230 BULLETIN 16 9, UNITED STATES NATIONAL MUSEUM 



by distinctions open to exception but fairly distinctive with tiie whole 

 group in mind, that they are different groups. Whether it be con- 

 sidered as a dichobunid or as a hyopsodontid, Phenacodaptes is a 

 peculiar form. Yet ail its characters known to me are either dupli- 

 cated or rather closely approached by various hyopsodontids, and its 

 reference to that group is at least as probable as any other view. The 

 fact that no artiodactyl, or no other artiodactyl, is known from the 

 Paleocene in itself carries no great weight as regards the affinities of 

 Phenacodaptes, except from the point of view of logical procedure in 

 the special case. If, as seems to me to be true, Phenacodaptes re- 

 sembles a group that is known to have been abundant and varied 

 when it lived at least as closely as it resembles another that has never 

 been found in deposits of that age, it seems preferable to refer it to 

 the former group pending discovery of decisive evidence. 



If Phenacodaptes should prove to be a hyopsodontid, it will not 

 very closely enter any of the categories of the key, not so much that 

 it has any nonhyopsodontid character as that it is a synthetic type. 

 The lower premolars are somewhat more suggestive of group /, 

 although they could well appear also in group //. The molar tri- 

 gonids, as far as I can judge by the data known to me, may be either 

 Ellipsodon- or Haplaletes-like, probably the former, but the entoconids 

 are more as in Haplaletes and its allies. 



Supergeneric grouping of these forms has always been based on a 

 separation of Hyopsodus from all other known forms. Historically it 

 is easy to see how this arose and that it was logical to the point of 

 being the only arrangement permitted by the data. Hyopsodus is, 

 within this group, an advanced genus with pronounced modifications, 

 tending to conceal its relationships to the very primitive forms. 

 Even within the genus, knowledge was principally based on relatively 

 late (especially middle Eocene) and specialized species. Further- 

 more the only Paleocene forms adequately known were from the 

 Lower and Middle Paleocene and were typified by such a form as 

 Mioclaenus turgidus, which lies rather far from the structural ancestry 

 of Hyopsodus. 



Even Matthew necessarily based his conception of the genus on 

 forms that suggest marked separation from Hyopsodus within the 

 family. Aside from Mioclaenus he knew only Oxyacodon, Ellipsodon , 

 Protoselene (with Haplomylus in latest Paleocene and early Eocene). 

 Oxyacodon is so ancient and primitive that intermediate stages were 

 necessary to show its probable phyletic position. Ellipsodon now 

 appears to lie nearer Mioclaenus, at least in its typical species, than 

 to the more Hyopsodus-like genera. Protoselene is curiously divergent 

 and not very near any other genus. He regarded Haplomylus as to 

 some extent intermediate between the earlier forms and Hyopsodus, 

 and probably this largely influenced him in uniting the Hyopsodon- 



