SIWALIK AND NARBADA CARNIVORA. 
27—204 
of post-parietal and mastoid foramina), which are essentially arctoid. Such cranial 
evidence as there is, points, therefore, in the direction of the amalgamation of the 
arctoid and canoid cranial types. Should future discoveries show that the crania of 
the former genera strictly conform either to the arctoid or the canoid type, then the 
question of the re-establishment of the TJrsidce and Oanidce as distinct families (if 
families have not to be abolished altogether) may be taken into consideration. In 
the meantime the author sees no other course but to unite them. 
It is at present impossible to give any definition of the family TJrsidce as thus 
extended. The upper true molars are very generally two in number, but there may 
be either (occasionally) one ( Icticyon ), three ( Amphicyon ), or four ( Otocyon ): the lower 
true molars are generally three, but may be only two {Gy on) or four {Otocyon). The 
first upper true molar is invariably placed directly behind the carnassial ; and in 
general the crowns of the true molars are well-developed : in the upper jaws they 
may vary in shape from oblong {TJrsus), through a square {Hycenarctos), to triangular 
{Cams). 
Although, as above said, it is logically impossible to draw any divisions of family 
value between the different genera of the TJrsidce as thus extended, yet the con- 
venience of having some division for working purposes among such a multitude of 
genera is so great, that it will be found advisable to rank the most bear-like genera 
under one arbitrary group, and the most dog-like genera under another. These two 
groups may respectively be termed TJrsince and Canince, and their most typical repre- 
sentatives will be the members of the modern families TJrsidce and Oanidce. The 
genus Dinocyon will be ranked under the first group, on account of at least one of its 
species having an upper carnassial of the Hycenarctos type, but it must be distinctly 
understood that its separation from the true dogs is a purely arbitary one, it being 
connected with that group through Cephalogale just as intimately as it is with the bears 
through Hycenarctos. 
Group A: UBSINHJ. 
Genera . — The existing bears are frequently divided into two or more genera, but 
for palaeontological purposes it seems best to include all the forms under one genus. 
Adopting this course the number of genera which may be ranked in the present 
arbitrary group are probably five , 1 viz.: fflluropus, TJrsus , Arctotherium i {Arctoidotherium), 
Hycenarctos , and Dinocyon. Of these the two first are alone existing, and the second 
is taken to include the whole of the true bears. The third is from the pleistocene of 
South America ; while the fourth and fifth are from the pliocene and miocene of 
North America. 
1 Leptarctus, Leidy, of the tertiaries of N. America though classed hy Prof. Leidy with the TJrsidce (“ Extinct Mammalian 
Fauna of Dakota and Nebraska,” p. 70,) is allied to Nasua, and, therefore, belongs to the Procyonidce. Metarctos of Pikermi 
probably belongs to a distinct family, unless it indicates the unity of the TJrsidce and Viverridce. Arctocyon (Palceocyon, Blain, 
not Lund), is the type of Prof. Cope’s family Arctocyonidce, classed under his order Creodonta. 
2 The history of this genus will he found below, under the head of Eycenarctos. 
