376 Mr. R. T. Pocock on some 



maritimus. The material examined consisted of a newly 

 born cnb of Thalarctos maritimus and of Ursus arctos and 

 adult examples of Mehirsus ursinus, Tremarctos thibetanus, 

 and of Ursus americanus , the conclusions regarding Hetarctos 

 malayanus being derived from an inspection of living 

 specimens and dried skins. Since that date I have had the 

 opportunity of examining in the Prosectorium of the 

 Zoological Society fresh dead specimens of Ursus horr'tbilis 

 and of Helarctos malayanus, thus adding two species to my 

 previous list, and also of Mehirsus ursinus, Tremarctos thibet- 

 anus, and of Ursus americanus, and an adult of Ursus arctos 

 from North Russia, enabling me to confirm and extend my 

 previous observations on these species. With regard to 

 Thalarctos I can add nothing to what I previously .stated. 



The examples of Ursus horribilis, two females 25 and 27 

 years old respectively, from the Missouri Brakes, Montana, 

 whence they were brought as cubs in 1890 by Mr. Ewen 

 Cameron, cause me to modify considerably my conception of 

 the genus Ursus. The feet, indeed, as explained below, 

 differ so markedly from those of Ursus americanus that I am 

 persuaded the two bears should be separated generically or 

 subgenerically, unless other examples of U. horribilis show 

 the characters to be variable, a conclusion not justified by the 

 analogy of other Carnivora. In certain respects the feet of 

 Ursus arctos bridge the interval between those of U. horribilis 

 and U. americanus. But, curiously enough, U. arctos is in 

 this character nearer U. americanus than U. horribilis, which 

 was hardly to be expected from the external appearance of 

 the species concerned. 



It may be recalled that Gray long ago gave superspecific 

 rank to these bears, applying the name Danis to horribilis, 

 Euarctos to americanus, and reserving the name Ursus for 

 arctos and its near allies. In the following pages the species 

 are recorded under those names. 



In 1917 (Ann. & Mag. Nat. Hist. ser. 8, vol. xx. p. 129) 

 I severed thibetanus from Tremarctos as a distinct genus 

 Arcticonus, based upon the cranial differences between the 

 Asiatic and South American forms. Hence it follows that 

 each of the well-marked types of living Bears takes generic 

 or subgeneric instead of specific rank, a conclusion which 

 many will deprecate, but which is more in accord with modern 

 schismatic treatment than the older conception. 



The existing genera and typical species of Ursidae will 

 therefore stand as follows: — Thalarctos maritimus, Ursus 

 arctos, Danis horribilis, Euarctos americanus, Arcticonus 

 thibetanus, Hetarctos malayanus, Mehirsus ursinus, Tremarctos 

 ornatus. Whether the various forms of Ursus, Danis, 



