272 



do u( j \j ass — poss r r, m amm a l f a 



As before stated, the molars have lour crescents with no fifth lobe, and the length 

 and width are nearly .equal. As nearly as I can judge from the illustrations of Falconer 

 and Cantley (1847, Part VII, PL LXII, Figs. 15 and 17), the last two molars are con- 

 siderably like those of Merycopotamus dissimilis, but here the close similarity apparently 

 ends. 



Molar 1 is so much worn that little can be said of it, except that it is much smaller 

 than M 2 and is square in section. The anterior and posterior worn surfaces are conflu- 

 ent medially, and only the last narrow traces of the transverse valley are preserved, one 

 as an external, the other as an internal projection, not quite meeting on the middle 

 transverse line. 



In the third and fourth molars the inner crescents embrace the inner bases of the 

 external crescents, so that the median transverse valleys are much deeper than the median 

 longitudinal ones. There are buttresses on the external horns of the external crescents ; 

 but the two median ones do not conjoin, forming a large one which cuts off the external 

 entrance to the median transverse valley as in I ' fyopotaitius. Owing to this and to the 

 depth of the median valleys, the anterior and posterior halves of tin; teeth seem much 

 more separated from each other than in that genus ; in fact the last molar, especially, is 

 almost cut in two. From the inner faces of the anterior inner crescents a short but 

 thick cingulum extends backward and slightly outward, abutting against the anterior face 

 of the posterior inner crescent. The anterior and posterior cingula are strong, and the 

 depressions or valleys which they enclose are divided by the horns of the inner crescents. 

 In the last molar the anterior horn of the posterior inner crescent continues to the ante- 

 rior outer crescent, thus dividing the bottom of the transverse valley by a thin partition. 

 The anterior roots of M 2 are connate, thus forming one broad root. This probably is 

 true of others, but they are not enough exposed to make it certain. 



Three molars on one side and the last two on the other have part of the jaws 

 attached. The incisor, canines and premolars were found separate. There are many 

 fragments of molar teeth of the same kind, so there must have been more than one 

 individual. 



A fragment of a mandible is present which has oik; imperfect tooth. It is doubt- 

 ful if this belongs to the same animal as the upper teeth. It looks more like the tooth 

 of some member of the Pecora than we would expect to see in this animal. It is pris- 

 matic and quite high. In section each half of the tooth is a three-sided prism, the outer 

 crescents being V-shaped at the top, but the outer angle becomes rounded below. On one 

 of the outer crescents or plates there is a, ridge that is not exactly median, as it begins 

 below at the base of the posterior buttress, passes diagonally upward and forward, 

 terminating in front of the apex of the crescent. This is probably M 2 . 



