NOTES ON THE CANID^ OF THE WHITE RIVER OLIGOCENE. 409 



On the assumption that the dogs and cats are thus quite closely connected, what can 

 be said concerning the relations of the other fissipede families with these groups and with 

 one another ? Of the derivation of the Procyonidce nothing is yet known ; the family 

 may be traced back into the Loup Fork without finding essential changes, but beyond 

 that period we lose track of it altogether. The position of the bears and hyaenas is rea- 

 sonably clear, the latter being late derivatives of the viverrines and the former of the 

 dogs, neither family making its appearance until long after the other fissipede groups 

 had become clearly differentiated. The Viverridce have a great many characters in com- 

 mon with both the early dogs and the early Machairodonts ; almost all the structural 

 features which are found in both Daphcenus and Dinictis recur also in the viverrines, 

 and the latter again have many points of similarity to Cynodictis, as has often been 

 remarked. That the viverrine features of Cynodictis are more numerous and apparent 

 than those of Daphcenus is largely due to the small size of the former, which agrees 

 much better with the stature usual in the recent viverrines. The viverrines thus seem 

 to be derivatives of the same Eocene stock as that which gave rise to both the dogs and 

 the cats, though, jierhaps, they are more nearly allied to the latter than to the former, 

 and apparently they have departed less from that primeval fissipede stem than has either 

 of the other families. Aside from the peculiar character of the auditory bulla and the 

 reduced number of the molar teeth, such a genus as Viverra would seem to differ but 

 little from the hypothetical Eocene ancestor of all the fissipede families. The Mustelidcs 

 represent a quite specialized branch of the fissipedes, but between its earlier and more 

 primitive members and the corresponding representatives of the viverrines are so many 

 structural resemblances that Schlosser does not hesitate to derive them from a common 

 stem. An interesting and significant example of this community of characters among 

 the early representatives of the different fissipede families is given by the os penis of 

 Cynodictis, which resembles that of the mustelines much more closely than that of the 

 modern clogs. This probably indicates that all of the earlier fissipedes had this bone 

 shaped very much as in the existing mustelines, which have thus retained the primitive 

 form, while in the other families it has become much modified in shape and size. This 

 would explain the apparent anomaly of the very large os penis of Cryptoprocta which is 

 so different from that of the. other viverrines. According to this way of looking at the 

 subject, there was a middle Eocene group of flesh-eaters, perhaps the creodont family 

 Miacidce, which rapidly diverged into four principal branches, the cats, dogs, viverrines 

 and mustelines, all of which families yrere established in the late Eocene or early Oligo- 

 cene, and to these should perhaps be added a fifth family, the Procyonidce, though of this 

 we know nothing definite. The Fissipedia are thus probably a monophyletic rather 

 than a polyphyletic group, which was derived from a single creodont family. 



