression, the inner end of which opens into the foramen for the first spinal nerve and extends to the atlanteo-diapophysial foramen for the ventral branch of the same nerve. The neural spine is quite prominent, thick and rugose. The inferior arch is very like the neural, but smoother and less convex, and on the hinder border has a hypapophysis which varies much in size, though the difference may be specific rather than indi- vidual ; in some specimens it is very prominent and heavy, in others it is much reduced. The neural canal is low and wide, of transversely oval shape and has prominent rugosities for the attachment of the transverse ligament ; ventral to these is the large, semicylindrical concavity for the odontoid process of the axis. The posterior cotyles are low, very wide, of irregularly oval shape and with outer margins raised and projecting quite prominently. The transverse processes are very long, but rather narrow from before backward and posteriorly are quite thick, being strengthened by a broad bar, which runs out from the inferior arch over the processes. The large vertebrarterial canal perforates the transverse process and opens anteriorly into a large fossa, into which also opens the atlanteo-diapophysial foramen. In Toxodon the atlas is very much like that of Nesodon , but is relatively longer antero-posteriorly and less depressed, with more strongly convex neural arch. The neural spine is rather more prominent and is bifid, and the transverse processes are broader antero-posteriorly and have more curved free borders. The foramina are the same, except that the pos- terior opening of the vertebrarterial canal is much smaller and has been displaced toward the median line. The posterior cotyles have a greater dorso-ventral diameter and a more irregular surface, with depressions at the mesial ends, which are hardly indicated in Nesodon. The axis (Pis. XXIII, fig. 8 ; XXIV, fig. 4) has a short centrum, which anteriorly is very broad and depressed, but grows narrower and thicker 146 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. behind the transverse processes, so that the posterior face is subcircular and slightly concave ; the ventral keel is very obscurely marked and the anterior cotyles for the atlas are very wide and low and feebly convex. The odontoid process is rather long, very stout and somewhat compressed laterally, giving a vertically oval section. The pedicles of the neural arch are narrow antero-posteriorly, but thick and heavy, and the neural canal is rather small. The neural spine is hatchet-shaped and not very large, projecting forward but little in advance of the pedicles of the arch ; the free border of the spine is thick and rugose, especially behind, where it forms quite a massive tuber. The postzygapophyses are large and prominent, presenting almost laterally, and are somewhat convex dorso-ventrally. The transverse processes are short, slender and thin, but diverge strongly from the sides of the centrum ; they are perforated by the very large canal for the vertebral artery. In Toxocion the axis differs from that of Nesodon in only a few details, except, of course, its very much greater size and massiveness. The an- terior cotyles are more saddle-shaped ; the odontoid process is relatively shorter and heavier; the neural canal is higher dorso-ventrally and the spine more massive ; the postzygapophyses are of a different shape, with the long diameter directed antero-posteriorly instead of dorso-ventrally, and presenting more obliquely downward ; the vertebrarterial canal is de- cidedly smaller. The third cervical has a short, heavy centrum and long, slender, de- pressed transverse processes, which are quite simple, showing no indica- tion of the inferior lamella, and are perforated at the base by the very large canal for the vertebral artery. The neural canal is quite small and the arch is broad and nearly flat on the dorsal surface. In all of the available specimens the neural spine is broken and its full height cannot be determined, but it evidently was unusually prominent. The prezyga- pophyses are large and present internally, with almost vertical faces, while the postzygapophyses present obliquely downward. The succeeding cervical vertebrae have short, heavy and slightly opis- thocoelous centra and prominent transverse processes. The inferior lamella first appears on the fourth vertebra, is considerably larger on the fifth and on the sixth becomes a great, hatchet-shaped plate, which is produced far posteriorly (PI. XXIV, fig. 3). In all, including the sixth, the vertebrarterial canal is very large and conspicuous ; the transverse scott: toxodonta OF THE SANTA CRUZ beds. 147 processes of the seventh are short, heavy and imperforate, but have dis- tinct inferior lamellae. The neural arches are quite broad antero-pos- teriorly, so that the intervertebral spaces are very narrow. The neural spines vary much in height, but, unfortunately, it is not practicable to determine whether these differences are to be regarded as specific. In some individuals the spines are short and weak, in others very much stouter and better developed ; they increase in length posteriorly and that of the seventh cervical is, in some instances, almost as long and heavy as that of the first thoracic. In Toxodon the cervical vertebrae posterior to the axis do not differ in any significant manner from those of Nesodon. They are considerably shorter proportionately and have very heavy centra ; the transverse pro- cesses resemble those of the Santa Cruz genus, but on the third vertebra they are broader, more compressed antero-posteriorly, curving more back- ward and less downward, and on the sixth the inferior lamella has no such antero-posterior extension as in Nesodon. The neural arches are much nar- rower, having considerable intervertebral spaces, and the neural canal is relatively much larger. The neural spines are shorter and much more slender than in some specimens of Nesodon , but in others there is not so great a difference from Toxodon in this respect. While not definitely known, the number of thoracic vertebrae in Neso- don could not have varied much from 16-17, forming a very long thorax. The first thoracic in some respects resembles the seventh cervical ; the centrum is somewhat longer and has a less decidedly convex anterior face than in the latter and its shape is changed by the addition of the very large anterior rib-facets ; the transverse processes are very promi- nent and massive and bear very large and deeply concave facets for the tubercles of the first pair of ribs ; behind each process is a deep groove for the passage of the spinal nerve. The prezygapophyses are, as usual, of the cervical type and the postzygapophyses, which are on the hinder part of the neural arch, present laterally, not ventrally, and are some- what convex. The neural canal is of triangular shape and notably small and the spine, the exact length of which cannot be determined from the material before me, is relatively much stouter and presumably much longer than in Toxodon. The succeeding vertebrae of the anterior thoracic region (PI. XXI, fig. 1 1 ) have short, heavy and slightly opisthocoelous centra, with large facets 148 PATAGONIAN EXPEDITIONS : PALEONTOLOGY, for the rib-heads, which extend upward upon the base of the trans- verse processes ; the latter are very prominent and heavy and carry large, deeply concave facets for the rib-tubercles. The second thoracic vertebra has prezygapophyses of the cervical type, arising from the dor- sal side of the transverse processes, but the postzygapophyses occupy the normal position on the ventral face of the overhanging neural arch, the pedicles of which are deeply incised for the passage of the spinal nerves, while the third and succeeding thoracic vertebrae have the pedicles per- forated by foramina for the transmission of the nerves. The neural spines of the first four or five thoracics are exceedingly elongate and have a strong backward inclination. The rapid diminution posteriorly in the length of the spines produces a decided hump at the shoulders, though this feature is less strongly marked than in Toxodon , in which it is very striking. In the middle thoracic region the vertebrae have centra which in size and form differ but little from those of the anterior region, though the facets for the heads of the ribs become smaller and less concave and have a more inferior position. It is perhaps unnecessary to say that the change is a gradual one. The transverse processes are shorter, but still massive, and bear much smaller and nearly plane facets for the rib tubercles, and prominent metapophyses arise from the dorsal border. The neural canal is remarkably small and the pedicles of the arch are perforated by foramina for the nerves ; the spines become much shorter and more slender and have a less decided backward inclination than in the anterior region. In the posterior thoracic region (PI. XXIII, fig. 9) the vertebrae gradu- ally take on the lumbar type ; the neural spines are short, broad and plate-like, with thickened and rugose tips. The anticlinal vertebra appears to be the antepenultimate thoracic and on the last two the spines have a slight forward inclination. The small neural canal and the per- forated pedicles of the arch are retained throughout the region. Only on the last two vertebrae are the zygapophyses like those of the lumbars and only imperfectly so, and the metapophyses are very prominent. There must have been at least four lumbar vertebrae (PI. XXIV, fig. 2) and in some individuals probably five ; the centra become broader and more depressed posteriorly and that of the last lumbar is very broad and low and is traversed by a pair of vascular canals, which open upon the scott: toxodonta OF THE SANTA CRUZ beds. 149 dorsal and ventral sides. The transverse processes are long, broad and depressed and extend straight out from the sides of the neural arch, with- out anterior or downward curvature, except on the last vertebra, which has quite peculiar transverse processes ; they are narrow and thick at the base, becoming wider and thinner toward the extremities and with a con- cave anterior border, which gives an appearance of a forward curvature ; on the posterior border of each process is a small facet for articulation with corresponding facets on the sacrum. The zygapophyses are of the cylindrical, interlocking kind, found in the Litopterna and Artiodactyla, the posterior pair being convex and the anterior strongly concave. From the outer margin of the latter arise the metapophyses, which are much more prominent than in the thoracic region. The neural canal remains small and the pedicles of the arch are not perforated for the nerves, which pass out through deep and narrow slits. The neural spines are rather short and weak and have a slight forward inclination. So far as I am aware, no sacral or caudal vertebrae of Nesodon have yet been found, but the sacrum is known in the very closely allied genus, Adinotherium, in which it consists of six vertebrae and the centrum of the last one is so large as to make it very probable that the tail was moder- ately elongate and stout. There is every reason to assume that the rela- tions were similar in Nesodon. The trunk-vertebrae of Toxodon differ in a number of details from those of Nesodon , differences which are for the most part referable to the greatly increased size and body weight of the Pampean genus. The vertebrae are of nearly the same relative length as in the Santa Cruz genus, but are much more massive. The first thoracic has a much shorter and more slender spine, but those of the four succeeding vertebrae are far longer and heavier than in Nesodon and the descent from them to the middle region is much more abrupt, making the hump at the shoulders far more promi- nent and conspicuous. In the posterior thoracic and lumbar regions the spines are low, but very broad antero-posteriorly and have greatly thick- ened and very rugose tips. All the neural spines of the trunk-vertebrae have a backward inclination, somewhat as in the Proboscidea, although chose of the last three thoracics are nearly erect. Throughout, the transverse processes are relatively shorter and the metapophyses some- what less prominent than in Nesodon. The ribs of Nesodon are not especially characteristic, but resemble those 150 PATAGONIAN EXPEDITIONS: PALAEONTOLOGY. of the medium-sized ungulates generally. The anterior ribs have larger heads, larger and more prominent tubercles, straighter, broader and more compressed shafts. In the middle region of the thorax the ribs are longer and thicker and have a stronger outward curvature, while the tubercles are smaller and much less prominent. Posteriorly, the ribs become much more slender and rod-like and all of the ribs are conspicuously more slender than those of Toxodon. As a whole, the thorax is long, deep and capacious. The sternum (PI. XXII, figs. 3, 4) is not completely preserved in any of the specimens, but what remains is very peculiar. The presternum, or manubrium, is long, narrow and laterally compressed ; the facets for the first pair of ribs form prominent projections and in front of these is a long, much compressed and keel-like process, somewhat as in the horses and tapirs, but more pointed and having a downward direction. Behind the rib-facets, the presternum is much broader on the dorsal face, narrowing ventrally to an edge or keel. Four of the mesosternal segments are pre- served, of which the first three are depressed and flattened, though very broad, the width and dorso-ventral thickness being nearly equal. The fourth segment is much narrower and more compressed, contracting toward the hinder end, and in this segment the thickness greatly ex- ceeds the breadth. In Toxodon the sternum is of similar type, but with considerable differ- ences of form ; the manubrium is relatively longer and the rib-facets are not prominences, but deep, irregular concavities. All of the presternum is more depressed and flattened than in Nesodon and the anterior process is broad and low, ending in a blunt point, and quite different from the sharply compressed and keel-like process of the Santa Cruz genus. The mesosternal segments are broader and more massive than in the latter and the fourth segment is not contracted. Appendicular Skeleton. — The scapula (PI. XXII, figs. 1, 2) is very large, of high, narrow and subquadrate shape ; the spine pursues a some- what oblique course, from above downward and backward, making the postscapular fossa much wider than the prescapular proximally, while dis- tally these proportions are reversed. The glenoid cavity is large and broadly oval in form and the neck is very wide, made so by the cora- coid, which, when seen from the outer side, appears to be merely a portion of the scapula, but in distal or internal view is shown to be a massive in- SCOTT : TOXODONTA OF THE SANTA CRUZ BEDS. 151 curved hook of unusual size. There is a distinct coraco-scapular notch, above which the blade widens abruptly. The coracoid border is gently convex and curves uninterruptedly into the suprascapular border, which Fig. 21. Toxodon burmeisteri : Left scapula, outer side, X T La Plata Museum. Fig. 22. Nesodon imbricatus : Left scapula, outer side, X f • is arched and joins the glenoid border by a pronounced angulation. The glenoid border, though slightly sinuous, is nearly straight and, above the neck, the blade is not far from being of uniform width throughout. The spine begins a little below the suprascapular border, rising gradually to its full i52 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. height, and is strongly recurved, so that the anterior surface is convex and the posterior concave. Distally, it ends in a short but distinct, over- hanging acromion, the spine here reaching its maximum height and prom- inence. There are two very large metacromial processes, which project prominently backward from the spine. Of these the upper, proximal one is smaller and the lower or distal one is much longer and is expanded at the free end ; both are quite thin and plate-like. In Toxodon the scapula differs from that of Nesodon in many respects ; the blade is somewhat narrower in proportion to its height and the coraco- scapular notch is less definitely marked ; the coracoid is reduced to a rugose nodule, which forms no incurved hook. The widening of the blade above the neck is more abrupt than in the Santa Cruz genus and the cora- coid border is nearly straight, while the glenoid border is convexly curved. The spine has no acromion, but dies away gradually upon the neck, and the metacromion is a broad, massive, roughened area, which projects backward but little behind the plane of the spine and is very obscurely divided into two parts. It is the change in the character of the spine and its processes which, more than any other single feature, makes the striking difference in the appearance of the scapula in the two genera. The humerus (PI. XXIII, figs. 3-5) is short and moderately heavy. The head is hemispherical and sessile and projects but little behind the plane of the shaft ; the external tuberosity is very large, rising high above the level of the head, but rather narrow, not extending across the entire proximal end and leaving the head partially visible from the front ; the internal tuberosity is broken away in all of the available specimens, but, judging from the analogy of the nearly allied Adinotherium, it was doubt- less very small. The division of the bicipital groove and the development of the bicipital tubercle appear to have already begun, part of the bicipital tubercle being plainly visible in one individual. The deltoid area is a mere roughened line and forms no process or hook. The shaft has the form common to nearly all of the heavier ungulates ; the proximal portion is laterally compressed and very thick antero-posteriorly, while the distal portion is antero-posteriorly compressed and broad transversely, and the middle region is subcylindrical. The supinator ridge is not strongly de- veloped, but the external epicondyle is massive and rugose ; the internal one is much smaller. The supratrochlear fossa is large and well-defined, though not very deep, while the anconeal fossa is small and profound. In scott: toxodonta OF THE SANTA CRUZ beds. 153 some individuals the two fossae communicate through a large perforation, but in others they are completely separated by a thin bony wall. The trochlea is wide and low and is divided into an external convexity for the head of the radius and a broad, saddle-shaped area for the ulna, which is bounded internally by a prominent flange. Fig. 23. Toxodon burmeisteri : Left humerus, dorsum, X T La Plata Museum. Fig. 24. Nesodon imbricatus : Left humerus, dorsum, xh In Toxodon the humerus has become enormously massive and is rela- tively very short ; the external tuberosity is proportionately lower than in Nesodon and is shifted more toward the outer side. The bicipital groove is divided into a very broad external and narrower internal portion and the former is again subdivided by a low, median bicipital tubercle. The deltoid area develops a low, but very massive and rugose, external hook. The supinator ridge is prominent and the external epicondyle very large and heavy ; the radial portion of the trochlea is high and narrow and i54 PATAGONIAN EXPEDITIONS : PALEONTOLOGY. more strongly convex than in Nesodon. In brief, the humerus of the Pampean genus is a far more massive bone and its processes for mus- cular and ligamentous attachment are better developed, heavier and more prominent than in the Santa Cruz animal, a change which corresponds to the general increase in size and bulk of the former. The fore-arm bones are always separate, not displaying the slightest tendency to coossification, and are crossed, the radius passing from the external to the internal side of the arm ; the two bones are in contact, not only at the proximal and distal ends, but also for some distance along the middle portion of the shafts, by means of interosseous crests. Thus, the radio-cubital arcade is divided into a shorter proximal portion, which is very narrow and slit-like, and a longer, wider distal portion. Relatively, the radius (PI. XXV, figs. 5-8) is elongate and quite slender ; the head is very small, and transversely oval in form and covers but a small part of the humeral trochlea ; the shape of the articular surface varies somewhat in different individuals, though possibly these differences should be regarded as specific rather than individual. In some specimens the surface is simply concave ; in others it is divided into a larger external concavity and a smaller, internal, saddle-shaped portion. The proximal facet for the ulna is a narrow band on the inner side of the head. On the external side of the head is an articular surface, to which is attached a large sesamoid bone, shaped like a small patella and apparently devel- oped in one of the extensor tendons of the digits. No bicipital tubercle is visible on the radius. The shaft is slender, irregular and with a forward curvature, increasing in size toward the distal end ; the interosseous crest is distinct and very rough, though of no great prominence, and runs for nearly the whole length of the shaft. The distal end is quite massive, both broad and thick ; the styloid process is long and prominent and con- tinues distally the facet for the scaphoid, which conspicuously notches the dorsal border of the distal end of the radius. The lunar surface is con- cave and is considerably deeper palmo-dorsally than that for the scaphoid. The distal facet for the ulna is very small and oblique, presenting proxi- mally almost as much as laterally. The ulna (PI. XXV, figs. 4, 5) is much heavier than the radius, except near the distal end. The olecranon, though long, projects but little behind the plane of the shaft, from which it slopes up very gradually, and the free end is only moderately thickened and rugose. The facets SCOTT : TOXODONTA OF THE SANTA CRUZ BEDS. 155 of the elbow-joint are very peculiar ; the sigmoid notch is deep, describ- ing nearly a semicircle, and the coronoid process is prominent. Aside from the facet for the anconeal fossa, the humeral surface is altogether internal — there is no .trace of any external facet, such as is found in almost all other ungulates. Thus, when viewed from the front or dorsal Fig. 25. Fig. 26. side, the articular surface pursues a very oblique course, from above downward and inward, and does not embrace the head of the radius, but merely touches it on the internal side. The shaft is heavy and of tri- hedral shape, with prominent interosseous crest on the anterior face, and tapers but slightly toward the distal end, though the very conspicuous styloid process is abruptly contracted ; it has a strongly convex facet for the pyramidal, which passes uninterruptedly on the palmar side into the surface for the pisiform. 156 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. In Toxodon the fore-arm bones are relatively shorter and very much more massive than in Nesodon, though of similar construction. The dis- proportion between the ulna and radius is even greater than in the Santa Cruz genus, the ulna having increased in thickness and weight more than the radius. The radius resembles that of Nesodon in form, but is pro- portionately shorter and stouter ; the distal end is especially massive and the scaphoid facet emarginates the dorsal border more deeply and in a more conspicuous way. The ulna is extraordinarily heavy ; the olecranon projects much more strongly behind the plane of the shaft and is far thicker at the free end than in Nesodon. The sigmoid notch is much the same in both genera, but the short shaft in Toxodon is not only very much heavier, but has acquired a subquadrate shape, somewhat as in the Proboscidea. The styloid process is relatively decidedly shorter than in Nesodon and is displaced to the extreme outer side of the distal end, which extends far internally, or radially, from it and thus gives quite a different appearance to the distal end. The manus (Pis. XXII, fig. 5 ; XXVI, figs. 4-6) is surprisingly small and weak in proportion to the size and weight of the animal. The carpus is very broad transversely and short proximo-distally and is already of completely interlocking type, as much so as in the Perisso- dactyla, the scaphoid articulating extensively with the magnum and the lunar resting almost equally upon the magnum and unciform. Aside from its much smaller size and lighter construction, this carpus is very similar to that of Toxodon , though there are many differences of detail, no doubt due to the readjustment necessitated by the greatly increased size and weight of the Pampean genus. The scaphoid is low, broad and irregular in form ; the dorsal face has the shape of an irregular, inverted and truncated triangle, broad proximally and narrowing distally ; the palmar face is narrow, strongly convex and rugose. Proximally, the articular surface for the radius is almost plane transversely and is made slightly concave palmo-dorsally by the elevation of the posterior or palmar border ; the whole proximal end is not covered by the radial facet, but a broad, rugose area extends around the palmar and internal sides. On the ulnar side is a facet for the lunar, which is continuous with that for the radius ; this lunar surface is very narrow proximo-distally, but extends for nearly the whole dorso-palmar diameter of the scaphoid and is slightly convex in that dimension. This is the SCOTT: TOXODONTA OF THE SANTA CRUZ BEDS. x57 only articulation between the scaphoid and lunar, for distally the two bones are not in contact at all. Distally, there are only two facets, for the trapezoid and magnum respectively, there is none for the trapezium ; Fig. 27. Fig. 28. Toxodoii burmeisteri: Left manus, dorsum, Nesodon imbricatus: Left manus, dorsum, X 5. La Plata Museum. X i- the trapezoid surface is quite large and deeply concave, while that for the magnum is narrow, oblique, nearly plane transversely and slightly con- cave palmo-dorsally. At this point the raised ulnar border of the mag- num is wedged in between the scaphoid and lunar, quite separating them. In Toxodon the scaphoid differs in several respects from that of Neso- don; being proportionately narrower transversely, more elongate proximo- distally and having a less oblique position in the carpus ; the dorsal face forms an irregular pentagon and the palmar protuberance has become much larger and more prominent. The proximal facet for the radius is more saddle-shaped, much more deeply concave palmo-dorsally, more convex transversely and covers the entire proximal end, which, as above mentioned, is not the case in Nesodon. The magnum facet is relatively wider and more directly distal and the trapezoid facet, which is very similar in form, does not extend quite so far proximally upon the radial side. 158 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. The lunar is very massive and larger than the scaphoid in almost every dimension, articulating with the latter only by a narrow band near the proximal end, a facet which is moderately concave in the dorso- palmar direction. The radial surface is in two connected portions, a broad anterior area, which is strongly convex palmo-dorsally and is reflected far over upon the dorsal face, and a very much narrower, oblique and somewhat concave posterior portion ; the palmar face is very low proximo-distally, very broad and much roughened. On the ulnar side there is no facet for the pyramidal, except a very small one at the palmo-distal angle ; dorsally there is no contact between the two carpals. Bistally, the almost equal facets for the magnum and unciform meet at an acute angle and form a sharp, ridge-like portion between the two carpals last mentioned. However, this ridge-shaped portion is confined to the dorsal moiety of the lunar and the facet for the magnum is in two quite distinct parts, though there is no break in the continuity of the articular surface ; the dorsal portion, which is on the wedge-like projec- tion, is convex and obliquely lateral, rather than distal in position, while the palmar portion presents distally and is concave. Much the same description may be given of the facet for the unciform, except that the dorsal is more concave and distal and the palmar portion is considerably narrower than the corresponding part for the magnum and is more oblique, less completely distal. In Toxodon the lunar differs in a number of particulars from that of Nesodon. The radial surface has a more decidedly convex anterior por- tion, which is reflected farther down upon the dorsal face, while the pal- mar portion is very much reduced in size. On the ulnar side, the lunar has a distinct contact with the pyramidal, a contact which is both dorsal and palmar, and on the distal end the faces for the magnum and un- ciform meet at a much more obtuse angle and, even dorsally, these facets are much more distal than lateral, so that there is hardly any ridge-like prolongation. The pyramidal is broad transversely and very low proximo-distally. On the dorsal face, near the radial border, is a prominent tubercle for ligamentous attachment. The articular surface for the ulna is large, transversely oval and deeply concave, and on the palmar side is quite an extensive facet for the pisiform, of irregularly semicircular shape. The distal end is covered by the large, somewhat saddle-shaped facet for the SCOTT: TOXODONTA OF THE SANTA CRUZ BEDS. 159 unciform and on the ulnar side is another facet, no doubt for the ves- tigial fifth metacarpal. In Toxodon the pyramidal is very similar to that of Nesodon, but is relatively broader and shorter, so that its proximal end is at a much lower level than that of the lunar and scaphoid, a difference which is much greater than in the Santa Cruz genus. On the ulnar side there is no facet for the rudiment of the fifth metacarpal. The pisiform is a short and heavy bone, not unlike a calcaneum in shape. The proximal end is transversely expanded and bears two irregularly tri- angular facets, for the pyramidal and ulna respectively; the latter is con- siderably the larger of the two and more concave. The body of the bone is short, heavy, laterally compressed, and has a strongly convex ventral border, while the distal or free end is much thickened and rugose. In Toxodon the pisiform is of quite a different shape, being much deeper dorso-ventrally, thinner and more compressed laterally, and is not nearly so much expanded at the proximal end. The trapezium is very small and has no direct articulation with the sca- phoid or even with the trapezoid. It is an irregular, nodular bone, with the principal diameter in the proximo-distal direction, and having rough- ened ends. On the ulnar side is a relatively large, subcircular and con- cave facet for me. II and on the palmar side is a very small facet, the pur- pose of which is doubtful, but may have been for the attachment of a sesamoid. Ameghino (’94*, 246, fig. 1) regards this bone as the vestigial me. I and it is quite possible that this is the correct interpretation, as is suggested by the absence of any articulation with the scaphoid. In that case, the minute facet on the palmar side might be for an even more reduced trapezium. From the analogy of other ungulates, however, it seems much more probable that the bone in question is the trapezium, which almost invariably persists as a relatively large element long after all trace of the pollex has disappeared and is itself suppressed only in a few highly spe- cialized types with monodactyl or didactyl feet. So far as can be judged from the cast of the La Plata skeleton, this problematical element, whether trapezium or first metacarpal, would seem to be lacking in Toxodon. The trapezoid is a relatively large and important bone, with subquadrate dorsal face. Toward the palmar side the bone narrows much and is thus wedge-shaped. Its position in the carpus is somewhat oblique, as it faces i6o PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. nearly as much toward the radial as toward the dorsal side of the manus. The proximal end forms a convex head, which fits into the concavity on the distal end of the scaphoid. The only other carpal with which the trape- zoid articulates is the magnum, for which there are two incompletely sepa- rated facets on the ulnar side, one proximal and the other distal. Distally, the large surface for me. II is saddle-shaped, concave palmo-dorsally and convex transversely. The trapezoid of Toxodon is very similar to that of the Santa Cruz genus, but proportionately larger and heavier. The magnum is quite large, especially in the proximo-distal dimension, exceeding the trapezoid in size, but is much smaller than the unciform. It has a very oblique position in the carpus, being inclined distally and toward the ulnar side. The ulnar face, which articulates with the lunar and unciform, is nearly plane and from the scaphoid to the head of me. IV there is a long, straight, oblique joint, between the magnum and me. Ill on the radial, and the lunar and unciform on the ulnar side. A some- what similar arrangement may be observed in the carpus of Ccenopus , a tridactyl rhinoceros from the Oligocene of North America. The magnum has a very complex arrangement of facets, articulating with all the other carpal elements except the trapezium, pyramidal and pisiform. The dorsal and palmar faces are strongly rugose and on the latter is a rough protuber- ance, but not the long hook so frequently seen in ungulates. As is usual, the magnum is much lower proximo-distally on the dorsal than on the palmar side, but the rise of the proximal surface is gradual and there is no well-defined head. On the radial side is the very large and complex surface which articulates with the process from the head of me. II ; this facet is quite strongly convex palmo-dorsally, except the much smaller palmar extension, which is concave. Above this facet is the small oblique, L-shaped surface for the trapezoid ; that for the scaphoid is large, transversely concave and dorsally occupies the whole width of the mag- num, but toward the palmar side this facet becomes narrower and more oblique. The lunar surface is quite distinctly divided into two parts ; the dorsal portion, which is slightly concave, is obliquely lateral in position and is continuous with the facet for the unciform, the two surfaces lying in nearly the same plane. The palmar portion of the lunar facet is convex and much more nearly proximal in position than the dorsal portion. The unciform surface is quite small dorsally and is continued as a very narrow band almost to the palmar border. Distally, the facet for me. Ill is large SCOTT : TOXODONTA OF THE SANTA CRUZ BEDS. 1 6 1 and very simple, almost plane transversely and concave palmo-dorsally. In Toxodon the magnum is of quite a different shape, being proportion- ately shorter proximo-distally and very much broader. The articulations are nearly the same as in Nesodon , except that the dorsal portion of the lunar facet is less lateral and more proximal and that, when the carpus is seen from the front, the facets for the lunar and unciform meet at a very obtuse, though distinct angle, while in Nesodon , as above described, they form a continuous line. The unciform is a very large bone, decidedly the largest of the carpals ; it is low and broad and has an irregularly pentagonal dorsal face. The proximal end is unequally divided between a narrow surface for the lunar and a very broad one for the pyramidal ; the former is oblique, dorsally concave, becoming convex toward the palmar side, while the pyramidal facet is more proximal and irregularly saddle-shaped. On the radial side the unciform extends over the magnum, upon which it rests, and the facet for the magnum and that for the lunar meet at an acute angle. Below the magnum facet and in nearly the same plane with it, indeed continuous with it, is a large, crescentic articular surface for the projection from the ulnar side of the head of me. III. On the distal end is the large, subquadrate and palmo-dorsally concave facet for me. IV, and on the ulnar side is the relatively large, triangular, concave surface for the vestigial me. V. In Toxodon the unciform is proportionately broader and more massive ; the proximal end is wider than the distal and on the ulnar side projects considerably beyond the head of me. IV. The facet for the rudi- ment of me. V is relatively smaller than in Nesodon and is displaced toward the palmar side. The metacarpus consists of three functional members, the second, third and fourth and a rudiment of the fifth. The symmetry is nearly, but not perfectly mesaxonic, for though me. Ill is symmetrical, me. II and IV do not quite form a pair. The three functional metacarpals are quite strongly divergent toward the distal ends, almost like the sticks of a fan, caus- ing a wide separation of the digits. As was noted above, Ameghino believes that a vestige of me. I is also present and in support of this view it may be urged that the bone in question does not articulate with the scaphoid or trapezoid, but only with me. II. On the other hand, it would be most exceptional for the trapezium to be suppressed while any part of the pollex remained. 162 PATAGONIAN EXPEDITIONS : PALEONTOLOGY. Metacarpal II is short and stout, with moderately expanded head, which overlaps that of me. III. On the radial side is a small, irregular, convex facet for the trapezium and the proximal end is occupied by the large, transversely concave surface for the trapezoid. The prominence from the ulnar side, which overlaps the head of me. Ill, abuts against the magnum by a very large facet, and on its distal side is a narrow, con- cave, articular band for me. III. The large facet for the magnum is extended for nearly the whole dorso-palmar diameter of the bone ; most of this surface is deeply concave, but near the palmar side is a small convex area, which fits into a corresponding concavity on the side of the magnum. The shaft is quite heavy and of irregularly trihedral form, with rounded angles, and has large rugose tuberosities upon it, notably on the ulnar side below the head (see PI. XXII, fig. 5) and on the palmar face somewhat below the middle ; it is moderately curved, with decidedly con- cave radial border, and from a point a little below the head increases in breadth toward the distal end. The distal trochlea is narrow and not very strongly convex, but is extended well upward upon the dorsal face ; the carina is low and inconspicuous and is confined to the palmar aspect of the bone, being quite invisible in front-view. Metacarpal III is the longest and heaviest of the series, though it does not much exceed me. II in length. The head, which is but little wider than the shaft, though of great dorso-palmar thickness, carries four artic- ular surfaces, for me. II and III and the magnum and unciform respec- tively. The facet for me. II is very narrow and presents proximally; that for the magnum, which is obscurely demarcated from the latter, is very large and quite strongly convex palmo-dorsally. The unciform surface, which is borne on a massive projection from the ulnar side of the head, is large, oblique and of crescentic shape, narrowing toward the palmar side. The surface for me. IV is divided into two almost separate facets, which are large and concave and placed, one near the dorsal and the other near the palmar border ; they are connected proximally by a very narrow band of articular surface. The shaft is heavy, antero-posteriorly compressed and broadening toward the distal end. Near the proximal end, on the anterior face, is a median pit, which probably served for the insertion of one of the extensor tendons. Distally, the trochlea is low and wide and of compressed semicylindrical form ; the carina is entirely palmar and is even less prominent than that of me. II. (PI. XXVI, %■ S-) SCOTT: TOXODONTA OF THE SANTA CRUZ BEDS. 1 63 Metacarpal IV is not an exact counterpart of me. II, even disregarding the different shape of the proximal ends, being shorter, heavier and straighter, but the difference of length is compensated for by the fact that the head of me. II rises much higher into the carpus than does that of me. IV, while their distal ends are in nearly the same horizontal line. In me. IV, the proximal end is overlapped by the head of me. Ill, with which it articulates by two convexities which fit into the concave, oval facets on the ulnar side of me. Ill already described and by a large, convex surface proximal to these. On the ulnar side of the head is quite a large facet for the vestigial me. V. The shaft is rather heavier and straighter than that of me. II, but is of similar trihedral cross-section, and the distal trochlea is narrower and its proximal border on the dorsal face is less regularly curved, being higher on the mesial and lower on the external side. Metacarpal V is not represented in any of the collections, but that it was normally present is demonstrated by the facets on the pyramidal, unciform and me. IV. Ameghino has figured it in the closely allied, if not identical, genus Xotoprodon (’94/, 246, fig. 1) as a small, irregular nodule, articulating with the three bones just mentioned. In Toxocion the metacarpals have assumed elephantine proportions, being relatively shorter and very much broader than those of Nesodon. Metacarpal II articulates extensively with the magnum, but does not overlap the head of me. Ill so much as in the Santa Cruz genus. The keel of the distal trochlea is much reduced and in me. Ill is hardly dis- tinguishable. The presence of the rudimentary me. V is indicated by the facets for it upon the adjoining bones, and, as in Nesodon , it is extended proximally to a contact with the pyramidal, but is shifted more to the palmar side than in that genus. The phalanges of the manus of Nesodon are very incompletely known and the few specimens which the collections contain may be most advan- tageously described in connection with those of the pes. The pelvis (PI. XXIV, fig. 1) is broad and heavy, but of no great antero-posterior length. The ilium is relatively rather small ; the neck is short and laterally compressed and has obscurely marked borders, so that the usual trihedral shape of the peduncle is hardly apparent ; there is no ilio-pectineal process. The anterior expansion or plate of the ilium is of only moderate width and is but slightly everted. The shape of the crista is not satisfactorily shown in any of the specimens, but apparently 164 PATAGONIAN EXPEDITIONS: PALAEONTOLOGY. there was no definite division of the plate into dorsal and ventral pro- cesses. The acetabulum is large, of nearly circular shape and has prom- inently projecting borders and is deeply invaded by the very large liga- mentous sulcus, which is out of all proportion to the size of the pit for the round ligament on the head of the femur. The ischium is proportionately elongate and, for most of its length, is slender and shallow dorso-ventrally, but quite thick transversely ; pos- teriorly, it expands into a thin and laterally compressed area, which has no definite tuberosity, but merely a low, rugose area. The pubis is quite broad, depressed and flattened and the symphysis is long. The large obturator foramen forms an almost regular oval, with the long axis directed antero-posteriorly. In Toxodon the pelvis is remarkably different from that of Nesodon and has acquired considerable resemblance to the pelvis of the elephants. The iliac plate is immensely expanded and very strongly everted and the crista has become very massive and rugose, especially near the antero-dorsal and postero-ventral angles. The large, circular acetabulum has a less promi- nent border and is not invaded by a sulcus for the round ligament ; there is merely a notch of the posterior border. All the post-acetabular region of the pelvis has been much reduced in length ; the ischium is relatively much shorter and thicker than in Nesodon and the tuberosity is a distinct ridge, though not at all prominent. The pubis is very heavy and the sym- physis is short, — actually but little longer than in Nesodon and therefore very much shorter proportionately. The very striking changes which the pelvis of Toxodon has undergone are all to be correlated with the great increase in stature and body- weight and in the mass and bulk of the viscera. Similar changes in the form of the pelvis, accompanying great increase in the size and bulk of the body, may be observed in several other ungulate phyla. Besides the Proboscidea, they are to be found in the Amblypoda and in the perisso- dactyl family of the Titanotheriidae and, in a less marked degree, in the larger rhinoceroses. In several of these groups the changes in the pelvis may be followed in successive stages, where they are plainly correlated with increasing body-weight and, as plainly, they are seen to be independently acquired in each phylum. To this list should now be added the Toxo- donta, the later and larger representatives of which have taken on many skeletal resemblances to the elephants, but, as the phylogenetic history SCOTT: TOXODONTA OF THE SANTA CRUZ BEDS. 165 shows, these likenesses are due merely to the mechanical readjustments necessitated by the greatly augmented bulk and weight. The femur of Nesodon (PI. XXIII, figs, i, 2) is much more perissodactyl in character than is that of Toxodon; it is of moderate length, a little shorter than the humerus, though the effective length, measured from the head to the condyles, is slightly greater than that of the humerus. The head is oval, with the long axis transverse, and is set upon a distinct neck ; the moderately large, though shallow, pit for the round ligament is placed upon the posterior aspect of the head, so as to be invisible in front view. A deep notch separates the head from the great trochanter, which is thick, massive and rugose, but low, not rising quite to the level of the head ; the digital fossa is fairly deep, but small, and the prominent linea intertrochanterica is short, not extending to the second trochanter. The latter is a long and moderately prominent ridge, which arises a short dis- tance below the head. The third trochanter is very conspicuous and is placed a little above the middle of the shalt. The proximal portion of the shaft is broad and antero-posteriorly com- pressed, becoming more rounded and transversely oval in section down- ward ; the posterior surface of the shaft is nearly smooth, owing to the absence of any distinct lineae asperae, but near the distal end there is a well-defined pit for the plantaris muscle, though the depth and distinct- ness of this pit vary much in different individuals. The rotular groove, which projects prominently in front of the shaft, is broad, shallow and asymmetrical, which is due to the greater thickness and prominence of the internal portion. Distally, its articular surface is continuous with the inner condyle, but separated by a narrow space from the outer; there is no definite suprapatellar fossa. The condyles are rather small and do not project strongly behind the plane of the shaft ; there is considerable in- equality in size between them, the inner one being broader and flatter, the outer one narrower and more convex. The external epicondyle is quite prominent and heavy, the internal one very much smaller. Like the pelvis, the femur of Toxodon is greatly modified from the con- dition in Nesodon and bears considerable resemblance to that of the ele- phant. It is relatively elongate and is much the longest of all the limb- bones ; the head is more hemispherical than in the Santa Cruz genus and the pit for the round ligament is merely a vestige, a shallow, trian- gular area, which emarginates the articular surface on the posterior side 1 66 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. of the head. There is no such notch between the head and the great trochanter as there is in Nesodon , and the trochanter itself, though very- heavy, is low, the head projecting considerably above it, while the digital fossa is smaller and deeper. The second trochanter is much reduced and Fig. 29. Toxodon burmeisteri: A, left femur, dorsum; B, the same, external view. Also left patella, dorsum. All figures X La Plata Museum. forms a mere rugose line, while the third trochanter has disappeared altogether. The shaft is long, broad and antero-posteriorly compressed, though not as strongly as in the elephants ; the posterior surface is quite SCOTT: TOXODONTA OF THE SANTA CRUZ BEDS. 1 67 flat and there is no pit for the plantaris muscle. The rotular trochlea differs from that of Nesodon in being broader, shallower and more asym- metrical, owing to the greater thickness of the internal portion, which is very broad and prominent and has the articular surface reflected over Fig. 30. Nesodon imbricatus : A, left femur, dorsum; B, the same, external view. Also left patella, dorsum. All figures X i- upon the internal side. The condyles are of more nearly equal size and similar shape than in Nesodon and are more nearly approximated, making the intercondylar fossa much narrower. Both condyles are sep- arated by narrow spaces from the articular surface of the rotular groove. That the great changes in the character of the femur between the Santa Cruz and the Pampean genus should be chiefly ascribed to the great increase in the weight of the latter animal, is made probable by the same considerations as in the case of the pelvis. Similar changes may be fol- lowed out in other phyla, notably in the Titanotheriidse and the Ambly- poda. In the late and well-nigh gigantic species of the latter group the femur has become most deceptively proboscidean in character. In each of the four phyla the modifications of the femur are as follows : ( i ) The pit for the round ligament is greatly diminished in size or altogether lost ; (2) the great trochanter becomes very thick and heavy, but its proximo- distal height is diminished ; (3) the third trochanter is greatly reduced or 1 68 PATAGONIAN EXPEDITIONS: PALAEONTOLOGY. entirely suppressed ; (4) the shaft loses its rounded form and becomes much compressed antero-posteriorly and very flat on the posterior face ; (5) the medullary cavity disappears and its place is taken by cancellous bone ; (6) the condyles become less prominent and more nearly alike in size and shape. The patella of Nesodon (PI. XXIII, figs. 6, 7) is a curious and char- acteristic bone and quite different from that of Toxodon. It is very short proximo-distally, quite broad and very thick, especially on the proximal and internal sides, thinning much to the external border. The distal border is concave and its internal portion is slightly produced downward. The anterior surface is very irregular and rugose and the posterior surface is obscurely divided by a low ridge into a narrow external, and much broader internal facet for the rotular groove of the femur. In Toxodon , on the other hand, the patella is short, of no great thickness, but im- mensely broad ; the internal portion is greatly extended and recurved, so as to cover the inner face of the rotular prominence, while the distal border is deeply notched in a very characteristic fashion. In Nesodon the leg-bones (PI. XXV, figs. 1-3) are ankylosed at the proximal, but not at the distal end, a very peculiar arrangement. Except at the ends, the two bones are separated by a very wide interosseous space. The tibia is a little shorter than the femur and has a broad thick head, which projects far external to the shaft. The outer condyle is oblique and nearly plane transversely, very slightly concave antero- posteriorly, while the inner condyle is more concave and its portion of the bifid spine is decidedly more prominent. The shaft is surprisingly slender and its proximal portion is so strongly compressed laterally as to lose the trihedral shape common to almost all terrestrial mammals. The cnemial crest is prominent and heavy, but short, and its proximal end is deeply impressed for the insertion of the patellar ligament. On the external face of the shaft is a long interosseous crest, which, however, is prominent and rugose only for a short distance about the middle of its course. The distal half of the shaft is less compressed and more rounded than the proximal moiety and the distal end is moderately expanded and heavy. The surface for the astragalus is nearly square and is obscurely divided into a broader and shallower external, and a narrower, deeper internal facet. The intercondylar ridge is very low, but forms a tongue on each border, that on the dorsal side being the narrower and more SCOTT: TOXODONTA OF THE SANTA CRUZ BEDS. 1 69 prominent of the two ; the internal malleolus is long, heavy and bluntly pointed and bears on its fibular side a facet for the inner side and neck of the astragalus. The proximal end of the fibula, which is indistinguishably fused with the tibia, is heavy, especially in the antero-posterior direction. The shaft is stout, of roughly trihedral shape, and strongly arched outward and forward, which gives it a very irregular appearance ; on the distal half of the shaft there is a prominent interosseous crest. The distal end is very large, both broad and thick, and forms a massive external malleolus, which bears two facets, one internal and slightly oblique for the astrag- alus, the other for the calcaneum, entirely distal in position, large and concave. Fig. 31. Fig. 32. Toxodon burmeisteri: Left tibia and fibula, Nesodon imbricatus: Left tibia and fibula, dorsum, X |. La Plata Museum. dorsum, X i- The leg-bones of Toxodon are very similar in type to those of Nesodon , but relatively much shorter and far more massive ; they are ankylosed in the same peculiar manner, that is, at the proximal end, but not at the distal. The tibia has an enormous proximal end and its condyles are very closely approximated, but the spine is formed only by the internal one ; the cnemial crest also is much more prominent and massive than in the 170 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. Santa Cruz genus. The shaft has a similar, laterally compressed shape, though it has no distal interosseous crest, but the distal expansion is pro- portionately much greater. The division of the astragalar surface is even more faintly marked, the intercondylar ridge lower, and there is no tongue on either the anterior or the posterior border; the internal malleolus is less prominent, but more extended antero-posteriorly. The fibula has a very large head and a short, heavy shaft, which differs from that of Nesodon in being nearly straight and laterally compressed, rather than trihedral in form, and is without an interosseous crest. The wide space between the tibia and fibula is due to the great expansion of the ends of the former, especially of the proximal end of the tibia and the distal ends of both bones. The distal end of the fibula is relatively very much heavier than in Nesodon , but otherwise of similar shape, though the facets for the cal- caneum and astragalus, which in the Santa Cruz genus are connected only near the dorsal angle, are continuous throughout in Toxodon. The pes (PI. XXV, figs. 9-12) is remarkably small in proportion to the length of the limb-bones and the size of the animal generally ; it is even much smaller than the manus. The tarsus is not interlocking, thus retain- ing a more primitive character than the carpus. The astragalus is both short and narrow and has but a feebly grooved trochlea, which is asymmetrical, the external condyle being higher and narrower and having a sharper and more angular border than the internal. On the outer condyle the articular surface passes uninterruptedly into the facet for the fibula and that of the inner condyle passes in the same way into the facet for the internal malleolus of the tibia, this latter facet extend- ing into a shallow concavity on the neck, into which the malleolus is re- ceived, when the foot is strongly flexed upon the leg. The neck is extremely short and extends obliquely inward, allowing the outer portion of the trochlea to overhang the distal end of the calcaneum, though with- out any contact with the latter. The head is moderately convex in both directions and articulates with the navicular only, being widely separated from the cuboid. The external facet for the calcaneum is long, narrow and concave ; for most of its length it is separate from the fibular facet, but becomes confluent with it distally ; separated from the external cal- caneal facet by a broad and very deep sulcus is the surface for the sus- tentaculum, which is nearly plane and is broad proximally, narrowing distally to its junction with the navicular facet on the head. scott: toxodonta OF THE SANTA CRUZ beds. 1 7 1 The calcaneum is short and massive. The tuber, which is very heavy, broadens gradually to the free end, which is roughened and has on the plantar surface a shallow tendinal sulcus, somewhat as in the Artiodac- tyla, though the groove is not median, as it is in the latter, but near the internal side. The fibular facet is very large and forms a broad, heavy prominence, on the tibial side of which is an oblique, convex facet for the astragalus ; this latter surface is produced proximally considerably beyond that for the fibula. While of only moderate size, the sustentaculum is very thick and prominent ; its astragalar facet is of irregularly oval shape and nearly plane. The distal end of the calcaneum is also very heavy and bears a broad, slightly saddle-shaped surface for the cuboid, which is concave planto-dorsally and convex transversely ; internal to this is a small facet for the navicular. The navicular is low proximo-distally and of subcircular shape ; its greatest proximo-distal height is through the facet for the large ectocunei- form, while the internal portion is hardly half as high. From the plantar side is given off a very low tuberosity, which appears to be the remnant of the usual navicular hook. On the proximal end, the astragalar surface is simply concave in both directions, narrowing toward the plantar side, and on the fibular side is a small, articular projection, formed by the con- joined surfaces for the calcaneum and cuboid, the former presenting later- ally and the latter obliquely distally. The distal end is occupied chiefly by the large facet for the ectocuneiform, which is oblique, sloping down toward the tibial side and contracting much toward the plantar side. The facet for the mesocuneiform, which is much smaller, is also oblique, slop- ing in the opposite direction and forming an obtuse angle with the ecto- cuneiform facet. The meso- and entocuneiforms are fused together, though the limits of each element may still be made out. The entocuneiform is a small, irreg- ular, nodular bone, which has no contact with the navicular and carries only one facet, a small, oblique surface on the distal end for the head of the second metatarsal. The mesocuneiform is much larger and bears a large, slightly convex surface for the navicular and interlocks with the ectocuneiform in a curiously intimate and complicated manner. Dorsally, there is a large concavity, into which fits a convexity on the tibial side of the ectocuneiform and on the plantar side the mesocuneiform sends a small shelf-like projection underneath the ectocuneiform and this shelf has a small, concave, obliquely proximal facet for the latter. 172 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. The ectocuneiform is very much larger and has much the same shape as in the perissodactyls, other than the horses. Its proximal surface meets that of the mesocuneiform at an open angle, the two bones together form- ing a shallow depression, into which the navicular fits. On the fibular side, in addition to the facets for the mesocuneiform already described, are two quite large and widely separated ones for the head of mt. II, of which that near the dorsal border is of complex shape, convex distally and concave proximally, while that near the plantar border is plane. On the fibular side is a large plane surface for the cuboid and the distal end is covered by the very large facet for mt. III. The cuboid is very short proximo-distally, but broad, thick and heavy ; it has no plantar hook, merely a low rugosity. There are four articular surfaces, a large, saddle-shaped, proximal one for the calcaneum ; an almost equally large, plane facet on the distal end for mt. IV, and on the tibial side an obliquely proximal surface for the navicular and, below this, a large, plane, lateral surface for the ectocuneiform. The metatarsus consists of three members, without any remaining ves- tiges of mt. I or V. The metatarsals are so arranged as to be strongly divergent distally, as is also the case in the metacarpals, and the mes- axonic symmetry is less complete than in the manus, since digits II and IV do not form an entirely symmetrical pair and even the symmetry of digit III is not perfect. Metatarsal II is the shortest and lightest of the series ; the head is narrow, but thick planto-dorsally and bears a trans- versely concave facet for the mesocuneiform, and on the tibial side is a small oblique surface for the entocuneiform, the head appearing to notch the compound cuneiform in a very characteristic way. This metatarsal rises above the head of mt. Ill and abuts laterally against the ectocuneiform, with which it articulates by means of two large facets, but no distinct sur- face for mt. Ill is visible. The shaft is stout and of trihedral section, becoming broad and more flattened toward the distal end. The trochlea is rather oblique and asymmetrical and the carina is very low and incon- spicuous. Metatarsal III is somewhat longer and much heavier than mt. II ; the head bears an oblique facet for the ectocuneiform, which is convex planto- dorsally and is incompletely divided by a constriction from each side into a larger dorsal and smaller plantar portion ; the latter is wider than the corresponding surface on the ectocuneiform. On the fibular side are two scott: toxodonta OF THE SANTA CRUZ beds. 173 facets for mt. IV, a large, deeply concave pit near the dorsal border and a much smaller, nearly plane surface on the plantar extension of the head. The shaft is short and heavy, broad and much compressed planto-dor- sally, widening distally. A slight degree of asymmetry is shown in the difference between the two lateral borders of the shaft, that on the tibial side being straight, while the fibular border is slightly concave. In the same way, the tuberosity above the pit for the lateral ligament is distinctly more prominent on the fibular than on the tibial side. The trochlea, which is low and wide, is strongly compressed planto-dorsally and the carina, though inconspicuous, is yet more prominent than that of mt. II. Metatarsal IV is nearly as long as mt. Ill and heavier, and is some- what longer and decidedly stouter than mt. II. The proximal end has a large, nearly plane surface for the cuboid, which is somewhat pear-shaped, contracting toward the postero-internal angle. On the tibial side of the head, near the dorsal border, is a very prominent convexity, which fits into the concavity on the fibular side of mt. Ill, already described ; the other facet for mt. Ill near the plantar side is very much smaller and slightly concave ; these two facets are separated by a deep sulcus. The shaft is rather narrow, but thicker planto-dorsally than that of mt. Ill and having a more rounded dorsal face ; the tibial border is nearly straight, while the fibular border is quite strongly concave. This form of asymmetry and the broader, heavier shaft distinguish this metatarsal decidedly from mt. II, of which it is not so nearly a counterpart as it is in the Perissodactyla. The tuberosity on the fibular side of the distal end is much larger and more prominent than that on the tibial side. The distal trochlea is much broader and more symmetrical, but lower, than that of mt. II, and the carina, which divides the plantar side of the trochlea into a broader and more concave external portion and a narrow, flatter internal part, is more distinct. The phalanges of the three digits are all different from one another ; those of digit III are the largest and those of digit II much the smallest of the series, as is also true of the manus, though the phalanges of each digit of the latter are considerably larger than those of the corresponding digit of the pes. In the hind foot the first phalanx of digit II is short, narrow and thick planto-dorsally and nearly symmetrical in form. The proximal surface, for articulation with the metatarsal trochlea, is slightly 174 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. concave and very shallow and is not grooved for the carina, but has merely a slight notch on the plantar border. Distally, the phalanx becomes somewhat narrower and thinner and the distal trochlea is nearly plane, with only a shallow emargination of the plantar border to indicate the usual division into two facets. The second phalanx is very small, especially in proximo-distal length, but is quite broad and thick in pro- portion to the length. The proximal articular surface is nearly plane, but the distal one is divided by a shallow, median groove into two parts and is reflected upon the dorsal side of the bone. The ungual is a small, asymmetrical hoof, not unlike that of the tapir ; the proximal trochlea is a very shallow concavity, divided by an obscurely marked, median ridge into two depressions. There is no subungual process. In the manus, the first phalanx of digit II is larger than that of the pes and broader relatively to its length ; the proximal trochlea is rather more concave, with plantar margin more distinctly notched for the metacarpal carina ; the distal trochlea is a little more convex and a very small area of it appears on the dorsal face. The second phalanx is very like that in the hind foot, except for its larger size and slightly more concave proximal articular surface. I have seen no example of the ungual phalanx. In the pes, the first phalanx of the third digit is very much larger than the corresponding one of the second digit and is proportionately broader and more depressed, and is also slightly asymmetrical. The very shallow concavity of the proximal end is not grooved for the carina, but the notch for it on the plantar margin is much more distinct and is continued dis- tally as a median depression between two rugose eminences on the plantar face. The distal trochlea is very nearly plane, but is a trifle more convex than in digit II. The second phalanx is likewise much larger than in digit II and proportionately broader and more depressed ; the distal articular surface is decidedly saddle-shaped and is reflected upon the dorsal side of the bone, but in a more symmetrical manner than in digit II. The ungual phalanx is large and quite symmetrical, and much depressed, with thin, rugose free border, but without any trace of the median cleft, which is replaced by a low ridge. In appearance, there is considerable resemblance to the median ungual of such Eocene Perissodactyla as Palceosyops. The trochlea has two continuous facets, of which that on the tibial side is distinctly the more concave. Of digit III in the manus, I have only the second phalanx. This is scott: toxodonta OF THE SANTA CRUZ beds. 175 relatively shorter and much broader than in the pes and has a more con- cave proximal surface and a similarly saddle-shaped distal trochlea, which is likewise more concave transversely. When the phalanges of the fourth digit of the pes are put together and in connection with the metatarsal, it is seen that they curve toward the median line in a peculiar and characteristic fashion, which is better shown in Sinclair’s figure (see p. 8, text-fig. 4, B) than in PI. XXV, fig. 9, because in the latter the phalanges are drawn as separated from one another. In this digit the phalanges resemble those of digit II in form, but are considerably larger and somewhat asymmetrical. The distal trochlea of the first phalanx is much more oblique to the long axis of the bone than in digit II or III and is slightly reflected upon the dorsal face, which is not true of the other digits. The ungual is relatively shorter and heavier than in digit II and has a more truncate free border, but is otherwise similar in form. I have before me only the second phalanx of digit IV in the fore foot and this, except for its larger size, is closely similar to the corresponding bone in the pes, but is not of itself sufficient to show whether the pha- langes of this digit had the same medial curvature as in the pes, or not. In both fore and hind foot a long and narrow sesamoid is attached to the metapodials on each side of the distal carina, but there appears to have been none between the second and third phalanges. In Toxodon the pes has undergone some remarkable modifications, most of which are obviously due to the great increase in the size and weight of the animal. The astragalus has an even more flattened trochlea than in Nesodon and a somewhat longer and more obliquely directed neck and, on the fibular side, the trochlea projects over and rests upon the distal end of the calcaneum, an additional articulation of an altogether exceptional character, though perhaps an incipient stage of it is present in Nesodon. The tuber calcis is very broad and massive, while the portion of the calcaneum distal to the fibular facet is very short ; the fibular facet itself is very large, though not so prominent as in the Santa Cruz genus. A very extraordinary feature of this calcaneum is the position of the cuboid facet , which is on the plantar side , at right angles to its normal position ; the tuber thus projects directly backward, nearly at a right angle with the long axis of the foot. No indication of this most unusual arrangement is to be seen in Nesodon, in which the calcaneo-cuboid articulation is of 176 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. the usual kind. The other tarsal bones resemble those of the Santa Cruz genus in character and arrangement, except that they are relatively shorter proximo-distally and much broader and more massive, a change which is particularly evident in the shape of the ectocuneiform, and in the overlap- ping of the latter by the cuboid. The metatarsals also are proportion- Toxodon burmeisteri : A, left pes, dorsum, Nesodon imbricatus : Left pes, dorsum, X J. X | B, the same, external view. La Plata Museum. ately shorter and much heavier and are of more nearly equal size, though mt. Ill is somewhat longer, broader and more compressed planto-dorsally than II and IV. The phalanges are shorter, broader and heavier than those of Nesodon and the unguals have more the appearance of those of the rhinoceroses. Restoration. (PI. XII, fig. 2.) The Santa Cruz fauna is largely made up of mammals which, from our modern standpoint, have a very curious and bizarre appearance. Among the more striking of these is Nesodon , which bears but a distant resemblance to any known ungulate of the northern hemisphere, recent or extinct. The short, deep and massive head has a decided suggestion of likeness to the rodent skull, though this likeness is not so strong as in the Typotheria. The neck is short and heavy and most of its vertebrae have neural spines of unusual height. scott: toxodonta OF THE SANTA CRUZ beds. !77 The thoracic region of the vertebral column is long and, anteriorly, its spines are very high, descending rapidly backward, so as to form a prom- inent hump at the withers, while the lumbar region is rather short and not very stout. The sacral and caudal vertebrae are not known, but the former may be restored from those of Adinoiherium, in which the sacrum is long and composed of no less than six vertebrae. The thorax is long, deep and very capacious. The limbs are short and moderately stout and the anterior pair are somewhat shorter than the posterior, though the difference in length is, to a great degree, compensated by the very large scapula, which has a highly characteristic form, with broad blade and prominent spine, which termi- nates in a distinct acromion and bears two very large and conspicuous metacromia. The humerus, which is short and heavy, is not especially characteristic, but the fore-arm is remarkable for the great size of the ulna, which, for most of its length, is heavier than the radius. In contrast with this, the tridactyl manus is extremely small in proportion to the size of the animal and the unguals of the lateral digits are almost claw-like. The pelvis is rather short and heavy, but the iliac plate is only moder- ately expanded and everted. The femur is relatively long and stout and has a small though very distinct third trochanter, and the patella is very thick and massive. Tibia and fibula are ankylosed at the proximal, but not at the distal end, and the fibula, though much compressed laterally, is still very thick antero-posteriorly. The pes is even smaller and weaker than the manus ; the calcaneum has a short and very heavy tuber and articu- lates in the normal manner with the cuboid. In one of his latest papers, Gaudry figures and describes the pes of Nesodon as thoroughly plantigrade and he gives certain reasons, the force of which must be admitted, for this conclusion (’o6, 28, fig. 53). A study of all the articulations in- volved, however, together with the shape of the ungual phalanges, ren- ders the digitigrade attitude more probable, in my judgment. The question is a notoriously difficult one in the case of animals which have no near analogues in the modern world. In size, N. imbricatus , which is the abundant Santa Cruz species, does not greatly exceed the American tapirs, though it is decidedly a heavier and stouter animal, with pro- portions not unlike those of the early rhinoceroses. A comparison of the skeletons of Nesodon and Toxodon (cf. PI. XII, fig. 2, with Lydekker’s plate, ’93) reveals many differences between the Santa 1 78 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. Cruz and the Pampean genus, not only in innumerable matters of detail, but also in the general proportions and whole appearance of the two animals. Toxodon is a far larger and, in particular, a much more massive animal and all the bones of its skeleton are very much heavier, not only actually, but proportionately as well. A very peculiar appearance is given to the skele- ton by the relative shortness of the fore-legs, depressing the whole anterior portion of the vertebral column and bringing the head much below the level of the back and loins. This peculiarity is especially striking and remark- able when the skeleton is seen from the front end. Something of the same sort is to be seen in Nesodon , but in no such pronounced degree, and this constitutes one of the principal differences in the appearance of the two skeletons. In Toxodon the skull is relatively longer and shallower dorso-ventrally and tapers more to the anterior end ; several minor modifications combine to alter its appearance very materially. Of these may be mentioned the more elevated orbits, the much shortened nasals, with the more elongated and sloping anterior nares, the more prominent crest on the dorsal side of the premaxillaries, and the long diastema behind the functional incisors. Even more characteristic is the shape of the mandible, which, with its com- pletely procumbent incisors, depressed and flattened chin, broad and shal- low symphysis, and very high, narrow ascending ramus, differs widely from that of Nesodon. The neck is relatively short and heavy, and has much less prominent spines than those of the Santa Cruz genus, while in the anterior thoracic region the spines are far higher and heavier and form a more prominent shoulder-hump. In the remainder of the thoracic and the lumbar regions the spines are lower, broader and more plate-like and all have a backward inclination, much as in the Proboscidea. The trunk is relatively shorter, not only having one or two less vertebrae, but the centra of most of the trunk-vertebrae are proportionately shorter and much heavier. The backbone, disregarding the spines, forms a regular arch, curving upward from the neck to the loins. The ribs are very long and broad and are strongly curved outward, forming a very capacious thorax. The sternum has quite a different appearance from that of Nesodon , owing to the depressed form of the manubrium, though the two are fundamen- tally similar. The scapula has proportions altogether different from that of the Santa Cruz genus, being much longer proximo-distally and narrower transversely. SCOTT: TOXODONTA OF THE SANTA CRUZ BEDS. 179 The prominent spine has no acromion and dies away gradually upon the neck ; there is only a remnant of the proximal metacromion about mid- way in the course of the spine, but no trace remains of the distal one, which is so very conspicuous in Nesodon. The pelvis is set at a more obtuse angle with the vertebral column than in the latter genus and is relatively shorter, but very much broader. The ilium has a short peduncle, but an immensely expanded and everted plate, causing a decided resemblance to the pelvis of the elephant. As already noted, the fore-limb is very short and heavy ; the humerus is extremely massive and has a greatly developed external tuberosity. The fore-arm bones, which are separate, are also very stout, especially the ulna, which much exceeds the radius in diameter. The hind-limb is much longer and the femur is by far the longest of all the limb-bones and, with its low, massive great trochanter, the antero-posterior compres- sion and flattening of the shaft and the suppression of the third trochanter, has considerable resemblance to that of the elephants. The leg-bones are ankylosed in the same exceptional manner as those of Nesodon , that is, fused at the proximal, but not at the distal end. They are relatively much shorter and far heavier than in Nesodon , but have a similar laterally compressed shape. The tridactyl feet resemble those of the Santa Cruz genus in structure and in the mutual relations of the various elements, but the metapodials are remarkably short and extremely stout and tend to an isodactyl ar- rangement, in which the three metapodials of each foot are of nearly equal weight. The phalanges also are very short, broad and thick. The altogether exceptional character of the calcaneo-cuboid articulation gives a peculiar position to the calcaneum and modifies the appearance of the pes, when seen from the side. Systematic Position of Nesodon. — That Nesodon is nearly related to Toxodon, and should be referred to the same suborder, is obvious from the most superficial comparison, but the exact relationship between the two genera offers a much more difficult problem. This problem can receive a definitive solution only when the ancestry of Toxodon can be traced back step by step to Santa Cruz times, for our knowledge of the evolutionary process is still too vague for us to determine positively the relation between two genera so widely separated in time and so different in structure. Not unnaturally, therefore, the question has been answered i8o PATAGONIAN EXPEDITIONS : PALEONTOLOGY. differently by different writers. Lydekker says of Nesodon : “It may, indeed, be regarded as the most generalized representative of the group with which we are at present acquainted, although it does not appear to have been the direct ancestor of Toxodon ” (’93, 25). If my memory does not deceive me, Dr. Roth expressed the same opinion to me in conversation, taking the ground that the more complex grinding teeth of Nesodon could not have given rise to the simpler molar-pattern of Toxo- don. On the other hand, Ameghino (’89*, 402; ’04, 219) appears to hold that the Pampean is directly descended from the Santa Cruz form. My own opinion is still undecided, though I am somewhat inclined to agree with the position taken by Ameghino. Dental evolution is not, as is often assumed to be the fact, always in the direction of increased com- plexity of pattern, but may result in simplification. Not to mention such groups as the Edentata and the Cetacea, we find among the Rodentia several clearly demonstrable instances, in which the acquisition or in- crease of hypsodontism is accompanied by a marked simplification of the molar-pattern, and I can see no insuperable difficulty in deriving the teeth of Toxodon from those of Nesodon. At all events, Nesodon very nearly represents the ancestor sought and may be used for all practical purposes of the comparison. Should it prove to be true that Nesodon is not the ancestor, it will, in all probability, be eventually shown that the ancestral and as yet unknown toxodont of Santa Cruz times very closely resembled Nesodon , differing from it chiefly in having a less developed post-tympanic region and possibly also in having simpler grinding teeth. In other regards, Nesodon is just what we should expect the ancestor of Toxodon to be. Species. — Though remains of Nesodon are individually among the most abundant of Santa Cruz fossils, the number of species represented is small. Indeed, only two species, of those hitherto described, can be definitely recognized, N imbricatus and N. conspurcatns (i. e. andium ), and these only by a constant difference of size. Many other names have been proposed, but they are either obvious synonyms, or of very doubtful propriety and until more abundant material has been gathered and espe- cially until the exact stratigraphic succession has been determined, it will hardly be practicable to make a satisfactory decision regarding some of these supposed species. scott: toxodonta OF THE SANTA CRUZ beds. 181 Nesodon imbricatus Owen. (Plates XII, Fig. 2; XIII-XVI; XVII, Figs 2, 3; XVIII, Figs. 1, 4, 6, 7; XIX; XX, Fig. 5; XXI, Figs. 1, 2, 11 ; XXII-XXV; XXVI, Figs. 4-6.) Nesodon imbricatus Ow.; Rept. Brit. Ass. Adv. Science, 1846, p. 66. Nesodon Sulivani Ow. ; Ibid., p. 67. ? Toxodon patagonensis Moreno ( nomen nudum ) ; Patagonia, resto de un continente hoy submerjido, 1882, p. 22. Colpodon propinquus Bunn, (in part) ; Anales del Mus. Nac. de Buenos Aires, T. Ill, 1885, p. 161. ? Protoxodon patagonensis (Mor.) Amegh.; Observ. gen. sobre el orden . . . Toxodontes, 1887, p. 62. Protoxodon marmoratus Amegh.; Enum. sistem., etc., 1887, p. 16. Protoxodon ob literatus Amegh.; Ibid. Adelphotherium ligatum Amegh.; Ibid. AcrotJierium rusticum Amegh.; Ibid., p. 17. ? Grouotherium decrepitum Amegh.; Ibid. Scopotherium cy clops Amegh.; Ibid., p. 18. Protoxodon Sidivani (Owen) Amegh.; Cont. al conoc. de Mam. fos. de la Repub. Argent., 1889, p. 443. AcrotJierium australe Mercerat; Rev. del Mus. de La Plata, T. I, 1891, P- 39i- AcrotJierium intermedium Merc.; Ibid., p. 392. AcrotJierium variagatum [sic] Merc.; Ibid. Nesodon Oweni Merc.; Ibid., p. 399. Nesodon cyclops (Amegh.) Merc.; Ibid., p. 400. Nesodon typicus Merc.; Ibid., p. 402. Nesotherium Studeri Merc.; Ibid., p. 413. NesotJierium rufum Merc.; Ibid., p. 416. ? NesotJierium patagonense (Moreno) Merc.; Ibid., p. 417. NesotJierium turgidum Merc.; Ibid., p. 419. NesotJierium rutilum Merc.; Ibid., p. 420. Nesotherium argentinum Merc.; Ibid., p. 421. NesotJierium NeJiringi Merc.; Ibid., p. 423. NesotJierium Burmeisteri Merc. ; Ibid., p. 425. Protoxodon demens Merc.; Ibid., p. 429. Protoxodon Trouessarti Merc. ; Ibid., p. 430. 1 82 PATAGONIAN EXPEDITIONS: PALAEONTOLOGY. Protoxodon americanus Merc.; Ibid., p. 431. Protoxodon decrepitus (Amegh.) Merc.; Ibid., p. 432. Protoxodon Henseli Merc.; Ibid., p. 435. Protoxodon speciosus Merc.; Ibid., p. 436. Adelphotherium lutarium Merc.; Ibid., p. 438. Adelphotherium trivium Merc.; Ibid., p. 438. Adelphotherium repandum Merc.; Ibid., p. 439. Adelphotherium Rothi Merc.; Ibid., p. 440. Adelphotherium pumilum Merc.; Ibid., p. 440. Nesodon marmoratus Amegh.; Rev. Argent, de Hist. Nat., T. I, 1891, P- 377- Nesodon ob literatus Amegh.; Ibid., p. 377. Nesodon cavifrons Amegh.; Enum. synopt., etc., 1894, p. 23. Nesodon brachycephalus Amegh.; Ibid., p. 24. “In the whole course of zoological literature there is, I think, nothing to compare with the appalling synonymy of this and the following species [N ovinus\ ” (Lydekker, ’93, 26). While this is true, the reason for the great multiplication of names is not far to seek ; it lies in the extraordinary changes in the character of the dentition during the course of individual development, which alter the whole appearance of the animal. This was first pointed out by Ameghino (’9 1 7/) , who showed that the initial con- fusion arose from the fact that the type of the species, which is an imma- ture animal with the milk-dentition, was described by Owen as an adult. In the long list of synonyms given above, it may well eventually prove that several of the names refer to species distinct from N imbricatus ; all that their inclusion in the list implies is, that their claim to separation has not been proved. In his revised lists of the species of Nesodon Ame- ghino (’94", 23-24; ’943, 244-5; ’98, 156) retains five, of which only two can be regarded as well established. Of the others, one or more may be distinct, but the available specimens are insufficient to demonstrate this. Both in this genus and in Adinotherium Ameghino employs the character of the fronto-nasal suture as a means of specific distinction, referring to N imbricatus most of those skulls in which the posterior ends of the nasals are separated by large triangular processes of the frontals and to N. mar- moratus those in which the nasal processes of the frontals are absent, the nasals are in contact with each other throughout their length and are received into a deep emargination of the frontals, as in Adinotherium. SCOTT: TOXODONTA OF THE SANTA CRUZ BEDS. 1 83 He adds that 95 per cent, of the specimens from the Santa Cruz beds are referable to N. imbricatus (’94*, 244). While it is perfectly true that differences in the form of the cranial bones may often be used for the distinction of species, yet this criterion is practicable only when the dif- ference is relatively constant. In the present instance, not only is the character a fluctuating one, but the two extremes are connected by inter- mediate gradations, the nasal processes of the frontals varying much in size to complete disappearance. On the other hand, it may be said that, as a whole, the genus Nesodon is characterized by the presence of large nasal processes of the frontals and Adinotherium by their absence, though in both genera exceptions are not rare. Much the same reasoning applies to the supposed species, N. obliteratus, which is defined by the absence of the lower canine, the crowded condi- tion of the lower teeth, the less procumbent inferior incisors and the more steeply inclined symphysis. (Ameghino, ’94*, 245.) If they were con- stant, these characteristics would be quite sufficient to distinguish the species, but, as a matter of fact, they are fluctuating. Between the two species of Nesodon which are here recognized, the only trustworthy distinction is that of size, N. imbricatus being the larger and N conspnrcatus the smaller. There is, of course, considerable varia- tion in size in both of these species, but they do not seem to intergrade, although there is as great a difference of size between the largest and the smallest individuals of N. imbricatus as between the latter and the larger examples of N conspnrcatus. In many cases these variations in size are due to differences of age and sex, but there are also marked differences between individuals of corresponding ages. In N. imbricatus the upper profile of the skull is typically of a moderate convexity in the antero- posterior direction, but not infrequently there is quite a steep ascent from the forehead to the occipital crest, as in the skull figured in Plate XIII. The forehead itself is nearly plane in the fore-and-aft dimension and is usually a little depressed below the level of the nasals. In the following table the measurements of a representative series of individuals are given and they bring out clearly the changes in the form and size of the teeth with advancing age, the columns being arranged with the younger individuals on the left and the older on the right, though Nos. 15,135 and 15,252 are of nearly the same age and the former is probably a female. 184 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. MEASUREMENTS. No. 15,252. Upper dentition, length. . . . .268 I—, length (i. e. ant. -post, diam.) . . .016 “ width (i. e. transv. diam.) . . .031 I-, length 017 “ width 0165 I-, length 015 “ width 0073 Upper canine, length 017 “ width .... .0083 Upper cheek-teeth series, length . . .182* “ premolar series, length . . .0825* P1, length 017 “ width 0125 P-2 , length 018 “ width ...... .016 P-, length 019 “ width 018 P-, length 027* “ width 029* Upper molar series, length . . . .108 M1, length 036 “ width ...... .028 M-, length ...... .046 “ width 029 M-, length ...... .038 “ width 0235 No. 15,492. Lower dentition, length. IT, length “ width I2, length “ width I5, length 023 1^, width ...... .014 Lower canine, length .... “ “ width .... No. No. No. No. No. 15,135 . 15,000. 15,336. 15,969. 15,256. ?.25o .272 .296 .279 .275 •0175 .017 .021 .018 .028 .029 .0285 .025 .016 •0155 .019 •0235 .020 .015 .016 .020 .026 .025 .014 .015 .015 •0135 •OO95 .0105 .010 .Oil .017 .017 .015 .009 .0125 .012 .179 .190 .191 .191 .082 .082 .080 .080 .015 .0165 .0155 .015 .0125 .013 .014 .019 .018 .022 .020 .021 .0195 .019 .0205 .020 .024 .025 •0235 .022 .025 .021 .0215 .022 .022 .023 .027 .027 .023 .024 .0245 •025 .024 .026 .024 .024 .032 •034 .104 .112 .104 .118 .119 .031 .036 .030 .032 .031 .025 .027 .029 .036 •037 .042 •045 •045 .040 .04I .026 .027 •033 .042 •045 .038 .046 .040 .060 .060 .021 .023 .028 .038 .046 No. 15,487. .262 •255 .262 .251 .269 .014 .012 .012 .013 .015 .015 .017 .017 .013 .Oil .013 .012 .010 .013 .0215 .020 .021 .OI4 .028 .025 .028 .026 .021 .019 .012 .016 .016 .OI4 .020 .018 .017 .OO85 .009 .010 * Dp4- still in place. scott: toxodonta of THE SANTA CRUZ beds. 185 No. No. No. No. No. No. 15492. 15,000. 15,336. 15,969. 15,256. 15,487- Lower cheek-teeth series, length . . .180 .200 .198 .203 “ premolar series, length . .069 .086 •0795 .074 Pt» length • -oi35 .017 •0155 .013 “ width . .009 .0105 .011 .0095 Pj, length • -015 .017 .020 .019 .017 “ width ..... . .013 .013 .013 .015 .014 length . .018 .021 .023 .0215 .022 .017 “ width ..... . .015 .0145 .0125 .016 .016 .014 Py, length . .021 .025 .025 .026 .023 .019 “ width ..... . .016 .017 .014 .018 .018 •oi55 Lower molar series, length . . .110 .127 .112 .120 .129 My, length . .027 .031 .031 .029 .029 “ width ..... . .015 .0165 .015 .0185 .018 M^, length • -034 .038 •034 .036 •0385 .032 “ width ..... . .0165 .017 .015 .016 .0195 .020 M^, length . .052 .060 .049 •055 .066 •073 “ width . .0175 .017 .014 .015 .020 .016 In the foregoing table No. 15,252 is the skull of a young animal, with dp— still in place. No. 15,135 is much the smallest adult skull in the collection, so that I was inclined to refer it to N. conspurcatus, but it is rather too large to be referable to that species ; it is a little older than the preceding individual and p- was already erupted, but not yet in use, as the crown is still much shorter than that of p- or m-. Though the total length of the upper dentition is so much less than in No. 15,252, it will be observed that the difference is chiefly in the anterior region and that the grinding teeth are of nearly the same size in the two skulls. No. 15,000 is a fully adult skull, but still in early maturity, and Nos. 15,969 and 15,256, which are of almost the same size, are beginning to show signs of age, while No. 15,487 is the mandible of a very old animal. The table brings out clearly the changes in the relative dimensions of the teeth, and especially of the last molar, which accompany increasing age. In the next table are given the measurements of the milk-teeth, taken from two individuals differing considerably in age and size. Measurements. Upper milk-dentition, length Di-, length (i. e. ant. -post, diameter) . “ width (i. e. transverse diameter) . No. 15,001. -?-i33 . .010 . .011 No. 16,010, PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. 1 86 Di1, length “ width .......... Upper milk-canine, length ....... “ “ “ width Upper milk-premolar series, length Dp-, length “ width .......... Dp-, length “ width .......... Dp-, length .......... “ width .......... Dp^, length .......... “ width .......... No. 15,001. . .013 . .006 . .014 . .006 . .087 . .017 . .009 . .019 . .017 . .0265 . .018 • -033 . .018 No. 16,010. Lower milk-dentition, length . .108 .118 Diy, length . .009 .007 “ width .......... • .011 Dig , length . .006 “ width .......... . .013 .013 Di¥, length .......... . .005 .008 “ width .......... . .014 .016 Lower milk-canine, length . . .015 .017 “ “ “ width . .0045 .006 Lower milk-premolar series, length • -079 .084 Dpy, length . .017 .018 “ width .......... . .009 .009 Dpg, length . .019 .019 “ width .......... . .009 .0095 Dp¥, length . .021 .023 “ width . . . . . . . . . . .010 Dpy, length . .026 .026 “ width .......... . .012 .Oil My, length . . .030 No. 15,001 is the skull of a very young animal, in which the first upper true molar (m-) is just beginning to erupt and shows no sign of abrasion, while the first lower molar (mT) was already in partial use, the anterior crescent being slightly worn. No. 16,010 is the left ramus mandibuli of an older and larger animal, in which mT is quite fully erupted and both crescents are worn, indicating that the corresponding upper tooth was in function. Curiously enough, no traces of the germs of the permanent pre- molars are yet visible. Doubtless, the germs were present in the living jaw, but not calcified. scott: toxodonta OF THE SANTA CRUZ beds. 187 So far as can be determined at present, there are no very marked skull- characters which are diagnostic of the present species. In the great majority of the skulls found the frontals have large, triangular nasal pro- cesses, but these processes vary in size and not infrequently are absent. Normally, the upper profile of the skull is nearly straight, or slightly con- vex antero-posteriorly, but in several specimens the sagittal crest rises quite steeply from the forehead to the occipital crest. In no individual of this species have I observed a descent at the forehead, such as is charac- teristic, though not invariable, in N. conspurcatus. There is considerable variation in the form of the chin and symphysis of the mandible, a varia- tion which does not appear to be correlated with age. In some individuals the symphysial region is much more depressed, in others more erect and abruptly rounded. Measurements. No. No. No. No. No. No. No. 15,252- 15,135- 15,000. 15,336. I5,i4i • 15,256- 15,001. Skull, length in median base line . .420 ?.36i •425 .448 •451 423 .258 “ fr. occ. condyles •447 •437 .472 .478 •442 “ length fr. occ. crest to end of nasals •354 •330 .366 •389 .227 Cranium, length fr. cond. to orbit. .242 .242 •245 .228 ?-i55 Face, length orbit to prmx. . .219 .202 •215 .246 .230 .123 Occiput, height .161 •155 .146 .076 width at base .... .216 .217 “ over condyles .098 .110 .127 .096 Sagittal crest, length .... .090 .066 .085 .101 Zygomatic arch, length fr. glen. cav. . .164 •153 .168 .165 .172 .167 .103 “ “ width. . . •075 .078 .087 .106 •0355 Cranium, width at constriction •055 •059 .062 .060 •055 zygomatic width . .251 .258 .225 depth at tymp. bulla .144 .121 .136 .068 Face, width over lachrymals .112 • 132 .127 •133 .091 " “ “ prmx. .065 •075 .084 .074 .048 “ depth at m1 .... •173 •153 .166 .188 a a a pi_ •145 .124 .136 .129 .140 .079 Nasals, length .184 .180 .202 .204 •095 Palate, length in med. line . .260 .229 .269 •293 .263 .150 “ width at p- .032 .041 •043 •043 .041 •035 11 ii 11 ^3. .081 .088 .098 .106 .098 PATAGONIAN EXPEDITIONS I PALEONTOLOGY. 1 88 No. No. No. No. No. I5A92. 15,000. 15,336. 15,256. 15,001. Mandible, length (excl. teeth) . .361 •425 .410 .412 .212 “ of symphysis . .092 .091 .105 •043 width at i^ . . .065 .076 .065 .042 depth at pj . • -073 .067 .080 •073 ■043 “ “ << m3- . . .077 .100 .080 .084 height of condyle . . .191 .203 .202 .203 coronoid . .200 .227 •239 width, angle to . .149 .192 •152 .168 The above measurements are taken, in nearly all cases, from the same individuals as those of which the tooth-measurements are given in the pre- ceding tables. No. 15,001 is a very young skull with all the milk- dentition in place and the first molar, above and below, in process of eruption ; this skull is not only very much smaller than the adult, but is differently proportioned, the cranium being relatively longer and broader and the face shorter. Of the six adult skulls measured, one, No. 15,135, is very much smaller than any of the others and may possibly belong to a different species, though the evidence is insufficient for such a distinction. The mean basal length of the other five skulls is .435 M. and the skull in question is only 17 per cent, shorter than this average length. Even from the largest of the series, No. 15,141, it differs by only 20 per cent, and such differences are well within the limits of fluctuating variability, as determined by the measurements of many existing species. On the other hand, what suggests the possibility of a specific difference, is the isolated position of the small skull, which is not connected with those of the usual size by numerous intergradations. Nor would this result be materially changed if the measurements of the numerous skulls in the Princeton and New York collections were cited, the series in the table being sufficiently representative. Bones of the axial and appendicular skeleton are adequately known only in the present species and the account of them in the generic description has been drawn entirely from N. imbricatus. In the abun- dant material there is no complete individual skeleton, though almost all the skeletal elements are represented in the collection. Among these elements there is considerable variation in size and also, in certain cases, some well marked differences in structural features, but, unfortunately, it is not yet practicable to associate these differences with peculiarities of scott: toxodonta OF THE SANTA CRUZ beds. 189 the skull and dentition. It is still an open question, therefore, whether the differences in vertebrae and limb-bones are specific or merely individual. In two specimens the neck is nearly complete and one of these (No. 15,000) is fortunately associated with a skull, the measurements of which are given in the preceding table. Measurements. No. 15,968. No. 15,000. No. 15,489. No. 15,967. No. 9192. Atlas, length ..... .069 .078 “ width . . . . . 7.246 “ height (dorso- ventral) •0 77 .082 Axis, length inch odontoid . .081 •073 .085 “ “ excl. “ •055 .051 •059 .060 “ width of anterior face. .101 •093 • 113 “ “ posterior face .046 .046 .046 •055 “ height, inch spine .097 .105 Cervical 3, length of centrum .031 •035 .038 “ width of anterior face .042 .042 .044 “ height excl. spine .063 .072 .074 “ width over transv. proc. . .125 .150 Cervical 4, length of centrum •033 •034 •035 .040 “ width of anterior face .040 .042 .047 .048 “ height, excl. spine .062 .068 .071 “ width over transv. proc. . .140 Cervical 5, length of centrum .031 •035 •035 .040 “ width of anterior face •039 •043 .046 .048 “ height, excl. spine .058 .067 .071 .079 “ inch spine .088 .090 .128 “ ant. -post, width of spine . .019 .023 .022 .025 “ width over transv. proc. . •115 .136 .153 Cervical 6, length of centrum •034 .030 •035 •033 .042 “ width of anterior face .041 .044 •043 .048 “ height, excl. spine .069 .061 .060 .070 .083 “ “ “ inch spine . .097 “ ant. -post, width of spine . .021 .023 .026 .024 .024 “ width over transv. proc. . .136 •153 “ ant. -post width inf. lamella .083 .100 7.098 Cervical 7, length of centrum .036 •034 .040 •039 .046 “ width of anterior face •043 .044 •047 .049 “ height, excl. spine .076 .068 .065 .076 .084 “ ant. -post, width of spine . .026 .028 .032 .030 .032 190 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. Thoracic 2, length of centrum width of anterior face “ over transv. proc. “ ant. -post, width of spine Thoracic 4, length of centrum width of anterior face “ over transv. proc. “ length of spine . Thoracic ?ll, length of centrum . “ width of anterior face “ over transv. proc “ length of spine Thoracic ?i6, length of centrum . “ width of anterior face i < H U il „ _ 1 i • spine at tip “ length of spine (fr. arch) No. 15,968. No. 15,000. No. 15,489. No. 15,967. No. 9192. •151 Cervical 7, width over transv. proc. . .148 .132 Neck, length in straight line .285 .300 Thoracic 1, length of centrum .044 •037 •045 “ width of anterior face .050 .041 .044 “ “ over transv. proc. . .115 .138 “ ant. -post, width of spine . •034 .032 .032 .044 .047 .088 .124 .046 .048 .120 •035 .048 .048 .101 .280 .048 .041 •057 .104 .050 •045 .132 .044 No. 15,492. Lumbar 2, length of centrum . . .050 .053 “ width of anterior face . .035 .046 “ length of spine (fr. arch) . .054 Last lumbar, length of centrum . . .047 width of anterior face . .055 “ “ posterior face . .057 No. 15,489 is remarkable for the massiveness of the cervical vertebrae and their thick spines ; unfortunately, these spines are all broken, so that their length is not determinable, but they were evidently much longer than in any of the other individuals. No. 15,967 is a large animal, which measures 128 mm. across the occipital condyles, and the third upper molar is 65 mm. in antero-posterior length by 44 mm. in trans- verse width. No. 15,489 is a very similar animal, of which the skull and teeth are not preserved, while No. 15,968, also a large individual, is represented by a considerable part of the skeleton. No. 15,492, on the contrary, is quite small, the lower series of grinding teeth measuring only 179 mm. in total length. SCOTT: TOXODONTA OF THE SANTA CRUZ BEDS. 1 9 1 With the exception of some phalanges of the manus, all of the bones of the fore-limb are represented in the collection, but not by any single individual. The proportions of the various elements cannot therefore, in all cases, be determined by a simple comparison of the measurements. Measurements. No. 15,489. No. 15,968. No. 15,967. Scapula, length ■ .411 •378 greatest width .... .205 width of neck .... .091 .080 ant. -post. diam. glen. cav. .063 .062 “ transv. “ “ “ . . .049 .050 Humerus, length from head . .298 “ “ “ ext. tuber. . •330 proximal thickness .122 No. 15,256. No. 15,982. Humerus, distal width over epicondyles .099 .109 width of trochlea . •0 77 .085 Ulna, length • .331 •349 •349 “ of olecranon . .113 .112 •115 “ width at sigmoid notch . .051 •057 .056 “ distal width .015 .020 .020 “ thickness .... .021 .029 •0255 Radius, length • -255 •257 .260 proximal width .046 .044 •045 “ “ thickness . .031 .028 .031 “ distal width .... • -055 .051 .056 “ thickness . .044 .042 The most complete manus in the collection, No. 1 5,460 (which is shown in Plate XXII, fig. 5), was found isolated, not in association with other bones, but enough is preserved in several other individuals to permit a comparison. Measurements. No. 15,460. No. 15,968. No. 15,967. Carpus, length in median line • -045 width of proximal row .084 “ “ “ distal row . .084 Pisiform, length ..... .050 .045 greatest dors.-vent. depth .032 .030 width of free end . .023 .022 .023 Metacarpal II, length .... . .117 .119 .120 192 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. No. 15,460. No. 15,968. No. 15,967 Metacarpal II, proximal width . .0295 .030 .031 thickness. .030 .029 •0315 “ dist. width of shaft • -033 •034 “ width of trochlea . .023 .023 Metacarpal III, length .123 .125 .128 “ proximal width . . .039 .036 “ thickness .038 .038 .036 “ dist. width of shaft .050 “ width of trochlea . • -033 .031 .032 Metacarpal IV, length . .110 •115 .120 “ proximal width .032 .028 •033 “ “ “ thickness .029 .0265 .028 “ dist. width of shaft • -0345 •037 “ width of trochlea . .026 .028 Phalanx I, digit II, length .030 “ “ “ “ prox. width . .024 Phalanx 2, digit II, length .024 “ “ “ “ prox. width . .021 Phalanx 2, digit III, length . .027 “ “ “ “ prox. width . . .034 Phalanx 2, digit IV, length . . .024 “ “ “ “ prox. width . .024 Only in No. 15,967 is the pelvis sufficiently well preserved to yield satisfactory measurements and, even in this case, some of the dimensions can be given only approximately, as, for example, the breadth of the iliac plate. Measurements. Pelvis, length. .... . .446 Acetabulum, ant. -post, diameter . . .066 Ilium, length . .260 “ transv. . .066 “ width of neck . • -057 Obturator foram., ant. -post. diam. . .098 “ “ “ plate (approx.) . • .185 “ “ transv. “ . .065 Ischium, length .... . .190 Symphysis, length . .142 posterior width • .151 Nearly all the femora are more or less damaged, in consequence of which measuring is difficult, but the subjoined table will suffice to show the more important variations in size and proportions. scott: toxodonta OF THE SANTA CRUZ beds. 193 Measurements. No. 15492. No. 15,968. No. 15,967- Femur, length from head • .311 •336 “ grt. trochanter . .309 •335 “ least width of shaft below 3d troch. . .042 •059 .041 “ thickness “ “ at same point . .033 •037 .038 “ proximal width . .120 .138 .123 “ distal width .082 •095 “ thickness. . . .096 .110 “ width of trochlea .048 •055 •059 Patella, length .... •053 width .... .058 thickness .... .042 Tibia, length (inch spine) .297 •3i5 “ proximal width . .085 thickness .086 “ least width of shaft .021 .026 “ thickness of shaft .031 .031 “ distal width. .048 “ thickness . •043 .048 Fibula, length .... .260 .266 “ proximal width . .029 •033 thickness . .048 “ least width of shaft . .015 .018 “ thickness of shaft .019 .020 “ distal width .032 •033 “ thickness. •039 .041 Pes, length in median line .216 Tarsus, length in median line •073 Astragalus, length .... .050 .063 width of trochlea. .028 •034 “ “ of head . .026 .030 Calcaneum, length. .086 .089 proximal width . •034 .040 distal width .031 •033 width over sustent. •045 .052 Metatarsal II, length •073 “ proximal width .017 “ “ “ thickness. .026 “ distal width of shaft .023 “ width of trochlea . .017 Metatarsal III, length . •075 “ proximal width .024 “ “ “ thickness .029 “ dist. width of shaft .030 194 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. No. 15,968. Metatarsal III, width of trochlea . .025 Metatarsal IV, length ..... .076 “ proximal width .023 “ “ “ thickness .027 “ dist. width of shaft .028 “ width of trochlea . .023 Phalanx I, digit II, length .... .020 “ “ “ proximal width .020 Phalanx 2, “ “ length .... .018 “ “ “ “ proximal width .017 Ungual, “ “ length .... .024 “ “ proximal width .016 Phalanx i, digit III, length .... .028 “ “ “ proximal width .026 Phalanx 2, “ “ length .... .023 “ “ “ proximal width .028 Ungual, “ “ length .... •033 “ “ proximal width .026 “ “ greatest width •033 Phalanx I, digit IV, length .... .027 “ “ “ proximal width .024 Phalanx 2, “ “ length .... .021 “ “ “ “ proximal width . .022 Ungual, “ “ length .... .024 “ “ proximal width .018 “ “ greatest width .020 A comparison of the dimensions of the pes with those of the manus of the same individual, as given in the table on p. 191, will show how very small the hind-foot is, not only in relation to the fore-foot, but still more in proportion to the size of the whole skeleton. Localities. — This species is one of the most abundant individually and also one of the most widely distributed elements of the Santa Cruz fauna and occurs in almost every fossiliferous locality. Skulls, unfortunately fragmentary, that cannot be distinguished from N imbricatus , are found at Lake Pueyrredon in association with another species, N. cornutus , described in a subsequent section. In the region of the plains and the Atlantic coast, the typical Santa Cruz beds have yielded a very large number of individuals and all of those cited in the preceding pages were found on the coast, at Coy Inlet, and 5, 8 and 10 miles south of that Inlet. This, however, is a mere accident of preservation, the better scott: toxodonta OF THE SANTA CRUZ beds. 195 specimens naturally being selected for the purposes of description and measurement. Other points on the coast, such as Cape Fairweather, and in the interior, such as Killik Aike, have produced numerous repre- sentatives of this abundant species. Nesodon conspurcatus Ameghino. (Text-figure 19, p. 142.) Protoxodon conspurcatus Amegh.; Enum. sistem., etc., 1887, p. 16. Contrib. al conoc. de Mam. fos. de la Repub. Argent., 1889, p. 445. ? Atryptherium bifurcatum Amegh.; Enum. sistem., etc., 1887, p. 18. Acrotherium patagonicum Merc.; Rev. del Mus. de La Plata, T. I, 1891. Nesodon bifurcatus Merc.; Ibid., p. 397. Nesodon Rutimeyeri Merc. ; Ibid., p. 401. Nesotherium carinatum Merc.; Ibid., p. 412. Nesotherium elegans Merc.; Ibid., p. 415. Protoxodon evidens Merc.; Ibid., p. 428. Nesodon conspurcatus Amegh.; Rev. Argent, de Hist. Nat., T. I, 1891, P- 377- Nesodon andium Amegh.; Ibid. Nesodon patagonicus (Merc.) Lydekker; Anales del Mus. de La Plata, T. II, 1893, p. 36. The type-specimen of N. conspurcatus , as originally described (Ame- ghino, ’87^, 16), has only the last two premolars and the molars of the upper jaw, the series of five teeth together measuring 120 mm. in length. From additional material Ameghino subsequently characterized this spe- cies by its small size and the rudimentary condition of the first premolar in each jaw (’91* 377), the latter feature alone distinguishing it from N. andium , as the two supposed species are identical in size. The reduction of the premolar is probably individual rather than specific, for the small teeth that follow the tusks (i-, c and px, c and py) are very variable and of little functional importance. The only character which distinguishes the present species with any degree of constancy is its small size and even this may prove to be fallacious, as there is much difference in this regard among the individuals of N imbricatus , the smallest of which approximate N. conspurcatus. 196 PATAGONIAN EXPEDITIONS: PALAEONTOLOGY. In the Princeton collection there is no specimen which can be confi- dently referred to the latter species, although one (No. 15,492) has an inferior dentition measuring only 225 mm. from L to mi, inclusive, while for N. conspuvcatus and N. andium Ameghino gives this measurement as 220 mm. On the other hand, No. 15,492 has a much larger and more robust jaw and probably a considerably larger skull. The skull figured by Lydekker (’93, PI. XIV) has a sharp descent at the forehead, but this is very much less marked in other individuals of similar size, especially in the type of N andium of the Ameghino collection. Localities. — No definite localities are given by Ameghino, Mercerat or Lydekker, although the first-named writer states that all of his specimens of N. andium were found “near the Cordillera, in the vicinity of Lake Argentino” (’94*, 239). Nesodon cornutus, sp. nov. (Text-figures 35, 36, 38.) The type, and, as yet the only known representative, of this species is a fairly well preserved skull, without mandible, and with most of the teeth Fig. 35. Nesodon cornutus, type: Skull, right side, X 3. (No. 16,012.) either damaged or quite destroyed. In size, it is rather smaller than the average example of N. imbricatus and larger than N. conspuvcatus. One scott: toxodonta OF THE SANTA CRUZ beds. 197 characteristic feature of this skull is the great vertical height of the cra- nium, which finds expression in the shape of the occiput. This surface, which in N. imbricatus is low and wide, the transverse much exceeding the dorso-ventral diameter, in the present species is relatively high and narrow and the two diameters are nearly equal, though the width is slightly greater than the height. The sagittal crest, which is much higher than in any example of N. imbricatus that I have seen, rises abruptly at the junction of the supraciliary ridges and thence pursues a nearly horizontal course to Fig. 36. Nesodon cornutus, type: Occiput, X b (No. 16,012.) the occipital crest, where there is a slight descent. The zygomatic arches arise lower down upon the occipital crest than in the preceding species and much more of the cranial wall is visible in side-view than in them. The mastoid process is broken on both sides, but appears to have been somewhat narrower than in N. imbricatus. 198 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. In all of the many skulls belonging to the latter species which I have examined, the triangular forehead is quite smooth, but in N. cornutus there is a low median protuberance, or boss, in the posterior portion of the forehead between the converging supraciliary ridges, very much as in Adinotherium. Very probably, this protuberance indicates the Fig. 37. Nesodon imbricatus: Occiput, X b (No. 15,437.) presence, in the living animal, of an incipient dermal horn. Behind and on each side of the protuberance the forehead is finely rugose and quite different from the smooth surface seen in the other species. Above the orbits, the frontals have an unusually swollen appearance, due, no doubt, to a large development of the sinuses, and the postorbital processes are longer and more pointed than in the preceding species. The fronto-nasal suture resembles that of Adinotherium in form and in the absence of nasal processes of the frontals, but no dependence can be placed upon this character without a very large suite of specimens. From a single individual it is not possible to say whether this suture is constant or fluc- tuating. SCOTT: TOXODONTA OF THE SANTA CRUZ BEDS. 199 The nasals, the upper profile of which is a nearly straight, hori- zontal line, are quite strongly convex in the transverse direction and of spatulate form ; they are widest near the anterior end of the premaxillary suture, from which point they narrow very slightly forward and end in very Fig. 38. Nesodon cornutus, type: Forehead, showing horn and fronto-nasal suture, X §• (No. 16,012.) bluntly rounded tips, which project but little in front of the free border of the premaxillaries. Posteriorly, the nasals contract very gradually to a point above the second molar and thence broaden slightly to the frontal border ; the hinder ends, which are enclosed by the frontals, are wedge- shaped. The lachrymal appears to have no spine, but, as this bone is slightly injured on each side of the head, this is somewhat uncertain. Measurements. Upper canine, length (i. e. ant. -post. diam.) 012 Upper premolar series, length . . .077 Upper canine, width (i. e. transv. diam.) .011 P1, length . 013 Upper cheek-teeth series, length . . .187 “ width 014 200 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. Measurements. Pa, length .... . .022 M-2-, length •055 “ width .... . .022 “ width ...... •045 P-, length .... . .024 Skull, length occip. crest to tip of nas. . •379 “ width .... . .024 Cranium, length cond. to orbit. . •235 PL length .... . .024 Occiput, height .179 “ width .... . .026 width at base .... •195 Upper molar series, length . . .117 Sagittal crest, length .... .106 M— , length .... • -035 Zygomatic arch, length .156 “ width .... • -033 Cranium, width at constriction . .050 M-, width .... . .042 Nasals, length. ..... .198 Localities. — The type specimen (No. 16,012) was found by Mr. Hatcher at Lake Pueyrredon. The teeth show that the animal was past maturity. ADINOTHERIUM Ameghino. (Plates XII, XVII, XVIII, XX, XXI, XXVI, XXVII.) Nesodon Owen (in part); Phil. Trans., 1853, p. 291. Adinotherium Amegh. (in part); Enum. sistemat., etc., 1887, P- I7- Acrotherium Amegh. (in part) ; Ibid. Nesodon Lydekker (in part); Anales del Mus. de La Plata, T. II, 1893, p. 25. Noaditherium Amegh.; Anales del Mus. Nac. de Buenos Aires, T. XVI, 1907, p. 84. Most writers on the subject have united this genus with Nesodon , but I am inclined to follow Ameghino in regarding it as distinct. The most characteristic features of difference maybe enumerated as follows: (1) The species of Adinotherium are all much smaller animals than those of Nesodon , with no transitions. In itself, this difference is not of generic value, but it is constant and not without significance. (2) The upper premolars are of somewhat simpler pattern, a distinction which can be observed only in nearly or quite unworn teeth. (3) The anterior lobe of the lower molars is relatively broader — the posterior lobe narrower. (4) The rostrum, including the anterior part of the mandible, is much more decidedly broadened, suggesting the form of this region in Toxodon. (5) In some species, at least, there is on each frontal a convex rugosity, which apparently indicates the presence of a small dermal horn. In addition, there are several constant and characteristic differences in the vertebrae SCOTT: TOXODONTA OF THE SANTA CRUZ BEDS. 201 and limb-bones, to which attention will be called in the succeeding pages. Of these differences, perhaps the most important are to be found in the calcaneum and astragalus, which, though of the same type, are yet quite distinct. Dentition (Pis. XVII, figs. 4, 5, 10; XVIII, fig. 5 ; XX, fig. 2; XXVII, fig. 5). — The dental formula is the same as in Nesodon , If, Cf, P i, Ml, and the individual teeth very closely resemble the corresponding ones of the latter, though with certain small and constant differences, which are not altogether without importance. A. Upper Jaw. — The median incisor is broad, heavy and chisel-shaped, contracting much to the root and therefore diminishing in size with age and the progress of abrasion, while i- is a large, trihedral and acutely pointed tusk, growing from a persistent pulp. Both teeth have the same form as in Nesodon and pass through the same remarkable series of changes in appearance as in that genus. (See p. 1 1 9). Ameghino states (’94/') 227) that in the fully adult animal the posterior part of this tooth becomes narrowed, ceases to form enamel and gradually develops a long, cylindrical root, but I have seen no example of this. The third incisor and the canine are very small, simple teeth, relatively even smaller and of less functional importance than in Nesodon. There is some individual, or possibly specific, difference with regard to the position of the canine ; it may either be almost in contact with the first premolar, or separated from it by a very short diastema, or, more rarely, by a considerable diastema. The second and third premolars are somewhat less complex than the corresponding teeth in Nesodon , though the difference is not shown in worn teeth, but only in those which are quite freshly erupted ; it consists in the much smaller size of the posterior valley in the teeth of Adino- therium and in the reduced number, or even absence, of the pits in the floor of that valley. In the worn state, the only noteworthy difference from the premolars of Nesodon is in the external wall of the crown, which in p- is less distinctly divided into two lobes, but much more so in p- and In Nesodon , it will be remembered, this division is almost obso- lete in p-. The molars are so precisely like those of Nesodon , save only in size, as to require no particular description. B. Lower Jaw. — The incisors and canine are almost exactly like those of Nesodon , but Ameghino states (loc. cit.) that the tusk-like i-3 eventually 202 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. forms a root and does not continue to grow throughout life. In the material before me I have found nothing confirmatory of this statement, though I am not in a position to question it. The molars differ quite constantly from those of Nesodon in a few particulars ; the anterior lobe or crescent is relatively somewhat broader and the posterior lobe nar- rower than in the larger animal, but the difference is not great. All of the lower grinding teeth have, in fact, a curiously narrow, slender appearance. Milk Dentition (Pis. XVII, figs. 7-9; XVIII, figs. 2, 3). — The tem- porary teeth differ but slightly from those of Nesodon and the functional milk-incisors are almost identical in the two genera, but di- and the upper canine are even smaller in Adinotherium , the former especially being a mere slender style ; the canine is larger than di-, but still very small. In the accessible material of Adinotherium , in which the milk- premolars are considerably abraded, I can detect no tangible difference from those of Nesodon , except that in dp- only two spurs, instead of three, project into the principal valley, and that the posterior valley is smaller, as is also the case in its permanent successor. Unworn teeth would probably show greater differences. Prelacteal Dentition. — As yet, I have seen no well-preserved examples of the prelacteal dentition in this genus, but one individual indicates that Fig. 39. Adinotherium ovinum : Left premaxilla with permanent incisors and prelacteal i— , X x- this dentition was present, at least in part. This animal, which is shown in text-fig. 39, retains the small, quill-like, prelacteal upper tusk, which is external to the permanent i- and cannot possibly be di-, as it is far too small, and of an entirely different form. Skull (Pis. XX, figs. 1-4; XXVI, fig. 2). — Aside from its smaller size and lighter construction, the skull of Adinotherium is so very closely like that of Nesodon , that it is difficult to find any tangible points of distinc- scott: toxodonta OF THE SANTA CRUZ beds. 203 tion. The most obvious and constant differences are (1) the palatine constriction at p1 and the forward expansion of the muzzle from that point, which is much more marked than in Nesodon and suggests the character of Toxodon ; (2) the presence of small rugosities on the frontals, which Ameghino regards as the bases of incipient horns ; (3) the greater length of the sagittal crest ; (4) the different character of the fronto-nasal suture. The upper profile of the skull is somewhat sinuous, but does not depart widely from a straight line and lies nearly in the same horizontal plane throughout. The sagittal crest is slightly convex in the antero- posterior direction, and the nasals are also somewhat convex in the same direction ; between the two, there is a moderate descent at the forehead. The occiput is even lower and wider proportionately than in Nesodon , though the difference is not a marked one, and the zygomatic arches rise rather higher upon the lambdoidal crest, giving the skull a some- what different appearance in side-view. (See PI. XII, figs. 1 and 2.) The remarkable peculiarities of the auditory region are identical in the two genera except that the mastoid and its process are proportionately less developed in Adinotherium. The parietals are relatively somewhat longer than in Nesodon and are contracted by a more definite postorbital constriction, and the brain-case is slender and of small capacity. The frontals differ from those of Nesodon in several respects ; the supra- ciliary ridges are rather better defined and extend nearer to the postorbital processes. Between the supraciliary ridges, there is, in a certain proportion of the skulls, a raised and more or less roughened area, which may be single, but is much more commonly divided by a median depres- sion. Ameghino (’07) regards this rugosity as the attachment for an incipient horn and this is very probably the correct explanation, but there is no such definitely rounded form as the dermal horns have in the Perissodactyla, even in the earliest stages, and the median depression would seem to indicate the presence of a pair of very small horns rather than of a single one. Further, there is some indication of a sexual difference in the development of the horns, as in some skulls the rugosities are much less prominent, or may be absent altogether, a difference which is not correlated with the age of the various individuals. Whatever its significance, this roughened area is an almost constant differ- ence from the smooth, plane or even concave surface, which occupies the 204 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. corresponding position on the forehead of Nesodon. Still another differ- ence from the latter is to be found in the character of the fronto-nasal suture. When the frontals possess nasal processes, which separate Fig. 40. Adinotherium ovinum: Series showing the variations^of the fronto-nasal suture and the frontal horn, X b A, No. 15,983; B, No. 15,482; C, No. 15,382; D, No. 9140, A. M. N. H. ; E, No. I5T59; F, No. 15,118; G, No. 9571, A. M. N. H.; H, No. 15,493- SCOTT: TOXODONTA OF THE SANTA CRUZ BEDS. 205 the hinder ends of the nasals, these processes are always much shorter and narrower than in Nesodon, though of similiar triangular shape. More commonly, however, there are no such processes and the anterior border of the frontals is deeply emarginated to receive the nasals in a concavity. According to Ameghino (’94*, 245) the skulls which have triangular pro- cesses of the frontals between the hinder ends of the nasals are the most abundant, but this is certainly not true of the Princeton collections. The nasals are long and narrow, slightly convex antero-posteriorly and strongly so in the transverse direction ; anteriorly, their projection in front of the premaxillae is rather shorter proportionately than in Nesodon , giv- ing a slightly different shape to the narial opening. Posteriorly, there are considerable differences among the various skulls in the form of the nasals, differences which Ameghino has employed for taxonomic purposes ('94/', 245; ’07, 65, ff. ), though the variability is such that the taxonomic importance of these characters is very doubtful. Four different types may be distinguished in the form of the hinder ends of the nasals and in the consequent character of the fronto-nasal suture. (1) The nasals are of almost uniform width from end to end and are simply rounded pos- teriorly; (2) they are suddenly narrowed at the contact with the frontals and make the fronto-nasal suture almost triangular ; (3) they gradually narrow backward from the frontal contact and the frontals are less deeply emarginated than in the preceding types. (4) In the first three types the two nasals are in contact with each other throughout their entire length, while in the fourth their posterior ends are separated by small nasal processes of the frontals, as above described. Other slight differences, probably individual, though perhaps specific, may be noted in the shape of the nasals. Usually, these bones are of nearly constant width, with a very shallow constriction near the middle of their length and broadening slightly in front of and behind the constriction. In some cases, however, the bones are widest near the middle, contracting toward the ends, but the difference is by no means striking. The premaxillae differ hardly at all from those of Nesodon; the ascend- ing ramus is somewhat broader antero-posteriorly in proportion to its height dorso-ventrally and its anterior free border is more vertical, form- ing a more definite angle with the horizontal ramus. The spine-like rugosities which arise from the antero-dorsal angles of the horizontal rami are rather less conspicuous than in the larger genus. The maxil- 206 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. laries have almost the same proportions as in the latter, while the lachry- mals are distinctly smaller, though of similar shape, and have quite prominent spines. Owing to the posterior height of the zygomatic arches, where they rise almost to the level of the sagittal crest, the descent of the arches forward is steeper than in Nesodon and the jugal has a more distinct postorbital process or angulation. The hard palate differs in some particulars from that of Nesodon , though the difference can be fully appreciated only when the comparison is made between skulls which are quite free from lateral crushing. In such skulls, the most obvious difference in the palate is its much more triangular form in Adinotherium, being relatively broader behind and narrowing forward more strongly to the line between the foremost premolars, thence widen- ing again to the muzzle, which, as already frequently mentioned, is decidedly broader than in Nesodon. Individually, the various elements of the bony palate differ very little from those of the last named genus, save as such difference is conditioned by the general form of the whole. The two rows of grinding teeth converge forward much more strongly than in Nesodon. The posterior nares, palatines and pterygoids are so nearly as in the latter that no particular description of these parts is required. The mandible differs in some respects from that of Nesodon. As in the latter, the two halves are, in the adult, indistinguishably fused into a single bone and form a long symphysis which is deeply channelled on the dorsal side. One of the most obvious differences from Nesodon is in the shape of the symphysial region, which, when viewed from above, is seen to be more distinctly constricted at the line of p2 and to widen more from this point to the incisive alveoli, in correlation with the similar broadening of the upper jaw. The ventral profile of the chin also differs in being somewhat less steeply inclined and more procumbent. The horizontal ramus has a relatively greater dorso-ventral depth than in Nesodon , though the difference is not a striking one, while the propor- tions of length and thickness are nearly identical in both genera. In Adinotherium, however, the ventral border is more strongly and simply convex, less sinuous, and the ascending ramus has a considerably greater antero-posterior breadth in proportion to the length of the tooth-row. The sigmoid notch is shallower than in Nesodon and the coronoid is reduced, not rising above the level of the condyle, which is much as in scott: toxodonta OF THE SANTA CRUZ beds. 207 the latter genus and has a similar dependent hook given off from the inner end. The anterior border of the ascending ramus is narrower than in Nesodon and the lineae obliquae are very obscurely marked, except dis- tally, where the internal border becomes very prominent, but the fossa behind the last tooth is much shallower than in the other genus. No part of the hyoid apparatus is preserved in connection with any of the skulls and I am therefore unable to say whether the place of attach- ment is the same highly exceptional one as in Nesodon , or not, though one decided difference from that genus should be noted. In Nesodon almost every well preserved adult skull shows the stylohyal ankylosed with the anterior end of the bulla, but I have seen no instance of such ankylosis in Adinotherium. Vertebral Column and Ribs. — So far as can be determined from the available material, the vertebral formula appears to be the same as in Nesodon , viz.: C. 7; Th. 16-17; L. 4-5; S. 6; Cd. ? The neck is short and rather weak and the individual vertebrae closely resemble those of Nesodon , with such differences of detail as would naturally occur in a much smaller and lighter animal. The atlas (PI. XXVI, figs. 7-9) is very short antero-posteriorly and very broad transversely ; the anterior and posterior cotyles are quite as in Nesodon , on a smaller scale, but the neural arch is narrower and more slender and has a more pronounced upward curvature, and the spine is quite obsolete, or very faintly indicated. The inferior branch of the first spinal nerve perforates the base of the transverse process, but the ventral opening of this canal is more widely separated from the anterior opening of the vertebrarterial canal than in Nesodon. — Except for its smaller size and lighter construction, the axis (PI. XXVI, figs. 10, n) is very like that of the latter genus ; it has a short centrum which is very broad anteriorly, contracting abruptly behind the transverse pro- cesses. The following small differences from the axis of the larger animal may be observed : (1) The odontoid process is more cylindrical and more abruptly truncated at the free end. (2) The anterior cotyles are propor- tionately narrower transversely, higher and more convex dorso-ventrally and they are separated more distinctly from the odontoid ; in some indi- viduals (or perhaps species) this separation is by means of a well-defined sulcus on each side of the process. (3) The centrum has a more distinct ventral keel and a rather more concave posterior face. (4) The neural 208 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. canal is relatively larger and the spine lower and less massive, but more produced anteriorly. (5) The postzygapophyses are more oblique, pre- senting more ventrally, less laterally, and their long diameter is antero- posterior, not dorso-ventral. (6) The transverse processes project less laterally and more posteriorly. The succeeding cervical vertebrae are smaller and lighter propor- tionately than those of Nesodon and have narrower neural arches, leaving wider intervertebral spaces ; the neural spines are very low and but feebly developed, except on the seventh vertebra. The transverse pro- cesses are more slender than in Nesodon and have no such development of the inferior lamella as in that genus. The thoracic vertebrae (PI. XXVI, fig. 12) differ from those of the last named genus chiefly in the form of the neural spines. In the anterior por- tion of the region the spines are long, laterally compressed and thin and have a strong backward inclination, but they do not have any such exag- gerated length as in Nesodon and Toxodon and there is no hump at the shoulders. Posteriorly, the spines decrease in length very gradually and, owing to the upward curvature of the back-bone in this region, the tips of the spines lie in nearly the same horizontal plane throughout, which is in marked contrast to the arrangement seen in the larger genera (see PI. XII, figs. 1 and 2). Behind the middle of the thoracic region the spines are quite low and weak, becoming broader and stronger in the posterior region, where they lose their backward inclination and become nearly erect. Cylindrical, interlocking zygapophyses are present only on the last two vertebrae and metapophyses only on the last. In all of the thoracic verte- brae, except the first and last, the neural arch is perforated on each side by a conspicuous foramen for the passage of the spinal nerve. Of the lumbar vertebrae (PI. XXVI, fig. 1) probably five were present normally, though no individual has yet been found in which all of the lumbars are preserved. As compared with those of Nesodon , these vertebrae have relatively lower and broader neural spines, which are either erect, or have a slight backward inclination, none of them inclining forward, while the transverse processes are longer, narrower and more antroverted. On the penultimate lumbar the transverse processes are broader than on the preceding vertebrae and on the last one they become very much broader and have articular facets for those of the penultimate lumbar and first sacral, as already described for Nesodon. In all SCOTT: TOXODONTA OF THE SANTA CRUZ BEDS. 209 of the lumbar vertebrae the spinal nerves pass out through deep and narrow notches, which emarginate the pedicles of the neural arch. The sacrum (PI. XXVI, fig. 1) resembles that of Toxodon on a very small scale, but is relatively longer and has one additional vertebra, six instead of five. The centra are broad and much depressed and that of the last vertebra is so large as to indicate a well developed tail, which was probably proportionately longer and more slender than in the Pam- pean genus. The surfaces for the attachment of the ilia are remarkably small, though relatively as large as in Toxodon , and, as in that genus, are formed almost entirely by the pleurapophyses of the second sacral. The transverse processes of the succeeding vertebrae are fused into a broad, very thin plate on each side, perforated by the foramina for the spinal nerves, which are relatively much smaller and less conspicuous than in Toxodon. In the latter the last sacral has on each transverse process a facet for articulation with the first caudal, but in Adinotherium the first caudal seems to be fused with the sacrum. The prezygapophyses of the first sacral are prominent and functional and bear well defined metapophyses ; those of the succeeding vertebrae are vestigial, but less reduced than in Toxodon. At the anterior end of the sacrum the neural canal is broad, but very low and depressed, and at the posterior end it is very small and nearly circular, the relative size of the canal being much the same as in Toxodon. The degree to which the neural spines of the sacrum are fused into a ridge differs in various specimens, though the material is insufficient to show whether the difference is specific or indi- vidual, or merely a matter of age. The spines of the first and last sacrals are isolated, while those of the intermediate vertebrae are ankylosed for more or less of their height, but the tips are always free. In Toxodon only the first spine is separate ; all the others are indistinguishably fused into a continuous ridge. No caudal vertebrae of this genus have as yet been identified, though the character of the tail may be inferred from that of the sacrum. The ribs are much more slender than those of Nesodon , but otherwise so like them as to require no particular description. The sternum is still unknown, but was doubtless essentially similar to that of Nesodon. Appendicidar Skeleton. — The scapula (PI. XXVII, figs. 8, 9) closely resembles that of Nesodon , but differs in a number of details, as may be readily seen on comparing Figs. 1 and 2 of Plate XII. The outline of 210 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. the blade is much the same in the two genera, but in Adinothevium the coraco-scapular notch is shallower and less distinct and the coracoid border has a more regularly convex curvature, passing without angula- tion into the suprascapular border. The spine arises farther below the suprascapular border and rises to its full height more abruptly. The two metacromial processes are of quite different shape and relative size from those of the larger animal ; the proximal one is smaller and of more triangular shape, while the distal one is much longer and narrower and has a rounded border at the free end, instead of the straight border seen in Nesodon. In all of the available specimens the acromion proper is broken away, but, from the thinness of the fractured bone, it is very probable that this process was decidedly shorter than in Nesodon and may have been almost absent. The coracoid also is very small and much more reduced than in the last-named genus. The humerus (PI. XXVII, fig. 4) differs little, save in size, from that of Nesodon. The external tuberosity is a high and very broad ridge, which extends across nearly the whole proximal end, concealing the head in front-view ; the free proximal border is more horizontal than in Neso- don. The internal tuberosity is very small and the broad bicipital groove is divided into two parts by a very inconspicuous bicipital tubercle, which is in hardly more than an incipient stage. The deltoid ridge is more prom- inent and ends more abruptly below than in Nesodon. The distal end of the humerus differs in a number of particulars from that of the latter genus ; the supinator ridge is longer and more prominent, but the epi- condyles, especially the external one, are much less so ; the trochlea is relatively narrower and has a more prominent internal flange for the ulna and a more convex facet for the head of the radius ; the supratrochlear fossa is shallow, but the anconeal fossa is very deep and a perforation connects the two. The fore-arm bones are almost exact replicas, on a smaller scale, of those of Nesodon , with hardly any differences that can be expressed in a description. The ulna (PI. XXVII, figs. 10, 11) has a very prominent and stout olecranon, which, however, is relatively somewhat shorter than in the last named genus. The sigmoid notch is almost identical in form in the two genera, but in Adinothevium the external border of the prox- imal portion is not flared upward so strongly, while the distal portion presents more anteriorly and less proximally. The shaft, which is rela- scott: toxodonta OF THE SANTA CRUZ beds. 21 I tively somewhat longer than in Nesodon, is rather less distinctly trihedral and has a more concave posterior border and the interosseous crest is much less prominent. The styloid process and carpal facet are similar to those of Nesodon , but are proportionately narrower and thicker palmo-dorsally. The radius (PI. XXVII, figs. 12-15) differs from that of Nesodon in a number of unimportant details. It is somewhat more slender in propor- tion to its length and the shaft is more irregular in shape, with less marked antero-posterior compression, and very feebly developed interosseous crest ; the proximal articular surface is more distinctly divided into two facets for the humeral trochlea and, near the internal side, the anterior border is raised into a more prominent and pointed process, and on the external side of the head is a large facet for the attachment of a sesamoid bone. On the distal end, the two carpal facets are more distinctly demarcated than in Nesodon and the external portion of the scaphoid facet does not notch the dorsal border so deeply, while the lunar facet is narrower palmo-dorsally. The manus (PI. XXI, figs. 4, 7) is of exactly the same type as in Neso- don and, save in size and in a few details of proportion, the differences are quite insignificant. The scaphoid is slightly lower and narrower, in proportion to its dorso-palmar thickness, than in Nesodon , and there are several differences in the shape of the facets. The proximal surface, for the radius, narrows more and is less elevated toward the palmar side ; the facet for the trapezoid is narrower and less deeply concave and that for the magnum contracts more toward the palmar border. There is no surface for the trapezium. On the ulnar side, in addition to the proximal facet for the lunar, there is a very small distal one for the same bone, which is not present in Nesodon. The lunar is of almost identical form in the two genera, but there are some differences in the facets. In Adinotherinm , the dorsal portion of the proximal surface is more convex and is reflected farther down upon the dorsal face of the lunar, while the palmar portion is decidedly narrower, as is also the palmar prolongation of the unciform facet, between which and the magnum facet there is no such emargination from the palmar border as occurs in Nesodon. On the ulnar side, as in the latter, there is no proximal contact with the pyramidal and the distal articulation is confined to a very narrow strip. The pyramidal is broad transversely, but very short proximo-distally, 2 12 PATAGONIAN EXPEDITIONS I PAL/EONTOLOGY. relatively shorter than that of Nesodon, but otherwise resembling it, though the proximal surface for the ulna is somewhat less deeply concave. There is more difference between the two genera in the shape of the pisiform than in any other carpal element. In Adinotherium, the proximal end of the pisiform is relatively broader and much more depressed and the facets for the ulna and pyramidal are relatively smaller ; the tuber is more slender and has a straighter inferior border, while the free, or distal, end is less thickened and rugose. It should be added, however, that there is considerable difference among various indi- viduals in the shape of the pisiform. Its character in some specimens approaches that of Nesodon , but I have no means of determining whether this variation is specific or merely individual. The trapezium is not preserved in any of the examples of the manus, but was no doubt present. The trapezoid differs in hardly any appreci- able way from that of Nesodon ; it is a little broader in proportion to its proximo-distal length and contracts to a narrow palmar edge, while the distal facet for me. II is more concave transversely, less so palmo- dorsally. The magnum differs from that of Nesodon chiefly in its reduced dorso-palmar diameter and in the very small size of the palmar rugosity. Like the trapezoid, the magnum also narrows more to a palmar edge and the facets are all correspondingly narrowed. The trapezoid facet is relatively larger and the sulcus which invades it from behind is much smaller. The unciform, like most of the other carpals, is proportionately lower and wider than in Nesodon and has a narrower palmar projection. Of the two proximal facets, that for the lunar is not reflected down upon the palmar side so far as in the latter genus and that for the pyramidal has no such palmar extension. The same is true of the facet on the radial side for the process from the head of me. Ill, this surface being confined to the dorsal moiety of the unciform, instead of extending over the whole dorso-palmar diameter of the bone, as it does in Nesodon. As in Nesodon , the metacarpus consists of three functional mem- bers, me. II, III and IV, and a vestigial representative of me. V. The functional elements, except for their very much smaller size, closely resemble those of Nesodon , but are proportionately somewhat more slender. Metacarpal II reproduces that of the last-named genus almost exactly, but the projection from the palmar side of the head is decidedly scott: toxodonta OF THE SANTA CRUZ beds. 213 more prominent, though less rugose. On the radial side, the facet for the trapezium is somewhat smaller and the proximal surface for the trapezoid is rather less concave ; the projection from the ulnar side, which abuts against the magnum, overlaps the head of me. Ill less extensively and the facet on the distal side of this projection is consider- ably narrower. The shaft, save for its smaller size and slightly more slender proportions, does not differ from that of Nesodon , but the distal trochlea is rather more symmetrical and the carina is even less prominent. Metacarpal III differs in only a few details from that of Nesodon. The proximal facet, which is overlapped by me. II, is narrower and, on the ulnar side, the process which articulates with the unciform is smaller and the facet for the unciform is narrower, especially toward the palmar side ; the two large, concave facets which articulate with cor- responding surfaces on me. IV, are more completely separated, not con- nected by a narrow, articular band. The shaft has a slightly dif- ferent shape, being of a more uniform width and not broadening distally so much. Metacarpal IV is slightly longer in relation to me. II, which it almost exactly equals in length, than is that of Nesodon , though, owing to the manner of articulation with the carpus, me. IV. extends distally below me. II. This outer metacarpal differs in a few respects, besides relative length, from that of Nesodon : The proximal facet which is covered by the head of me. Ill is narrower and is confined to the dorsal moiety of the head ; the two projections on the radial side, which fit into correspond- ing depressions on the ulnar side of me. Ill, are more completely separated and the palmar one is more prominent and, on the ulnar side, the facet for me. V is narrower. The shaft is somewhat straighter and less irregular than in the larger animal. A small and very irregular nodule represents the vestigial remnant of me. V ; it has three well- defined facets, for me. IV, the unciform and pyramidal respectively, for it has the remarkable peculiarity of articulating with the proximal carpal as well as with the distal one. I have not seen this bone in Nesodon , though there can be no doubt as to its presence in that genus. Judging from Ameghino’s figure of the manus in Xotopvodon (’94^, 246, fig. 1) it is relatively longer and narrower than in Adinotherium. The phalanges are very much like those of Nesodon , but smaller and relatively lighter. Those of the lateral digits (II and IV) are of almost 214 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. exactly the same size, forming digits of nearly equal length. Of the median digit (III) the proximal phalanx is slightly shorter than in the other digits, but broader, more depressed and symmetrical ; the second and ungual phalanges are both longer and broader than in the lateral digits, but also more depressed and flattened. The pelvis (Pis. XXVI, fig. i ; XXVII, fig. 17) displays a number of differences from that of Nesodon. The ilium has a somewhat longer neck, which is thin and laterally compressed and has no distinct ilio-pectineal process or pubic border, having lost, as in the other genus also, the original trihedral shape. The iliac expansion or plate is relatively nar- row and but little everted and the gluteal surface is nearly flat ; the crista shows a slight tendency to a division into dorsal and ventral portions, which, however, are much less distinct than in the Litopterna. The acetabulum is of a slightly depressed, oval shape and the sulcus for the round ligament is relatively much larger than in Nesodon. The ischium is slender and compressed and has no definite tuberosity or sciatic notch ; the pubis also is slender and enters into a long symphysis with its fellow. The obturator foramen differs in shape from that of Nesodon in being broader and less regularly oval. The femur (PI. XXVII, figs. 1-3) besides being smaller and lighter than that of Nesodon , differs from it in a number of details. The head is set upon a more distinct neck and is lower, not rising above the level of the great trochanter ; the pit for the round ligament is even smaller, though a trifle deeper, than in the last named genus and the notch between head and great trochanter is deeper and narrower. The pos- terior face of the proximal part of the shaft is very flat and smooth and there is no intertrochanteric ridge. The second trochanter is more prom- inent than in Nesodon , but not so elongate proximo-distally, while the third trochanter is distinctly shorter and less prominent than in the latter. The shaft has a shape similar to that of Nesodon , but is more compressed antero-posteriorly and has a less convex anterior face, and the pit for the plantaris muscle is very inconspicuous. The rotular groove and femoral condyles do not differ in any significant features from those of Nesodon. The patella (PI. XXVII, fig. 16) is very similar to that of the latter, but is relatively more elongate proximo-distally, which is chiefly due to the greater extension of the external part of the distal border, and it has the same great antero-posterior thickness ; the articular surface for the SCOTT: TOXODONTA OF THE SANTA CRUZ BEDS. 215 femoral trochlea is less unequally divided into outer and inner facets and the dividing ridge is even less distinct. As in Nesodon , the leg-bones (PI. XXVII, figs. 6, 7) are ankylosed at the proximal, but not at the distal end. Except in size, there is hardly any tangible difference between these bones and those of Nesodon , at least so far as the damaged specimens at my disposal admit of a com- parison. The tibia has a somewhat less prominent and less rugose cnemial crest, which ends rather more abruptly below. A more important, though not at all striking, difference between the two genera is in the character of the surface for the astragalus, which in Adinotherium has a better defined intercondylar ridge, covers the distal end of the bone more completely and is more broadly continuous with the facet on the internal malleolus. This latter facet is also larger and extends over the entire fibular side of the process. The fibula lacks the interosseous crest, which is so conspicuous on the distal half of the bone in Nesodon , and the distal end is relatively narrower, especially on the posterior side. There is a close resemblance to Nesodon in the character of the hind- foot (PI. XXI, figs. 3, 5, 6, 8, 9), though accompanied with more differ- ence of detail than is the case in most of the other skeletal structures. A full comparison of the two genera in regard to the pes is not, however, practicable, as I have seen no quite perfect example of the hind-foot of Adinotherium. The astragalus is proportionately longer and narrower than in Nesodon and, on the tibial side, is considerably depressed planto- dorsally ; the trochlea is somewhat, though not very much, more deeply and narrowly grooved ; the external condyle is relatively larger and rises higher above the level of the internal one, and the whole trochlea is more convexly arched proximo-distally, while the facet for the fibula is broader planto-dorsally. The rugose tuberosity on the tibial side of the trochlea, near the proximal end, which is so conspicuous in Nesodon , is much reduced and hardly perceptible in Adinotherium. In the latter genus the neck of the astragalus is considerably longer and more oblique, projecting more strongly toward the tibial side. On the plantar side, the two calcaneal facets are more widely separated and the pit between them is much shallower, and the sustentacular facet, indeed the whole tibial side of the bone, projects much more strongly plantarwards. The ex- ternal calcaneal facet is of more uniform width and the sustentacular facet is much smaller, more oval in shape, and partly or completely sepa- 2l6 PATAGONIAN EXPEDITIONS : PALEONTOLOGY. rated from the surface for the navicular, which, owing to the obliquity of the neck, it touches near the fibular border, while in Nesodon it becomes confluent with the navicular facet near the tibial border. The navicular facet itself is relatively narrower and of more quadrate outline in A dinotherium. The calcaneum has a somewhat more elongate, slender and laterally compressed tuber than in Nesodon, with a more nearly straight plantar border ; the fibular facet is shorter proximo-distally, but more prominent and rises more abruptly from the dorsal side, and the external facet for the astragalus is less oblique and is not extended so far upon the tuber. The sustentaculum is less produced proximo-distally and its two diameters are more nearly equal ; it is also relatively thinner planto-dorsally and has a deeper and better defined sulcus on the plantar face. The distal facet for the cuboid is rather more concave. Ameghino, who has given a very full comparison of the calcaneum and astragalus of Adinotherium and Nesodon, says of them: “Ces differences dans la forme du calcaneum et de l’astragale sont aussi importantes que celles qui existent entre les memes os des Paleoth£res et des chevaux, animaux qui Ton place dans deux families differentes” (’94/, 243). In my judgment, however, this greatly overstates the importance of the differences above described. These differences are merely such as one would expect to find between a relatively large and heavy animal and a small and light one of the same group. The remaining elements of the tarsus in Adinotherium are less differ- ent from those of Nesodon than are the calcaneum and astragalus. The navicular and cuboid differ hardly at all in shape and proportions from those of the latter, but the plantar hook of the navicular is even more reduced and on the cuboid also this hook is very inconspicuous, and the proximal surface of the cuboid for the calcaneum is somewhat more convex. The ento- and mesocuneiforms are ankylosed and the compound bone agrees in every respect, save size, with that of Nesodon , but the ectocuneiform differs in several details. It is relatively narrower and slightly higher proximo-distally and contracts more rapidly toward the plantar side ; the proximal facet for the navicular is more nearly concave and the facet on the fibular side for the cuboid is much more extended planto-dorsally. The metatarsals are very much smaller than the metacarpals and show the same lack of symmetry as in Nesodon . Metatarsal II is the shortest SCOTT: TOXODONTA OF THE SANTA CRUZ BEDS. 217 and most slender of the series, though there is no great difference in length between any of the three members, and is considerably more slender in proportion to its length than in Nesodon , which it otherwise re- sembles in shape and connections. Metatarsal III is much heavier and somewhat longer than mt. II, and in shape and proportions is very similar to that of Nesodon , except that the distal broadening of the shaft is more gradual. Metatarsal IV is also very stout, much heavier than mt. II ; in- deed, the proximal end has as great a dorso-plantar diameter as that of mt. Ill, but is a little narrower transversely. The phalanges of digit II are much more reduced than those in the corresponding digit of the manus and are considerably smaller proportion- ately than those of Nesodon ; they are short, narrow and thick, not de- pressed, and the small ungual is not cleft. In digit III the phalanges are very much larger than in either of the lateral digits and thus make the median toe greatly exceed both the laterals in length. Nearly all of this excess is due to the phalanges, for the three metatarsals differ but little in regard to length and, relatively, these phalanges are longer and narrower than in Nesodon. Compared with those of the median digit of the manus, they are slightly longer, broader and more depressed and flattened. The phalanges of digit IV, as in Nesodon , considerably exceed those of digit II in size. Restoration (PI. XII, fig. i). — In its general appearance and propor- tions, the skeleton of Adinotherium differs very considerably from that of Nesodon , being not only a much smaller, but also a much lighter animal, and lacking entirely the massiveness of structure which characterizes Nesodon. The proportions of the skull, as to length and dorso-ventral height, do not differ materially in the two genera, but in Adinotherium the neck is slightly shorter, the length of the skull being taken as a stand- ard, and much lighter, the processes of the cervical vertebrae being rela- tively shorter and more slender. The trunk, on the other hand, is pro- portionately longer and the vertebral column has quite a different appear- ance, owing partly to the lesser degree of curvature in the anterior region of the thorax and still more to the shorter and more slender neural spines. The changing length of the spines quite accurately compensates for the curvature of the column, so that their free ends reach nearly the same horizontal plane and the profile of the back must have been almost straight in the living animal. This is in very marked contrast to the 2l8 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. arrangement seen in Nesodon , in which the great length of the anterior thoracic spines produces a considerable hump at the shoulders. In Adi- notherium also the posterior thoracic and lumbar spines are broader and lower. The thorax is shallower and less capacious in the latter and the ribs more slender and less strongly arched outward. The scapula differs more from that of Nesodon in appearance than does any other bone of the skeleton ; the blade is narrower and more recurved and the spine also pursues a more curved course, running nearly parallel with the coracoid border. But the chief difference lies in the position and shape of the two very conspicuous metacromial processes, which in Adinotherium are closer together and the proximal one is smaller and more triangular, while the distal one is very much longer and narrower than in Nesodon. The pelvis also is longer and narrower than in the latter, the ilium having a longer neck and less expanded plate and the ischium being more elongate and slender. The limbs are relatively shorter, so that Adinotherium was proportionately a longer and lower animal, but the limb- and foot-bones are almost copies, on a smaller and lighter scale, of those of Nesodon , though there are several differences in the details of structure. For the most part, however, these minor differences are such as may properly be referred to the merely mechanical conditions of a rather massive and a small, light animal, and are hardly visible in a general view of the two skeletons, though obvious enough when the bones are separately compared. A not inconspicuous difference is in the character of the femur, which in Adinotherium is more slender, less flat- tened, and has a more prominent third trochanter ; the calcaneum also has a longer and less massive tuber, which lends a somewhat different appearance to the pes. This comparison has reference entirely to the two commonest species of their respective genera, A. ovinum and N. imbricatus. No doubt, other and greater differences might be enumerated, were the comparison extended to several species of each genus, but for this purpose the material now available is not sufficient. Species. — The lack of accurate knowledge of the stratigraphy of the Santa Cruz beds is an insuperable obstacle to any satisfactory discrimina- tion of the species of almost any genus of the Santa Cruz fauna and the toxodonts offer nearly as difficult a problem in this regard as do the gravigrade edentates. At present, we have but very inadequate means SCOTT: TOXODONTA OF THE SANTA CRUZ BEDS. 219 of distinguishing contemporary and fluctuating variations from successive and relatively constant mutations and, until that distinction can be made, nothing more than a tentative arrangement of species is practicable. Ameghino has, it is true, given some valuable stratigraphical notes upon the vertical range of the species, but these notes are far from exhaustive and they do not entirely agree with the results obtained by Messrs. Hatcher, Peterson and Brown. Before full and trustworthy evidence upon these points can be obtained, it will be necessary to make very much larger collections than have yet been gathered and, above all, to make these collections together with a minutely accurate stratigraphical survey. Until then, the problem of species will be incapable of satis- factory solution. In his latest publication on this genus (’07, 67-89) Dr. Ameghino recognizes six species of Adinotherium, A. magister , A. ovinum, A. niti- dnm , A. corriguenense, A.ferum and A. robustiim, to which should be added two species referred to other genera, the so-called Acrotherium karaikense and Noaditherium splendidwn. A. haplodontoides Amegh. (’ gib, 129; gih, 376; 94", 25) is not mentioned, even in the synonymy. The six species enumerated are arranged in two series, in accordance with the presence or absence of the incipient frontal horn. An obvious suggestion, which almost forces itself upon the observer, is that this minute horn is a sexual rather than a specific character, especially as the horned skulls are almost always larger and more robust and have more prominent and rugose crests and processes, all of which are characteristically male features. Of course, this suggestion does not necessarily imply that some of the species may not have been hornless in both sexes, for that would be quite compatible with the presence of horns in the males of other species, as appears to have been the case in Nesodon. In that genus only one of the known species, N. cornutus, distinguishable as such upon other grounds, shows any indication of the frontal horn, but, within the limits of the species, the horn may well have been a sexual character. At present, however, there is no way of deciding whether this is true or not. Ameghino attaches much importance to the form of the fronto-nasal suture for the discrimination of species in both Nesodon and Adinotherium , but I have found this character, in the latter genus as well as in the former, too fluctuating to be trustworthy. Broadly speaking, there is a distinction between the two genera in regard to the form of this suture ; in Adinothe- 220 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. rium the frontals usually have no nasal processes and, when present, these processes are always very small, while in Nesodon they are commonly large and conspicuous, but may be entirely absent, and every gradation between the two extremes may be found in any large series of skulls. In the original description of most of the species (Amegh., ’87*, 17, 18) they are distinguished almost entirely by differences of size, but in the latest paper (’07) the diagnoses and descriptions are much more full and complete and deal largely with the characters of the skull, distinguishing the species in what appears, at first sight, to be a very satisfactory man- ner. I have, however, found the greatest difficulty in applying these definitions to the material contained in the Princeton and New York col- lections. Not only does Ameghino take no account of individual and possible sexual differences, but the characters upon which he lays the chief stress are combined in a very different way in the material before me from the association of features found in his specimens. To adopt his method would involve a great increase in the number of species, for hardly any of the skulls at my disposal agree at all closely with his descriptions and figures and there is such variation among them that any constant characters are very difficult to find. It is just here that the lack of stratigraphic information makes itself so painfully felt and it may very well turn out, when that information shall have been secured, that several of the species, here regarded as synonyms, are entitled to recognition. The arrangement of species which follows is, it must again be empha- sized, purely tentative and is founded upon the assumption that the females were hornless and that the males, of most of the species at least, possessed a small frontal horn, or possibly a pair of such horns. The Systematic Position of Adinotherium. In the paper already so often cited (’07, 57-59) Ameghino reports the extraordinarily interesting and unexpected fact that the genus Trigodon (or Eutvigodon, as it was formerly called) from the Monte Hermoso beds, possesses a very large and conspicuous, median, frontal boss, which must have served in life to support an unpaired dermal horn, like that of cer- tain rhinoceroses, especially, though on a far smaller scale, like that of Elasmotherium. “Cette protuberance s’eleve graduellement du bout anterieur du plan sagittal vers l’avant, mais dans la partie anterieure elle descend brusquement, de sorte que la bosse entire semble penchde en SCOTT: TOXODONTA OF THE SANTA CRUZ BEDS. 221 avant. La surface de cette protuberance est tres rugueuse, la partie se trouvant dans cette condition parfaitement delimitee de la partie lisse, de sorte qu’il ne peut rester aucun doute que cette bosse servait a supporter une forte corne qui etait inclinee en avant” (t. c., p. 57). Another remarkable peculiarity of Trigodon is that in the mandible the number of Fig. 41. Trigodon gaudryi. Skull, left side, about ^ nat. size. From a photograph of a specimen in the Ameghino collection. Monte Hermoso beds. incisors is uneven, 5 instead of 6, the unpaired one standing in the median line of the symphysis. So many individuals have been found, all of which show the same curious arrangement, that the chance of its being an abnormality is very small ; both Lydekker (’93, 20) and Ame- ghino (’07, 55) agree in regarding it as normal and constant in the genus. Unusual as the condition is, the fact of its constant occurrence may be accepted without difficulty, though few, I think, will be inclined to follow Ameghino in believing that the unpaired tooth has been formed by the coalescence of the two median incisors. 222 PATAGONIAN EXPEDITIONS : PALEONTOLOGY. Certain mandibles of Nesodon imbvicatus appear to show a similar arrangement of the incisors as an individual peculiarity, but the imperfec- tion of the specimens precludes any positive statement regarding them. Ameghino regards Adinotherium as the ancestor of the Haplodonther- iidae, the family to which he assigns Trigodon. “ Aussi bien les Haplodon- theriidae que les Toxodontidae, ils prennent leur origine dans les Nesodon- tidae. Les Haplodontheriidae descendent du genre Adinotherium , tandis que les Toxodontidae doivent descendre d’une esp£ce du genre Nesodon ” (’07, 91). Not at all improbably, this may prove to be the correct inter- pretation of the relationships of Adinotherium , but, on the other hand, the discovery of a species of Nesodon , N cornutus , which appears to have the same small dermal horn as Adinotherium , leads to the conjecture that Nesodon may perhaps have been the ancestor of Trigodon. This conjec- ture is somewhat strengthened by the fact that in the latter genus the shape of the muzzle and symphysial region of the mandible is more as in Nesodon , Adinotherium agreeing better in this regard with Toxodon. In short, it is possible that Ameghino’s suggestion should be reversed, but only the identification of the successive steps of change which led up to Toxodon , on the one hand, and Trigodon , on the other, will enable us to decide this question. Adinotherium ovinum (Owen). (Plates XII, Fig. 1 ; XVII, Figs. 4, 5, 7-10; XVIII, Figs. 2, 3, 5; XX, Figs. 1-4; XXI, Figs. 3-10; XXVI, Figs. 1-3, 7-12; XXVII.) Nesodon ovinus Owen ; Phil. Trans., 1853, p. 291. ? Adinotherium inagister Amegh.; Enum. sistem., etc., 1887, p. 17. Adinotherium proximum Amegh.; Ibid. Adinotherium ferum Amegh.; Ibid., p. 18. ? Acrotherium mutabile Mercerat ; Rev. del Mus. de La Plata, T. I, 1891, P- 393- ? Nesodon typicus Mercerat; Ibid., p. 402. ? Adinotherium haplodontoides Amegh.; Rev. Argent, de Hist. Nat., T. I, 1891, p. 129. Adinotherium ovinum (Owen) Amegh.; Ib’d., p. 376. Adinotherium Kobyi Mercerat; Rev. del Mus. de La Plata, T. I, 1891, p. 410. scott: toxodonta OF THE SANTA CRUZ beds. 223 As here regarded, this species is of medium size, horned in the male (A. ferum Amegh.) and hornless in the female and in the former the cranium is lower and wider and the zygomatic arches are much more ex- panded. A. magistev , as figured by Ameghino (’07, 70, fig. 9) differs so much from his specimen of A. ovinum (ibid., p. 73, fig. 10) in the form of the cranium and especially of the nasals, that it should perhaps be admitted as distinct, but, as already pointed out, dependence upon such characters would increase the number of species beyond all reasonable probability. As the description of the genus is drawn principally from individuals re- ferred to A. ovinum , it remains merely to give the dimensions. Upper dentition, length I1, length (i. e. ant. -post, diam.) “ width (i. e. transv. diam.) I-, length .... 14 width .... I-2-, length .... 44 width .... Upper canine, length 4 4 4 4 width . Upper cheek-teeth series, length “ premolar series, length Pb length .... “ width .... Pb length .... “ width .... Pb length .... “ width .... P-, length .... “ width .... Upper molar series, length M1, length .... “ width .... Mb length .... “ width .... Mb length .... “ width .... Lower dentition, length IT, length Measurements. No. 15,131. •143 •0055 .008 .013 • -0135 .005 .004 .006 .004 .096 .041 .0105 .0105 .011 .0125 .012 .014 •057 .015 .019 .020 .0205 .028 .019 •135 .006 No. 15,118. .147 .007 .018 .014 .013 .0045 •0035 .094 •0435 .008 .006 .0105 .0115 .012 .012 .013 •oi55 •0555 .017 .021 .022 •0235 .026 .021 No. 15,136. .143 .006 No. 15,382. .0125 .014 .096 .0405 .006 .006 .011 .014 .013 .016 .059 •0135 .0215 .020 .024 .030 •0235 No. 15,003. .051 .016 .017 .019 .0195 .0225 .017 .124 .004 224 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. No. 15,131. No. 15,136. No. 15,003. Iy, width ..... .008 .010 .008 h > length .008 •0055 .005 “ width ..... .0105 .013 .010 I¥, length .014 .014 .012 “ width .010 .010 .007 Lower canine, length .010 .009 .0085 “ “ width . . .0045 .004 .0045 “ cheek-teeth series, length . . .105 .102 .091 “ premolar series, length • .043 .044 .038 PT, length .008 .007 .0075 “ width .005 .0045 .0043 P^, length .009 .010 .010 “ width .006 .006 •0055 Py, length .Oil .012 .010 “ width .007 .0075 .007 Py, length .0125 •0135 .0115 “ width .008 .007 .0073 Lower molar series, length .062 •059 •053 My, length .014 .016 .0145 “ width .008 .008 .007 Mj, length ..... .0185 .019 .017 My, length ..... . .030 .025 •0235 “ width ..... .008 .008 .007 Aside from the effects of age and wear, these measurements, it will be observed, agree quite closely. The effects of abrasion are clearly shown, for example, in the very small size of I1 in No, 15,131, which is quite an old female and in which the median upper incisors are worn to mere stumps. In the old animals also the third molar is relatively larger, as appears from a comparison of the measurements of Nos. 15,131 and 15,003, the latter a young and quite small female. The other indi- viduals are males and No. 15,136 is especially large and robust ; the skull has the frontal protuberance, but the upper cheek-teeth are unfortunately all missing. Several young individuals enable me to give the dimensions of the milk-teeth, but the skulls in all cases but one have lost the lower jaw. Measurements. No. 15,159. No. 15,945. No. 16,011. No. 15,114. Upper milk-dentition, length .085 .082 •0835 Di1, length (i. e. ant. -post, diameter) .004 .004 .0043 “ width (i. e. transv. diameter) .Oil .Oil .013 SCOTT: TOXODONTA OF THE SANTA CRUZ BEDS. 225 No. I5J59 ■ No. 15,945. No. 16,011. No. 1 5, 1 1 4. Di-, length ...... • -0055 •0055 .006 “width .008 .008 .010 Di5-, length ...... . .005 .005 .0045 “ width ...... .004 .003 .003 Upper milk-canine, length. .0065 .006 .0066 “ “ “ width .005 •0035 .0036 Upper milk-premolar series, length • -0455 .051 •045 Dp-, length ...... .008 .009 .009 “ width ...... .005 .005 •0055 Dp-, length .012 .Oil .0103 “ width ...... . .0095 .009 .009 .0095 Dp-, length ...... • -oi35 .017 .0155 .014 “ width ...... .012 .012 .Oil .012 Dp-, length ...... .017 .019 .018 .017 “ width ...... . .015 •oi35 •0135 .014 M-, length ...... .021 .0215 .0205 “ width . .0145 .012 .0125 .013 Of these four specimens, No. 16,011 is the youngest and in it m-, though almost fully erupted, shows no signs of wear ; next in point of age comes No. 15,945, in which the foremost part of m1 is somewhat abraded, while in' No. 15,1 14 this tooth was in full use and is considerably worn. No. 15,159 is the oldest of the series. The dimensions of the lower milk teeth are all taken from a single individual, No. 15,159, in which mT is in quite an advanced stage of wear. Lower milk-dentition, length . . .073 DpT, length 006 Dif, length. .... . .004 “ width. 004 “ width ..... . .007 Dp^, length 0105 Dij, length. .... • -0035 “ width. 006 “ width ..... . .0095 Dps, length 013 Dig, length . .004 “ width. 0065 “ width ..... . .008 Dpj, length 015 Lower milk-canine, length . . .0075 “ width. 007 “ “ “ width . . .003 M-j, length 017 Lower milk-premolar series, length . .0445 “ width . 0075 From the considerable number of skulls which are probably referable to this species it is necessary to make a selection for measuring and the choice is made so as to display the different proportions which are believed to characterize the two sexes and also to bring out the range of individual variation in size. Nos. 15,003 and 15,131 are presumably females, while the others are supposedly males with frontal horn. 226 PATAGONIAN EXPEDITIONS 1 PALAEONTOLOGY. Measurements. No. 15,003. No. 15,131 • No. 15,118. No. 15,382. No. 15,983- Skull, length in med. basal line . .236 •245 .238 “ fr. occ. condyles •253 .256 .258 “ fr. occ. crest to end of nasals ...... .224 Cranium, length fr. cond. to orbit .148 •145 .146 .151 Face, length fr. orbit to prmx. . •115 .116 .117 .117 Occiput, height .101 .085 width at base .104 ?.I2I •133 “ over condyles .051 •054 •059 .061 Sagittal crest, length .... .087 .082 Zygomatic arch, length fr. glen. cav. . .088 .100 •093 .096 .08l “ width (dorso-vent.) . •039 .048 .048 Cranium, width at constriction . .030 •034 .042 •039 zygomatic width . •139 •154 .l6l .l80 depth at tympanic bulla .107 .110 .102 Face, width over lachrymals .070 .080 .084 prmx. . .049 .047 “ depth at m- .... .099 .098 .083 •095 “ “ “ p* .080 .079 .06l .079 Nasals, length ..... .101 .092 .105 .101 Palate, length in median line •153 .152 •143 •145 “ width at p1 .018 .023 .022 .024 “ “ 14 m^ . •054 .040 .052 .048 Mandible, length (excl. teeth) .202 .212 “ of symphysis . •054 width at i^. .031 .042 depth at p^ .041 •039 U ll ll __ ms- •045 .050 height of coronoid •125 “ “ “ condyle .124 width, angle to m3 .084 .084 In some respects these measurements are misleading, owing to the dis- tortion of the skulls through pressure. Nos. 15,003 and 15,982 are almost free from distortion, though in the latter the great zygomatic breadth may have been somewhat exaggerated. No. 15,118 is slightly asymmetrical, but in Nos. 15,181 and 15,382 material changes in the proportions of the skull have been caused by pressure ; the former has suffered from a lateral compression, which has notably diminished the transverse dimensions, while in the latter the compression was vertical, increasing the transverse and decreasing the vertical dimensions, except in the case of the palate, which appears to be normal. scott: toxodonta OF THE SANTA CRUZ beds. 227 The measurements of the skeleton are nearly all taken from a single individual, No. 15,131, missing parts supplied from other specimens. Measurements. Atlas, length .... •037 Penult, thoracic, length of centrum . .029 “ height (dors. -vent.) . •043 width of ant. face . .027 Axis, length inch odontoid . .044 height, inch spine . .046 “ “ excl. .030 Lumbar 1, length of centrum . .028 “ width of anterior face .056 “ width of ant. face . .022 “ of posterior face .022 “ height, inch spine . .050 “ height, inch spine .049 Lumbar 2, length of centrum . .031 Cervical 3, length of centrum .018 “ width of ant. face . .023 “ width of ant. face .019 “ height, inch spine • .050 “ height, excl. spine •034 Lumbar 3, length of centrum . .032 “ “ “ inch “ .040 “ width of ant. face • -025 Cervical 4, length of centrum .018 Sacrum, length in med. line . .184* “ width of anterior face .022 Sacral 1, width of ant. face. • -035* “ height, inch spine .041 “ “ “ over pleurap. • .103* Cervical 5, length of centrum .019 Sacral 6, width of post, face . .030* “ width of anterior face .022 “ “ depth (dors. -vent). .015* “ height, excl. spine •033 Scapula, length . .I99f, .190 “ “ “ inch “ ■043 greatest width .io8f Cervical 6, length of centrum .018 width of neck .046 f, .040 “ width of anterior face .022 “ glen. cav. (trans.). .025f, .024 “ height, inch spine •043 Humerus, length from head. .165*, .163 Cervical 7, length of centrum .020 “ “ “ ext. tuber. .180*, .175 “ width of anterior face .022 proximal width . .050* “ height, excl. spine .031 “ “ thickness .059*, .061 Thoracic 1, length of centrum .019 distal width over epi- “ width of anterior face .023 condyles .051*, .051 “ “ over trans. proc. .067 Humerus, distal width of trochlea .039*, .040 Thoracic 4, length of centrum . .025 Ulna, length .... . .198 “ width of anterior face .020 “ of olecranon . • .065 “ “ over trans. proc. .044 “ width at sigmoid notch . .024 “ height, inch spine .103 “ distal width . .010 Thoracic 7, length of centrum . .028 “ thickness. • -0135 “ width of anterior face .025 Radius, length . . . . . .146 “ “ over trans. proc. P.042 proximal width . .024 “ height, inch spine .088 thickness. . .016 Thoracic 9, length of centrum . .025 “ distal width • -033 “ width of anterior face .021 “ thickness . .019 “ height, inch spine .068 * No. 15,966. f No. 15,004. 228 PATAGONIAN EXPEDITIONS : PAL/EONTOLOGY. The dimensions of the manus are taken from two individuals, one of which, No. 15,158, is a very young animal with metacarpal epiphyses still separate. No. 15,131. No. 15,158. Carpus, length in median line . width of proximal row . Pisiform, length. greatest dors.-vent. depth Metacarpal II, length “ proximal width . “ distal width of shaft “ width of trochlea Metacarpal III, length “ proximal width. “ distal width of shaft “ width of trochlea Metacarpal IV, length “ proximal width “ distal width of shaft “ width of trochlea Phalanx 1, digit II, length. “ “ “ proximal width Phalanx 2, digit II, length. “ “ “ proximal width Ungual, digit II, length “ “ proximal width Phalanx 1, digit III, length “ “ “ proximal width Phalanx 2, digit III, length “ “ “ proximal width Ungual, digit III, length . “ “ proximal width Phalanx I, digit IV, length “ “ “ “ proximal width Phalanx 2, digit IV, length “ “ “ “ proximal width Ungual, digit IV, proximal width .024 .041 .023 .013 ?-oi55 ?.on .060 .016 .019 •oi35 .050 .014 .015 .0105 .013 .012 .009 .0105 .010 .010 .013 .015 .0105 .013 .Oil .0125 .014 .012 .OIO .0105 .OO95 .025 .038 .020 .013 .052 ■013 .015 .Oil .055 .017 .017 .013 .050 .015 .014 .0115 .Oil .0125 The pelvis is not sufficiently well preserved in any of the specimens to render satisfactory measurements practicable, and in the most complete skeleton (15,131) the bones of the hind leg and foot are all more or less injured and distorted, so that measurements of other individuals are added for comparison. SCOTT : TOXODONTA OF THE SANTA CRUZ BEDS. 229 No. 15,131 • No. 15,127. Femur, length from head . .187 .185 “ “ “ grt. trochanter . . .187 .187 “ least width of shaft below 3rd troch. . • .026 “ thickness at same point. . . .016 “ proximal width . . .058 .061 thickness .027 “ distal width ...... . .044 .047 “ thickness .048 “ width of trochlea .026 Patella, length width thickness No. 15,480. Tibia, length inch spine . *.165 ?.l86 “ proximal width . .042 .051 thickness .049 “ least width of shaft .013 “ thickness of shaft . .018 .017 “ distal width . .023 .025 “ thickness . . .0225 .0265 Fibula, length .140 .156 proximal width . . . . . . .015 .018 thickness. . . . . .027 “ least width of shaft . . . . .009 “ thickness of shaft . .011 “ distal width ...... .017 “ thickness .023 * Considerably reduced by distortion. Unfortunately, the pes of No. 15, 1 31 is quite incomplete, lacking the calcaneum and astragalus, which are supplied from No. 15,978, and most of the fourth digit. In the American Museum collection, however, is a nearly perfect hind-foot, No. 9275, with only the astragalus missing. All these individuals agree well in size. No. 15,131. No. 9275. Tarsus, length in median line . . . . . .038 Astragalus, length . .030 width of trochlea .... .014 “ “ “ head . .013 Calcaneum, length • .053 •053 proximal width . . . . .018 distal width .017 .016 width over sustentaculum • .025 .024 230 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. No. 15,131 • No. 9275. Metatarsal II, length. .040 .041 “ proximal width . .0085 .008 “ “ “ thickness . • .015 •015 “ distal width of shaft . .010 .011 “ width of trochlea .009 .010 Digit II, length . .068 .071 Metatarsal III, length . .044 .044 “ proximal width . • .0135 .015 “ distal width of shaft . . .015 .016 “ width of trochlea • -013 .014 Digit III, length .... •095 Metatarsal IV, length .041 “ proximal width . . . .014 .014 “ “ “ thickness . .016 “ distal width of shaft . .015 “ width of trochlea .0125 Digit IV, length .079 Phalanx I, digit II, length. . .0105 .Oil “ “ “ “ proximal width . . .0095 .0105 Phalanx 2, “ “ length. .008 .009 “ “ “ “ proximal width . .009 .0095 Ungual, digit II, length . .013 .012 Ungual, digit II, proximal width .008 .008 Phalanx I, digit III, length • -0155 •0175 “ “ “ “ proximal width . .014 .015 Phalanx 2, “ “ length . . .012 .013 “ “ “ proximal width . . .0115 .012 Ungual, digit III, length . .021 “ “ proximal width .014 Phalanx I, digit IV, length .015 “ “ “ proximal width . .013 Phalanx 2, “ “ length •Oil “ proximal width . .012 Ungual, digit IV, proximal width .0105 Localities. — Nearly all the representatives of this species in the collec- tions of Princeton University and the American Museum, collected by Messrs. Hatcher, Peterson and Brown, were obtained on the Atlantic coast near Coy Inlet and from five to twenty miles south of that point; a few only were found at Killik Aike. SCOTT: TOXODONTA OF THE SANTA CRUZ BEDS. 23I Adinotherium nitidum Ameghino. Adinotherium nitidum Amegh.; Enum. sistem., etc., 1887, p. 18. ? Adinotherium corriguenense Amegh.; Anales del Mus. Nac. de Buenos Aires, T. XVI, 1907, p. 77. The only definite character which distinguishes this species is its small size. In the original description (loc. cit.) the length of the last five lower teeth (pit-m-j) is given as 52 mm. and the depth of the jaw below mi as 37 mm., dimensions which in a small individual of A. ovinum are 74 and 46 mm. respectively. The skull-characters which Ameghino cites in his last paper (’07, 74, 76, fig. 12) are not trustworthy, because the speci- men figured is a very young animal and agrees closely with young skulls of A. ovinum. None of these small skulls has yet been found having any indications of horns and it may well be that this was a hornless species in both sexes. Localities. — The type of A. nitidum was found in the cliffs of the Rio Santa Cruz and that of A. corriguenense at Corriguen Aike on the Atlantic coast. Adinotherium robustum Ameghino. Adinotherium robustum Amegh.; Rev. Argent, de Hist. Nat., T. I, 1891, p- 376. Among the material now accessible to me, this probably valid species is represented by a photograph of the type in the Ameghino collection, a skull (No. 9532) and an incomplete fore-foot (No. 9184), both belonging to the American Museum. The skull, which is in a good state of preser- vation, though much cracked and somewhat distorted by vertical down- crushing, agrees well with the type in size and proportions (cf. Amegh., ’07, 83, fig. 16), but has quite a different fronto-nasal suture, which is more regularly curved than in the type and the nasals narrow less toward the posterior end. The skull in question is without frontal boss and is therefore probably a female. A. robustum is distinguished by its larger size and heavier and more robust proportions than in any of the other known species of the genus ; the cranium is unusually broad and capacious and the zygomatic arches are greatly expanded, even in the female skull. The auditory chamber in the post-tympanic portion of the squamosal is inflated so as to form a pro- tuberance on the occipital surface, which is more conspicuous than in any 232 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. of the other species, indeed, than in any other known member of the sub- order. The sagittal crest is very high and descends quite steeply to the forehead. The dimensions in the following table are taken from No. 9532, A. M. N. H., an adult, still rather young, as is shown by the relatively small size of m-. Measurements. Upper dentition, length . .160 M-, length .030 I-, length (i. e. ant. -post, diam.) . . .007 “ width .022 “ width (i. e. transv. diam.) . .018 Mj, length .021 I-, length ..... . .0125 M-j, length .031 “ width ..... . .014 “ width ...... .009 I-, length . .005 Skull, length in med. basal line . .246 ‘ ‘ width . .003 “ fr. occ. cond. to prmx. . .270 Upper canine, length . . .007 Cranium, length fr. cond. to orb. . •157 “ “ width . . .005 width at constriction . •059 Upper cheek-teeth series, length . .107 Occiput, width at base •145 Upper premolar series, length . .043 “ over cond. . .065 P-, length . .008 Zygomatic arch, length fr. glen. cav. . .097 “ width ..... . .006 “ width (dors.-vent.) . .052 P-, length ..... . .Oil Skull, zygomatic width .190 “ width ..... . .011 Face, length fr. orb. to prmx. .132 P5-, length ..... . .0115 “ width over lachrymals •095 “ width . .0125 “ “ “ prmx. •059 P-, length . .013 “ depth at m- .... .094 “ width ..... . .016 4 4 4 4 P1, • • • • .069 Upper molar series, length . . .064 Palate, length in median line •155 M-, length .... . .018 “ width at p- . .026 “ width . .0225 “ < < < < m 3. .066 M-, length .... . .023 Mandible, depth below m-g-. .056 “ width ..... . .024 The fore-limb and foot differ from those of A. ovinum even more than the comparative measurements would indicate. The distal end of the radius is much more massive and has a far larger and more rugose ex- pansion toward the ulnar side, and on the inner side of the dorsal face is a conspicuous tendinal sulcus not present in A. ovinum. The carpals are broader and heavier than in the latter and have more roughened dorsal faces. The metacarpals are longer and relatively much heavier and the tubercles for muscular attachment are decidedly more prominent. The dimensions are taken from No. 9184, A. M. N. H. SCOTT: TOXODONTA OF THE SANTA CRUZ BEDS. 233 Measurements. Radius, distal width .... .041 Metacarpal III, length .065 “ thickness .028 “ proximal width . .019 Carpus, length in median line .030 “ distal width of shaft . .0225 width of proximal row . •043 “ width of trochlea .016 “ “ distal row •0445 Metacarpal IV, proximal width . .016 Metacarpal II, length. .060 Phalanx I, digit II, length . .013 “ proximal width . .017 “ prox. width. .0145 “ distal width of shaft . .0185 Phalanx 2, “ “ length . .Oil “ width of trochlea .0115 “ “ “ “ prox. width. .012 Localities. — The type-specimen in the Ameghino collection is from Kar Aike. Nos. 9184 and 9532 were collected by Mr. Barnum Brown, the former 15 miles south of Monte Leon and the latter 5 miles north of Coy Inlet. Adinotherium karaikense (Ameghino). Acrotherium karaikense Amegh.; Rev. Argent, de Hist. Nat., T. I, 1891, p. 131. Nesodon imbricatus Lydekker; Anales del Mus. de La Plata, T. II, 1893, p. 26. This species, which is known only from the type-specimen, appears to be merely an Adinotherium with a supernumerary premolar, just as the so-called Acrotherium rusticum would seem to be an example of Nesodon imbricatus with an additional premolar, such as Bateson has noted in many recent mammals. That A. karaikense is specifically distinct is made extremely probable by the following facts: ( 1 ) In size, this species exceeds any other known example of the genus, the length of the skull, which in A. robustum is 27 cm., being 29 cm. (2) In proportion to its length, the skull is narrow, the zygomatic arches curving outwardly but little, which is in strong contrast to the broad head and greatly expanded arches of A. robustum. (3) The facial region narrows anteriorly much more gradually than in the latter. My photograph shows a pair of rough frontal ridges, which may or may not represent the horn-bosses ; it is difficult to determine this from the photograph. .Locality. — Kar Aike. 234 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. Adinotherium splendidum (Ameghino). Adinotherium splendidum Amegh.; Enum. sistem., etc., 1887, p. 17. Adinotherium pule hrumMtrctYdii ; Rev. del Mus. de La Plata, T. I, 1891, p. 407 (. fide Ameghino). Adinotherium silvaticum Merc., in part; Ibid., p. 408 [fide Ameghino). Adinotherium antiquum Merc.; Ibid., p. 410 [fide Ameghino). Nesodon ovinus Lydekker, in part ; An. del Mus. de La Plata, T. Ill, 1893. p- 25. Noaditherium splendidum Amegh.; An. del Mus. Nat. de Buenos Aires, T. XVI, 1907, p. 84. No example of this well-defined species is contained in either the Prince- ton or the New York collection. Ameghino’s account of it is, briefly, as follows : The cranium is notably higher dorso-ventrally than in the pre- ceding species. The forehead is short and triangular and the temporal ridges are very thick and prominent and describe a sigmoid course in con- verging into the sagittal crest ; the postorbital processes have an unusually anterior position and are long and extremely broad and massive, projecting almost directly outward, they conceal the orbits when the skull is viewed from above. These processes are very rugose and Ameghino believes that they supported small horns. “Tout parait indiquer que ces apophyses portaient une paire de petites cornes super-orbitaires ” (’07, 88-9). This is a most improbable suggestion, and such rugose postorbital processes occur in old males of the other species, not to mention the various rugosi- ties of the rhinoceros skull which have no relation to horns. The frontal horn-bosses of A. splendidum are better defined and form a more distinct pair than in any of the preceding species. Though specifically well distinguished, there is nothing in the known structure of A. splendidum which would justify its generic separation from Adinotherium. Localities. — The type-specimen was obtained from the cliffs of the Rio Santa Cruz (Amegh., ’89^, 453) and, in general, it is stated that the spe- cies occurs in the Notohippus Beds and the base of the Santa Cruz. PHOBEREOTHERIUM Ameghino. Phobereotherium Amegh.; Enum. sistem., etc., 1887, p. 18. Adinotherium Mercerat, in part; Rev. del Mus. de La Plata, T. I, 1891, p. 408. SCOTT: TOXODONTA OF THE SANTA CRUZ BEDS. 235 Nesodon Lydekker, in part; An. del Mus. de La Plata, T. II, 1893, p. 25. This is a small animal which differs from Adinotherium by the absence of the median upper incisors, giving the dental formula : If, Cf, P i, Ml. Phobereotherium silvaticum Ameghino. PJiobereotherium silvaticum Amegh.; Enum. sistem., etc., 1887, p. 18. Adinotherium silvaticum Mercerat, in part ; Rev. del Mus. de La Plata, T. I, 1891, p. 408. The only specific description yet given is that the length of “the five upper molars” (presumably p—m-) is 105 mm. (’87 b, 18). Locality. — Cliffs of the Rio Santa Cruz. STENOTEPHANOS Ameghino. Stenotephanos Amegh.; Bol. de la Acad. Nac. de Cienc. de Cordoba, T. IX, p. 107, Stenostephanus Lydekker; An. del Mus. de La Plata, T. II, 1893, p. 24. No example of this interesting, but imperfectly known, genus is con- tained in the collections at hand and I shall therefore quote Lydekker’s account of it, which is the most complete that has yet appeared. “The upper molars are characterized by their narrowness, and by the strong curvature of the outer wall of the third ; that tooth having the middle column very largely developed and consequently two well marked inner clefts. . . . The first and second molars . . . seem to have had but one cleft; and as this did not extend to the base of the crown, when much worn these teeth show a central island of enamel. In the lower jaw, the molars are almost if not quite indistinguishable from those of Xotodon; but the premolars have complex internal and external folds, and are thus quite different from the simple premolars of the latter” (’93, 24). Stenotephanos speciosus Ameghino. Stenotephanos speciosus Amegh.; Enum. sistem., etc., 1887, p. 14. Stenostephanus speciosus Lydekker; An. del Mus. de La Plata, T. II, 1893, p. 24. Lydekker’s account of the species is as follows: “The smaller size of the third upper molar serves to distinguish the specimen from the corre- sponding tooth of S. plicidens from the Parana.” The upper molars 236 PATAGONIAN EXPEDITIONS: PALAEONTOLOGY. “present the characters already noticed, while the lower molars are almost indistinguishable from those of Xotodon ; the premolars being of Fig. 42. Stenotephanos speciosus. Fragment of right upper maxillary with the molars in place. About natural size. Type specimen, La Plata Museum. a very complex type. The length of the space occupied by the five lower teeth [p3-m-3] is three inches ” (’93, 24-5). Locality. — Cliffs of the Rio Santa Cruz. GENERA INCERTaE SEDIS. Ameghino has named a number of genera, concerning which very little is known and which will therefore be merely listed, as none of them is represented in the collections at my disposal and I have no information to add to the little already published. I am not in a position to express an opinion as to the validity of these species. RHADINOTHERIUM Ameghino. Rhadinotherium Amegh.; Enum. sistem., etc., 1887, P- 1%- Nesodon Mercerat, in part; Rev. del Mus. de La Plata, T. I, 1891, p. 394. Rhadinotherium limitatum Ameghino. Rhadinotherium limitatum Amegh.; Enum. sistem., etc., 1887, p. 18. Nesodon limitatum Mercerat; Rev. del Mus. de La Plata, T. I, 1891, p. 403. PALzEOLITHOPS Ameghino. Lithops Amegh.; Enum. sistem., etc., 1887, p. 15. (Preoccupied.) Palceolithops Amegh.; Rev. Argent, de Hist. Nat., T. I, 1891, p. 240. scott: toxodonta OF THE SANTA CRUZ beds. 237 Pal/EOLIthops pr^evius Ameghino. Litliops prcevius Amegh.; Enum. sistem., etc., 1887, p. 15. Palceolithops prcevius Amegh.; Rev. Argent, de Hist. Nat., T. I, 1891, p. 241. ACROTHERIUM Ameghino. (See pp. 181, 233.) Acrotherium stygium Ameghino. Acrotherium stygium Amegh.; Rev. Argent, de Hist. Nat., T. I, 1891, p. 133. XOTOPRODON Ameghino. Xotoprodon Amegh.; Rev. Argent, de Hist. Nat., T. I, 1891, p. 241. Xotoprodon solidus Ameghino. Xotoprodon solidus Amegh.; Rev. Argent, de Hist. Nat., T. I, 1891, p. 241. Xotoprodon maximus Ameghino. Xotoprodon maximus Amegh.; Rev. Argent, de Hist. Nat., T. I, 1891, P- 375- NO TOHIPPIDAi. NOTOHIPPUS Ameghino. Notohippus Amegh.; Rev. Argent, de Hist. Nat., T. I, 1891, p. 63 ( nomen nudum), Ibid., p. 135. Nesodon Burmeister, in part; Anal, del Mus. Nac. de Buenos Aires, T. Ill, fasc. XVIII, 1892. Notohippus toxodontoides Ameghino. Notohippus toxodontoides Amegh.; Rev. Argent, de Hist. Nat., T. I, 1891, p. 63 ( nomen nudum), Ibid., p. 135. Nesodon ovinus Burmeister, in part ; Anal, del Mus. Nac. de Buenos Aires, T. Ill, fasc. XVIII, 1892. This very interesting, but imperfectly known, species is remarkable for having the lower molars surrounded by an uninterrupted wall of enamel and having the crowns covered by a thick layer of cement (“por un grueso deposito de cemento”). Originally, Ameghino referred the genus to the 238 PATAGONIAN EXPEDITIONS: PALAEONTOLOGY. Litopterna, making it the type of a family, at first called Protequidae (’9P, 135), subsequently Notohippidae (’94*, 27). Later, he assigned the family to his new order Hippoidea and regards it as closely related to the true horses (’04, 33). With this view I am not able to agree, being of Roth’s opinion that the Notohippidae are of toxodont affinities (’03, 33). The molar pattern is typically toxodont and, as Roth points out, the structure of the auditory region, at least in the pre-Santa Cruzian genera, is also that of the toxodonts. Whether the family is referable to the suborder Toxodonta, or not, is not yet entirely clear, though I am inclined to believe that it should be so referred, as has been done by Schlosser (T 1, 519). PART III. ENTELONYCHIA. The members of this suborder are relatively rare in the Santa Cruz beds and the Princeton and New York collections contain but few and unsatis- factory specimens. For this reason, the following account is largely drawn from the collections of Dr. Ameghino and the La Plata Museum. I am under great obligations to Dr. Ameghino and to Sr. Moreno for the very liberal manner in which the material in their charge was placed at my disposal. It is much to be regretted that I was not permitted to see the skeletal material in the Paris Museum on the occasion of my last visit, August, 19 1 1, the absence of M. Boule rendering the Patagonian collec- tions of M. Tournouer inaccessible to me. In Santa Cruz times the Entelonychia were evidently verging toward extinction, for the group is not known to be represented in any later formation and their rarity is in striking contrast to the abundance and variety of them in the more ancient Deseado stage ( Pyrotherium Beds). From this latter formation, Ameghino has described no less than twelve genera, which are distributed in three families, while but one family and, at most, two genera are known from the Santa Cruz. Individual abun- dance corresponds to this variety of form and members of the Entelo- nychia are among the commoner fossils of the Pyrotherium fauna, which thus records the culmination of the group, followed by swift decline and extinction. The interval of time between the Deseado and the Santa Cruz Beds, which is occupied by the marine Patagonian formation and its presumable equivalents, the Colpodon , Astrapothericulus and Noto hip- pus faunas of Ameghino, cannot have been very great, as measured by the degree of structural change in the mammalian groups common to both, but the difference in regard to the Entelonychia in the two formations is a very marked one. Obviously, the Entelonychia of the Santa Cruz are the reduced and vanishing remnants of a previously dominant group. Any description of them which leaves their predecessors out of account must necessarily be partial and even misleading. It will therefore be advisable to depart from the method followed in the other sections of 239 240 PATAGONIAN EXPEDITIONS I PAL/EONTOLOGY. these reports on Santa Cruz mammals, so far as to consider more fully the pre-Santa Cruzian representatives of the suborder, with particular reference to those of the Pyrotherium Beds, for which Gaudry’s geo- graphical term, the Deseado stage , will be employed as more convenient (Gaudry, ’06, 107). Throughout the suborder the teeth are brachyodont and rooted, never prismatic, and most of the genera have them in unreduced number, though in the early and prematurely specialized family of the Notostylopidae there is considerable reduction in the anterior teeth, the second and third lower incisors and the canine being usually absent and the first upper premolar is frequently wanting ; Diorotherium of the Santa Cruz lacks p1. Typically, there is a gradual transition in form from the incisors to the molars, as is implied in the name of the genus first described, but there are some ex- ceptions to this rule. In Notostylops the first incisor is enlarged and, though not growing from a permanent pulp, is yet scalpriform in appear- ance. In Leontinia , of the Deseado stage (the reference of which to the Entelonychia I strongly question) we find the toxodont feature of prom- inent tusks formed by the enlargement of the second upper and third lower incisor. Should this genus, when its structure is fully known, prove to be a rightful member of the Entelonychia, it would show that this curious transference of function had occurred separately and inde- pendently in three groups of South American ungulates, the Toxodonta, the Entelonychia and in the Proterotheriidae of the Litopterna. In the Isotemnidae, though the canines are enlarged in some instances, the transition in pattern from one tooth to the next succeeding one is very gradual throughout the dental series, while in the Homalodonto- theriidae there is considerable variation in regard to the development of the canines. In the most ancient known genera of this family, as Thomas- huxleya, for example, the general appearance of the dentition recalls that of the Eocene Perissodactyla of the northern hemisphere, the canines forming stout and conspicuous, though not very long, tusks and the lower canine passing in front of the upper one and received into a diastema between the latter and the external incisor. Much the same description applies to Asmodeus of the Deseado. Subsequently, these tusks diminish in relative size and in the Santa Cruz forms the lower canine is hardly larger or more prominent than an incisor ; but the upper one is still fairly large in what are presumed to be the male animals, much smaller in the supposed females. SCOTT: ENTELONYCHIA OF THE SANTA CRUZ BEDS. 241 The premolars are always less complex than the molars, though the transition from one to the other is gradual, except in some of the earlier genera, in which the molars are extremely simple. Probably even in those cases, however, perfectly unworn teeth would show a difference between the two classes. Ameghino states (’02, 34-5) that in Prochalico- therium , of the Colpodon Beds, the upper premolars have isolated and conical inner cusps, which are not connected with the external wall till a very advanced age. I have not seen these specimens. The molar-pattern is essentially that of the Toxodonta. In the upper molars the external wall is completely formed, even in the most ancient genera, and is nearly smooth, the lobes being very obscurely, or not at all, indicated ; there are two oblique transverse crests, of which the posterior one is much the shorter, and a varying number of spurs or “combing plates” from the outer wall, which are much less prominently developed than in the Toxodonta. When well worn, these teeth have as a conspicuous feature an elongate and narrow enamel-lake, the remnant of the main valley, which has a very oblique direction, more antero-posterior than transverse. The lower molars are made up of two crescents, the hinder one of which is much more elongate antero-posteriorly and has in its internal valley a spur or pillar, such as is found in all of the groups of indigenous South American ungulates. That the grinding teeth of the Entelonychia have a certaiu resemblance to those of the rhinoceroses is not to be denied, but perfectly fresh and unworn examples show this likeness to be but superficial, the approximation to the Toxodonta being much closer and more fundamental. As Roth has pointed out (’03), all of the Toxodontia, or Notoungulata, agree in the unusual structure of the auditory region, but each of the three suborders has its particular modification of the general plan. In the Entelonychia the inflated post-tympanic region of the squamosal makes up less of the occipital surface than in the Toxodonta; the mastoid portion of the periotic is excluded from the surface of the skull, by the junction of the post-tympanic and postglenoid processes, and the external auditory meatus is tubular in the Notostylopidae, and a mere hole in the Homalo- dontotheriidae. Only in these two families is the skull known, if, as I think should be done, the Leontiniidae are excluded from this suborder. In the Notostylopidae the skull is depressed and flattened and has a very short, pointed muzzle, which, with the enlarged and scalpriform incisors, gives it quite a rodent-like appearance. Little is known of the skull in 242 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. the earlier and more primitive genera, though it may be seen that in Thomas hux ley a and Asmodeus the premaxillae are relatively very large ; there is a considerable diastema between the canine and external incisors and all of the incisors are arranged in nearly the same fore-and-aft line as the grinding teeth, making the muzzle long and pointed. In the Santa Cruz genera the premaxillae are greatly reduced, the incisors are in a transverse row and the muzzle and chin are abruptly rounded. The nasals are much shortened and the anterior nares greatly enlarged and are no longer ter- minal in position. The zygomatic arch does not rise so high upon the side of the cranium and the tympanic bulla is much larger than in the Toxodonta. Hardly anything is known of the vertebral column, and limb-bones have been found in association with only a very few of the genera. There is no coossification of the fore-arm or leg bones, nor any loss or fusion among the bones of the feet, so far as these elements are known. In Asmodeus of the Deseado stage, and in Homalodontotherium the humerus is extremely massive and is remarkable for the great development of the deltoid crest, the external epicondyle a‘nd the supinator ridge, and in Asmodeus and Diorotherium an entepicondylar foramen is present. The fore-arm bones are very long and stout, quite as in the Chalicotheres, but without any tendency to coossification ; the radius has a relatively small head, with more or less power of rotation upon the humerus, and very heavy distal end. The ulna is also very stout, especially in Asmodeus , in which it does not taper so much distally as it does in the Santa Cruz genus. The femur, which is known only in Homalodontotherium , is some- what proboscidean in appearance, owing to the antero-posterior com- pression and flattering of the shaft, but the third trochanter is large and conspicuous, though not very prominent. The leg-bones, likewise known only in the same genus, are proportionately short. The tibia is very heavy and broadens distally, where it extends over and rests upon the fibula. The latter has a stout, subcylindrical shaft and very massive distal end, with large, oblique facets for the calcaneum and astragalus. Aside from Leontinia , which will be considered below, the very curious feet are not at all fully known except in the Santa Cruz genus. The manus is pentadactyl, with interlocking carpus and relatively very long metacarpals, of which the fifth is considerably the stoutest ; the distal trochlea is curiously recessed and thrown back of the axis of the shaft. scott: ENTELONYCHIA OF THE SANTA CRUZ BEDS. 243 The phalanges show an unusual degree of mobility and the unguals are converted into large, pointed and deeply cleft claws. The pes, also penta- dactyl, has an astragalus with almost ungrooved trochlea, elongate neck and small, strongly convex head. The tuber calcis is very much longer and more compressed than in the Toxodonta, in which the tuber is short and massive ; the fibular facet of the calcaneum is very large, forming a prominent external projection, and very oblique, presenting distally almost as much as dorsally. The metatarsals are extremely short, hardly more than one-third as long as the corresponding metacarpals ; the fifth (mt. V) has a great, hook-like projection from the fibular side of the proximal end, much as in the Santa Cruz genera of Gravigrade Edentates. Phalanges of the pes have not yet been obtained. A few isolated bones, including an ungual phalanx, show that in Asmodens, of the Deseado stage, the feet were of the same character as in Homalodontotherium, but in the very small Trimerostephanns (of the same stage) which Ameghino refers to the family Isotemnidse, the calcaneum and astragalus are much more primitive, the former having a slender, elongate tuber and a narrow, non-projecting fibular facet. Under the name of Colpodon Gaudry has given a brief description of the pes of Leontinia and a figure of the astragalus and calcaneum (’o6rt, 28, fig. 46, 30). According to this writer, Leontinia has a tridactyl pes, of which the calcaneum and astragalus are closely similar to those of Nesodon and altogether different from those of Homalodontotherium. The calcaneum has a short and very heavy tuber and large, but not projecting, fibular facet ; the astragalus has a narrow and elongate trochlea, a very short neck and but moderately convex head. Schlosser states (Ti, 523, apparently on the authority of Gaudry) that in this family the ungual phalanges are broad hoofs. If there is no error in attributing these foot- bones to Leontinia , it is perfectly evident that this genus is not referable to the Entelonychia at all, but to the Toxodonta, of which it forms a some- what aberrant, but perfectly characteristic member. So far as the structure of these animals is known, they resemble the Entelonychia only in having rooted molars, while, on the other hand, the peculiar development of tusks from the second upper and third lower incisors, the character of the skull and the hind-foot are typically Toxodont. 244 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. THE SYSTEMATIC POSITION OF THE ENTELONYCHIA. The early view of Flower (’84, 181) and of Lydekker (’86, 160) that Homalodontotherium was allied to the rhinoceroses, which was sug- gested when nothing was known of these animals save the teeth, may be disregarded, especially as Lydekker subsequently adopted a very different opinion (’96, 77). Of late years, but two conflicting opinions, each in somewhat varying forms, concerning the systematic position and relationships of the Entelonychia have been expressed, (1) that of Ame- ghino, that the Entelonychia are to be referred to the Ancylopoda, and (2) Roth’s view (’03) which makes them a part of his “ Notoungulata,” a term equivalent to the “Toxodontia” as here employed. Lydek- ker’s opinion, which may be regarded as essentially like that of Roth, assigns the Homalodontotheriidae to the suborder “ Astrapotheria,” but holds that the Toxodonta, Typotheria, Astrapotheria and Litopterna “have originated from a common ancestral stock” (’96, 77), which is merely adding the Astrapotheria and Litopterna to the Notoungulata, a procedure for which much may be said. Ameghino’s views on this subject may best be made clear by a series of quotations: “Cette apparence lophodonte des molaires et premolaires de X Homalodontotherium est due k l’age tres avance des individus figures. Cette denture est en realite buno-lophodont, les deux lobules internes des molaires superieures (protocone et hypocone) restant longtemps separes en forme de tubercules pointus” (’940, 56). “Les relations de parente entre les Homalodontotheridce et les C. ha lie other idee se manifestent d’une maniere tr£s evidente par la forme crochue des doigts, par la disposition des sur- faces articulaires distales des metacarpiens et des metatarsiens, par la forme des articulations proximales des premieres phalanges, par les pha- langes ongueales qui ont une forme semblable et sont fendues perpen- diculairement k leurs extremites, par le caractere tout particulier d’avoir le doigt externe de chaque pied plus developp£, et enfin par le caractere encore plus singulier d’etre les doigts externes des pieds ceux qui sup- portaient le poids principal du corps. “Tous les caracteres par lesquels les Homalodontotheridce s’eloignent des Chalicotheridce . . . indiquent un degre devolution peu avancee” (pp. 60, 61 ). “ Les caracteres qui distinguent les Chalicotheridce indiquent au contraire un degre d’ evolution tr£s avancee. Parmi ces caracteres, scott: ENTELONYCHIA OF THE SANTA CRUZ BEDS. 245 celui dii diplarthrisme du pied est peut-etre le plus notable ; il consiste dans l’articulation de l’astragale avec le scapho'ide et le cubo'ide a la fois comme chez les perissodactyles stereopternes. Mais, il s’agit certainement d’un diplarthrisme acquis independamment de celui des perissodactyles, par un modification graduelle de l’astragale taxeopode des Homalodonto- theridce ” (p. 61). In a later publication he suppresses the term Entelo- nychia, merging it with the Ancylopoda, and distinctly recognizes a rela- tionship with the Toxodonta. “Par la disposition de la denture dans son ensemble, par la forme g£nerale du crane ainsi que par les parties commes du squelette, ces animaux \Leontiniidce\ paraissent constituer une transition entre les Homalodontotheriidce et les Nesodontidce ” (’97, 65). Roth’s position is very briefly, but sufficiently, stated in his summary. “Das Vorhandensein der Pars mastoidea bei den Homalodontotheriden zeigt uns aber, dass diese Familie zu den Notoungulaten gehort, wahrend bei den Astrapotheriden keine Spur eines solchen vorhanden ist und sie folglich nicht zu dieser Gruppe gehoren konnen. In der That zeigt auch eine eingehende Untersuchung der Entwicklung des Gebisses der beiden Familien, dass dasjenige von Homalodontotherium auf clem Grundplan des Zahnsystems der Toxodonten und das von Astrapotherium auf dem der Rhinoceriden basirt ist” (’03, 33). My own opinion is entirely in agreement with that of Roth, so far at least as the relationship of the Entelonychia to the Toxodonta is con- cerned. Every known part of the skeleton confirms this view and testifies against any connection with the Ancylopoda of the northern hemisphere. The points of resemblance to the Ancylopoda upon which Ameghino lays stress are essentially but three in number, (1) the bunolophodont teeth, (2) the form of the phalanges, (3) the enlargement of the external digit of both manus and pes to carry the chief part of the body-weight. (1) The difference in pattern between the upper molars of the chalico- theres and the homalodontotheres is not accurately expressed by calling one bunolophodont and the other lophodont. As a matter of fact, they are both lophodont, having well-defined transverse crests, though in the former family the antero-internal cusp (protocone) is sometimes, not always, isolated. The important difference is in the character of the external wall, which in the chalicotheres has two deeply concave cusps separated by a prominent mesostyle, as in the titanotheres, palasotheres, horses and Litopterna, while the South American family has the almost 246 PATAGONIAN EXPEDITIONS: PALAEONTOLOGY. smooth, continuous outer wall, in which the distinction of cusps is nearly obliterated. So far as the development of tooth-patterns is understood, it would be quite impossible to derive either type from the other. (2) That the phalanges have a resemblance in the two families, though not an especially close one, is undoubtedly true, but this of itself forms a very insecure foundation for any theory of relationship. The distal trochleae of the metapodials are very peculiar in the Santa Cruz genus, and quite different from the convex, hemispherical knobs of the chalicotheres, which are more like those of the clawed oreodont Agriochocrus. (3) The discovery of nearly complete skeletons of the chalicotherian genus Moropus in the lower Miocene of North America, and the descrip- tions given by Mr. Peterson (’07) and Professor Barbour (’08) clearly indicate that in the Chalicotheriidse the greater development of the external digit was acquired within the limits of the family, not inherited from any Santa Cruz or earlier South American ancestry. Moropus is much less extremely specialized than the genera found in the middle and upper Miocene of Europe and its feet are nearly mesaxonic, the third metapodial being the longest. In the manus the second digit is much the heaviest of the series and has by far the largest claw, while in the pes the third and fourth digits are of nearly equal size. (See Peterson, ’07, 746, fig. 26.) That the so-called Ancylopoda are aberrant perissodactyls, is the con- clusion reached by all who have examined Moropus and in my judgment, it is equally clear that the Entelonychia represent an analogous variant of the toxodonts. The likeness between these two aberrant groups is really a remote one and is confined to the phalanges of the fore-foot, while the resemblance between the chalicotheres and the perissodactyls, on the one hand, and between the homalodontotheres and the toxodonts, on the other, is fundamental, involving all parts of the dentition and skeleton. Whichever conclusion be adopted, it will be impossible to avoid the admission of a convergent development, either of the Ancylopoda toward the Perissodactyla, or. of the Entelonychia toward the Ancylopoda. HOMALODONTOTHERIUM Flower. (Plates XXVIII-XXX.) Homalodotherium Huxley; Quart. Journ. Geolog. Soc. London, Vol. XXVI, 1870, p. lvii ( nomen nudum). SCOTT: ENTELONYCHIA OF THE SANTA CRUZ BEDS. 247 Homalodontotherium Flower ; Phil. Trans. Roy. Soc. London, Vol. CLXIV, 1884, p. 173. Homalodon Burmeister; Anales del Mus. Nac. de Buenos Aires, T. Ill, 1891, p. 389. Homalotherium Ameghino ; Ibid., T. XV, 1906, p. 317 [err ore). It is customary to attribute this genus to Huxley, but it is much more properly referable to Flower. The name proposed by Huxley is the baldest nomen nudum , as becomes perfectly clear when the original ac- count (it cannot be called a description) is read. “Still more perplexing are the strange and interesting forms Toxodon, Macrauchenia , and Typo- therium , and a new Anoplotherioid mammal ( Homalodotherium ) which Dr. Cunningham sent over to me some time ago from Patagonia.” (Hux- ley, loc. cit.) The first description of the genus was given by Flower (/oc. cit.) who adopted Huxley’s name in emended form. I am unable to comprehend the course followed by Dr. Palmer in his admirable “ Index Generum Mammalium”; he attributes the genus to Flower and yet re- tains Huxley’s name, which is marked “nomen nudum.” This is a self- contradictory procedure, since a nomen nudum can have no rights of priority (Palmer, ’04, 330). Homalodontotherium is a comparatively rare member of the Santa Cruz fauna and the Princeton and New York collections contain but little material illustrating it. In the La Plata Museum are the valuable speci- mens which have been figured and briefly described by Lydekker (’93^ 44-47) and the genus is very well represented in the collection of Dr. Ameghino. It is from this material, as well as from the publications of Flower, Ameghino, Lydekker and Gaudry, that the following account has been chiefly drawn, through some of the specimens collected by Messrs. Hatcher and Peterson have proved very useful. Thanks principally to the work of Ameghino, the structure of these very curious animals is now fairly well known and their systematic position may be made reasonably clear. Dentition. — The dental formula is unreduced ; If, Cf, PI, Mf ; as Flower has pointed out, the teeth, which in both jaws are arranged in unbroken series, “ in their configuration present a remarkable and gradual transition from the first incisor to the last molar, easily traced in both jaws, and more even and regular than in any other known heterodont mammal” (’84, 175). At the time when this description was written, the dentition 248 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. of Theosodon, a genus of Santa Cruz Litopterna, was not yet known ; in this animal the transition in the form of the teeth, from the first incisor to the last molar, is even more regular and gradual than it is in Homalo- dontotherium. The teeth are of nearly even height, except that in some individuals, probably males, the canines project above and below the level of the lower and upper series respectively. Roots are formed at an early stage of tooth-development, but the crowns are relatively high, though I cannot follow Flower in calling them “ hypsodont,” a term which should be reserved for the prismatic tooth, which develops roots only in advanced age. A. Upper Jaw (Pis. XXVIII, figs. 1, 2; XXIX, figs. 2, 3). — The two straight and slightly divergent dental series are connected in front by the short and moderately curved, transverse row of incisors. These teeth are quite simple and conical, increasing somewhat in size from i- to i- and have very strongly developed cingula, both internal and external. The first incisor, r1, has a long, bluntly pointed crown, with a median vertical thickening, thinning to an edge on each border. The crowns of i- and -, when viewed from below, have an acutely pointed, hastate shape, broadening abruptly from the root. The median thickening produces quite a distinct convexity upon the external face of the tooth and the edges are trenchant. I- is considerably broader than i-, but otherwise of similar shape ; it has, sometimes at least, a small internal cusp arising from the cingulum. The canine differs considerably in the various individuals in relative size and prominence. Presumably, this is a sexual rather than a specific character, but the materials are insufficient for a definite determination of this question. In Flower’s specimen, the type of H. cunninghami, the canine “only differs from the posterior incisor in being somewhat larger and in trifling details of configuration. The apex is rather more pointed and conical, being supported by a median vertical ridge, not only on the outer, but also on the concave inner surface of the crown ; the inner tubercle is relatively smaller to the principal cusp ” (Flower, loc. cit., P. 176). In his original description of H. segovice, Ameghino distinguishes this species from H. cunninghami , among other characteristics, by “the largely developed upper canine” (’91, 295). Much more probably, however, this scott: ENTELONYCHIA OF THE SANTA CRUZ BEDS. 249 difference is sexual rather than specific, for a palate in the La Plata Museum, obviously referable to H. cunninghami, has relatively quite large and tusk-like upper canines, which are much more prominent than in the type ; presumably, therefore, the latter individual was a female. The palate in question has been figured by Lydekker (’93, PI. XIX). The premolars, which increase gradually in size posteriorly, have a decided resemblance in pattern to the molars, yet are easily distinguish- able from them and none is altogether molariform. Except in p1, in which the transverse and antero-posterior diameters are nearly equal, the transverse width of the premolar crowns considerably exceeds the antero- posterior length. A strong cingulum is developed all around the crown, being especially prominent and conspicuous on the internal and external sides ; on the latter it is most prominent, rugose and tuberculated. The first premolar (p-1) varies considerably in size, both individually and spe- cifically, though always the smallest of the series ; it is much smaller in H. segovice and relatively quite large in H. cunninghami. In the type of the latter species this tooth is somewhat displaced toward the inner side and overlapped externally by the canine, but this is no doubt an indi- vidual peculiarity, as I have not observed it in any other specimen. The crown is supported, according to Flower, by two roots, one internal and the other external, and is not very different from that of the canine. It is, however, less pointed and conical and lower dorso-ventrally, longer antero-posteriorly and broader transversely ; the external face is more quadrate and the vertical ridge is displaced to a position which is some- what behind the middle. The internal cusp, or deuterocone, is much better developed than that of the canine and forms a prominent ridge ; the broad valley opens anteriorly and, at least in the moderately worn condition, is closed posteriorly by a raised border, which connects the internal cusp with the external wall of the crown. From this border a large spur projects forward, incompletely dividing the valley into external and internal portions, the anterior part of the valley being undivided. The remaining premolars (p-, -, A) have more nearly acquired the molar pattern and differ merely in size. Each is implanted by three roots, two external and one internal, the latter being very large in p- and A, quite slender in p-. I am quite at a loss to understand Ameghino’s statement that the upper premolars have “the roots all coalesced into a single one” (’98, 172), for I have seen no specimen that would even suggest such a 250 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. description. The external face of the crown, in the unworn or moderately abraded condition, is short antero-posteriorly, and quite high dorso- ventrally, but their appearance changes much with advancing abrasion, which steadily reduces the height of the teeth. This outer face seems to be convex, but there is a well-defined vertical ridge near the anterior border and another, much less distinct, near the posterior border, with a very shallow concavity between the two ridges. Two short, transverse crests are given off from the anterior and posterior borders of the external wall and curve towards each other so as to meet internally. The two internal cusps (deutero- and tetartocones) with which the transverse crests are indistinguishably fused, are very incompletely separated by a shallow vertical groove and only in perfectly unworn teeth is there an internal opening of the valley between the apices of the inner cusps, the valley being almost immediately converted into a closed enamel-lake. I have actually observed this internal opening of the valley only in pA, no perfectly unworn examples of p- and - being available. If present in them at all, the opening must be exceedingly shallow. Flower states that p-, “in addition to the principal oval fossa [/. e., valley], has a second smaller one behind and to the outer side of it” (p. 177). This is very early removed by attrition ; it appears in one of the skulls of the La Plata Museum collection, but not in any other individual which I have examined. The molars have considerable resemblance to those of the rhinoceroses, but the likeness is seen to be a superficial one, when the freshly erupted, unabraded teeth are examined. While all of the upper molars are of the same essential pattern, they differ appreciably from one another in con- struction. Here again, it is necessary to compare unworn teeth, for the differences are much less clear after abrasion has made much progress. The molars are notably larger than the premolars, the change from p- to mA in this respect being quite striking, and not only is there a difference of size, but also one of proportions. In the premolars the transverse diameter of the crown distinctly exceeds the antero-posterior, while in the molars the two diameters are nearly equal, or the antero-posterior is the greater one. The fore and aft dimension increases posteriorly and is greatest in m-, though the width of m- makes that tooth the largest of the series. All the molars are so closely appressed that the antero-external angle of each tooth overlaps the one in front of it, another difference from the premolars, which is also found in the Toxodonta. SCOTT : ENTELONYCHIA OF THE SANTA CRUZ BEDS. 251 The disposition of enamel upon the molar-crowns is very peculiar. In unworn teeth the enamel is not reflected over upon the inner side of the external wall, or upon the hinder side of the anterior crest, or the front face of the posterior crest. The valley is thus enclosed in walls of enamel- free dentine, except near the base of the crown, where it reappears. In the abraded tooth, however, the enamel is again displayed as completely lining the valley and the resulting pattern is the same as though the unabraded crown had been entirely covered with enamel. The development of the cingulum varies greatly, but the differences are fluctuating and are not correlated with other characters ; apparently they are individual, not sexual, and have not the taxonomic significance which has sometimes been attributed to them. In Flower’s type: “All the teeth have crowns . . . with a well-marked cingulum around their base” (p. 175), but his figures seem to show that only m- has it on the outer face and even in this tooth the external cingulum is much less prominent than in the premolars, while all of the molars have it strongly developed on the internal, anterior and posterior sides of the crown. Ameghino states that the molars “have a cingulum on the inner, but not on the outer side” (’89^, 551), a conclusion which agrees with my own observations, for in none of the individuals which I have had an oppor- tunity of examining is there an external cingulum on any of the upper molars and it is variable on the other faces. In both of the La Plata specimens it is prominent on all sides except the outer one, while in one individual of the Princeton collection (No. 16,016), which agrees quite closely in size with the type of H. cunninghami , the cingulum is present only on the anterior and posterior faces and there is no trace of it on the inner side. In another of the Princeton specimens (No. 16,014), a smaller species, probably referable to H. segovice, the internal cingulum is partially and faintly developed on m-, and entirely lacking on m- and -. On the other hand, Ameghino’s type of H. segovice has a prominent internal cingulum. While invariably present, so far as I have observed, on the anterior face of all of the molars and on the posterior face of m1- and -, it varies even here in degree of prominence and tuberculation. The facts just cited make it clear that the development of the cingulum of the upper molars is subject to great fluctuation in specimens which, on every other ground, are referable to the same species. That the varia- tion is individual and not sexual is made extremely probable by the type 252 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. of H. cunninghami, which has an exceptional development of the cingulum and yet would appear to be a female, as indicated by the small size of the canines. Like the premolars, the molars are each carried upon three very long roots, of which the two external ones are comparatively slender and the internal one is very large, extending for the whole antero-posterior length of the crown. This inner root, however, is obviously formed by the coalescence of two roots and on the buccal side the fusion is incomplete, a ridge of bone in the alveolus partially separating them. The first molar (m-1) is considerably larger than p- and has a sub- quadrate crown, the antero-posterior and transverse diameters of which are nearly equal. The external face is nearly smooth and convex, but there is a low vertical ridge near the anterior border, which is much less conspicuous than in the premolars. This ridge is followed by an extremely shallow concavity and, opposite the posterior transverse crest (meta- loph), by a second very faintly marked ridge. The outer face is thus feebly sinuous. From the external wall two transverse crests are directed inward, ending in the two internal cusps, the proto- and hypocones, with which the crests are indistinguishably fused. Even in quite well worn teeth these transverse crests are widely separated internally by a broad and deep valley, which is the most obvious difference from the premolar pattern. The crests are nearly transverse in direction and are less oblique than in the rhinoceros molar, but, owing to the shape of the valley, the posterior crest is much shorter than the anterior. The external wall of the crown extends well behind the metaloph, with which it partially encloses a shallow fossa. As the valley is much deeper toward the external side, its internal opening is obliterated with the progress of abrasion and the valley is converted into a narrow, elongate lake, which has an oblique course, running backward and inward. At this stage of wear the difference in pattern between molars and premolars is far less obvious than in the unworn condition. The second molar differs from m1 chiefly in its greater size, especially in the antero-posterior diameter, and in a slightly greater complexity of pattern. In the fresh and unworn state the two internal cusps are very widely separated by a broad, V-shaped notch, the inner opening of the valley. The anterior crest is simple and nearly straight and a shallow notch demarcates it from the external wall of the crown, but this notch SCOTT: ENTELONYCHIA OF THE SANTA CRUZ BEDS. 253 speedily disappears with wear. The posterior crest is much shorter and the notch between it and the outer wall is much deeper, but it also is soon destroyed by abrasion. Presumably, it is to the presence of these notches that Ameghino refers when he says: “Cette denture est en realite buno-lophodonte, les deux lobules internes des molaires superieures (protocone et hypocone) restant longtemps separes en forme de tubercules pointues” (’94*, 56). This statement is quite erroneous, as immediately appears when the upper molars of Homalodontotherium are compared with those of the titano- theres, which family has truly buno-lophodont teeth. Disregarding the peculiar character of the external wall in this family, which it shares with the horses and the palaeotheres, we find that there in no indication of a transverse crest in the titanotheres, the internal cusps being altogether distinct and of conical shape. In some of the Eocene genera more or less definite remnants of the conules may be found, but there is no tendency for these conules to coalesce with the inner cusps. In the Santa Cruz genus, on the contrary, there are well defined transverse crests, which are partially separated from the outer wall, in the unworn state of the teeth, by shallow notches, which are soon obliterated by wear. Fig. 43. Homalodontotherium segovice : Second and third upper molars of left side, X y- Oblique view of grinding surface. M- slightly worn, m- not yet erupted. (No. 16,014.) In m- a thin, almost plate-like ridge runs obliquely forward from the isolated apex of the posterior crest, dividing the valley into two por- tions, a deep internal and shallow external part. Besides, two short spurs are given off inward from the external wall and, with the ridge just mentioned, enclose two shallow fossae. After a short period of use, how- 254 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. ever, these complications disappear ; the outer portion of the valley, the oblique ridge, spurs and fossae are no longer visible, the external wall is much thicker and the valley is a mere oblique slit, which eventually is converted into a closed lake. The external wall has the same ridges and sinuosities as in m1, but extends farther behind the posterior crest, en- closing a larger fossa. In the unworn condition this wall has a tren- chant edge and behind the metaloph curves upward abruptly to the base of the crown. When abrasion is well advanced, hardly any difference from nr1 remains, except the greater antero-posterior elongation of the crown. When fully protruded, m- is still more elongate antero-posteriorly than m- and is of quite a different shape, being less quadrate and more tri- angular, somewhat as in the rhinoceroses, but the external wall is not fused with the metaloph, as it is in the latter. In the freshly erupted tooth the crown-pattern differs in several details from that of m-. The anterior crest is well developed and at its inner end, behind the protocone, a con- spicuous conical cusp is added, no trace of which is visible in m1 or the anterior crest is thus L-shaped, and its transverse and antero-posterior limbs are of nearly equal length. The notch between the anterior crest and the outer wall is much shallower than in m-, nothing more than a slight depression of the free margin. The external wall has a trenchant edge, where the thick enamel ends abruptly, and, seen from the side, the tooth has somewhat the appearance of a carnivorous sectorial ; the upward curve of the trenchant edge begins a little behind the anterior transverse crest. The sinuosity of the outer wall is very much as in m-, but the anterior vertical ridge is even less distinct. The posterior transverse crest is greatly reduced and is very short and very low ; to this reduction is chiefly due the triangular shape of the crown ; the crest is widely separated from the external wall, a striking difference from m1 and -, but it is, nevertheless, a crest and not a conical cusp. The oblique ridge which in m- passes forward from the isolated apex of the posterior crest, in m- originates in the middle of the crest and is shorter, lower and less distinct than in m-, not dividing the valley into internal and external portions and extending only to the third of the four spurs which project from the outer wall. These spurs are shorter, thinner and more closely crowded together than the two which are found in m-; a small fossa is enclosed between the third and fourth spurs and the SCOTT: ENTELONYCHIA OF THE SANTA CRUZ BEDS. 255 oblique ridge and a much larger one by the fourth spur, the ridge, the metaloph and the outer wall, but the latter fossa opens posteriorly between the metaloph and the wall. The inner face of the wall is very rough and given a wrinkled appearance by numerous irregular vertical ridges of dentine. The third upper molar, in the unworn condition, thus differs from m1 and m- in several particulars, which maybe thus summarized: (1) the triangular shape of the crown; (2) the presence of a large, conical, acces- sory tubercle on the posterior side of the protocone ; (3) the great reduc- tion of the posterior transverse crest and its wide separation from the external wall ; (4) the presence of four separate spurs projecting inward from the outer wall. The appearance of the tooth is much changed by abrasion, though the characteristic triangular shape of the crown is always retained. The accessory tubercle, the oblique ridge and the spurs are all worn away and the metaloph becomes connected with the external wall, but remains conspicuously short ; the valley is converted into a narrow, oblique slit and, in old animals, into a closed lake, while the enamel lining of the valley is displayed. In this abraded condition the resemblance to the other molars is much closer. B. Lower Jaw (PI. XXIX, figs. 1, 1 a). — Mandibles are even more rare in the collections than skulls and such as are known belong in nearly all cases to old animals. I am therefore unable to give as complete an account of the inferior dentition as of the superior, having seen no examples of quite fresh and unworn teeth. The youngest specimen known to me is Flower’s type of H. cunninghami. The incisors are smaller than those of the upper series, but quite similar to them in form ; they are comparatively simple, compressed-conical teeth, with acute points, trenchant edges and prominent cingulum, and grow larger from iy to iy. The third incisor is figured by Flower {t. c., PI. XVI, figs. 2, 3) in an almost perfect state and shows on the internal face “a vertical ridge running from base to apex of the crown, rather nearer the anterior than the posterior edge of the tooth” (pp. 177-8). The canine in the type of H. cunninghami , and in a mandible of the Ameghino collection referred to the same species, is larger than iy, but considerably smaller than the upper canine. Whether in the supposed males with enlarged upper canine, the lower tooth is also tusk-like, can- 256 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. not be determined from the available material. In Flower’s specimen the lower canine is very similar to h in form, but is less regularly conical, the crown is more elongate antero-posteriorly and thicker transversely, and the vertical ridge on the inner side is more prominent and massive. The premolars are all different from one another and from the molars, though pi approximates the latter in structure and p3 does so also, but in a less marked degree, the transition from py to py being very gradual. The first premolar is subject to considerable variation in size, which may be specific, or only individual, and to some fluctuation in form. In the type of H. cunninghami py on the left side has a smooth, convex and undivided external surface, while on the right side there is an incipient division into two lobes visible on the outer face. (Flower, t. c., p. 178, PI. XVI, fig. 2.) This division into two crescents becomes more con- spicuous and the posterior one grows larger on each successive tooth of the premolar series. In a specimen of the Princeton collection (No. 16,015) which agrees almost exactly with the type in size, py on both sides of the jaw has a simple and rounded outer surface without trace of a division. Internally, the valley is divided into anterior and posterior portions by a vertical, wedge-shaped ridge, the deuteroconid, which is broad at the base and narrows regularly to the apex. In the unworn tooth, of which I have seen no examples, this apex no doubt forms a dis- tinct and separate cusp, but after a very brief period of wear it becomes confluent with the external wall. The broad and shallow posterior valley has a tubercle and one or two short, vertical ridges of enamel, which differ in number and prominence on the two sides of the jaw. The cingulum is very prominent all around the crown. The first premolar is implanted by a single root, the others have each two roots, which Ameghino states (’89*, 552) are again divided, each into two, near the tip. I am unable to confirm or dispute this statement. The second premolar is plainly divided by a deep external groove into two imperfect crescents, of which the anterior one is the larger ; its internal valley is shallow, but distinctly marked, while that of the posterior crescent is small but deep, and has an extremely narrow internal opening, owing to the backward extension of the internal pillar (deuteroconid) which arises at the junction of the two crescents, and also to the forward curvature of the hinder horn of the posterior crescent. The valley is thus soon con- verted into a lake by abrasion. The cingulum is prominent all around scott: ENTELONYCHIA OF THE SANTA CRUZ BEDS. 257 the tooth, except on the anterior and posterior faces, where it rises high upon the crown and dies away in the median line. The third premolar is closely appressed to po, so that it has a concave anterior face, where the enamel becomes very thin, or fails altogether ; and the cingulum is wanting, as it also is on the posterior face, at least in some individuals. The posterior crescent is both actually and relatively larger than in p^, equalling or slightly exceeding the anterior one in fore- and-aft diameter. The anterior valley is extremely shallow, so shallow that it can hardly be said to be present at all in moderately worn teeth, while the posterior valley is well-defined, considerably larger than in p^ and more widely open internally. The internal pillar (deuteroconid) forms a backward curving ridge, with convex inner face and, in conjunction with the posterior crescent, makes up a C-shaped masticating surface, which is more regular in outline than that of p ¥. In pT the hinder crescent is still larger and more distinct than in pg and its posterior horn curves forward more sharply, while the valley of the anterior crescent is still extremely shallow. The internal pillar is a more elongate ridge and more nearly closes the posterior valley, so that the masticating surface formed by the ridge and the posterior crescent is of an almost complete, but very irregular oval shape. The cingulum is heavy and prominent on the internal and external faces of the crown, but incom- plete on the forward and hinder ends. All of the lower molars are very nearly alike, there being no such dif- ferences among them as are found in the upper series. In size, there is a moderate increase from mT to m¥, though m^ has the greatest transverse breadth. There is no talon or heel on m-3, as is generally the case through- out the order. Each of the molars is carried upon two broad roots, which, according to Ameghino (’89*, 552), are bifurcated near the free ends, but I have seen no specimens which make this point clear and therefore can make no comment upon Ameghino’s statement. All of the molars have the bicrescentic pattern common to all known Santa Cruz ungulates, but the anterior crescent is very short and the posterior one very elongate, the latter having more than twice the fore- and-aft length of the former. The external groove which demarcates the crescents is less deep and distinct than in the hinder premolars ; especially is this true of mi. While the forward crescent is far shorter antero-posteriorly than the hinder one, it is more complete, its posterior 258 PATAGONIAN EXPEDITIONS! PALAEONTOLOGY. horn curving inward and backward and taking the place of the internal pillar of the premolars, with which it is perhaps homologous, though probably it is not. The valley of the anterior crescent is narrow, but deeply impressed and thus differs strongly from the same structure in the premolars ; that of the hinder crescent is much broader, but relatively shallower. In the valley of the posterior crescent is a vertical pillar, or spur, such as is found in almost all Santa Cruz ungulates and is highly characteristic of the South American types generally. In the rare cases, such as the horse-like Thoatherium , in which this pillar is lacking, its absence is obviously due to a secondary reduction. In Homalodontotherium the pillar is very conspicuous, and, after a short period of abrasion, becomes connected with the hinder horn of the anterior crescent, enclosing an enamel lake, which is worn away in old teeth. The cingulum is quite strongly developed on the internal and external sides of the crown, but not on the anterior and posterior ends. The three teeth of the series are very closely appressed, the anterior face of each being flattened or concave and the posterior face convex. On these approximate surfaces the enamel is in- complete, being reflected over from the inner and outer faces, but becomes very thin and dies away in the middle. Needless to say, the hinder end of m3, which is freely exposed, is completely covered with enamel. Nothing has yet been learned of the milk-teeth in this genus. The dentition of Homalodontotherium is obviously of the same plan as that of the Toxodonta, but is in a somewhat simpler and more primitive stage of development. That there is a certain resemblance, especially in the grinding teeth, to the dentition of the rhinoceroses, is not to be denied and Flower goes so far as to say: “On comparing a lower molar of Homalodontotherium with an equally worn tooth of Rhinoceros , it will be seen that they are formed on precisely the same type.” “The molar and premolar teeth of both upper and lower jaws thus without question show strongly marked Rhinocerotic characters” (/. c., p. 180). A study of perfectly fresh and unworn teeth, however, leads to the conclusion that the rhinocerotic resemblances are superficial and without special signifi- cance, the toxodont affinities being far stronger, and, as has been shown in a preceding section (p. 112), the toxodont dentition has certain distinct likenesses to that of the rhinoceroses. In making the comparison, the peculiar and simplified pattern of the grinding teeth in the Pampean toxodonts should be omitted from con- scott: ENTELONYCHIA OF THE SANTA CRUZ BEDS. 259 sideration and the Santa Cruz forms, such as Nesodon and Adinotherium , employed, since in them the essential characteristics of the toxodont den- tition are much more clearly displayed. From the dentition of the Santa Cruz toxodonts that of Homalodontotherium differs: (1) in the relatively unmodified form of the incisors and canines; (2) in the brachyodont character of the dentition ; (3) in the presence of the cingulum on all of the teeth, though very incomplete in some cases ; (4) in the much greater transverse width and consequently less compressed form of the grinding teeth; (5) in the somewhat less complex pattern of these teeth, due espe- cially, in the upper molars, to the shortening of the posterior crest and to the less prominence and early disappearance of the spurs (“combing plates”) from the external wall of the crown. In Nesodon the principal spur is very prominent and even in well-worn teeth divides the valley into two parts and give it its characteristic Y-like shape (see Pis. XV, XVII). The skull (Pis. XXVIII, figs. 1, 1 a\ XXIX, fig. 2) is quite small in proportion to the size of the trunk and the length of the limbs. The two specimens which I have had an opportunity to study have both suf- fered from distortion by pressure, though in different directions. That belonging to the type of H. segovice in the Ameghino collection which is shown in Plates XXVIII and XXIX, is very complete, but has been depressed and flattened by vertical pressure, making it appear to be unduly flat and low in the dorso-ventral dimension, as is clearly indi- cated by the shape of the orbit. A very young and imperfect skull in the Princeton collection (No. 16,014) lacking the preorbital portion, has undergone a lateral compression, which has diminished the transverse width and increased the dorso-ventral diameter. On this account the two skulls, at first sight, seem to have very different appearance and pro- portions. The skull of Homalodontotherium is essentially toxodont in character, differing only in a number of relatively unimportant details ; the very unusual and peculiar structure of the auditory region, which is common to all of the Toxodontia, is repeated in this genus, though in somewhat less pronounced fashion. The upper profile of the skull is nearly straight and horizontal through- out its length, but there may be present a slight descent at the forehead. The orbits have a forward position, with the anterior rim over m~, and the preorbital portion of the face is somewhat shorter than in Nesodon , 26o PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. making the cranium relatively longer, though the brain-case proper, meas- ured from the postorbital constriction, is quite short. The whole skull is lower dorso-ventrally than in the Toxodonta, and the greatly shortened nasals, the much larger and obliquely placed anterior nares and the reduced premaxillaries, as well as the absence of large tusks, give a very different appearance to the entire facial region. The zygomatic arch, which is very broad dorso-ventrally, making it seem to be very heavy and massive in side view, does not rise so high or so steeply backward and conceals less of the cranial wall, while the dorsal border of the pos- terior portion descends quite strongly and unites with the occipital crest at a much lower level than in Nesodon or Adinotherium or the Santa Cruz typotheres, in all of which types the dorsal border of the zygomatic process at its origin is nearly on a level with the sagittal crest. The sagittal crest is quite long and very prominent, its development varying with age and perhaps with sex also. A very characteristic difference from the skull of the Toxodonta and Typotheria is the backward prolongation of the dorsal portion of the occiput, which extends well behind the plane of the condyles and, in side view, makes the occipital surface appear to be deeply concave, while in Nesodon and Adinotherium this surface is either almost vertical, or slopes forward and upward from the condyles. Though the structure of the auditory region is essentially the same as in Nesodon , the appearance of this region, when seen from the side, is very different. The very large tympanic bulla, with its great antero-posterior extension ; the long and freely projecting postglenoid process ; the small size, inferior position and non-projecting form of the external auditory meatus ; the very long, slender and antroverted paroccipital process, and, apparently, at least, the absence of any surface exposure of the mastoid portion of the periotic and of the mastoid process, are all highly characteristic features of this skull and clearly differentiate it from that of the Toxodonta and Typotheria. The upper view of the skull (PI. XXVIII, fig. id) has not the truncated triangular outline seen in the corresponding view of Nesodon and Adino- therium (cf. PI. XIV and PI. XX, fig. 3), but rather that of an elongate, somewhat narrow oval, with the maximum transverse width across the zygomatic arches a little behind the orbits, not at the glenoid cavity, as it is in the Santa Cruz Toxodonta. The backward projection of the scott: ENTELONYCHIA OF THE SANTA CRUZ BEDS. 261 dorsal portion of the occiput conceals the condyles from sight in this view, as is also the case in Adinotherium , but not in Nesodon. A con- spicuous feature, at least in H. segovice , is the notch which emarginates the occipital crest on each side and interrupts the connection between the crest and the dorsal border of the zygomatic arch, the continuity of which with the crest and the elevated position give such a peculiar appearance to the skulls of all the Santa Cruz Typotheria and Toxodonta. Another striking difference from these suborders is the inferior position and lack of prominence of the external auditory meatus and the post-tympanic process of the squamosal, so that no trace of them is visible when the skull is seen from above, while in the other two suborders the projecting tubular meatus and the swollen post-tympanic form the postero-external angles of the skull. The brain-case, though quite narrow, is more capacious than in Neso- don, but contracts rapidly forward to the postorbital constriction, making the temporal openings very wide, but quite short, while in Nesodon and Adinotherium the antero-posterior diameter much exceeds the transverse. The sagittal crest differs much in the two skulls, a difference which is obviously in part a matter of age, though possibly both do not belong to the same species. In the type of H. segovice , which is quite an old animal with well worn teeth, the crest is long and very high and prominent, while the young skull (No. 16,014) in which m5- is just beginning to come into use, it is lower and much shorter, broadening anteriorly into a narrower, elongate sagittal area. In neither specimen are there any definite tem- poral ridges, extending upon the forehead. The latter is very wide, expanding more than in Nesodon , but remarkably short antero-posteriorly, which is due to the backward displacement of the nasals. The nasals are far shorter and flatter than in the Santa Cruz Toxodonta, making the anterior narial opening much larger and of entirely different shape ; in- stead of being terminal, it presents more dorsally than anteriorly. The dorsal narrowing of the face is very much less marked than in Adino- therium or Nesodon and the border of the maxillaries receives a curiously fluted appearance from the prominence of the alveoli. The very small premaxillaries differ greatly in appearance from the large and massive ones of Nesodon. Seen from below (PI. XXVIII, fig. 1) the skull presents corresponding differences from that of the Santa Cruz Toxodonta, though the resem- 262 PATAGONIAN EXPEDITIONS I PAL/EONTOLOGY. blances are more fundamental and significant. The different character of the incisors and canines is one of the most striking features of the base- view and, in correlation with this, is the reduction of the premaxillaries, which form a much smaller portion of the palate than in the last named suborder. The muzzle narrows regularly to the anterior end, without the broadening forward so characteristic of nearly all the known Toxodonta, and the whole palate is relatively shorter, wider and less deeply concave. The peculiar form of the posterior nares, which obtains in all of the Santa Cruz Toxodonta and Typotheria, with its prominent and divergent side-walls, formed by the palatines, pterygoids and alisphenoids, is re- peated in the present genus, but the divergence of the walls and conse- quent breadth of the opening are even more striking than in Nesodon. This base-view also displays the shortening and strong outward curvature of the zygomatic arch, reducing the distance between m- and the glenoid cavity, as compared with the last named genus. The tympanic bulla is much larger than in the contemporary Toxodonta, especially in the antero-posterior dimension. The lack of prominence in the region of the external auditory meatus and the swollen post-tympanic process is even better displayed than in the superior view of the skull, the broadening of the cranium external to the paroccipital processes being much less than in Adinothevium and Nesodon. The much narrower occiput and the differ- ent shape of the condyles are also noteworthy features. The condition of the material at my disposal is such that no very detailed account of the various elements of the skull can be given, yet a certain number of important facts may be made out. The basioccipital is quite long and heavy, though narrowing between the large tympanic bullae, broadening considerably behind them ; it has a short, but quite well-defined ventral keel. The condylar foramina occupy a conspicuous position on each side of the posterior expansion and are not hidden by the condyles. The latter are rather small, but quite prominent, so as to be fully visible in the side-view of the skull. On the ventral side, the articular surface of each condyle, near the outer end, is emarginated by a deep notch, which is not shown in any of the Santa Cruz Toxodonta, and the articular surface is continued over upon the basioccipital in a median, shield-shaped area, which is peculiar to the present genus. I am unable to determine with certainly the limits of the exoccipitals, but they would appear to be much higher dorso-ventrally and to form SCOTT: ENTELONYCHIA OF THE SANTA CRUZ BEDS. 263 a larger proportion of the occipital surface than in Adinotherium or Ne- sodon; they also are of quite a different shape, not being so broad at the base nor so much constricted above. The paroccipital process is longer than in the last named genera and differently formed : the proximal por- tion is almost as massive as in Nesodon, but distally it becomes much more slender, laterally compressed and of trihedral cross-section. The process has a strong inclination downward and forward, and ends distally in a bluntly rounded point. The supraoccipital is quite narrow at the suture with the exoccipitals, but widens dorsally and is alone concerned in the formation of the occipital crest : this dorsal portion is also extended and recurved, so as to overhang the occipital plane and project behind the condyles, but is not so thickened and heavy as it is in Nesodon. The broad median convexity of the exoccipitals above the foramen magnum becomes, on the supraoccipital, divided into two ridges by a median groove. The occipital crest is thin and prominent and, as has been mentioned before, is demarcated on each side from the dorsal border of the zygomatic arch by a well-defined notch, which, however, does not prevent the junction of the two crests. As a whole, the occipital surface, though constructed on the same peculiar plan as in the Toxodonta and Typotheria, has yet a very different appearance, being higher and narrower without any such broadening of the base, while the inferior position of the zygomatic arches and the very long and relatively slender paroccipital processes are important factors in the difference, as are also the more pronounced median convexity and lateral concavities, the surface in the Toxodonta and Typotheria being more nearly plane. As in those suborders, the occiput proper is sharply constricted above the foramen magnum, which is low and wide, of trans- versely oval shape. The constriction leaves on each side a broad space, which is filled by the convex, inflated, hinder surface of the post-tympanic process of the squamosal, which is not so large and, in particular, not so broad as in Nesodon , the supraoccipital expanding dorsally and extending somewhat over the post-tympanics. The whole structure of the occipital surface is rather less extremely aberrant than in the Toxodonta. Nothing can be determined as to the extent and relations of any of the sphenoidal elements, because in the Ameghino specimen they are con- cealed by the matrix and in the Princeton skull they are so badly crushed that no identification is possible. 264 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. The parietals are long and narrow, forming but little of the side-walls of the cranium. So far as I can make out, these bones are thin through- out their length and do not have the posterior thickness and massiveness which they have in Nesodon. In the fully adult animal the sagittal crest, which is very prominent, continues for the entire length of the parietals, but in the young skull it passes anteriorly into a narrow, triangular area, which broadens to the frontal suture. The anterior ends of the bones diverge less than in Nesodon and the frontals do not project nearly so far backward between these divergent ends. Along the squamosal suture on each side is a row of four or five large and conspicuous vascular foramina. As in Nesodon , the squamosals are very large and form the greater part of the cranial side-walls, but there are many differences of detail, especially in the shape and size of the various processes. The glenoid cavity is a saddle-shaped surface, the direction of which is rather oblique ; it is narrow and very slightly convex antero-posteriorly, broad and some- what concave transversely. Externally, this articular surface forms a low and barely perceptible protuberance, which is very much lower and less conspicuous than in the Santa Cruz Toxodonta. The space between the glenoid cavity and the postglenoid process is much narrower than in the latter and the postglenoid process is of an entirely different shape ; it is a long, freely projecting and rather massive process, broad and thick proximally, where it is extensively applied to the post-tympanic process, completely excluding the mastoid from the surface of the skull, but taper- ing and becoming styliform distally. In Nesodon and Adinotherium this process is a broad, heavy and swollen-looking plate, no part of which is free, but is closely united with the mastoid, which separates it from the post-tympanic. The glenoid foramen is a larger and more conspicuous opening than in the genera last named, and channels the postglenoid process. The post-tympanic process is of the same curious and anomalous char- acter as in all of the other Toxodontia, though each of the three suborders of this group has its own particular modification of the general plan. In Homalodontotherium the lateral exposure of the post-tympanic is much greater than in Adinotherium or Nesodon , forming the posterior portion of the zygomatic arch and its dorsal border being raised into a ridge, which is continuous with that of the zygomatic process on one side and SCOTT : ENTELONYCHIA OF THE SANTA CRUZ BEDS. 265 with the occipital crest on the other. This lateral portion is inflated and, partially at least, filled with cancellous bone. The posterior portion, which makes up part of the occipital surface, is also much inflated and contains a sinus, which no doubt communicates with the cavity of the tympanic bulla, as in Nesodon , though I have not been able to make sure of this communication. This occipital exposure is relatively narrower and the occiput proper wider than in the Toxodonta, while in the Typo- theria the true occiput is much broader and the post-tympanic narrower than in either of the other suborders. The zygomatic process also differs in certain details from that of Neso- don, though, as in that genus, it has considerable resemblance to the same structure in the rhinoceroses, being rather thin and compressed, but very deep dorso-ventrally, giving it quite a massive appearance when seen from the side. It does not have the steep anterior descent seen in the Santa Cruz Toxodonta, but the dorsal border rises moderately for- ward to a point about the middle of its course and thence, in one of the skulls, descends anteriorly, while in the other it rises gently almost to the boundary of the orbit. In Adinotherium and Nesodon the process ends abruptly in front and is received into a shallow notch of the jugal, forming little or none of the orbital boundary, while in Homalodonto- therium it is relatively longer, does not notch the jugal, but rests upon it and narrows anteriorly to a blunt point, forming the posterior and much of the inferior boundary of the orbit. The dorsal border descends very abruptly in a concave curve to make the posterior orbital boundary. The jugal is a large and heavy bone, with great dorso-ventral depth and with its ventral border projecting downward freely, so as to conceal, in side view, the alveolus of m- and, in the young skull, of m- as well. In the Toxodonta there is no such prominent projection of the ventral border. As already mentioned, the jugal is not excavated to receive the anterior end of the zygomatic process, but extends beneath that process almost to the glenoid cavity, where it terminates in a rounded end. The jugal forms some of the inferior and nearly the whole anterior margin of the orbit. Though agreeing in essentials with that of Nesodon , the tympanic has a very different appearance, whether viewed from the side or from below. The inflated bulla is much larger than in that genus, especially in antero- posterior diameter and, in ventral view, forms an elongate, narrow oval. 266 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. The walls of the bulla are thin and the interior is hollow and free from cancellous tissue. The external auditory meatus is far removed from the bulla and yet there is no visible connection between them and it is uncer- tain whether there is any tube other than that formed by the junction of the postglenoid and post-tympanic processes of the squamosal. The meatus itself is a small, irregularly circular opening, without any project- ing edge or lip, which has a far lower position, level with the upper part of the occipital condyle, than in the Santa Cruz Toxodonta or Typotheria. The lachrymal, the limits of which are not easy to make out in either of the skulls, is, to all appearance, a very small triangular bone, without spine, which is exposed at the supero-anterior margin of the orbit; the foramen is not visible from the side. The frontals have a decidedly curious and exceptional form and their appearance changes considerably with age. They are very short and broad and anteriorly are very deeply emarginated to receive the nasals, the emargination extending backward nearly or quite to the line joining the two postorbital processes. Behind the latter the frontals are very much contracted and pass into the notch formed by the divergent anterior ends of the parietals, but these posterior extensions are decidedly shorter than in Nesodon. In the young animal the postorbital processes are quite short and the orbits have hardly any bony roof, but in the fully adult skull the processes are very prominent, though not so long or so much decurved as in Nesodon , and the frontals are extended well out- ward over the orbits, thus greatly changing their appearance. The fore- head is very broad across the postorbital processes, narrowing anteriorly, and moderately convex transversely, showing the low protuberances caused by the frontal sinuses. There are no distinct temporal, or supra- orbital ridges, but in the type of H. segovice a shallow depression (which seems to be too symmetrical to be the result of the down crushing that the skull has undergone) makes the posterior part of each frontal stand out as a broad, flattened ridge. The anterior emargination, which lodges the nasals, is very deep and greatly shortens the frontals in the median line. The sutures with the posterior ends of the nasals form two nearly straight lines, which meet at an angle considerably greater than a right angle, though interrupted, it may be, by a very short, asymmetrical nasal pro- cess from one frontal only. The nasals are altogether different from those of the Santa Cruz Typo- SCOTT: ENTELONYCHIA OF THE SANTA CRUZ BEDS. 267 theria and Toxodonta, and are one of the most characteristic features in the skull of the present genus. They are very short, of moderate width and almost plane both antero-posteriorly and transversely. In shape, each nasal is trapezoidal, the external border being much shorter than the mesial, for at both ends the bone narrows rapidly to a blunt point and thus anterior and posterior ends are much alike in shape. Owing to the depth of the median excavation of the frontals, the nasals have a long lateral suture with these bones, for more than half their length, in fact, while the maxillary suture is short, the anterior ends of the nasals pro- jecting freely. There is no suture with the premaxillaries, from which the nasals are far removed, one of the most obvious differences from the skulls of the contemporary Typotheria and Toxodonta. In correlation with the reduction of the incisors, the premaxillaries are very small and hardly visible when the skull is seen from the side, which is in striking contrast to the arrangement found in the other two suborders, in which these bones have great lateral extension upon the face. They are short, narrow transversely and low, with little or no distinction of horizontal and ascending ramus. The palatine processes are extremely small and contribute almost nothing to the long palate, but the spines are stout and quite elongate, separating the large and conspicuous incisive foramina. The maxillaries differ considerably in their proportions from those of Nesodon and Adinotherium , being relatively longer, on account of the reduction of the premaxillae, and lower, in correlation with the more brachyodont character of the teeth, and there is no such dorsal narrowing of the face. The maxillaries are also produced somewhat farther pos- teriorly over the orbits than in the latter and have quite extensive oblique sutures with the frontals. The infraorbital foramen occupies nearly the same relative position as in Adinotherium , over m- and well in advance of the orbit. The zygomatic process of the maxilla is rather better devel- oped than in Nesodon and extends farther back beneath the jugal. The palatine processes are very large and form much the greater part of the hard palate, but are deeply incised in front by the premaxillae and behind by the palatines. As a whole, the bony palate is much less concave trans- versely than in the Toxodonta, in which the strongly curved teeth, hypso- dont or semi-hypsodont in character, produce very prominent alveolar borders and a profoundly concave palate. 268 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. The palatines extend forward to about the middle of m- as narrow, wedge-shaped plates ; their median contact ceases a little behind m- and is not carried back so far as in Adinotherium and Nesodon , so that the front border of the much larger posterior nares is farther forward than in those genera. The ventral border of that portion of the palatines which forms the side-walls of the posterior nares is thick and rounded, but not nearly so broad as in the Toxodonta, and these walls diverge more rapidly than in the latter and form a much wider opening. The limits of the pterygoids and the descending processes of the alisphenoids cannot be determined in either skull. No complete specimen of the mandible is known to me, the few which are to be found in the collections all lacking the ascending ramus and angle. It is obvious, however, from the position of the glenoid cavity with reference to the level of the upper teeth that the ascending ramus cannot have been so high as in Nesodon and that the condyle was much less elevated above the teeth than in that genus. Enough remains in one mandible to show that the linea obliqua externa is not at all promi- nent and the interna much more so, though far less developed than in Nesodon and enclosing a much shallower fossa behind m-. The hori- zontal ramus is of only moderate dorso-ventral depth, but very thick and heavy ; the ventral border is nearly straight, with but slight sinuosity, at least as far back as 1113. The two rami are firmly coossified in a long, oblique symphysis, which extends to p4 and is very deeply concave on the dorsal side, but the chin is abruptly rounded, rising steeply from the ventral border and with decidedly convex profile, which is very different from the gently inclined and more or less procumbent chin of most of the Toxodonta and Typotheria. As yet, nothing has been found of the hyoid apparatus, nor of its place or mode of attachment to the skull. Little is known concerning the vertebrae of Homalodontotherium. I did not have an opportunity to examine those in the Ameghino and La Plata Museum collections and neither Lydekker nor Ameghino give any descrip- tion of them, except of the axis, which is figured and very briefly described by Lydekker (’93, 45, PI. XX, fig. 5). This axis is of the same type as that of Nesodon and other Toxodonta, but is somewhat longer in propor- tion to its breadth. The centrum is much depressed and very wide anteriorly and the surfaces for the atlas have much the same shape as in SCOTT: ENTELONYCHIA OF THE SANTA CRUZ BEDS. 269 Nesodon; the odontoid process is a short and heavy peg and the neural spine is a hatchet-like plate, similar in shape to that of the genus last named, but longer antero-posteriorly and lower dorso-ventrally. Of the other vertebrse, Ameghino says merely that the centra have flat faces with a small fossette in the middle (’94a, 59). The scapula is not known, but a well-preserved pelvis in the La Plata Museum, referred to this genus, differs much from that of Nesodon , owing to the great expansion and eversion of the ilia, which, with the shortness of the ischia, produce quite a Proboscidean appearance and are points of resemblance to Toxodon. The limb-bones are mostly, so far as they are known, quite elongate and very heavy. They are much larger in proportion to the size of the skull than are those of the contemporary Toxodonta and indicate animals which were massive and slow-moving, but with small heads. The humerus, which has been described and figured by Lydekker [op. cit ., PI. XX, fig. 1) is very peculiar. It is a large bone, measuring in H. cunninghami 16^ inches in length, and very heavy; the head is low and the internal tuberosity is much reduced, while the external one is very large. The shaft is stout, rounded for most of its length, becoming very broad and antero-posteriorly compressed distally. “Its chief char- acteristic is to be found in the enormous development of the deltoid crest, which extends three fourths down the shaft, where it terminates in a bold prominence, standing some three inches above the general level of the surface. ... In the great development of this crest the humerus of Homalodontotherium is approached by that of Phascolomys and the conformation in question suggests fossorial powers” (Lydekker, op. cit., pp. 45, 46). The posterior surface of the distal broadening of the shaft is very flat and smooth and the anconeal fossa must be variable, for Lydekker describes it as small and shallow and Ameghino as “profonde” (’94", 59). The trochlea is broad, very low and simple, saddle-shaped and undivided by any indication of an intercondylar ridge, but the external portion for the head of the radius is quite strongly convex. The internal epicondyle is large and prominent and not perforated by a foramen, while the external one is enormously developed and, with the extremely large supinator ridge, adds greatly to the unusual breadth of the distal end. While this humerus differs strongly in appearance from that of Neso- 270 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. don , there is no material difference of structure ; in Homalodontotherium the great development of the deltoid crest, the epicondyles and the supi- nator ridge is what gives the characteristic form. The fore-arm bones are separate and show no tendency to coossify. Of the radius, only the distal end is known, but, from the form of the capitellum of the humeral trochlea and the proximal end of the ulna, it may be inferred that the head of the radius was discoidal in shape and allowed considerable freedom of rotation. Further, the length of the ulna makes it very obvious that the radius must likewise have been very elongate. The distal end is extremely heavy and of broad, transversely oval shape ; on the outer side is a large, nearly plane facet for articula- tion with the ulna. The scaphoid facet is very broad transversely, narrow antero-posteriorly, and behind its internal moiety is a very large and deep pit completely encircled by bone. The lunar facet, which is very obscurely demarcated from that for the scaphoid, is as broad transversely as the latter and very much more so antero-posteriorly, covering the entire thickness of the distal end ; it is decidedly concave in both directions. The missing shaft may be reconstructed from these indications ; the proximal portion was evidently slender and crossed over the front of the ulna ; the shaft became heavier downward and was actually massive at the lower end. Of the ulna, I have examined a distal epiphysis in the Ameghino col- lection and an entire isolated bone in the Princeton Museum (No. 15,747), the lower end of which agrees closely, save in one unimportant particular, with the Ameghino specimen. I am inclined to believe that there is some mistake in the identification of the very fragmentary ulna which is figured by Lydekker (’93, PI. XX, fig. 4). The complete ulna shows that the fore-arm was extremely elongate and far longer than the leg, the disproportion being much as in Macrotherium , as figured by Filhol (sub Chalicotherium , ’91, PI. XLIII) but this is true only in very much less marked degree of the North American genus Movopus. In its whole character the ulna of Homalodontotherium is closely similar to that of Nesodon, but on a much larger scale and relatively far more elongate. The olecranon is long and is continued upward nearly in the line of the shaft, projecting backward even less than in Nesodon; the posterior border of the process is broadened and flattened in a way not seen in the latter genus, but the proximal end is only moderately thickened and SCOTT: ENTELONYCIIIA OF THE SANTA CRUZ BEDS. 27I rugose. The sigmoid notch displays the same remarkable peculiarity as that of Nesodon, though with some modifications. The coronoid process is less prominent and the facet for the posterior part of the humeral trochlea is much broader and less convex transversely, not being reflected upward so strongly on the outer and inner sides. The anterior humeral surface is entirely internal, as in Nesodon , but is much broader than in that genus and is so extended as to project inward beyond the plane of the shaft in a shelf-like overhang. Thus there is the same strange obliquity of the articular surface, when seen from the front, as in the Toxodonta, inclining downward and inward. The head of the radius is not embraced by the ulna, but the two are merely in contact, as is shown by a narrow articular surface on the outer side of the internal pro- jection for the humerus, above described. The shaft has quite a different shape from that of Nesodon , being less trihedral and more compressed laterally ; it contracts distally but little, so that it is stout and heavy throughout. The interosseous crest is short, only the proximal portion being present and that not very prominent. The styloid process is formed by a sudden contraction of the distal end to hardly more than half its width, so that the facet for the pyramidal and that on the radius for the lunar are widely separated. The pyramidal facet is simply convex and forms an obscure angulation with the large surface for the pisiform. In Ameghino’s specimen there is a deep, irregular pit on the distal end, internal to the styloid process, but this may be merely an abnormality, as may indeed be the pit on the distal end of the radius of the same indi- vidual, as mentioned in the preceding description. The very remarkable and curious manus (PI. XXX, fig. 1) has been quite fully described and figured by Ameghino (’94“, 57, ’94^, figs. 3, 4) who emphasizes its many likenesses to that of the Ancylopoda. These likenesses are, however, less fundamental than the similarities to the toxodont manus, though the fore-foot of Homalodontothermm is at once the more primitive and the more highly specialized. The carpus is inter- locking or “ diplarthrous,” not serial; the bones of the proximal row are relatively lower and those of the distal row higher proximo-distally than in Nesodon. The scaphoid is not known, but evidently it was broad and rested chiefly upon the trapezoid and, in much less degree, upon the magnum ; it appears to have had no contact with the trapezium. The lunar resembles that of Nesodon , but is much wider transversely 272 PATAGONIAN EXPEDITIONS ! PALAEONTOLOGY. and shorter proximo-distally ; the proximal end has a strongly convex facet for the radius, which projects dorsally and overhangs the anterior face of the carpal. The contact with the pyramidal is entirely distal, the two bones being quite widely apart at the proximal end, as might be expected from the corresponding surfaces on the ulna and radius. Dis- tally, the lunar rests upon the magnum and unciform, more extensively upon the former, and the magnum facet differs from that in Nesodon in being more distal and less lateral, in consequence of which the two sur- faces meet at a more obtuse angle. The pyramidal is very much like that of Nesodon in form, except that the interno-distal angle is more pro- duced and tends to extend underneath the lunar ; it has quite an oblique position in the carpus, diverging proximally from the lunar, in accordance with the space between the distal facets of the ulna and radius. The surfaces for the ulna, pisiform and unciform are not particularly characteristic. The trapezium is a small, laterally compressed and scale-like bone, and, though it supports a functional pollex, it would seem to have had no connection with the scaphoid, as is also the case in Nesodon , in which me. I has completely disappeared. The trapezoid is quite different in appearance from that of the contemporary Toxodonta, being much higher proximo-distally in proportion to its width, and is closely interlocked with the magnum. The proximal end has an oblique facet for the sca- phoid and the distal portion of the radial side articulates with the trape- zium. The long axis of the magnum is oblique to that of me. Ill, inclin- ing downward and toward the ulnar side. The bone is much longer and narrower than in Nesodon and contracts proximally, so that the dorsal face has almost the shape of an inverted wedge, the truncation of which forms the narrow facet for the scaphoid. The lunar facet is far larger than in Nesodon and less lateral, more obliquely proximal in position and is quite convex. The magnum is not overlapped by the unciform, the facet for which on the fibular side is vertical and quite narrow. The distal end is also oblique, but sloping in the opposite direction, toward the radial side. The unciform, which is much the largest of the carpal elements, is higher and narrower than in Nesodon , with its principal diameter proximo-distal. The proximal end is unequally divided between the narrow surface for the lunar and the much broader one for the pyramidal ; the latter slopes steeply downward, making the external, SCOTT: ENTELONYCHIA OF THE SANTA CRUZ BEDS. 273 or ulnar, border of the bone very much shorter than the internal, reversing the proportions seen in Nesodon, in which the internal border is a mere angle. The distal end is occupied chiefly by the facet for me. IV, that for the larger and heavier me. V is more lateral than distal. The internal border is nearly straight, the facets for the magnum and me. Ill lying in almost the same vertical plane. The metacarpus is incompletely known, inasmuch as only me. Ill is represented by an entire specimen, but parts of all the others are also preserved. This metacarpus is totally different in character from that of any other known Santa Cruz ungulate: (1) It is pentadactyl, all the digits complete and functional ; (2) the symmetry is not mesaxonic, for, though me. Ill is symmetrical in itself, the two parts of the manus result- ing from the vertical bisection of digit III are by no means equal, the outer part preponderating, which is chiefly owing to the large size of me. V, the heaviest and, very possibly, the longest of the series ; (3) there is very little interlocking of the metacarpals ; (4) the distal trochlea is altogether exceptional. The first metacarpal, represented only by the proximal half, is the most slender of the series. As figured by Ameghino (’94^, 254, fig. 3) the head is placed considerably below that of me. II, with which the trapezium articulates laterally. There is a possibility of error in this arrangement, but unfortunately, my attention was drawn to this question too late for a re-examination of the originals. The second metacarpal, likewise known only from the proximal portion, differs notably from the shape seen in Nesodon and in the three- and four-toed perissodactyls, in having no pro- cess which overlaps the head of me. Ill to articulate with the magnum. This articulation is present, but it is vertical and lateral, the head of me. II rising above that of me. Ill, but not extending over it. The third metacarpal is remarkably long and straight and relatively slender; the proximal end is narrow, but little wider than the shaft, and bears an oblique, saddle-shaped facet for the magnum, which is slightly concave transversely and strongly convex palmo-dorsally. On the radial side of the head is a narrow, curved facet for me. II, which extends over the entire dorso-palmar thickness, and on the ulnar side a very large, vertical and nearly plane surface for the unciform, but this connection is lateral only and there is no process extending over me. IV, such as occurs in most polydactyl ungulates. On the dorsal face near the ulnar border is 274 PATAGONIAN EXPEDITIONS: _ PALAEONTOLOGY. a large tubercle for tendinal attachment. The shaft is long and slender, broadening but slightly to the distal end, and subcylindrical in shape, with nearly equal transverse and dorso-palmar diameters. The distal trochlea is most peculiar ; instead of projecting forward from the dorsal face of the shaft, or being coincident with the latter, it is displaced far toward the palmar side, its proximo-distal axis making an angle of nearly 450 with that of the shaft. Dorsally, the articular surface is at first con- cave, where it is reflected upon the overhanging distal end of the shaft, while the trochlea proper is a low, transverse semicylinder. The carina is quite prominent, but so completely palmar in position as to be invisible in front view. The proximal end of me. IV is somewhat broader than that of me. Ill, which is due to a short, heavy process on the radial side, which carries the facet for articulation with me. III. The shaft, so far as it is preserved, is rather more slender than that of the latter and of more angular, sub- quadrate cross-section. The distal end has the articular surface carried up much higher proximally than that of me. Ill, which gives it quite a different appearance. The fifth metacarpal is the heaviest of them all. The head is made somewhat wider than the shaft by the presence on the radial side of a heavy, but very short projection, which articulates by a large facet with me. IV. The surface for the unciform is very oblique and concavely curved, the head of the metacarpal extending around the outer side of the unciform so as almost to reach the pyramidal. The stout shaft, the length of which cannot be determined, is trihedral in shape, with broad radial side, thinning to a rounded border on the ulnar side. As much of the shaft as has been preserved is quite straight. The distal articular surface is extended upward much more than in me. IV and receives an asymmetrical appearance from the distal production of the internal portion of the trochlea. The phalanges (PI. XXX, fig. 2) are as peculiar as the metacarpals would lead one to expect. Only in digit III are they all preserved, but apparently those of the other digits differ merely in size and relative width. The first phalanx is broad proximally, narrowing much to the distal end. The proximal facet is concave and very oblique, in conse- quence of which the palmar face considerably exceeds the dorsal in length. The distal facet is divided by a median groove into two condyles, but the articular surface is not carried far over upon the dorsal face and indeed is SCOTT: ENTELONYCHIA OF THE SANTA CRUZ BEDS. 275 hardly visible from that side. The second phalanx is much more slender and but little shorter than the first, in marked contrast to Nesoclon, in which this bone is very short and wide. The proximal trochlea is divided by a median ridge into two shallow pits, and the distal one, which is feebly grooved in the median line, is remarkably extended, being reflected far over upon the dorsal side and, when seen in profile, describes nearly two-thirds of a circle. This gives to the phalanx quite the appearance of the corresponding bone in certain Carnivora. The ungual is much modified and has assumed the shape of a claw, somewhat as in Macro- therium and Chalicotherium of the European Miocene and in the North American Moropus , but the modification is less extreme and the departure from the ordinary ungulate type not so great. The phalanx is long, almost equalling the combined first and second in length, very rugose and cleft at the distal end, the cleft continuing as a deep groove of the dorsal surface for nearly half the length of the bone. The proximal portion of the ungual is very thick palmo-dorsally, this dimension exceeding the transverse, but the dorsal face is strongly curved and the thickness diminishes rapidly to the distal end. Proximally, there is a beak-like extension of the median dorsal line, but no conspicuous sub- ungual process. In the other digits the ungual phalanx differs somewhat in size and form. The arrangement made by Ameghino (which is followed in PI. XXX, fig. 1) gives the smallest of the unguals to digit II, but this is more probably referable to the pollex, while the one assigned to the latter, the broadest and heaviest of the entire series, should probably be trans- ferred to the fifth digit. In none of the digits is there visible any ten- dency to coossification of the phalanges, such as constantly occurs between the proximal and second phalanges of digit II in Moropus. The phalanges have an unusual degree of mobility upon the metacarpals and apparently could move through an arc of 180°. This but adds to the difficulty of understanding how the manus was used. The fact that in three continents three different orders of ungulates, the Artiodactyla, Perissodactyla and Entelonychia, should have acquired clawed feet, cer- tainly would seem to indicate a response to some general need, but no one has yet made any plausible suggestion as to what that need was, nor does any existing herbivorous mammal throw any light upon the problem. From the character of all the articulations involved, it is probable that 276 PATAGONIAN EXPEDITIONS: PALAEONTOLOGY. this manus was digitigrade, with the phalanges of each digit extending forward nearly at a right angle with the metacarpal, as shown in Ame- ghino’s Fig. 4, A (’94*, 255) and as Peterson (’07, 746, fig. 26) and Barbour (’08, PI. X) have figured the fore-foot of Moropus. Yet the manus was capable of some rotation upon the arm and the phalanges could be strongly flexed, whatever the use that was made of these movements. Filhol, on the other hand, restores Macrotherium ( Chalicotherium ) as strictly plantigrade both before and behind (’91, PI. XLIII) and Ame- ghino expresses still another view regarding the gait of Homalodonto- therium: “Ces animaux etaient plantigrades parfaits, dans ce sens, que le tarse et la carpe reposaient sur le sol, mais le poids du corps etait porte par la partie externe des pieds d’une maniere aussi accentuee que chez les edentes gravigrades, ou chez les fourmilliers actuels” (’94“, 60). This suggestion does not appear to be a very happy one, for its only support is the somewhat larger size of the external digit in both manus and pes, and all the articulations involved are radically different in the anteaters and the Gravigrada from those in Homalodontotherium. So far at least as the pes is concerned, the foot could not be brought to rest upon its external border without complete dislocation of the ankle. The femur preserved in the La Plata Museum (PI. XXX, figs. 3, 3 a) is of somewhat uncertain reference, for, though the evidence for the iden- tification as presented by Lydekker (’93, 46) seems very convincing, some doubt is thrown upon it by Gaudry’s figures of the leg of Astrcipo- therium (’06, 19, fig. 29). Yet the La Plata specimen is by no means identical with that figured by Gaudry, despite the strong general resem- blance, differing in the following particulars. In A strapotherinni (1) the head of the femur is on an even shorter neck and is directed more internally; (2) the great trochanter is higher and the second trochanter is wanting; (3) the shaft, especially the distal half, is much broader and more flattened ; (4) the third trochanter has a more elevated position, is shorter proximo-distally and decidedly less prominent; (5) the trochlea is extended farther up the shaft. Ameghino’s brief description: “Le femur est un os, court, tres large, plat et presque rectangulaire, resemblant a celui des edentes gravigrades” (’94*, 60) does not seem to fit the La Plata specimen at all and yet Dr. Ameghino pronounced the photographs of the latter which I showed him as properly referable to the present SCOTT: ENTELONYCHIA OF THE SANTA CRUZ BEDS. 277 genus, and, owing to the circumstances of the moment, I was unable to see his own specimens and compare them with the photographs. Assuming, then, that the La Plata femur has been correctly identified, it may be described as follows. The head is large, hemispherical and set upon a very short neck, which is directed obliquely upward and inward. The great trochanter, though massive and rugose, is low, not rising to the level of the head and there is scarcely any notch between them. The proximal third of the shaft is very broad, narrowing to a point below the third trochanter and thence broadening very gradually to the distal end. The shaft is much compressed antero-posteriorly and flattened, as in the very large ungulates generally, and the internal border describes a strongly concave curve, interrupted only by the second trochanter. This is a low, elongate ridge, with convex free border, placed very far down on the inner side of the shaft, so far indeed as to make it somewhat doubtful whether this ridge be actually homologous with the second tro- chanter or not. The third trochanter also has a very distal position and is elongate, but rather low, and, while very conspicuous, is yet not very prominent. The distal end of the femur is very heavy, especially in breadth, which exceeds the antero-posterior dimension inclusive of the trochlea. The latter is wide, deeply grooved, with prominent borders and nearly symmetrical, the internal border projecting but little in advance of the external ; proximally, the trochlea ends abruptly in a nearly straight line, and distally, the articular surface is separate from that of the inner condyle, connected with that of the outer. The condyles, which are sep- arated by a very wide intercondylar space, project but slightly behind the shaft, and differ considerably from each other in form and size. The external condyle is of nearly uniform width, with straight mesial border ; while the internal one is more prominent, both posteriorly and laterally, and more strongly convex and widens backward, having a very concave mesial border. The leg-bones, like those of the fore-arm, are entirely separate and free. The tibia (PI. XXX, figs. 6, 6a) is relatively short and heavy, much shorter than the femur and distantly resembles that of the Gravigrada, with which Ameghino has compared it. The proximal end is broad and massive, projecting far externally over the head ol the fibula, and bears a prominent spine. The shaft is also very heavy near the proximal end, where it is trihedral, and has anteriorly a massive, rugose and T-shaped 278 PATAGONIAN EXPEDITIONS PALAEONTOLOGY. cnemial crest, which, after a short course, suddenly narrows and is con- tinued as a mere linea aspera for nearly two-thirds of the length of the bone. The shaft rapidly narrows and becomes very much compressed laterally ; the middle portion is stout, trihedral in section, with concave posterior face, owing to the elevation of the lateral borders. Distally, it widens very much, attaining nearly the width of the proximal end, but is also greatly compressed antero-posteriorly. The internal border of the shaft, except just below the head, is nearly straight, while the external border is deeply concave, making a very wide interosseous space. At the distal end the inner border is thick and rounded and terminates in a short, but very heavy internal malleolus, on the inner side of which is a massive and rugose prominence ; the outer border, on the other hand, is a thin edge. The astragalar surface is broad transversely, narrow antero-pos- teriorly, especially the external portion ; it is very obscurely divided into a larger and moderately concave surface for the inner condyle of the astragalus, which is produced and slightly reflected upward upon the posterior side of the tibia, and a much narrower and nearly flat surface for the external condyle. On the anterior side there is no intercondylar tongue, but a low and inconspicuous one appears on the posterior side. Connected with this external astragalar surface and demarcated by slight difference of level, is the small, D-shaped fibular facet, which presents distally, the tibia here resting directly upon the fibula. Of the fibula only the distal portion is known. The shaft is very heavy, bowed outward and of subcylindrical shape, while the distal end is very massive both in breadth and antero-posterior thickness. A short, but prominent interosseous crest arises near the bottom of the shaft. Internally, the fibula extends beneath the tibia, which rests upon it and the surface for which faces almost directly upward. On the external side, near the anterior border, is a large and prominent rugosity, which corre- sponds to that on the postero-internal angle of the tibia. The astragalar facet is large, slightly concave in both directions and has a moderate inclination outward and downward. The facet for the calcaneum is very large, slightly concave and very oblique, presenting backward almost as much as distally. In this most peculiar animal no structure is more remarkable than the pes, the manner of using which is highly problematical. The astragalus (PI. XXX, fig. 4), though fundamentally similar to that of the other SCOTT: ENTELONYCHIA OF THE SANTA CRUZ BEDS. 279 Toxodontia and the Litopterna as well, is very characteristic. Aside from the neck, the bone has an almost square outline, with proximo- distal and transverse diameters nearly equal. The trochlea is but feebly grooved, almost flat, and is divided into a much broader internal and a narrower, but elevated and prominent external condyle, somewhat as in Nesodon , but a very exceptional feature is the proximal prolongation of the condyles as narrow tongues, which are separated from each other by a broad V-shaped surface. On the plantar side are two narrow, elongate and parallel surfaces for the calcaneum, separated by a straight, deep sulcus, which is much narrower than in Nesodon. Of these, the external facet is oblique and concave, more directly plantar and of more uniform width than in the last named genus ; the sustentacular facet is slightly convex proximo-distally and is longer, broader and of more uniform width than in Nesodon and, as in that genus, extends to a junction with the navicular facet, but is not produced so far proximally. The neck of the astragalus is decidedly longer and more constricted than in theToxo- donta and ends in a narrower, more convex head, which, as in that sub- order, articulates with the navicular only. The calcaneum (PI. XXX, figs. 5-5$, 7) is likewise highly characteristic and differs in many respects from that of the Santa Cruz Toxodonta. The tuber is much longer proportionately than in that group and is more compressed, with thinner and more prominent dorsal border and the ventral border concave proximo-distally, instead of convexly arched. The tuber broadens to the free or proximal end, which is immensely expanded, thickened, rugose and club-shaped ; the internal portion of the free end is still further elongated as an additional process, the length of which varies in the different individuals and species and is most developed in the very large II. crassum. The fibular facet is peculiar in several respects: (1) it is very large and projects prominently on the external side as an overhanging shelf, which is not to be found in the Toxodonta; (2) it is but moderately convex proximo-distally and very slightly concave transversely ; (3) it presents distally rather than dorsally, a very excep- tional arrangement, which is in correlation with the unusual obliquity of the corresponding surface on the fibula. The external astragalar facet is large and has a more dorsal, less lateral presentation than in Nesodon; another difference from the latter is that the proximal border of this facet is sharply defined and raised above the level of the tuber, not continued 28o PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. over upon it, as is the case in the Santa Cruz Toxodonta. The susten- taculum is prominent and bears a slightly concave facet for the astragalus ; it has a distinctly more distal position than in Nesodon , is much thicker in the dorso-plantar dimension and its articular surface is continuous with the facet for the navicular on the inner side of the distal end of the Fig. 44. Homalodontotherium segovice. Left pes, dorsum, X After Ameghino, except the calcaneumf astragalus and cuboid. calcaneum, while in the latter genus the two are separated. In corre- spondence with the longer neck of the astragalus, the portion of the cal- caneum distal to the fibular facet is longer than in Nesodon. The cuboid SCOTT: ENTELONYCHIA OF THE SANTA CRUZ BEDS. 281 facet is much smaller than in the latter and does not cover the whole of the distal end, massive rugosities projecting beyond it on the external and plantar sides. The navicular is very large, especially in transverse breadth, projecting internally much beyond the astragalar surface, which is rather small and quite deeply concave. On the fibular side, the navicular articulates with both calcaneum and cuboid by distinct facets. Of the cuneiforms only the ectocuneiform is known. This I have not seen, but, if Ameghino’s figure (’94*, 256, fig. 5) is correct, it is a relatively large bone, with nearly square dorsal face; it not only covers the head of mt. Ill, but overlaps it on each side and articulates by narrow facets with mt. II and IV, a feature not found in any other known mammal. The cuboid differs much from the short, nearly square bone of Adinotherium and Nesodon; it is much higher proximo-distally and the calcaneal facet is oblique, sloping steeply downward and outward. On the tibial side are two quite large facets, the proximal one for the navicular and the distal one for the ectocunei- form. Most of the distal end is taken up by the facet for the head of mt. IV and, making an obtuse angle with this, is another very large surface for mt. V, which, however, is more lateral than distal. Like the metacarpus, the metatarsus consists of five members. These show no interlocking or overlapping, but are so arranged that their prox- imal ends describe a regular, concave curve, which rises much higher on the fibular side, while distally the bones diverge like the sticks of a fan. Mt. II, III and IV have a narrow, projecting facet on each side of the head, which is largest between mt. IV and V, and the prominence of these facets makes wide interspaces between the metatarsals. The shafts are grotesquely short, mt. Ill being but little more than one-third as long as the corresponding metacarpal. So far as can be determined from the imperfect material, this pes is isodactyl, mt. II, III and IV being of nearly the same length and thickness, though III is a little more slender than the others. The fifth metatarsal is strikingly different in being much broader and heavier and in having a massive and rugose prominence, which is given off from the fibular side of the proximal end, much as in the Santa Cruz Gravigrada. Its facet for the cuboid is very concave and this meta- tarsal encircles the cuboid much as me. V does the unciform. The phalanges of the pes are entirely unknown. All the articulations, from the femur to the metatarsals, make it prob- 282 PATAGONIAN EXPEDITIONS ! PALAEONTOLOGY. able that the pes was plantigrade, and the peculiar position of the fibulo- calcaneal articulation suggests that only the massive and club-like free end of the calcaneum rested on the ground and that the pes was arched upward somewhat in the manner of the human foot. The improbability of Ameghino’s view, that the pes rested its weight chiefly on the external border, has been pointed out in a previous page (p. 276). Species . — As the material is so scanty and incomplete, it would be premature to attempt any revision of the species, of which Ameghino recognizes four. These are distinguished almost entirely by differences of size, a notoriously untrustworthy criterion, yet so great is the range in this respect that more than one species must certainly be admitted. Much importance has been attached to the development of the cingulum of the upper molars, but this is subject to such fluctuation that it cannot be regarded as even a sexual character, though it is quite possible that it may be constant in some of the species. Homalodontotherium cunninghami Flower. (Plates XXIX, Figs. I, la, 3; XXX, Figs. 3, 3 a, 6, 6a.) Homalodontotherium cunninghami Flower; Phil. Trans., Vol. CLXIV, 1884, p. 173. This, the type-species, is the commonest, or more accurately, the least uncommon of the Santa Cruz forms. It can at present be characterized merely by its large size, less than that of H. crassum , greater than in H. segovice and H. excursum. Measurements. Type. No. X. No. ip, 4 Upper dentition, length C to m-, inch .238 Canine, length (i. e. ant. -post, diam.) . .017 .026 Upper cheek-teeth series, length .223 Upper premolar series, length . . .087 .090 PT length ...... . .020 .020 “ width ...... . .023 .0225 P-, length ...... . .022 .0225 “ width ...... . .0295 •0375 P-, length ...... . .025 .025 .026 “ width ...... . .029 •0375 .029 P-C length ...... . .027 .030 .027 “ width ...... • -035 •043 •035 Upper molar series, length •133 .121 SCOTT : ENTELONYCHIA OF THE SANTA CRUZ BEDS. 283 Type. No. X. No. 15,406. M-, length ..... • -039 •0445 .039 “ width ..... . .042 .049 .041 M-, length ..... •0525 .049 “ width ..... •054 .047 M-, length ..... .056 .047 “ width ..... .051 •045 No. Y. No. 16,015. Lower dentition, length C to nif, inch . .207 .221 Lower canine, length . .022 .018 “ width . .017 .021 Lower cheek-teeth series, length • -195 .1845 Lower premolar series, length . . .089 .085 Pr, length . .020 .0 16 .018 “ width ..... . .017 .019 .016 Pg-, length ..... . .021 .0195 .0205 “ width ..... . .019 .020 .015 P3 , length . .024 .0245 “ width ..... . .019 .023 .017 Pj, length ..... . .027 .026 “ width 7 . .022 .026 Lower molar series, length . .108 •0995 My, length ..... . .034 .028 “ width ..... . .024 .022 M-g-, length ..... . .040 •0325 “ width ..... . .026 .023 Mg, length . .044 .041 “ width ..... . .024 .025 In this table the measurements of the type are taken from Flower’s plate, but there is some inaccuracy involved, for the side and crown views do not exactly agree in size. No. X is a palate in the La Plata Museum and No. Y a mandible in the Ameghino collection ; in both instances the dimensions are from photographs which I made and are therefore not exact, but the error is probably small. No. X is a somewhat older animal than the type-specimen and No. Y is a very old individual with teeth much worn down. The table illustrates the tendency of the grinding teeth to shorten in antero-posterior length with advancing age. The isolated ulna (No. 15,747), described in a preceding page, has the following dimensions : Ulna, length. ..... .585 Ulna, prox. width at radial facet. . .081 “ of olecranon . . . .128 “ distal width .... .0655 “ width of coron. process . . .054 “ “ “ of styloid process . .035 284 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. The femur of somewhat doubtful reference, belonging to the La Plata Museum and described above, yields the following measurements : Femur, length fr. head . . . .536 Femur, width over 3rd trochanter . .128 “ “ “ gr’t trochanter . . .518 “ distal width 12 1 proximal width. . . . .173 “ “ thickness . . . .101 In the Ameghino collection are the distal ends of a tibia and fibula, of which the approximate dimensions, measured from photographs, are here- with given and, for comparison, those of an isolated distal moiety of a tibia (No. 15,435) in the Princeton Museum. No. 15435. Amegh. Tibia, distal width, inch tubercle ....... .099 .088 “ “ “ excl. “ 089 .079 “ thickness, int. border ....... .046 .042 “ “ “ ext. “ 015 Fibula, width of shaft ......... .024 “ distal width, inch tubercle .058 “ “ “ excl. “ .052 “ thickness ......... .055 Localities. — The type of the species was obtained by Dr. R. O. Cun- ningham from the cliffs of the Rio Gallegos. The specimens collected by Messrs. Hatcher and Peterson were found on the Atlantic coast in the neighborhood of Coy Inlet. Ameghino and Lydekker assign no localities for the individuals described by them. Homalodontotherium segoviae Ameghino. (Plates XXVIII, XXIX, Fig. 2; XXX, Figs. 1, 2, 4, 7.) Homalodontotherium segovice Amegh.; Rev. Argent, de Hist. Nat., T. I, 1891, p. 295. This probably valid species differs from H. cunninghami chiefly in the narrowing of the muzzle, the smaller incisors and much reduced first upper premolar (px). The development of the cingulum and the larger canines, upon which stress was laid in the original description, are too fluctuating to have taxonomic significance. In a later publication Ame- ghino mentions merely the difference of size, saying that H. segovice is “one-third smaller than” H. cunninghami (’98, 172). It is difficult to know precisely what is meant by this vague expression. In linear dimen- sions the difference between the types of the two species is only about SCOTT: ENTELONYCHIA OF THE SANTA CRUZ BEDS. 285 one-eighth, which is well within the ordinary limits of individual varia- tion. It is probable, therefore, that there is no substantial difference in size between the two species. In the original description the following dimensions are given : length of skull from incisors to occipital condyles, inch, .400; maximum width, .250; length of upper dental series, .210; breadth of palate at m-, .061. In the imperfect young skull (No. 16,014) which I have, with some doubt, referred to this species, the almost unworn m- measures 45 mm. in antero-posterior, by 40 mm. in transverse diameter. Localities. — For the type, none is given, and the field-label of No. 16,014 is unfortunately lost, but the specimen almost certainly came from the Atlantic coast near Coy Inlet. Homalodontotherium excursum Ameghino. Homalodontotherium excursum Amegh.; Enum. synopt. des especes de Mamm. Foss, des Formations Eocenes de Patagonie, 1894, p. 64. This is the smallest of the known Santa Cruz species, as appears from the comparative measurements which Ameghino gives in the original de- scription and may be thus tabulated : H. segovice. H. excursum. Astragalus, length .......... .084 width .......... .062 Tibia, proximal width 135 .074 •054 .100 Localities. — Not given. Homalodontotherium crassum Ameghino. (Plate XXX, Figs. 5, 5a, 56.) Homalodontotherium crassum Amegh.; Enum. synopt., etc., p. 64. Ameghino has distingushed this species on the ground of its very large size, “taille gigantesque,” though only a few foot-bones have been found. His measurement are as follows : H. segovice. H. crassum. Metacarpal III, distal width ........ .036 .042 Unciform, length .......... .049 .066 “ width 057 .072 thickness .......... .034 .050 286 PATAGONIAN EXPEDITIONS : PALEONTOLOGY. A calcaneum in the Princeton collection (No. 15,213), which I refer to this species, has the following dimensions : Calcaneum, length .... .187 Calcaneum, distal thickness . . . .062 proximal width . . . .076 “ cuboid facet, transv. diam. .034 thickness . . .067 “ “ “ dors. -plant, distal width . . . .046 diam. ...... .049 Localities. — Not given for the type specimen. No. 15,213 was found by Mr. Hatcher on the coast, ten miles south of Coy Inlet. DIOROTHERIUM Ameghino. Diorotherium Amegh.; Rev. Arg. de Hist. Nat., T. I, 1891, p. 296. The original description, in which there is no discrimination of generic and specific characters, may be reproduced in somewhat abbreviated form. “Distinguished from Homalodontotherium by the absence of p-; a diastema between the canine and p- ; canine very large and with strong vertical striations ; true molars with strong external and internal cingulum. Occipital and parietal crests more elevated ; skull of more slender form, considerably narrower and higher than in Homalodontothe- rium; forehead and nasal opening much narrower.” To this was subsequently added the statement that the humerus possessed an entepicondylar foramen. (Amegh., ’94", 67.) Obviously this genus is of very doubtful validity, but it may be allowed to stand pending the discovery of more adequate material. Diorotherium egregium Ameghino. Diorotherium egregium Amegh.; Rev. Argent, de Hist. Nat., T. I, 1891, p. 296. Length of skull from canine to occipital condyles, inclusive, .420 ; height of skull from the masticating surface of m- to dorsal side of nasal, .170. The height and narrowness are partially, at least, due to lateral compression. Localities. — Not given. Order TOXODONTIA. The exigencies of serial publication have made it necessary to defer the account of the order until the descriptions of the three suborders had been completed. As here employed, the term Toxodontia is equivalent to Roth’s Notoungulata, which has found a wide acceptance, but I prefer to use the older term and to regard as suborders its three very distinctly marked subdivisions. It may well be that the number of subordinal groups should be increased for the reception of certain of the more ancient genera and families, such as the Notohippidae, but it would be premature to do so until much more complete material shall have been obtained. While the essential plan of structure is the same throughout the order, each of the three suborders, Toxodonta, Typotheria and Entelonychia, has its well defined and characteristic modifications, which affect all parts of the organism. The Toxodonta are large and moderate-sized animals, becoming extremely massive in the Pampean and immediately preceding formations, but there are quite small animals of this group in the Santa Cruz. In the Deseado stage, when they reached their culmination, the Entelonychia had some gigantic members and in the Santa Cruz most of the species are large animals. The Santa Cruz Typotheria, on the contrary, are all small creatures, some, like Pachyrnkhos, very small, and even in the Pampean they attain to no great size. Dentition. — In the Santa Cruz and subsequent formations the Toxo- donta and Typotheria have a strong tendency to hypsodontism, which in the latter is already fully established in Santa Cruz times, and the upper molars are very strongly curved, almost meeting in the median line of the palate. The Santa Cruz Toxodonta have high-crowned but rooted premolars and even the molars develop roots in old age ; in the subse- quent geological stages the teeth are completely hypsodont, but the Leon- tiniidae, of the older Deseado stage, which in my judgment should be referred to the Toxodonta, have brachyodont teeth. The Entelonychia are all brachyodont, though an increase in the height of the crowns may 287 288 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. be noted between the more ancient genera, like Tliomashuxleya , and the Santa Cruz forms, the latter stage marking the end of the series. In the Toxodonta the canines are greatly reduced and have lost their functional importance, but the second upper and third lower incisors develop into tusks, which grow from permanent pulps. The Typotheria show great variety in the character of the incisors ; in some, e. g. Pvo- typotherimn , all of the anterior teeth are of moderately large size and very much alike in form ; in others, such as Interatherium and the Hegeto- theriidae, there is a marked tendency for the first upper incisor to enlarge greatly and become scalpriform and, in the latter family, the tooth is root- less and grows from a permanent pulp. Coincident with this enlargement of the median incisor there is great reduction and ultimate loss of the other incisors, the canine and one or more of the anterior premolars. In the mandible the median and, in less degree, the second incisor become scalpriform, while the succeeding teeth in a varying number atrophy and disappear. In the larger and later genera, such as Typotherium itself, which has the formula, ii, c§, pf, mf, the scalpriform incisors are very large and the rodent-like appearance of the dentition is so strongly marked that these animals have actually been referred to the Rodentia. Protypotherium displays the exceptional peculiarity of having the lower incisors deeply cleft and fork-like, somewhat resembling those of the Recent Procavia (Hyrax). In the Entelonychia the incisors and canines have undergone less modification from the primitive type than in the other suborders. Aside from the early family of the Notostylopidae, in which the median incisors are scalpriform, the incisors are all retained and have simple, conical or hastate crowns, and the canines are quite large, though not very prominent tusks. The canines are relatively reduced in size in the Santa Cruz genera, Homalodontotherium and Diorotherium , especially in certain individuals which are presumably females. The premolars are almost always smaller and simpler than the molars, though the hinder ones approximate the molar pattern so closely, espe- cially in the Santa Cruz Typotheria, that only in unworn teeth is the dif- ference appreciable. The first premolar, particularly the upper one, is generally very small and not infrequently wanting and in Typotherium the premolars are reduced to f. The pattern of the upper molars is some- what rhinocerotic in character, a feature which is least marked in the Typotheria and is lost in the Pampean Toxodonta. The outer cusps SCOTT: TOXODONTIA OF THE SANTA CRUZ BEDS. 289 have coalesced to form a smooth continuous outer wall or ectoloph, in which the indication of the constituent elements is but feebly shown or entirely absent. The two oblique, transverse crests, proto- and metalophs, are fused with the external wall and the internal cusps, except in some of the more ancient genera of the Entelonychia, in which the inner cusps are described as being more or less separate from the crest, and doubtless this was the primitive condition in all of the suborders. The external wall is continued well behind the posterior crest, enclosing with it a fossa, which varies in size and distinctness. In addition to these principal crests, a variable number of cristse or spurs project inward from the ectoloph, which are most prominent and important in the commoner Santa Cruz Toxo- donta, in which one of these cristas is especially large, persisting even in well worn teeth and giving to the valley its characteristic Y-shape, while in Stenotephanos , which is a rare and imperfectly known genus, m1 and - have a simple undivided valley, a character which recurs in all the molars of the post-Santa-Cruzian genera Xotodon and Toxodontherimn. In Toxodon , on the contrary, this principal crista is very prominent and larger every way than in Nesodon and Adinotherium. The enamel is not continuous around the whole periphery of the crown, but covers the external side and forms vertical bands on the other sides. The Leontini- idae of the Deseado stage have a slit-like, undivided valley, agreeing in this respect with the Entelonychia, in which the spurs are short and are speedily removed by wear. In the Typotheria the valley is very shallow and the crests low, so that even in young adults the grinding surface is a smooth expanse of dentine with an enclosing border of enamel and, in some genera (e. g. Hegetotherium ), a thin coating of cement. On the inner face of the crown, however, the vertical groove which marks the sep- aration of the two internal cusps is persistent. Skull. — The unity of the order is nowhere better displayed than in the skull, especially in the remarkable structure of the auditory region, to which Roth (’03) was the first to call attention. The peculiarity consists in the relatively great size of an element which Roth regards as homologous with the mastoid in man, but not with the element which is so called in other ungulates and which he designates as the pvotuberancia petrosa. The occiput proper is very broad at the base, is sharply constricted above the foramen magnum and then widens again moderately to the lamb- doidal crest. The large area on each side which would be left vacant by 290 PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. the constriction is filled by the element which Roth identifies with the pars mastoidea of the human skull. In the Toxodonta and Entelonychia and in some of Typotheria (e. g. Pachyrukhos ) this element contains a large cavity, which communicates with the cavity of the tympanic bulla by a small canal, but in most of the Santa Cruz members of this suborder is filled with cancellous bone, obviously a secondary condition. Roth gives no convincing reason for regarding this element as homologous with the human mastoid, which is a part of the periotic cartilage, while that so named in the Toxodontia is according to Roth the ossification of a membranous pouch, a difference which is not easy to harmonize. Rambaud and Renault (see Roth, p. 17) have shown that the human squamosal ossifies from three centres, for the squamous portion, the zygomatic process and the auditory portion (pars Serrialis ) respectively, and Roth figures skulls of Toxodon, in which these elements remain sep- arate in the adult stage. He does not make clear his conception of the relation between the pars Serrialis of the squamosal and the pars mas- toidea of the periotic. As a matter of fact, it is the pars Serrialis , or post- tympanic portion of the squamosal which is inflated, as clearly appears in those Typotheres in which the two are separate. Sinclair has figured (this volume, p. 71, fig. 14) a skull of Hegetotherium (A. M. N. H., No. 9223) in which the element in question is divided by a transverse suture (fig. 14, mp.) into dorsal and ventral portions ; the former is the inflated pars Serrialis of the squamosal and the latter a thin, compressed plate ankylosed with the tympanic, which is the homologue of the similar plate in the skull of Nesodon which is called the mastoid or protuberancia pe- trosa. Attention should be called to the fact that my interpretation of these parts differs entirely from that of Sinclair, who has followed Roth. The occiput of the pig (Text-fig. 45) presents at least a very similar ap- pearance to that of the Toxodonta. Here the occiput proper, as in Neso- don, is wide at the base and gives off very long paroccipital processes. Above the foramen magnum is a very sudden and sharp constriction, the two exoccipitals being very much broader than the ventral portion of the supraoccipital. The posterior surface of the skull is completed on each side by a large area of the squamosal (pars Serrialis ) quite as in the Toxodonta, but with the notable difference that there is no cavity in this region. In the very young pig-skull the supra- and exoccipitals are rela- tively larger and the part taken by the squamosals in the formation of the SCOTT: TOXODONTIA OF THE SANTA CRUZ BEDS. 291 posterior surface is much smaller, though very distinct, and increases with advancing age to the fully adult condition. The position of the external auditory meatus is much the same as in Nesodon, but the mastoid is ex- cluded from the surface of the skull and the postglenoid and post-tym- panic processes of the squamosal are fused together. Even if Roth’s identification of the mastoid in the Toxodonta should prove to be correct, it would not indicate any radical difference from the Fig. 45. Sus scrofa. Occiput, X S.oc., supraoccipital ; Exoc., exoccipital; P.oc., paroccipital ; M.a.e., exter- nal auditory meatus; Sq., post-tympanic portion of squamosal; Zy., zygomatic process. Fig. 46. Procavia capensis. Occiput, X f. S.oc., supra- occipital; Exoc., exoccipital; P.oc., paroccipital process; P.g., post-glenoid process; Mas., mastoid portion of periotic. Hyracoidea (see Text-fig. 46) in which the posterior surface of the skull is completed by the large, inflated mastoids, the cavities of which are filled with cancelli. The mastoids are ankylosed with the tympanies, but usually not with the squamosals, though in some species the latter fusion also occurs. Roth’s view of a near relationship between the “ Notoungulata” and the Primates, and radical distinction from all other hoofed animals, involves such a degree of convergence as staggers belief. It further involves, as 292 PATAGONIAN EXPEDITIONS ’ PALAEONTOLOGY. he has pointed out, the polyphyletic origin of the mammals from reptiles and is thus made extremely improbable from the standpoint of present knowledge. So far as I am aware, his theoretical conceptions have not found any support even among those who accept his homologies of the toxodont skull. Each of the three suborders has its peculiarities in the structure of the auditory region. In the Santa Cruz Typotheria the tympanic bulla has a long, tubular, auditory meatus, the opening of which has a relatively low position and is directed backward rather than outward. The pars Serrialis of the squamosal and even the root of the zygomatic process have a very inflated appearance externally and are either hollow, as in Pachyvuklios , or, much more commonly, filled with cancelli. The mastoid (post- tympanicus of Roth) has no process, the distal end being closely applied to the paroccipital ; it is ankylosed with the tympanic and usually with the squamosal, in which case it appears to form the post-tympanic pro- cess. The hyoid arch is loosely attached to the skull and is inserted in a deep pit on the outer side of the bulla, external to the paroccipital process. The comparison of the Typotheria and Toxodonta, as regards the audi- tory region, is by no means clear. What would seem to represent the mastoid (the protuberancia petrosa ) of other ungulates is conspicuous in the Toxodonta as a thin plate which ends in a well defined process ; it is ankylosed with the bulla and may, in fact, be an outgrowth of the tympanic. If it is really the mastoid, it occupies a highly exceptional position, being separated from the exoccipital by another element, which appears to be the post-tympanic process (or pars Serrialis ) of the squa- mosal, and is almost invariably continuous with the squamous portion, though I have detected traces of a sutural connection with the latter. There is no visible tube leading to the external ear-opening, the passage being entirely concealed by the structures just described. The opening has a very elevated position at the postero-external angle of the zygo- matic arch, much as in the pig and, indeed, the whole appearance of the occipital surface is suggestively like that seen in the latter animal. The tympanic bulla is, as in both of the other suborders, large, completely ossified and hollow, free from cancellous bone ; in shape it is mammillate, with the principal diameter dorso-ventral. A most exceptional character, at least in the Santa Cruz genera of the suborder, is the position of the scott: TOXODONTIA OF THE SANTA CRUZ BEDS. 293 hyoid arch, which is ankylosed with the anterior end of the bulla in the adult skull. While very few of the indigenous South American ungulates were horned, all the known genera which display any indications of horns are referable to the Toxodonta. In the Pyrotherium Beds the Leontiniidae, which are usually assigned to the Entelonychia, seem to have had a nasal horn, and in the Santa Cruz one species of Nesodon (N cornutus ) and nearly or quite all the species of Adinotherium apparently had a very small frontal horn, which attains large proportions in Trigodon of the Monte Hermoso horizon. In the Entelonychia the skull is of the same type as in the Toxodonta and Typotheria, with several peculiar modifications. Except in the very ancient family of the Notostylopidae, the skull has less likeness to that of the rodents than in either of the other suborders and in the homalodon- totheres, which are the most specialized family of the Entelonychia, the skull has become very characteristic. The questionable element ( pars Serrialis), which I have regarded as belonging to the squamosal, is less extensively exposed on the occipital surface, and the mastoid portion of the periotic is not visible externally, so that there is no mastoid process. The external auditory meatus, which is an irregular hole, has a more inferior position than in either of the other groups, while the tympanic bulla is very large, with its principal diameter in the antero-posterior direction. The nasals are very short and the anterior nares present ob- liquely upward and forward ; the premaxillae are very small, far smaller than in the Toxodonta, though they are quite large in the more ancient and primitive genera, such as Thoniashuxleya and Asmodeus. The muzzle and chin are short and abruptly rounded. Vertebral column. — This varies considerably, as would of course be ex- pected from the very different stature and bulk of the three suborders, but there is, nevertheless, a general uniformity of structure. In all the neck is short, or of only moderate length, and the axis always has the conical, peg-like odontoid and large, hatchet-shaped neural spine. The trunk is long, with 20-22 vertebrae in the Toxodonta and Typotheria and in the larger genera of the former, such as Nesodon and Toxodon , the neural spines are very elongate in the anterior thoracic region and form a decided hump at the shoulders. The tail is stout and of moderate length in the Toxodonta, very long in the Santa Cruz Typotheria, except in 294 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. Pachyrnkhos. Very little is known concerning the vertebral column of the Entelonychia. The Limb-girdles display characteristic modifications in each of the three suborders. In some, if not all, of the Typotheria there is a clavicle and the scapula has a prominent acromion and a single large meta- cromion, given off from the spine near the distal end. The pelvis is unguiculate rather than ungulate in form, retaining the narrow, more or less trihedral ilium. There is very considerable change in the limb-girdles between the Santa Cruz and the later genera of the Toxodonta ; in the former the scapula has a well-defined acromion and two very prominent metacromia, one arising near the middle of the spine and the other near the distal end, while in the Pampean Toxodon there is no acromion or distal metacromion and the proximal one is so reduced as to be hardly more than a thickening of the spine. No trace of the clavicle has yet been found in any member of the suborder and it is highly improbable that the bone was present in the post-Miocene genera, but the distinct acromion in Nesodon and Adinotherium indicates the possible presence of the clavicle in more or less reduced and vestigial form. In the Santa Cruz genera the pelvis is not greatly modified, the iliac plate being but moderately expanded and very little everted, but in Toxodon the ilium is so broadened and flared outward as to give the pelvis quite a probos- cidean appearance, and the same is true, in less pronounced degree, of Homalodontotherium ; otherwise, nothing is known of the limb-girdles of the Entelonychia. The Limb-bones vary much, though chiefly in their proportions, accord- ing to the size and weight of the animal. The fibula articulates largely with the calcaneum in all members of the order, except a few of the Typo- theria, in which this articulation has been secondarily lost. In this group, which is made up almost entirely of small animals, the limb-bones are slender and weak and the processes for muscular attachment are mostly low and inconspicuous. The ulna and radius are always separate and the tibia and fibula are sometimes so, but in most of the genera they are coossified at both proximal and distal ends. The femur has the third trochanter, the humerus has an epicondylar foramen. The Santa Cruz Toxodonta have limb-bones very like those of the Typotheria except for their larger size and greater relative stoutness ; the epicondylar foramen is lost ; the tibia and fibula are ankylosed at the proximal end only, not SCOTT : TOXODONTIA OF THE SANTA CRUZ BEDS. 295 at the distal, and the fore and hind limbs are of approximately equal length. In Toxodon, probably also in its contemporary genera, whose limb-bones have not yet been found, the hind-limbs are very much longer than the fore, a disproportion due chiefly to the great elongation of the femur, which has lost the third trochanter and in other ways has acquired a resemblance to the femur of the Proboscidea. The other limb- bones are extremely heavy and relatively short. In the Entelonychia the limb-bones, though of the same general type as in the other suborders, are much modified in accordance with the very curiously specialized feet. The humerus has remarkably prominent del- toid and supinator crests and epicondyles and in one genus ( Diorothe - riuni) there is an epicondylar foramen. The fore-arm is much elongated and the ulna is, for most of its length, heavier than the radius, which probably had considerable freedom of rotation. The femur is also long, and antero-posteriorly flattened and has a reduced third trochanter. The bones of the lower leg are relatively short, so that the fore-limb is longer than the hind, somewhat as in the European Macrotheriwn. The tibia and fibula are separate and the former has a distal end of peculiar shape, being broad transversely and much compressed antero-posteriorly. The fibula is very stout, especially the distal end, which forms a massive external malleolus. The Feet differ more among the three suborders than does any other part of the skeleton, though the unity of plan is made evident by the uni- formity of the carpus and tarsus. Throughout the order, at least in all the genera of which the feet are known, the carpus is arranged in alter- nating series and there is no free central, that element being probably ankylosed with the scaphoid. Otherwise, there is no coossification among the carpals and the trapezium is always present, except probably in the Pampean toxodonts. In the tarsus the astragalus is grooved in varying degree, but never deeply, and always has a rounded, convex head, which rests only on the navicular and is widely removed from the short cuboid. In the known toxodonts the meso- and entocuneiforms are coossified, but in the other suborders all of the tarsals are free and none is lost. Except in a few of the Typotheria, the calcaneum has a large and prominent facet for articulation with the fibula. While the fundamental character of carpus and tarsus is thus uniform, there are many varieties in details; thus, in the Toxodonta the astragalus 296 PATAGONIAN EXPEDITIONS: PALAEONTOLOGY. has a very short neck and a short trochlea, which is but little or hardly at all grooved, and the calcaneum has a short, heavy tuber. In the Entel- onychia the astragalus has a very elongate trochlea, with parallel sides, a long neck and very convex head, and the tuber calcis is long, laterally com- pressed and with much thickened free end. The Typotheria have a more deeply grooved astragalus, which is quite different in shape from that of either of the other suborders and the same is true of the calcaneum, as is made clear by a comparison of the plates in this volume. It is in the metapodials and phalanges that the most striking differences between the three groups of the order are to be noted. All of the known Toxodonta are tridactyl and the feet have a mesaxonic symmetry, which, however, is not perfect, and a vestigial fifth metacarpal is retained in all of the genera, even the latest, but no trace remains of the fifth digit in the pes. The phalanges are short and heavy and the unguals are broad and hoof-like, changing their proportions with the bulk of the animal. In general appearance, the feet of the Santa Cruz genera bear a decided resemblance to those of Eocene perissodactyls, but they are remarkably small and weak in proportion to the size of the skeleton. In the Typo- theria the number of digits is more variable ; the formula is usually IV— IV, but in one series, including the Pampean Typotherium , the pollex is retained. The symmetry of the digits is not very strongly marked and may be either mesaxonic or paraxonic. The phalanges are slender and elongate, the unguals narrow and pointed, nail-like, rather than hoof-like, except in the comparatively large Typotherium , in which they are short and wide. The Entelonychia have the most curious and aberrant type of feet, which in some measure resemble those of the Ancylopoda of the northern hemi- sphere, though it should be noted that these parts are not fully known in any of the genera and only approximately so in a single genus, Homalo- dontotherium. There are five fully developed and functional metacarpals, which are very long and slender ; me. V is much the stoutest and per- haps the longest of the series ; the distal trochleae are very unusual in form and indicate a remarkable degree of mobility on the part of the phalanges. The unguals are large heavy claws, cleft at the free ends. The pes is likewise pentadactyl and isodactyl in symmetry, and the meta- tarsals are extremely short, hardly more than one-third as long as the metacarpals ; mt. V is the heaviest and from the fibular border, near the proximal end, is given off a large process, much like that seen in the Santa Cruz Gravigrada. The phalanges of the pes are unknown. SCOTT: TOXODONTIA OF THE SANTA CRUZ BEDS. 297 The Systematic Position of the Toxodontia. This difficult problem can be dealt with here only in a brief and tenta- tive manner, reserving the full discussion to a subsequent volume, after the description of the Astrapotheria shall have appeared, for the problem of the Toxodontia is that of all the indigenous groups of South American ungulates. It is not surprising that different observers should have reached very divergent conclusions, for the difficulty of giving the true valuation to the long lists of likenesses and unlikenesses among the various groups is exceedingly great. Indeed, a definitive solution of the problem is hardly to be hoped for until much more is learned regarding the skeletal struc- ture of the pre-Santa Cruzian genera, especially those of the Notostylops Beds. In his later writings, Ameghino attempted to bring nearly all the pecu- liarly South American groups of hoofed animals into relation with the orders of the northern hemisphere ; the Litopterna were united with the Perissodactyla, the Astrapotheria with the Amblypoda, the Pyrotheria with the Proboscidea and the Entelonychia with the Ancylopoda, except the Notostylopidae, which are referred to the Tillodonta. The Notohippidae are removed from the toxodonts and united with the horses to form a new order, the “ Hippoidea.” The Toxodontia are retained as a distinct order and no very definite relationship is suggested. An almost diamet- rically opposite view is that of Lydekker (’96) who considers that all the South American ungulates, except perhaps the Pyrotheria, are more or less closely related and derived from a common stock. In particular, the astrapotheres and homalodontotheres are regarded as nearly related and are included in the same order. Roth, in the remarkable paper so often quoted (’03), refers the Litopterna to the perissodactyls and apparently, though not explicitly, is inclined to make the same reference of the Astra- potheria. The toxodonts, in the narrow sense, the typotheres and the homalodontotheres are included in a new order, the Notoungulata, which Roth believes to be widely separated from all other hoofed animals and related to the Primates. Gregory (To) extends Roth’s term to include all the indigenous South American forms, even the Pyrotheria, and raises the Notoungulata to a super-order. Schlosser’s (Ti) classification of the hoofed mammals is an unusual one, admitting only three orders and rele- gating the other groups to subordinal rank. The three orders are I Ungu- 298 PATAGONIAN EXPEDITIONS! PAL/EONTOLOGY. lata , with the suborders, 1 Condylarthra, 2 Litopterna, 3 Perissodactyla, 4 Artiodactyla, 5 Amblypoda: II Notoungulata , with the suborders, 1 Typo- theria, 2 Toxodontia, 3 Entelonychia, 4 Astrapotherioidea, 5 Pyrotheria: III Subungulata , with the suborders, 1 Embrithopoda, 2 Hyracoidea, 3 Proboscidea, 4 Sirenia. Except for the position assigned to the Litop- terna, this scheme, so far as it deals with South American forms, is sub- stantially the same as that of Gregory and Lydekker, with whose general views I am disposed to agree. Leaving aside for the present the imperfectly known Pyrotheria and Astrapotheria, the indigenous South American groups of hoofed animals are, in my judgment, all more closely interrelated than they are to any of the northern orders, which is substantially the conclusion reached by Lydekker. That the toxodonts, typotheres and homalodontotheres together form a natural assemblage of allied forms is a matter of practically uni- versal agreement ; it is the position of the Litopterna that gives rise to the greatest difference of opinion. My own views on this question and the reasons for them are fully explained in Vol. VII of these Reports. The four groups, whatever rank be assigned to them, agree in the fol- lowing significant particulars : (1) Though some of the incisors may be enlarged and tusk-like, there is a gradual transition in form from the inci- sors to the grinding teeth, except in the Proterotheriidae. (2) The premo- lars are different from the molars. (3) The grinding teeth are lophodont. (4) In the upper molars the posterior crest is short and more or less imper- fectly developed. (5) The lower molars are bicrescentic and, except in the Litopterna, the anterior crescent is much smaller than the posterior. (6) A pillar, or spur, is present in the inner valley of the posterior crescent this has been secondarily lost in a few of the proterotheres. (7) rarely has a talon and it is never large. (8) The odontoid process of the axis is always conical, even in the long-necked forms. (9) The scapula frequently has two prominent metacromia. (10) The ulna and radius are always separate, except in the later Macrauchenidae. (11) Save in the latest toxodonts, the femur retains the third trochanter. (12) The carpal bones are arranged in alternating series and there is no free central. (13) The number of digits in manus and pes ranges from five to one and the symmetry is usually mesaxonic ; digital reduction is “ inadaptive” in char- acter. (14) The astragalus has a narrow trochlea and convex head, which articulates only with the navicular and is far removed from the short SCOTT: TOXODONTIA OF THE SANTA CRUZ BEDS. 299 cuboid. (15) There is an extensive articulation between the calcaneum and the fibula, except in a few typotheres which have suppressed this connection. While some of these resemblances are trivial and without taxonomic importance, it is difficult to believe that so many correspondences are due to a purely convergent mode of development and imply no community of origin. Indeed, there is but one very important structural difference between the Toxodontia (Notoungulata) and the Litopterna and that is the auditory region of the skull. Equally striking differences in the skull of the Rodentia are not held to justify the breaking up of that order. BIBLIOGRAPHY. Ameghino, F. 18876. Enumeracion sistematica de las Especies de Mamiferos Fdsiles coleccionados en los Terrenos Eocenos de la Patagonia Austral.— Buenos Aires, 1887. 18896. Contribucion al Conocimento de los Mamiferos Fosiles de la Republica Argentina. — Buenos Aires, 1889. 1891*. Observaciones criticas sobre los mamiferos eocenos de la Patagonia Austral. — Rev. Argent, de Historia Natural, T. I., 1891, p. 328. 1894“. Enumeration synoptique des Espdces de Mammifeires Fossiles des Formations Eocenes de Patagonie. — Boletin de la Acad. Nac. de Ciencias, T. XIII, p. 259. — In the citations for this paper given in the text of this volume, the paging is that of the separate edition. 18946. Sur les Ongules fossiles de l’Argentine. — Revista del Jardin Zodlogico de Buenos Ayres, T. II, 1894, p. 194. 1895. Premiere Contribution & la Connaissance de la Faune Mammalogique des Couches a Pyrotherium. — Boletin del Instituto Geogrdfico Argentino, T. XV, 1895. 1897. Mammiferes cr6tac£s de l’Argentine. — Ibid., Vol. XVIII, 1897. 1898. Segundo Censo de la Reptiblica Argentina, T. I, Territorio.— Buenos Aires, 1898. 1902. Premiere Contribution & la Connaissance de la Faune Mammalogique des Couches cl Colpodon. — Boletin de la Academia Nacional de Ciencias de Cordoba, T. XVII, 1902, p. 71. 1904. Recherches de Morphologie phylog6netique sur les molaires supdrieures des Ongules. — Anales d. Museo Nac. de Buenos Aires, ser. 3, T. Ill, 1904, p. 1. 1907. Les Toxodontes k Cornes. — Ibid., T. IX, 1907, p. 49. Barbour, E. H. 1908. Skeletal Parts of Moropus. — Nebraska Geological Survey, Vol. 3, pt. 3. Filhol, H. 1891. Etudes sur les Mammiferes Fossiles de Sansan. — Paris, 1891. Flower, W. H. 1884. On a Newly Discovered Extinct Ungulate Mammal from Patagonia. Phil. Trans., Vol. 164, 1884, p. 173. 1885. Osteology of the Mammalia, 3rd Ed. — London and New York, 1885. 300 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. Gaudry, A. 1906. Fossiles de Patagonie. — Attitudes de quelques Animaux.— Annales de Pal£ontologie, T. I, p. 1. Gregory, W. K. 1910. The Orders of Mammals. — Bull. Amer. Mus. Nat. Hist., Vol. XXVII, p. 1. Lydekker, R. 1886. Catalogue of the Fossil Mammalia in the British Museum, pt. Ill, 1886. 1893. A Study of the Extinct Ungulates of Argentina. — Anales d. Mus. de La Plata, T. II, 1893. 1896. A Geographical History of Mammals. — Cambridge, 1896. Osborn, H. F. 1910. The Age of Mammals. — New York, 1910. Palmer, T. S. 1904. Index Generum Mammalium. — Washington, 1904. Peterson, O. A. 1907. Notes on some American Chalicotheres. — American Naturalist, Vol. XLI, 1907, p. 733. Roth, S. 1903. Los Ungulados Sudamericanos. — Anales d. Mus. de La Plata, T. V, 1903, p. 5. Schlosser, M. 1911. Mammalia, in Zittel’s Grundztige der Palaeontologie, 2te AufL, II Abth., p. 325. — Munich and Berlin, 1911. Scott, W. B. 1894. The Mammalia of the Deep River Beds. — Trans. Amer. Philosoph. Society, Vol. XVII, p. 55. 1910. The Litopterna of the Santa Cruz Beds.— These Reports, Vol. VII, p. 1. Sinclair, W. J. 1909. The Typotheria of the Santa Cruz Beds. — These Reports, Vol. VI, p. 1. PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. EXPLANATION OF PLATE XII. PAGE Fig. i. Adinotherium ovinum : Restoration of the skeleton, X abt. A: (7:38). Mostly from one individual, No. 15,131. 21 7 Fig. 2. Nesodon imbricatus : Restoration of the skeleton, X abt. -§■ (1:6.2). 176 (vol. vr.) PATAGONIi Bruce Horsfall Patagonian Expeditions Vol.vi. Plate xii. Werner 4 Winter. Frankforr°M, lith, ' - [Si 11 Ml l\ ill Nesodon PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. EXPLANATION OF PLATE XIII. PAGE Nesodon imbricatus : Skull (No. 15,000) and mandible (No. 15,260) from the right side, X T M.a.e. = External auditory meatus. Pty. — Post-tympanic process of the squamosal. Mp. — Mastoid process. . . . . . 132 (vol. VI.) Wm*"- VimSk -■> 'M . «,A S5*p^ . pjMr BBJfe. A ;'■ Patagonian Expeditions Vol.vi. Plate xiii Bruce Horsfall del. werners wmrer, rrarmrorr /m. Nesod on PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. EXPLANATION OF PLATE XIV. PAGE Nesodon imbricatus : Skull (No. 15,000) from above, X i- M.a.e. = Ex- ternal auditory meatus. . . . . . 133 (vol. VI.) Patagonian Expeditions Vol.vi Plate xrv. M.a.e. Bruce Horsfall del. Werner & Winter, Frankforl °/W Nesodon PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. EXPLANATION OF PLATE XV. PAGE Nesodon imbricatus : Skull (No. 15,000) from below, X i M.a.e. = Ex- ternal auditory meatus. P.oc. = Paroccipital pro- cess. Pty. — Post-tympanic process of squamosal. St.h. = Fragment of stylohyal attached to anterior border of tympanic bulla. . . . . . 134 (VOL. VI.) Patagonian Expeditions Vol.vi. Plate xv Bruce Horsfall del. Werner & Winter FrankforlAM . Poc. Pt\: M.ae. Nesod on. V PATAGONIAN EXPEDITIONS ! PALEONTOLOGY. Figs. I- Fig. 6. EXPLANATION OF PLATE XVI. PAGE . Nesodon imbricatus : Stages in development of the upper in- cisors, from specimens in the Ameghino collection. Anterior view. . . . 119 “ “ Upper incisors of an animal past maturity. Type of JV. marmoratus, La Plata Museum. . . . . . . 138 All figures natural size. (VOL. VI.) Patagonian Expeditions Vol.vi Plate xvi Bruce Horsfall del Werner <& Winter; Frankforf°/M. N E S O D ON. PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. EXPLANATION OF PLATE XVII. Fig. i. Nesodon imbricatus : Upper milk-dentition, crown view, X 1 (No. 15,354). M- = First molar. . Fig. i. “ “ Upper permanent dentition, crown view, X 2" (No. 15,135). P- — Fourth premolar. . Fig. 3. “ “ Upper permanent dentition, crown view, X i (No. 15,000). Fig. 4. Adinotherium ovinum : Upper dentition, crown view, X t (No. 15,493). P- = Third premolar. Dp- — Fourth milk-premolar. Fig. 5. “ “ Upper permanent dentition, crown view, X t (No. 15,118) Fig. 6. Nesodon sp. : Lower milk-incisors from above, X t- Ameghino collection. ........ Fig. 7. Adinotherium ovinum : Permanent upper incisors beginning to erupt, anterior view, X t (No. 15,114). I- — Median incisor. Di- = Second milk-incisor. . . . . . Fig. 8. “ “ Upper milk-dentition, crown view, X t (No. 15,945). Dpk =■ Fourth milk- premolar. M- — First molar. . Fig. 9. “ “ Upper milk-dentition, left side, external view, X t (No. 1 5,945). Dp- = Fourth milk-premolar. . Fig. 10. “ “ Lower incisors, ventral view, X t (No. 1 5,539), showing vestiges of 3d milk- incisors. ...... PAGE 126 I 19 201 127 201 202 201 (VOL. VI.) Patagonian Expeditions Vol.vi Plate xvn Bruce Horsfall del Werner & Winter, Frankfort °/M Adinotherium, Nesodon. PATAGONIAN EXPEDITIONS ! PALAEONTOLOGY. Fig. i. Fig. 2. Fig- 3- Fig. 4. Fig. 5- Fig. 6. Fig. 7. EXPLANATION OF PLATE XVIII. Nesodon imbricatus : Skull, occipital view, X t (A. M. N. H., No. 9234). Pty. = Post-tympanic portion of squamosal. Exoc. — Exoccipital. Mp. — Mastoid process. P.oc. — Paroccipital pro- cess. M.a.e. — External auditory meatus. Adinotherium ovinum : Lower milk-dentition, crown view, X t (A. M. N. H., No. 9517). Dp 4 = Fourth milk-premolar. MT — First molar. “ “ Upper dentition, crown view, X t (No. 15,235). First permanent incisor, sec- ond and third milk-incisors. Dp- — Fourth milk-premolar. .... Nesodon imbricatus : Lower permanent dentition, crown view, X i (No. 15,215) Adinotherium ovinum : Lower permanent dentition, crown view, Xi (No. 131) Nesodon imbricatus : Lower milk-dentition, crown view, X t (No. 15.354) “ “ Lower dentition, crown view, X 1 (No. 15,- 115). Permanent incisors erupting. Dp ± — Fourth lower milk-premolar. = First molar. ...... PAGE 133 202 201 124 201 127 1 24 (VOL. VI.) Patagonian Expeditions Vol.vi. Plate xviii. Bruce Horsfall del Werner & Winter. FrankforlAM Adinotherium, Nesodon. PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. Fig. I. Fig. 2. Fig- 3- Fig. 4. Fig. 5- Fig. 6. Fig. 7. Fig. 8. EXPLANATION OF PLATE XIX. PAGE Nesodon imbricatus : Left lower milk-dentition, inner side, X t (No. 15,400) 127 “ “ The same, crown view, X t- “ “ Right lower milk-dentition, outer side, X t (No. 15,354) “ “ Fourth lower premolar, inner side, X t (No. I5.II5)- • 125 “ “ The same, crown view, Xi- “ “ Young mandible, from above. Milk-incisors erupting, roots of prelacteal teeth retained. DiT = First milk-incisor. PliT = Root of first prelacteal incisor. Amegh. coll. . 129 “ “ Right ramus mandibuli, showing the fourth premolar (7^4) displacing its predecessor {Dp-z), X l (No. 15,115). . . . 129 “ “ Upper milk-dentition of right side, external view, X 1 (No. 15,354) 126 (vol. VI.) Patagonian Expeditions Vol vi Plate xrx Bruce Horsfall del. Werner & Winter, FrankforlAM. Nesod on PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. EXPLANATION OF PLATE XX. Fig. i. Adinotherium ovinum : Skull (No. 15,382) and mandible (No. 15,- PAGE I3I)> Xi 202 Fig. 2. U (( Skull, from below, X i (No. 15,382). 203 Fig- 3- i( << The same, from above, X i- Fig. 4. << (( Forehead, showing rugosities, X x (No. 15-382) . , 203 Fig. 5- Nesodon imbricatus : Right third upper molar, from inner side, showing the open pulp-cavity, X 2 1 masti- cating surface upward (No. 15,526). 122 (VOL. VI.) Patagonian Expeditions Yol.vt Plate xx. Bruce Horsfall del. Werner & Winter, FranUfort0/M Adinotherium, Nesodon. PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. EXPLANATION OF PLATE XXI. PAGE Fig. i. Nesodon imbricatus : Premaxillae from the front, X t- Ameghino collection. Di- — First milk-incisor. 126 Fig. 2. “ “ Symphysis mandibuli, ventral view, X x- Amegh. coll. Di — Second milk-incisor. 1 27 Fig. 3- Adinotherium ovinum : Left pes, dorsum, X x (Nos. 15,966 and 15,131). Cn. 1 -j- 2 = Ankylosed ento- and mesocuneiforms. Cb. = Cuboid. ...... 215 Pig- 4- “ “ Right manus, dorsum, Xi (Nos. 15,158 and 15,131). Sc. = Scaphoid. Td. — Trapezoid. P. = Pisiform. V. = Ves- tigial fifth metacarpal. 2 1 1 Fig. 5- “ “ Left calcaneum, from tibial side, X t (No. 15,131). . . ... 216 Fig. 6. “ “ Left astragalus, plantar side, X t (No. 15,130 215 Fig. 7- “ “ Left carpus, proximal end, X t (Nos. 15,158 and 15,131). Sc. = Scaphoid. P. — Pisiform. ..... 2 1 1 Fig. 8. “ “ Metatarsals of left pes, proximal ends, X t (No. 15,966) 216 Fig. 9- “ “ Metatarsals III and IV, right pes, side view, X t (No. 15,966). 2 1 7 Fig. IO. “ “ Metatarsal IV, right pes, distal end, X x (No. 15,966) 217 Fig. ii. Nesodon imbricatus : Anterior thoracic vertebra, from the left side, X i (No. 15,967) 147 (VOL. VI.) Patagonian Expeditions Volga Plate xxl Cn. 1+2- V A M Bruce Horsfall del. Werner & Winter. Frankfortf-M Adinotherium, Nesodon PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. EXPLANATION OF PLATE XXII. Fig. i. Nesodon imbricatus : Left scapula, outer side, X i (No. 15,489). . PAGE 150 Fig. 2. < < “ The same, distal end. ..... Fig- 3- << “ Presternum, left side, X \ (No. 15,968). 150 Fig. 4. i < “ The same, ventral side. .... Fig. 5- < < “ Left manus, dorsum, X 2 (No. 15,460). Sc. = Scaphoid. Tm. = Trapezium (placed somewhat too high). P. = Pisiform. 156 (VOL. VI.) Patagonian Expeditions Vol.vi Plate xxii Bruce Horsfall del. Werner & Winter. FrankfortAM Nesodon. PATAGONIAN EXPEDITIONS ! PALAEONTOLOGY. EXPLANATION OF PLATE XXIII. Fig. i. Nesodon imbricatus : Right femur, dorsum (No. 15,492). PAGE 165 Fig. 2. 4 4 44 “ “ distal end (No. 15,968). . Fig- 3- 44 i( Right humerus, anterior side (No. 15,256). . 152 Fig. 4. 4 4 4 4 “ “ posterior side (No. 15,968). Fig- 5- 44 4 4 The same, distal end. ..... Fig. 6. 44 4 4 Right patella, anterior face (No. 15,968). 168 Fig. 7. i 4 4 4 The same, posterior face. .... Fig. 8. 4 4 44 Axis, right side (No. 15,489). 145 Fig. 9. a 44 Posterior thoracic vertebra, left side. . All figures one half natural size. I48 (VOL. VI.) Patagonian Expeditions Vol.vi Plate xxiii. 5 Bruce Horsfall del. Werner & Winter; Fra nkfort°/M N E S O D O N PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. EXPLANATION OF PLATE XXIV. Fig. i. Nesodon imbricatus : Pelvis, right os innominatum (No. 15,967). . PAGE I63 Fig. 2. a “ Lumbar vertebra, anterior face (No. 15,492). I48 Fig- 3- i ( “ Fifth and sixth cervical vertebrae, left side (No. 15,489) I46 Fig. 4. i i “ Axis, anterior face (No. 15,489). . H5 Fig- 5- i i “ Atlas, dorsal side (No. 15,489). H5 Fig. 6. i t “ The same, anterior face. .... All figures one half natural size. (VOL. VI.) Patagonian Expeditions Vol.vi Plate xxiv Bruce Horsfall del. Werner & Winter Frankforh^M. Ne s od on. PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. Fig. I Fig. 2 Fig- 3 Fig. 4 Fig. 5 Fig. 6 Fig. 7 Fig. 8 Fig. 9 Fig. io Fig- ii Fig. 12 EXPLANATION OF PLATE XXV. Nesodon imbricatus : Right tibia and fibula, distal end (No. 1 5>968) “ “ The same, dorsum. . . . . . “ “ The same, proximal end. . . . . “ “ Left ulna, anterior side (No. 15,256). “ “ Left fore-arm bones, oblique view (No. 15.256). “ “ Right radius, distal end (No. 15,660). “ “ Left radius, proximal end (No. 15,256). “ “ Right radius, proximal end, with sesamoid, s. (No. 15,968) “ “ Right pes, dorsum (No. 15,968). Cn. 1 2 = Ankylosed ento- and meso- cuneiforms. Cb. — Cuboid. . “ “ Right astragalus, plantar side (No. 15,968). “ “ Right calcaneum, dorsal side (No. 15,968). “ “ Right metatarsus, proximal end (No. 15.968). All figures one half natural size. PAGE 168 154 154 154 154 154 170 170 171 172 (VOL. VI., Patagonian Expeditions Vol.vi Plate xxv Bruce Horsfall del. Werner & Winter, Frankfort0^ Nesod on PATAGONIAN EXPEDITIONS : PALEONTOLOGY. EXPLANATION OF PLATE XXVI. Fig. i. Adinotherium ovinum : Lumbar vertebrae, sacrum and pelvis, dor- PAGE sal side (No. 15,963). .... 209 Fig. 2. “ “ Skull, occipital view. P.oc. = Paroccipital process. Pty. = Post-tympanic portion of squamosal. M.a.e. = External audi- tory meatus (No. 15,382). 203 Fig. 3* “ “ Right metacarpus, proximal end, X t (No. 15,158). 212 Fig. 4- Nesodon imbricatus: Left carpus, proximal view (No. 15,460). 156 Fig. 5- Adinotherium ovinum : Third metacarpal, distal end, X t (No. I5.9b3) 213 Fig. 6. Nesodon imbricatus: Left metacarpus, proximal end (No. 15,460). l6l Fig. 7- Adinotherium ovinum : Atlas, from above (No. 15,131). 207 Fig. 8. “ “ The same, from below. .... Fig. 9- “ “ The same, from behind. .... Fig. IO. “ “ Axis and third cervical vertebra, left side (No. 15,131) 207 Fig. 1 1. “ “ Axis, from below (No. 15,131). Fig. I 2. “ “ Two anterior thoracic vertebrae, left side (No. 15,131) 208 All figures, except figs. 3 and 5, one half natural size. (VOL. VI.) Patagonian Expeditions Vol.vl Plate xxvi Bruce Horsfall del. Werner & Winter, Frankforf°/M --■Poc. Adinotherium, Nesodon PATAGONIAN EXPEDITIONS I PALAEONTOLOGY, EXPLANATION OF PLATE XXVII. Fig. 1. Adinotherium OVINUM Left femur, posterior face (No. 15,131). . PAGE 2 14 Fig. 2. a 44 The same, anterior face. Fig- 3- 4 4 4 4 Left femur, distal end (No. 15,963). Fig. 4. 4 4 4 4 Left humerus, anterior face (No. 15,966). 2 IO Fig. 5- 4 4 4 4 Left ramus mandibuli, crown view (No. 15,131). . . ... 201 Fig. 6. 4 4 4 4 Left tibia and fibula, anterior face (No. r5>i3 0- . 215 Fig. 7. 4 4 4 4 The same, distal end. .... Fig. 8. 44 4 4 Right scapula, outer side (Nos. 15,004 and 15,131) 209 Fig. 9. 4 4 4 4 The same, distal end. .... Fig. 10. 4 4 4 4 Right ulna, outer side (No. 15,004). 2 10 Fig. 1 1. 4 4 4 4 The same, anterior face. .... Fig. 12. 4 4 44 Right radius, anterior face (No. 15,004). 2 I I Fig- 13- 4 4 4 4 The same, external side. Fig. 14. 4 4 44 The same, proximal end. Fig. 15- 4 4 4 4 The same, distal end. .... Fig. 16. 4 4 4 4 Left patella, three views (No. 15,131). 214 Fig. 17- 4 4 4 4 Pelvis, right os innominatum (Nos. 15,963 and 15,016). ..... 214 All figures one half natural size. (vol. VI.) Patagonian Expeditions Vol.vl Plate xxvri. Bruce Horsfall del. Werner & Winter, Frankforl-°M. Adinotherium. PATAGONIAN EXPEDITIONS : PALAEONTOLOGY. EXPLANATION OF PLATE XXVIII. Fig. 1. Homalodontotherium Segovia: Skull of type-specimen, from below. Ameghino collection. Fig. 1 a. “ “ The same, from above. Fig. 2. “ “ Right first upper molar, unworn (No. 16,014). . All figures one half natural size. PAGE 259 260 252 (VOL. VI.) Patagonian Expeditions Vol.vl Plate xxviii. Bruce Horsfall del Werner& Winrec FrankforlAM Homalodontotherium PATAGONIAN EXPEDITIONS! PALAEONTOLOGY. Fig. I. Fig. i a, Fig. 2. Fig- 3- EXPLANATION OF PLATE XXIX. Homalodontotherium cunninghami : Lower dentition, left side ; approximately natural size. Ameghino collection. “ “ The same, crown view. Homalodontotherium segovi^e, type: Skull, left side, X i- Amegh. coll. . ... . Homalodontotherium cunninghami : Left upper teeth, crown view ; approximately natural size. La Plata Museum. . PAGE 255 259 248 (VOL. VI.) Patagonian Expeditions Vol.vi. Plate xxix. Bruce Horsfall del. Werner & Winter, Frankfort°/M. Homalodontotherium PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. Fig. I. Fig. 2. Fig- 3- Fig- 3a- Fig. 4. Fig. 5- Fig. $a. Fig- Sb- Fig. 6. Fig. 6a. Fig. 7- EXPLANATION OF PLATE XXX. Homalodontotherium segoviae, type: Right manus, X T Amegh. coll. Outline restoration after Ameghino. Td. = Trapezoid. L. — Lunar. Py. = Pyramidal. “ “ Phalanges of digit III, radial side, XT Homalodontotherium cunninghami : Left femur, dorsum, X T La Plata Museum. . “ “ The same, distal end, X \- Homalodontotherium sp. : Right astragalus, plantar side, X T Amegh. coll. ..... Homalodontotherium crassum : Left calcaneum, dorsum, X i (No. 15.435)- • “ “ The same, tibial side, X T • “ “ The same, distal end, X T • Homalodontotherium cunninghami : Left tibia, dorsum, X i (No. 15.435) “ “ The same, distal end, X T • “ “ Right calcaneum, dorsum (A. M. N. H.). PAGE 271 274 276 278 285 2 77 278 279 (VOL. VI.) "*■» lift*. Bruce Horsfall del. WernerS Winter. Frankfort°/M. Homalodontotherium ... - .