14 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 1 28 



separated even generically from the complex that includes the large 

 Protoreodon pumilus annectens. 



Diplobunops, the larger of the two Eocene agriochoerids, has much 

 enlarged and widely separated caniniform teeth but is further charac- 

 terized by cheek teeth rather more primitive appearing than in con- 

 temporary protoreodonts. As an aberrant line Diplobunops would 

 appear to have separated from Protoreodon not much earlier than 

 lower Uinta time, or possibly even as late as Uinta B. It is known 

 only from Uinta B, C, and Randlett horizons. 



Removed from inclusion with the poebrotherines, in the Camelidae, 

 are a group of genera closely related to the Chadronian Eotylopus, 

 here distinguished as the new family Oromerycidae. Their camelid 

 resemblances, I believe, are largely adaptive, as differences in molar 

 tooth structure from that in the poebrotherines seem fundamental 

 and indicative of a rather remote differentiation, suggesting a hereto- 

 fore unnatural family grouping. I find no close relationship indicated 

 between the oromerycids and any of the other families, and their 

 origin in the bunodont forms, while uncertain in position, may have 

 been quite independent of other selenodonts. Eotylopus is evidently 

 the last of the line. 



The poebrotherine forms, like the leptotragulids, show early acquisi- 

 tion of a distinctly selenodont tooth structure, suggesting perhaps a 

 closer relationship with the latter group than with oromerycids, al- 

 though this is not demonstrated, inasmuch as the early development 

 of Pi as a caniniform tooth in the leptotragulids is a significant differ- 

 ence. In the absence of any distinctive lower caniniform tooth the 

 poebrotherines are most like the Eocene oromerycids ; however, this 

 also characterizes the homacodonts. Although like others of the upper 

 Eocene selenodonts in having obscure pre-upper Eocene antecedents, 

 the poebrotherines nevertheless distinguished themselves by the well- 

 documented diversity and geologic longevity of their descendant forms. 



The Leptomerycidae as represented by the leptotragulids would, 

 together with the poebrotherines, appear to be the most advanced and 

 specialized in the attainment of crescentic or selenodont molars among 

 the upper Eocene fornis. The Oligocene forms were surely derived 

 from the leptotragulines, but it is not clear that either Leptoreodon or 

 Leptotragulus gave rise to Leptomeryx. Its origin may have been 

 from an intermediate and perhaps earlier but closely related form. 

 There is, however, an even better case for Protoceras and the Proto- 

 ceratidae, and I find that Le\ptotragulus, as far as it is known, com- 

 pletely meets the requirements. In addition to the similar molar 

 structure, the premolars in Protoceras are entirely similar to those 



