164° 



160° 



Figure i. 



156° 152° 148° 



-Geographic subdivisions of Gulf of Alaska study area. 



144° 



considering marine mammal populations as well 

 as fishery resources. 



Methods 



A total of 548 harbor seals were collected by rifle 

 throughout the Gulf of Alaska from 1973 through 

 1978 (Table 1). Reasonably complete seasonal 

 coverage was obtained for Prince William Sound 

 and the Kodiak area. Stomach contents were 

 removed in the field, wrapped in muslin, and 

 preserved in 10% Formalin.^ In the laboratory the 

 volumes and number of occurrences (number of 



Reference to trade names does not imply endorsement by 

 the National Marine Fisheries Service, NOAA. 



Table l. — Geographic and seasonal distribution of harbor seals 

 collected in the Gulf of Alaska. 



stomachs in which a prey species was found) were 

 determined for prey species. Because digestion 

 was often advanced, skeletal materials, partic- 

 ularly fish otoliths and cephalopod mandibles 

 (beaks), were the primary criteria for identifica- 

 tion (Fitch and Brownell 1968; Pinkas et al. 1971). 



Otoliths and other skeletal components from 

 fish were tentatively identified to the lowest 

 taxonomic level possible by comparison with ref- 

 erence materials. Otolith identifications were ver- 

 ified by John E. Fitch, California Department of 

 Fish and Game, Long Beach. Cephalopod beaks 

 were classified as either squid or octopus with the 

 aid of Pinkas et al. (1971), and squid beaks were 

 identified to family by Clifford H. Fiscus, National 

 Marine Fisheries Service, NOAA, Seattle, Wash. 

 Decapod crustaceans were identified by Kathryn 

 J. Frost and Lloyd F. Lowry, Alaska Department 

 of Fish and Game, Fairbanks. 



In order to integrate data on both frequency of 

 occurrence and prey volumes into a single ranking 

 of prey utilization I used a modified form of the 

 Index of Relative Importance (IRI)^ devised by 



^Original Index of Relative Importance as derived by Pinkas 

 et al. (19711 was calculated by summing the numerical and 

 volumetric percentage values and multiplying by the frequency 

 of occurrence jjercentage value. 



545 



