1922] Kellogg: Pinnipeds from Miocene and Pleistocene Deposits 85 



not only for the Phocinae but of the other subfamilies as well. Captain 

 Barrett-Hamilton 171 has figured a remarkable series of variations in 

 the teeth of Ommatophoca rossi. 



One is at once impressed, in studying the dentition of the Phocidae, 

 by the wide range of variation both in the number of the teeth in the 

 series and in the number of the cusps possessed by the premolars and 

 molars. Either the dentition of this family is in a very plastic state 

 or else the acquisition of additional cusps is very irregular. Recently, 

 Gregory 172 has advanced the theory that the molariform teeth of Phoca 

 gichigemis may readily be derived from those of Cyanarctus saxatilis. 

 The molariform teeth of the existing Phocidae have so many secondary 

 modifications that little reliance can be placed on attempts to homolo- 

 gize their teeth with those of other fossil carnivora. Until we know 

 the dentition of the Phocidae from forms older than the Upper 

 Miocene it will be hopeless to attempt to trace their derivation. 



As already remarked, there are a number of facts that lead one 

 to believe that all the Pinnipedia may not have a common ancestor. 

 It is at least conceivable that the otarids may have been derived from 

 an ursid type while the phocids may have been derived from another, 

 possibly from an aeluroid type. On the basis of osteological char- 

 acters, Sir George Mivart 173 has adduced a series of relationships that 

 go far toward showing that the phocids are considerably unlike the 

 otarids. His summary is as follows : 



1. In the Phocidae, as in Lutra, there is no alisphenoid canal, while in both 



Otaria and Vrsus it is present. 



2. In the Phocidae and Lutra the paroccipital and mastoid processes are not 



united by a prominent ridge of bone, while in Otaria and Ursus they are 

 so united. 



3. In the Phocidae and Lutra the mastoid process does not much depend ; in 



Otaria and Ursus it depends considerably. 



4. The bulla of Lutra could be easily made to resemble that of Phoca by giving 



a rounded form to the mastoid ; in both genera there is the same sort of 

 groove between the mastoid, and the tympanic. The bulla of Otaria, on 

 the contrary, is exceedingly like that of Ursus, and in both these genera 

 the sort of groove which exists between the mastoid and tympanic in 

 Lutra and Phoca is absent. 



5. The angle of the mandible is large in Otaria and Ursus, while in Lutra and 



Phoca it is smaller. 



171 Barrett-Hamilton, G. E. H., Eeport on the collections of natural history 

 made in the Antarctic regions during the voyage of the Southern Cross. Publ. 

 Brit. Mus., pi. 1. London, 1002. 



172 Gregory, W. K., The orders of mammalsL Bull. Am. Mus. Nat. Hist., vol. 27, 

 p. 314:. New York, 1910. 



173 Mivart, G., Notes on the Pinnipedia. Proe. Zool. Soc. London, p. 498. 

 1885. 



