99 
NOTES ON THE CANID2 OF THE WHITE RIVER OLIGOCENE, ow” 
all three facets are confluent. The neural arch is low and broad, considerably elongated 
from before backward, and without ridges of any kind, save an inconspicuous tubercle, 
which represents the neural spine. Near its anterior border the arch is perforated by 
the usual foramina for the first pair of spinal nerves. The inferior arch is very slender, 
forming a more curved bar and has a much less antero-posterior extension than in Canis. 
Wortman (794, p. 137) has pointed out that the foramina of the atlas display certain 
characteristic features in the various carnivorous families. ‘In all of the Felidae which 
I have had the opportunity of studying, the [vertebrarterial] canal pierces the transverse 
process at its extreme posterior edge, where it is thickened and joins the body of the 
bone. The superior edge of this posterior border slightly overhangs the inferior edge. 
.... This character appears to be very constant in the Melide and so far as we know 
the structure of the atlas in the more generalized Nimravide [Machairodonts], it is true 
of them also. In the Canide, upon the other hand, the foramen for the vertebral artery 
is situated well in advance of the posterior border of the process, and instead of having 
a fore-and-aft direction, as in the cats, pierces the process almost vertically from above. 
In the Viverride and Hyenide the position of the foramen is very much as in the cats. 
There is, however, an important difference between these two families and the felines 
where the artery enters the suboccipital foramen in the anterior part of the atlas. The 
difference consists in the formation of a bony bridge in this situation, which gives to the 
suboccipital foramen a double opening in the hyzenas and civets, whereas it is single in 
the cats.” 
In Daphenus, it is interesting to observe, the foramina of the atlas are in all respects 
like those characteristic of the cats and thus depart in a very marked way from the 
arrangement found in the recent Canide. The transverse processes are broken away, so 
that their shape is not determinable, but enough remains to show that the atlanteo-diapo- 
physial notch is not converted into a foramen, thus agreeing with the canines and felines 
and differing from most of the hyzenas and civets. 
The azis is likewise feline rather than canine in its general character and appear- 
ance. The centrum is elongate, narrow and depressed, with a thin and inconspicuous 
hypapophysial keel, running along the ventral surface, and has a slightly concave posterior 
face. The articular facets for the atlas are convex and rise higher upon the sides of the 
neural canal than in Canis, and on the ventral side they project below the level of the 
centrum, so that they are separated by a broad notch, which is not present in the modern 
dogs, and is not well marked in the cats. The odontoid process is a long, slender, bluntly 
pointed peg, with a heavy, rounded ridge upon its dorsal surface, which is continued 
back along the floor of the neural canal. The transverse processes are quite long and 
relatively very stout; they are shorter and heavier than in Canis, and keep more nearly 
