82 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I49 



therium, clearly the most distinctly herbivorous of the condylarths, 

 is rather remote. Nevertheless, this relationship is evidently less 

 remote than that to the periptychids on the basis of teeth as well as 

 feet, although the latter are now usually included within the Condy- 

 larthra. The periptychid foot structure as represented by Ectoconus 

 is not condylarthran but of the taligrade type, as observed by Matthew, 

 more closely allied to that of the pantodonts. A foot structure of 

 this type was also observed for the tillodonts, whose teeth incidentally 

 are basically more P ant olamb da-like than arctocyonid. A slight 

 resemblance between lower molars of Meniscotherium and Esthonyx 

 deserves comment, although the two patterns are truly different and 

 any thought of relationship is discredited by the upper molar structure 

 as well as feet. 



A relationship to the perissodactyls was early predicated on tooth 

 structure as outlined in a foregoing section, but this was abandoned as 

 more became known of the foot structure in Meniscotherium. The 

 tooth structure somewhat less closely resembles that of the chali- 

 cotheres than perhaps certain of the Old World palaeotheres, although 

 discrepancies in detail may be observed such as the somewhat more 

 crescentic character of the protoconule in the upper molars and the 

 very different third lower molar in Meniscotherium. Resemblances 

 in molar structure are no doubt due to parallelism in development of 

 the selenodont or crescentic tooth pattern, inasmuch as the inter- 

 locking arrangement of the elements of the carpus in the early peris- 

 sodactyls readily excludes Meniscotherium from a close relationship. 



The artiodactyls are even more remote, and although a roughly 

 similar selenodonty has occurred in certain forms, as far as I can 

 determine, the structures are not all homologous. Generally the 

 position of the hypocone has been taken by the metaconule in early 

 development of the selenodont pattern in upper molars of arti- 

 odactyls (or by the protocone in the cainotheres), but these would 

 both appear to be definable in Meniscotherium. The artiodactyl foot 

 structure, of course, is quite unlike that in M enisc other ium. 



A resemblance which may be more than casual involves the late 

 Paleocene and early Eocene litopterns of South America. The rela- 

 tive age of the Sao Jose de Itaborai, as well as the Casamayor 

 assemblage, however, would seem to preclude the possibility of Menis- 

 cotherium having given rise to these earlier proterotheriids, as would 

 be implied in Wortman's postulation. Ameghino was, no doubt, 

 closer to the truth in suggesting a common but as yet undiscovered 

 ancestry. Perhaps the closest resemblance is seen in a comparison 



