NOTES ON THE CANIDH OF THE WHITE RIVER OLIGOCENE. 405 
tion of the condylar foramen from the foramen lacerum posterius, ete.; the femur has a 
third trochanter and the humerus an extremely prominent deltoid ridge; the feet are 
plantigrade and pentadactyl and, like those of many of the viverrines, they are supplied 
with partially retractile and very incompletely hooded claws. In all probability these 
structural characters also occurred in the ancestral Feline, but what distinguishes even 
the earliest Machairodonts is the elongation and compression of the upper canines, the 
reduction in size of the inferior ones and the development of bony flanges from the ven- 
tral border of the mandible for the protection of the superior tusks. From such begin- 
nings the sabre-tooth series may be traced, with various divagations and side branches, to 
the Pleistocene Smilodon, which in all parts of its structure is extraordinarily like Felis, 
the only important differences consisting in the dentition (which is of similar type) and 
in the modifications of the skull, which are necessarily correlated with the enormous 
enlargement of the upper canine tusks. | 
Seeing, therefore, that the machairodont series is well-nigh complete and that none 
of its known members is at all likely to prove ancestral to the true eats, there can be 
little reasonable doubt that the remarkably close resemblance which we observe between 
Felis and Smilodon is not directly due to their relationship, but has been independently 
acquired in the two series and is the outcome of a parallel course of development, con- 
tinued from the Oligocene to the Pleistocene. If this be true, there can be no @ priori 
ground for denying that the same phenomena may have been repeated in the dogs and 
that Boule’s suggestion concerning the derivation of the alopecoids from Cynodictis may 
possibly prove to be correct. In this case, however, the final identity of the two series is 
even more striking than in the cats and Machairodonts ; to verify the suggestion, it will 
be necessary to recover the missing links of the alopecoid phylogeny and to show that it 
has followed a course parallel to but independent of that of the thooids. 
Another alternative possibility is that the foxes became separated from the principal 
canine phylum at a comparatively late date, and that, consequently, Cynodictis and its 
allies represent but an abortive side-branch from the main stem. That the separation is 
of considerable antiquity is shown by the parallel arrangement of the two series to which 
Huxley has called attention. In both wolves and foxes we find species with microdont 
and macrodont dentition, with sagittal crests and lyrate sagittal areas, with lobate and 
non-lobate mandibles. So far, at least, we are almost certainly dealing with indepen- 
dently acquired characters. From the standpoint of present actual knowledge it is more 
probable that the separation did not take place before the end of the Miocene than that 
it had already been accomplished in the Eocene, though this conclusion involves the 
admission that Cynodictis had anticipated the foxes in quite a remarkable way. While 
very far from denying the possibility of such convergence as is implied in Boule’s 
