203—26 INDIAN TERTIARY AND POST-TERTIARY YERTEBRATA. 
there is an almost complete transition from the true bears through Bycenarctos (which 
undoubtedly is essentially a bear) to Dinocyon , and so through Gephalogale to Ganis. 
Gephalogale is classed by M. Filhol as closely related to the dogs, and also to Dinocyon , 
and it seems impossible to separate it from the modern Canidce by any character of 
more than generic value. 
In another article in the same memoir, 1 M. Filhol has shown the intimate 
relationship exhibited by Amphicyon and Gynodictis to the dogs on the one hand, and 
to the bears on the other, and comes to the conclusion that these two genera probably 
took origin from a common stock; the former being not improbably the direct 
ancestor of the dogs, and the latter of the Viverridce. It is at the same time 
suggested that the former, which is described as having the head of a dog and the 
limbs of a bear, may have been on the ancestral line of the bears, though there was 
not sufficient evidence to fully establish this point. Such evidence it is here 
submitted is afforded by the relationship of the dentition of Bycenarctos to Dinocyon 
and thus to Gephalogale and Amphicyon. 
On this evidence it appears to the present writer, at all events for palaeonto- 
logical purposes, to be impracticable to continue to refer the bears and the dogs to 
separate families ; as if this be done it is absolutely impossible to say with which of 
the two Dinocyon should be classed ; since from its upper true molars it should be 
referred to the dogs, from its upper carnassial to the hysenarctoid bears, and from its 
lower carnassial to both. 
For this provisional conjoint family is seems best to adopt the name Drsidce, 
making it to comprehend the two modern families of the Drsidce and the Canidce. In 
respect of other anatomical characters, it is already known that while the modern 
bears are plantigrade and pentedactylate, and the modern dogs digitigrade with only 
four anterior digits, there is good reason to believe that the extinct dog-like genera 
Amphicyon and Cynodictis were plantigrade and pentedactylate. The limb-bones of 
Amphicyon are described by M. Filhol in the memoir quoted as being precisely 
intermediate between those of the dogs and the bears. In their cranial 
characters Prof. Flower remarks 2 that the modern Cynoidea ( Canidce ) are intermediate 
between the modern Arctoidea ( Drsidce , Frocyonidce , and Mustelidce ) and the 
AEluroidea ( Viverridce, Hycenidce and Felidae ), and there is therefore no improbability 
that fossil forms should exhibit a complete transition from the canoid to the arctoid 
type of cranium. 3 It unfortunately happens that in many of the transitional forms 
(e.g. Bycenarctos , and Dinocyon ,) the base of the cranium when known at all, is too 
imperfect for its characteristic points to be determined. In Amphicyon , however, M. 
Filhol has shown 4 that there are certain characters of the base of the cranium (notably 
the minute size, or total absence, of a septum in the auditory bulla, 5 and the presence 
1 Op. cit., p. 70, “ Observations relatives aux Chiens Actuels, et aux Carnassiers Fossiles s’en Kapprochant le plus.” 
2 ‘Pro. Zool. Soc.,’ 1869, p. 24. 
3 The difficulty of distinguishing the Canidce from the Viverridce will be mentioned below. 4 Op. cit., pp. 84-5. 
5 This septum is fully developed in the cats, absent in the bears, and half developed in the dogs. 
