FISHERY BULLETIN: VOL. 81, NO. 3 



cavation morphology. Each excavation was linked to 

 a particular species of clam, and a larger region (dive 

 station) was characterized both by the primary type 

 of excavation and by the most abundant shells. Be- 

 cause each of two divers traveled at least 50 m and 

 often over 100 m per dive, the primary type of exca- 

 vation was easily assessed, albeit qualitatively, de- 

 spite the poor water clarity. Benthic feeding records 

 were located on 18 of 33 dive sites, but were only 

 well quantified at 6 of the 18 sites (Table 1). 



Furrows 



The most extensive and distinct furrows were found 

 near Sledge Island (Table 1). Water clarity in this 

 area was relatively poor (about 1 m), but was ade- 

 quate to see furrow widths and to trace lengths. By 

 swimming rapidly over a long distance (>50 m), we 

 estimated qualitatively that at least 40% of the bot- 

 tom was furrowed at one Sledge Island dive site. 

 Furrows generally formed a complex maze of ex- 

 cavations, but discarded shells always were abun- 

 dant (as many as 5-10/10 m 2 ) within and along the 

 furrow edges and were rare (<1/10 m 2 ) on undis- 

 turbed bottoms between furrows. The average fur- 

 row width was 45 cm and depth was 1 7 cm (Table 2) . 

 Macoma spp. were primarily excavated from fur- 

 rowed bottoms (Table 1). 



but lacked the central shaft. They were relatively 

 rare, as furrows were the primary excavations as- 

 sociated with Macoma shells (Table I). Macoma spp. 

 generally lives <20 cm into the sediment, and has a 

 shorter siphon than Mya truncata. 



The largest Serripes pits were much smaller than 

 theMya andMacoma pits (Fig. 2). Small Serripes pits 

 were impossible to distinguish from sea star pits and 

 surface irregularities. Some large Serripes-type pits 

 may be made by larger sea stars (Lethasterias and As- 

 terias). This bivalve has a short siphon, is a shallow 

 burrower, and commonly occurs at the sediment sur- 

 face. Serripes shells were conspicuous on the sedi- 

 ment surface. Therefore, although either the shells or 

 pits of Mya truncata andMacoma spp. could be count- 

 ed to estimate prey consumption, only the shells pro- 

 vided an adequate estimate of the number of S. 

 groenlandicus eaten by walrus. 



Pit-Furrows 



Pit- furrow systems consisted of a series of pits con- 

 nected by a shallow, continuous, and distinct furrow 

 (Fig. 3). These systems were less common than the 

 isolated pits or deeper furrows, and were found only 

 at Cape Nome. Species excavated in the pit-furrow 

 systems were primarily Mya truncata and 5. groen- 

 landicus. 



Table 2. — Morphological differences between the 

 three major types of excavations of the walrus. Means 

 and 95" confidence limits (sample size). 



Diameter 



or width 



(cm) 



Depth 

 (cm) 



Mya truncata pit 30+1(30) 32+3(10) 



Serripes groenlandicus pit 14+2 (9) 11 ±3 (14) 



Macoma spp. furrow 45±3 (7) 1 7±2 (7) 



Pits 



Three distinct types of pits were made in excavating 

 Mya truncata, Macoma spp., and Serripes groenlan- 

 dicus. The pits differed in diameter or depth (Fig. 2, 

 Table 2), reflecting a species position in the sedi- 

 ment. 



Mya pits had a deep central shaft (Fig. 2). Divers 

 readily identified these pits by thrusting a fist into a 

 shaft. Mya truncata has a long, tough siphon and lives 

 deep in the sediment (about 30 cm). Eighty-nine per- 

 cent (n = 190) of the Mya pits contained only a Mya 

 shell within 1 m of the pit. 



Macoma pits were similar in diameter to Mya pits, 



Shells 



The shells of primarily three groups of bivalves, 

 Mya truncata, Serripes groenlandicus, and Macoma 

 spp., were observed on the sea floor (Fig. 4). Macoma 

 and Serripes shells were commonly attached at the 

 umbus. About 6 to 8% of the shells from these groups 

 were broken, while 187c of the Mya shells were bro- 

 ken (Table 3). Greater breakage of Mya shells ap- 

 peared to be related to shell hardness (Table 3), and 

 not necessarily to a different feeding method. The 

 outer lining of the siphon (the periostracum) was at- 

 tached to 83 % of the Mya shells (n = 65) , and the di s- 

 tal end of the siphon commonly was intact. 



TABLE 3. — Percentage of broken shells found in major prey 

 species (based on number of reconstructed whole shells), 

 and an index of shell hardness. 



No of 

 bivalves 



Percent 

 broken 



Shell 

 hardness 1 



Mya truncata 96 



Macoma spp. 32 



Serripes groenlandicus 1 54 



78 

 6 

 8 



4.911.7 

 7.2±2.3 

 7.2+.1.2 



1 Pounds of pressure required to break a single valve Means and 95% con- 

 fidence limits in 10 trials. 



504 



