OF THE CANIDE AND URSID. 94] 
justifying the admission of Danis (type horribilis) and Luarctos 
(type americanus), unless the more backward position of the first 
digit in the fore paw of americanus and the higher position of 
the carpal pad be given generic value—in my opinion, an 
exaggerated view of their importance*. 
As regards the genealogical position of the genera judged from 
their feet, analogy justifies the opinion that the naked-footed 
forms with free digital pads, like Helarctos and Tremarctos, ave 
the more primitive 7. From a stock probably resembling these 
in the particulars named, Melwrsus is specialised on one side by 
the fusion of the digital pads and Ursus on another side by the 
growth of hair over the carpal region. Thalarctos appears to me 
to be nothing but a specialised type of Ursus, adapted for 
swimming and movement on ice, its longer and more powerful 
canine teeth being developed for the seizing and slaying of 
seals. 
* Merriam admits Evarctos as a subgenus of Ursus on account of certain cranial 
and dental differences. Danis, however, appears to be undefinable (Proc. Biol. Soc. 
Wash. x. pp. 65, 83, 1896). 
+ In the bears themselves this view finds support in the scantiness and shortness 
of the hairs clothing the areas behind the digital and carpal pads in the newly-born 
cubs of Thalarctos maritimus and Ursus arctos. 
