FISHERY BULLETIN: VOL. 79, NO. 3 



FIGURE 1.— Portion of the Gulf of Alaska 

 where Steller sea lion prey utilization 

 studies were conducted. 



TABLE 1. — Comparison of seasonal and geographical distributions of 548 harbor seals and 250 Steller sea lions collected in the Gulf of 

 Alaska. Numbers in the table are percentages of the samples taken during a particular season or in a geographical area. 



' Includes 45 harbor seals from the Copper River Delta. 



Baines 1966). For all other samples, volumes and 

 occurrences of the various prey categories were 

 determined in the laboratory. Prey identifications 

 were based primarily on skeletal materials, par- 

 ticularly fish otoliths and cephalopod mandibles 

 (beaks) (Fitch and Brownell 1968; Pinkas et al. 

 1971). Otoliths and other skeletal components 

 from fish were identified to the lowest taxon possi- 

 ble by comparison with reference materials. 



Food habit data were organized and examined 

 as percentage of occurrence (number of stomachs 

 in which a prey item occurred/total number of 

 stomachs with food) and percentage of total vol- 

 ume (total volume of a prey item/total volume of 

 all stomach contents). Confidence intervals for 

 percentages of occurrence were calculated from 

 tables presented by Rohlf and Sokal (1969). In the 

 percent occurrence analysis, unimportant small 

 but numerous organisms may be dispropor- 

 tionately evident (Perrin et al. 1973) as may 

 species which have hard parts which resist diges- 

 tion (Fiscus and Baines 1966). Volumetric analyses 

 are distorting because various organisms are di- 

 gested at different rates and because contents are 

 at different stages of digestion when the collec- 

 tions are made (Perrin et al. 1973). 



A combination rank index (CRD was devised to 

 integrate volumetric and occurrence data into a 

 single indicator of prey use. Each prey category 

 was ranked in descending order of percentage of 

 occurrence and percentage of volume. The two 

 rankings for each prey category were multiplied 

 together to produce the CRI. 



To compare sizes of walleye pollock, Theragra 

 chalcogramma, eaten by Steller sea lions and har- 

 bor seals, random samples of walleye pollock 

 otoliths recovered from stomachs were measured 

 (total length). Fork lengths of the fish were then 

 estimated, using a formula derived from regres- 

 sion analysis of otolith length and fish length 

 (Frost and Lowry 1981). Estimated mean fork 

 lengths are given in Table 2. 



Table 2. — Estimated mean fork lengths based on otolith 

 lengths, of walleye pollock eaten by Steller sea lions and harbor 

 seals. 



Item 



Sea lions Harbor seals 



Otolith measurements, no. 

 Estimated mean fork length, cm 

 standard deviation, cm 

 Range, cm 



2,030 



29.8 



11.6 



5.6-62.9 



2,180 



19.2 



9.6 



4.2-53.2 



468 



