408 NOTES ON THE CANIDE OF THE WHITE RIVER OLIGOCENE. 
eral of which, like the shape of the radius, reeur in Cynodictis. Tending to the same 
conclusion is the fact that what little is known of the structure of the ecreodont Miacis 
is of similar composite canine-feline character and it is to that creodont family to which 
most of the lines of fissipede Carnivora appear to lead back. It may be hoped that the 
problem will receive its definite solution when we shall have recovered the as yet miss- 
Ing or very imperfectly known dogs from the Uinta, uppermost White River and lowest 
John Day formations, and are thus enabled to trace the successive changes step by step. 
Assuming, then, as probable that Daphenus should have a place in the direct canine 
phylum, the larger question at once arises: What was the relation between the early 
members of the Caunide and Felide, and of both of these groups to the other fissipede 
families? It seems to be a comparatively rare phenomenon among the mammals that 
parallelism or convergence of development should be manifested in all parts of the struc- 
ture of two independent lines, though that this may happen is shown by the case of the 
Machairodonts and felines, to which reference has already been made. Usually, however, 
parallelism is displayed in a few structures only, such as the dentition, or the feet, or the 
vertebrae, and the more widely separated any two phyla are at their point of origin, the 
less likely are they to develop along similar lines. It will be sufficiently clear from the 
foregoing descriptions that the resemblances between Daphenus and the more primitive 
Machairodonts, such as Dinictis, are not only exceedingly close, but that they recur in 
all parts of the skeleton. The skull, the vertebral column, the limbs and the feet are 
all so much alike in the two series that, in the absence of teeth, it is often very difficult 
to decide to which of the two a given specimen should be referred. Such close and gen- 
eral resemblance is prima facie evidence of relationship, even though it should have been 
independently acquired, because parallelism is much more frequent between nearly allied 
than between distantly related groups. In the present instance, however, there is no rea- 
son to infer that the resemblances were separately attained ; on the contrary, the evidence 
now available seems to favor the conclusion that the dogs and cats are derivatives of the 
same Eocene stock. It cannot be pretended that this conclusion is, as yet, a well-estab- 
lished one, nor can it be so established until we recover the missing links of the canine 
and feline genealogies. Daphenus may eventually prove to be merely an abortive side- 
branch without phylogenetic significance, though this seems unlikely in view of its rela- 
tionship to the John Day dogs. On the other hand, when we have learned more of the 
Uinta dogs, it may appear that all the many resemblances of Daphenus to the Machai- 
rodonts haye been separately attained ; but existing evidence does not favor this sug- 
gestion either. It seems exceedingly likely that the dogs and cats are more closely 
related than has hitherto been believed and that they were derived from a common mid- 
dle or late Eocene progenitor. 
