338 NOTES ON THE CANID.E of the white river oligocene. 



parallel with the centrum, not diverging so much posteriorly. As in the felines, the ver- 

 tebrarterial canal is longer than in the modern dogs, and its posterior opening is not vis- 

 ible when the vertebra is seen from the side ; the anterior opening is larger and is placed 

 farther forward than in the recent Canidce. The neural canal is proportionately larger 

 than in the latter, both vertically and transversely, nor does it contract so much toward 

 the hinder end. The neural spine forms the great, hatchet-shaped plate usual among the 

 Carnivora, and in its details of structure it is feline rather than canine. In the latter 

 group, the spine is not continued back of the postzygapophyses into a distinct process, but 

 its hinder borders curve gently into them. In Daphcenus, as in nearly all the cats and 

 viverrines, the spine is drawn out into a blunt and thickened process behind the zyga- 

 pophyses, from which it is separated by a deep notch. The zygapophyses are rather 

 small and do not project so prominently from the sides of the neural arch as they do in 

 Canis. 



The other cervical vertebrae are more slender and lightly constructed than in the 

 existing Canidce of corresponding stature. The centra are long, narrow, depressed and 

 very feebly keeled in the ventral median line ; in most of the species this keel does not 

 terminate in a posterior hypapophysial tubercle, such as is found in the existing dogs. 

 In the largest species, however, D.felinus, the keels are more prominent, especially on the 

 third and fourth vertebrae, and there is some indication of the tubercle. The centra are 

 slightly opisthoccelous and the faces are somewhat oblique in position. In very few of the 

 specimens are the transverse processes sufficiently well preserved to require description, 

 and in such cases as they are present (as, for example, on the fifth and seventh cervicals 

 of one individual of D. hartshomianus) they display no noteworthy differences from the 

 corresponding processes of Canis. The vertebrarterial canal is, however, somewhat longer 

 than in the latter. 



The neural arches are very different from those seen in the modern representatives 

 of the family. In them the dorsal surface of the neural arch is very broad and on each 

 side projects outward as an overhanging ledge, which connects the prezygapophysis with 

 the postzygapophysis of the same side ; ridges and rugosities for muscular attachment are 

 well marked and in the large species often very prominent; the zygapophyses, and 

 especially the posterior pair, project but little in front of and behind the arches, and those 

 of each pair arc separated by notches of only moderate depth. In consequence of this 

 arrangement, there are but small interspaces visible between the successive arches, when 

 the vertebra 1 are in position. In Daphmnus, on the other hand, the dorsal surface of the 

 neural arch is relatively narrow, somewhat convex transversely and usually smooth, with- 

 out ridges or tubercles ; the overhanging ledge which gives such an appearance of breadth 

 to the arch in Canis is little developed : the zygapophyses project far in advance of and 



