126 PROCEEDINGS OF THE ACADEMY OF [1894. 



amples are to be found approximating to the narrow front usual in 

 americanus. In arctos and horribilis the fourth lower premolar has 

 but one cusp, on the posterior, inner base of which a small projec- 

 tion or accessory cusp is usually found in the latter, with two longi- 

 tudinal ridges running back to end of the tooth. According to 

 Busk" the tubercle and its ridge are commonly wanting in arctos 

 and when present are very small. In two specimens in the British 

 Museum, both tubercle and ridge are absent, in all the others it is 

 present, though variable and small — but in one specimen of Ursus 

 arctos isahellhius from Cashmir, it is quite as well developed as in 

 ordinary cases of horribilis. Dr. C. Hart Merriam '' appears to have 

 found this tubercle in two skulls of americanus from Prairie Mer 

 Rouge, Louisiana, in the National Museum. In the skull of cinna,- 

 momeus (3,308) the ridge is unquestionably present and in all prob- 

 ability the tubercle as well. 



A tricuspid crown on this tooth appears to be diagnostic of ameri- 

 canus when it is present, but a large majority of this species examin- 

 ed by me, show but one cusp. 



A critical survey of the whole field of cranial and dental charac- 

 ters among the species here treated, reveals little that is constant but 

 variation, and absolutely forces the conviction that among them there 

 is not one, sufficiently stable and uniform to be of specific value. 

 The European bear and the American grizzly run into each 

 other so regularly that except in extreme cases there is little 

 possibility of distinguishing them certainly, or in many cases doubt- 

 fully, apart from geographical considerations, and this even is not a 

 safe guide, one skull in the British Museum, marked "Barren 

 Ground Bear," 321 mm. in extreme length, 229 zygomatic breadth 

 and 32 length of molar, being in all respects an ordinary skull of 

 arctos, which might just as well have been collected in Scandinavia 

 or Kamtchatka. Mr. J. A. Allen* does, in fact, regard this bear 

 as indistinguishable from arctos^. There is less difficulty in separat- 

 ing americanus from arctos, but even here it has been shown how 

 much their proportions overlap and one specimen from Transylvania, 



GTrans. Zool. Society of Loiidou, Vol. X, pp. 60-69. 

 Troc. Biol. Soc. of Washington. Vol. VIII, p. 150. 

 8 Bull. U. S. Geo. & Geo. Survey, Vol. II, p. 336. 



^ Since tlie above was written Dr. Merriam has kindly shown me several 

 skulls from the barren grounds which have distinct characters of their own. 



