THE MAMMALIA OF THE DEEP RIVER BEDS. 73 



referred provisionally to Temnocyon by Cope, under the name of T. josephi ; but 

 until the mandibular dentition of that species is discovered, its place in the cynoid 

 series must remain indeterminate. In spite, however, of the considerable gap between 

 Cynodesmus and Baphcenus, their relationship is indicated by nearly every detail of 

 known structure in the two genera. The skull characters are closely similar in both ; 

 e. g , the long, narrow cranium, with postorbital constriction placed far back of the 

 orbits, and the short, rapidly tapering face. Cynodesmus shows advance over Daph- 

 cenus in the following particulars : (1) The auditory bullae are enlarged, fully ossified 

 and the posterior chamber indistinguishably fused with the anterior, while in the 

 White River genus the posterior chamber remains cartilaginous, or, at all events, is 

 separate from the anterior. In all the skulls which I have had the opportunity of 

 examining, the posterior chamber is lost, exposing the periotic from below, and the 

 anterior chamber is very small. (2) The cranium is somewhat more rounded and 

 capacious, and it, together with the zygomatic arches and ascending ramus of the 

 mandible, has become somewhat shortened. (3) The sectorials are rather more mod- 

 ernized and efficient shearing blades, the cusps being more compressed and extended 

 and less conical in form ; in M the deuterocone is reduced. (4) The third upper 

 molar has disappeared. (5) The first upper molar has become smaller and the outer 

 cusps moved nearer to the edge of the crown. 



Temnocyon represents a slightly modified side branch, in which the inferior sec- 

 torial has developed a trenchant talon, through the reduction or suppression of the 

 entoconid. In all other respects, the true canine character of Temnocyon and its 

 close resemblance in skull structure to Cynodesmus are very striking. Whether 

 T coryphaeus and T altigenis are properly placed in the same genus, or whether, as 

 Schlosser suggests, they belong to widely separated phyla, need not be discussed 

 here, nor can we determine at present whether Temnocyon and Cynodesmus have any 

 common ancestor nearer than Daphcenus. Icticyon alone, among recent dogs, shares 

 with Temnocyon the character of the trenchant talon on the inferior sectorial. As 

 this character is a rare one, both in fossil and recent cynoids, we may, perhaps, 

 expect that the existing South American genus will prove to be derived from the 

 John Day type. If so, many intermediate forms remain to be discovered. 



Daphcenus, in the structure of its skull, dentition and limbs, approximates 

 closely to the creodonts. This approximation is seen in the character of the secto- 

 rials, which are very like those of the Miacidce, in the primitive form of the cranium, 

 in the low humeral trochlea with its epicondylar foramen, in the third trochanter of 

 the femur, the creodont-like calcaneum and the relatively weak plantigrade feet. The 

 cynoids of the Uinta formation are, unfortunately, not sufficiently well known for 



