106 CABN1YOEA. 



M. 1374. The greater portion of the right ramus of the mandible, 

 containing all the cheek-teeth except pmTT; from the 

 Upper Eocene of Caylux. This specimen is precisely 

 similar to the last. Purchased, 1884. 



GENUS non det. (serial position uncertain). 

 Hob. England. 



36791. Portion of the cranium; from the Headon beds (Upper 

 Eocene) of Hordwell, Hampshire. This specimen has been 

 described by Davies in the Geol. Mag. dec. 3, vol. i. p. 437 

 (1884), who was unable to determine its affinities. It 

 indicates an animal of considerably larger dimensions than 

 Viverra hastingsice (p. 100). It shows a canine and three 

 premolars in fair preservation; and one premolar and a 

 carnassial, with their crowns broken, are present in their 

 respective alveoli. Presented by Samuel Laing, Esq., 1862. 



Family URSID^E. 



In the * Palseontologia Indica,' ser. 10, vol. ii. pp. 202-204, the 

 present writer has shown that it is impossible to draw any distinc- 

 tion of more than generic value between the Ganoids and the Ursoids. 

 and he accordingly united the two modern families of the Canidce 

 and Ursidce, adopting the latter name as the family one 1 . The same 

 arrangement has been adopted in the present work. It may, how- 

 ever, be observed that there are almost equally strong reasons for 

 uniting the Ganoids and Viverroids ; but as it is absolutely necessary 

 to have a certain number of divisions, the family Viverridce is 

 maintained. 



Although, as above said, it is logically impossible to draw any 

 divisions of family value between the different genera of the Ursidce 

 as thus extended, yet the convenience 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 under another. These 

 two groups may respectively be termed Ursince and Canince, and 

 their most typical representatives will be the members of the modern 

 families Ursidce and Canidce. The genus Dinocyon will be ranked 

 under the first group, on account of at least one of the species 



1 Were it not for the inadvisability of introducing new terms, and the in- 

 trinsic objection to the name Hycenarctos, the name Hycenarctidcs would be in 

 many respects more advisable for the family name. 



