166 



NORTH AMERICAN FAUNA 74 



pass larger, hard objects through its digestive tract. Zoo and aquarium officials 

 have been aware of this for some time (see Pournelle 1962; Hagenbeck 1963). 

 Several captive specimens have died in recent years as a result of intestinal ob- 

 struction by objects no larger than 3 or 4 cm in diameter (R. M. McClung and 

 E. D. Asper, personal communication) as well as from gastric ulcers caused by 

 retention of such objects in the stomach (G. C. Ray, personal communication). 

 The claim by Elliott (1886:451) that many of the whole, unbroken clams in the 

 one stomach that he examined were "large as my clinched hands" must have been 

 an error in judgment or recollection, for the walrus' maximum gape (about 8 x 

 8 cm) is not large enough to allow passage of a man's fist, nor is the esophagus 

 capable of passing such a large, hard object. 



In general, only the siphon or foot of the larger clams, cockles, mussels, and 

 whelks are ingested, as noted previously by several investigators and described in 

 greatest detail by Vibe (1950). Notwithstanding Elliott's (1875:162) statement to 

 the contrary, whole, crushed clams, "shell and all," do not occur in the ingesta, 

 unless they are very small, and then infrequently. Even such small clams as 

 Astarte horealis and Hiatella arctica, most of which in the Bering-Chukchi 

 region are less than 3 cm in greatest diameter, ordinarily are not swallowed 

 whole; instead, the foot or siphon usually is removed and the shell rejected. Of 

 some 50,000 specimens of Hiatella that were present in the ingesta examined by 

 Fay et al. (1977) from walruses taken in the St. Lawrence Island-to-Bering Strait 

 area, not more than three still had most of the broken shells attached; nearly all 

 of the rest were represented only by the siphon. 



The question of how the siphons and feet are selectively removed and ingested 

 without the shells has been the object of much speculation. Malmgren's (1864 in 

 Allen 1880:136) hypothesis, that the shells are crushed with the cheek teeth, then 

 expelled, and the soft parts swallowed, has been most widely accepted. Con- 

 versely, Freuchen (1936 in Mohr 1952:254-255) and Freuchen and Salomonsen 

 (1958) hypothesized that the moUusks (along with sediments and stones) are car- 

 ried upward with the forelimbs, while allowing the stones and sediments to fall 

 away and sink to the bottom. Then, the moUuscan shells are crushed between the 

 palms, and the semi-suspended soft parts are ingested after the heavier shells 

 settle to the bottom. This hypothesis never has been widely accepted, for as Mohr 

 (1952:255) pointed out, some shells and especially stones are found in the ingesta 

 frequently enough to indicate that this is not the usual method employed. 



The Malmgren hypothesis also is untenable, for as I showed in the section on 

 Dentition, the patterns of abrasion and attrition on the cheek teeth indicate that 

 large, hard objects such as molluscan shells ordinarily are not taken into the 

 mouth beyond the incisive area. None of the postcanine teeth shows any signs of 

 frequent contact with such objects; only the vestigial incisors and, occasionally, 

 the anteriormost surfaces of P and C^ are harshly abraded, presumably from 

 frequent contact with molluscan shells. The shape and musculature of the 

 mandible provide further evidence in support of that contention. 



The walrus' mandible is different in form from that of the sea otter, Enhydra 

 lutris, the only marine mammal that is definitely known to crush shells of marine 

 invertebrates, including mollusks, with its postcanine teeth (Kenyon 1969). In 

 the sea otter, the mandible is comparable to that of other carnivores, in that the 

 temporal and masseter-pterygoid muscles have large, nearly equal moment 



