Mammalia. 55 



ii i)iii-ticulaiiy wide augle. They are sliglitly smaller than the 

 remainder of the series. All the teeth of this specimen, except the 

 tirst premolars, both upper and lower, possess well-developed double 



roots. P.m. -J is a remarkable tooth. Its crown is, for Oinmatophoca, 



particularly well formed and carries three cusps. The central cusp 

 is far more prominent than either of the others, and is partially split 

 into two by a groove. This seems to me to afford an exact parallel 

 to the " cloven " cusp of a specimen of Phoca groenlandica, as 

 described and figured by Mr. Bateson at p. 239 of his work. But 

 in this instance the " cloven " tooth is not opposite to a reduplicated 

 tooth on the corresponding side of the same jaw, as in the case of 

 Ommato'phoca. Dr. Torsyth Major has been good enough to draw 

 my attention to a similar occurrence in Didclphys marsuinalis. In a 

 specimen of this species (No. 93.12.24.1) in the British Museum, 



o 



i. — on the left side is an exceptionally well-developed tooth, and 



is cloven in an exactly similar manner. 



In No. 1 we come to a completely new phase of the question. 

 This specimen, which, although fully adult and the smallest known 

 example of the species, is, as regards its post-canine dentition, 

 absolutely toothless. Not only have all the cheek-teeth dis- 

 appeared, but even the alveoli have partially or wholly vanished 

 with them. 



We have then in Ommatophoca a highly remarkable instance of 

 variation in the teeth — an instance which, I believe, is unique 

 amongst the mammalia. It may be, perhaps, well to summarise 

 the main points : — 



(I.) Disregarding the toothless skull (No. 1) we have seven 



specimens. Of these only two (Nos. 43. 11. 25. 4 and 897) have 



5x5 

 the normal Phocid formula of p.m. and m. _. -z:. For four of the 



^ 5x5 



remainder the formula is p.m. and m. -p ^, and for one (No. 324&) 



, 6 X 5 



p.m. and m. -^ zr. 



^ 5x5 



(II.) In all cases where six cheek-teeth occur in the upper jaw, 



two adjoining individuals of the series are smaller than the 



remainder, and are sometimes also so closely associated, that it may 



reasonably be supposed (as in the case of the original specimens) 



that they represent a corresponding single tooth of an individual 



possessing only five upper cheek-teeth. In one case only (No. 3) 



