1922] Kellogg: Pinnipeds from Miocene and Pleistocene Deposits 55 



From the stage which is represented by Prorosmarus, the Odoben- 

 idae became more and more specialized, finally acquiring the heavy 

 bones which mark them as bottom feeders. In correlation with the 

 increase in size of the canines as tusks, there was a corresponding 

 crowding backward of the rostral region until, in the present forms, 

 Odobenus rosmarus and Odobenus divergent, it abuts against the post- 

 orbital processes. Furthermore, possibly in response to this need for 

 a more effective crushing- apparatus, the distal portion of the jaw 

 became more massive, as is illustrated by Prorosmarus, Alachtherium, 

 and Trichecodon, culminating in Odobenus, if they may be taken 

 as respective stages. Finally, the symphysial portion of the ramus 

 became ankylosed as it is in the latter. With time the enormous 

 growth of the upper canines has resulted in the disappearance of the 

 incisors, as they became functionless, and also in the transference of 

 the lower canine to the molariform series. 



While the tusklike canines were developing, the length of the skull 

 was shortening, becoming more massive as time went on. There was 

 no need of an increase in surface area for the attachment of muscles 

 and ligaments since water does not require the development of as 

 heavy muscles as are needed by a terrestrial animal. At first the 

 upper canine was no doubt small, very similar to that possessed by 

 the otarids. In the existing walruses it has increased to an enormous 

 size and has become greatly specialized. This required that the skull 

 become denser to bear its weight, and that the rostral region become 

 heavier and upturned in order to serve as a more effective socket for 

 its insertion. 



The statement is commonly met with that the evolution and 

 specialization of the tusks of the walrus have been in a large measure 

 brought about through the use made of them. Several writers have 

 erroneously stated that walruses use their tusks to enable them to 

 crawl out of the water upon the ice, but, according to Elliott, 52 this is 

 not true. Lankester 53 accounts for the large size of the tusks of the 

 Crag walrus as due to the absence of hard rocks against which the 

 tusks could be worn down. The observations of Brown 54 and of 

 Johansen 55 have shown that the food of the walrus is composed largely 



•"Elliott, H. W., The Seal Islands of Alaska, 10th U. S. Census, pp. 94-95. 

 Washington, D. C, 1881. 



03 Lankester, B., On the tusk of the fossil walrus found in the Bed Crag of 

 Suffolk. Jour. Linn. Soe. London, vol. 15, p. 145. 1881. 



a* Brown, B., Proc. Zool. Soc. London, pp. 429-430. 1868. 



5r ' Johansen, F., Observations on seals (Pinnipedia) and whales (Cetaceae) on 

 the "Danmark expedition," 1906-1908. Danmark-Expeditionen til Gr0nlands 

 Nord0stkyst, 1906-1908, vol. 5, no. 2, pp. 214-215. Copenhagen, 1910. 



