340 Wortman — Studies of Eocene Mammalia in the 



tinctions between the two genera are seen in the following 

 characters : in Vulpavus the jaw is relatively slender, the heel 

 of the inferior sectorial is comparatively small and basin- 

 shaped, the second and third lower molars have sharp cusps, 

 the main internal cusp of the first superior molar is large, 

 lunate, and connected with the outer cusps by an anterior and 

 posterior ridge upon which intermediates are developed, and 

 there is a postero-internal cusp. In Uintacyon, on the other 

 hand, the jaw is thicker and more robust, the heel of the 

 inferior sectorial is small and cutting, the cusps of the suc- 

 ceeding molars are low and obtuse, the main internal cusp of 

 the first superior molar is more conic and connected with the 

 external cusps by a ridge in front only, and there is no pos- 

 terior internal cusp. 



A third group having the jaw characters of Uintacyon, as 

 far as known, is represented by a few scattering fragments. 

 In this group the premolars are much reduced in size and the 

 canine is laterally flattened. Members of this latter group are 

 found in the Wind River and Wasatch deposits, which carries 

 them well back towards the base of the Eocene. That they 

 represent a distinct and independent phylum there can be little 

 doubt, and one, moreover, that is known to have left descendants 

 in the succeeding Uinta beds. 



If, upon further investigation, it should be found that the 

 scaphoid, lunar, and centrale of the carpus are separate, which 

 is not altogether improbable, the question will then arise as to 

 whether they should be properly classed in the Canidse or 

 yiverravidge or whether they should be placed in a distinct 

 family by themselves. According to the previous and perhaps 

 more common acceptation of the limits of the Creodonta they 

 would without doubt be placed in this suborder, but I believe 

 with Schlosser that they stand much nearer to the Carnassi- 

 dentia and, as I will presently attempt to show, were the 

 immediate progenitors of the two and probably three main 

 branches of the canine phylum. If we range them with 

 Viverravus in the Yiverravidge we shall fail to express 

 by such a classification the distinctive and important posi- 

 tions which these two genera hold with respect to the 

 canine and viverrine branches of the Carnassidentia. There 

 now appears to be little doubt that when the evidence is more 

 complete than it is at present, it will be found that the 

 Canidse and Viverravidse were derived from a common ances- 

 tor. That the separation of these two genera took place, 

 however, previous to the deposition of the Torrejon beds is 

 evidenced by the fact that a species of Viverravus, with all 

 the distinctive dental characteristics of this genus, is found 

 therein, and we must therefore look to an earlier date for the 



