OLIVER ET AL.: WALRUS FEEDING IN BERING SEA 



400 



200 



m 

 h- 

 o 

 o 

 d 



Q. 



CO 



_l 

 < 



g 

 > 



Q 



2 







100 



50 







20 i 



10 



Polycnaetes 







Bivalves 



Cape 

 Rodney 



Sledge Sledge 

 Island Island 



N(in)= 3 

 N(out)= 3 



6 

 4 



3 

 3 



Nome 



3 

 3 



Cape 



Nome 



7 

 3 



Cape 

 Nome 



6 

 6 



FIGURE 6. — Abundances of major infaunal groups inside and out- 

 side the furrows dug for Macoma spp. (Sledge Island), and the pits 

 excavated for Mya truncata (all other areas). Means and standard 

 errors in N cores. 



n = 26 cores). Lysianassids respond to various dis- 

 turbances and are voracious scavengers (pers. obs.) 

 that probably were attracted to the tissue on a discard- 

 ed bivalve shell. 



DISCUSSION 



Walrus Feeding Behavior 



Walrus are highly specialized for feeding on benthic 

 infauna, especially bivalve molluscs (Fay 1982). Of 

 other marine mammals, only the diet of bearded seals 

 overlaps with the bivalve prey of walrus near Nome, 

 but bearded seals have a much broader diet than do 

 walrus (Lowry et al. 1980). Because bearded seals eat 

 certain shallow-burxo wing clams (e.g., Serripes groen- 

 landicus) and rarely eat deep burrowers (e.g., Myc 

 truncata), they cannot account for the diverse feeding 

 records observed near Nome. No other biological or 

 physical process can account for the record of ex- 

 cavations and discarded shells. While some large sea 

 stars can make pits as large as the larger Serripes pits, 



400- 



Rodney 



N(in) = 



N(out) = 



3 



3 



6 

 4 



3 

 3 



3 

 3 



7 

 3 



6 

 6 



FIGURE 7.— Abundances of three numerically dominant infaunal 

 species inside and outside walrus excavations. The polychaete worm 

 Myriochele oculata, the ophiuroid Amphiodia craterodmeta, and the 

 crustacean Protomedeia fasciata. See Figure 6 legend. 



none of the Mya and Macoma pits or the furrowed 

 bottoms are produced by sea stars. 



The low water clarity in the Bering and Chukchi 

 Seas and the poorly developed eyes of walrus suggest 

 that prey are not located by sight (Fay 1982). How- 

 ever, the benthic feeding records suggest that walrus 

 often search for certain bivalves by sight. The most 

 important evidence was the presence of many dis- 

 tinct, isolated pits with no indication of bottom dis- 

 turbance between pits. These pits were made in 

 excavating Mya truncata, the same species that 

 divers routinely located by sight because of the large, 

 conspicuous siphons. Apparently, walrus used the 

 snout and vibrissae to search for prey without con- 

 spicuous siphons or shells (Fay 1982), as extensive 

 furrowing only occurred in excavating Macoma 

 spp., clams with small and cryptic siphons. These 

 "rooting" activities clearly disturbed surface sedi- 

 ments and infaunal communities. Even the move- 



509 



