FISHERY BULLETIN; VOL. 87, NO. 4, 1989 



July. There was no noticeable trend over years 

 with regard to areas where legs were consumed 

 versus areas where whole red king crab were 

 eaten. The percentage by weight of legs in the 

 diet of cod at each station was generally less than 

 25% for all 3 years. The percentage of whole king 

 crab eaten at each station seemed to decrease 

 from 1981 to 1985. 



The geogr-aphic distribution of C. opilio and C. 

 bairdi in Pacific cod stomachs during 1981, 1984, 

 and 1985 are shown in Figure 3. The figure also 

 shows the approximate ice edge location before 

 the final ice retreat for each year. In 1981, small 

 percentages (<25% by weight) of C. bairdi were 

 found in Pacific cod stomachs around the Pribilof 

 Islands and also at bottom depths of 50-100 m in 

 the area southeast of the Pribilofs. In contrast, 

 this species was consumed over a much broader 

 area of the southeastern Bering Sea shelf during 

 1984 and 1985. Chionoecetes bairdi also ap- 

 peared in stomach contents of cod caught near 

 the shelf edge at 200 m, even in areas northwest 

 of the Pribilof Islands. This species was en- 

 countered in stomach contents throughout the 

 sampling period of May through September. 



In general, consumption of C. opilio did not 

 overlap much geographically with the areas 

 where C. bairdi were consumed, except near the 

 Pribilof Islands. Most C. opilio were eaten north 

 of the Pribilof Islands in a broad band encom- 

 passing depths from 35 to 200 m, although the 

 highest percentages by weight in Pacific cod 



stomachs seemed to be in the middle shelf area 

 with bottom depths of 50-100 m. In 1981, high 

 percentages by weight of C. opilio appeared in 

 cod diet north of 59°00'N, corresponding with 

 the location of the ice edge before its retreat in 

 that year. While the southward extension of pre- 

 dation appeared to go down to 57°30'N in 1984, 

 the percentages by weight in the diet were not 

 as high as in 1981. In 1985, cod diets were com- 

 posed of fairly high percentages by weight of C. 

 opilio as far south as 56°30'N; the ice edge in 

 that year was at approximately the same lati- 

 tude. 



Differences in Diet Composition Within 

 Areas by Year and Cod Size 



Results of logistic regi'ession of the frequen- 

 cies of occurrence for each crab species in the 

 two Pacific cod size groups for 1981, 1984, and 

 1985 are presented in Table 2. The most signifi- 

 cant relationship for describing Pacific cod con- 

 sumption of whole red king crab was Pacific cod 

 size: cod larger than 60 cm contained whole red 

 king crab more frequently than cod 30-59 cm in 

 length. Interannual differences in frequency of 

 occurrence of red king crab in stomachs were 

 also significant, showing a decrease in occur- 

 rence from 1981 to 1985. Percentages by weight 

 and frequency of occurrence of red king crab in 

 cod stomachs followed similar year and size 

 trends (Fig. 4). 



Table 2. — Results from logistic regression of frequencies of occurrence of each crab 

 species against year (1981, 1984, and 1985) and Pacific cod size group (Size 1 = 30-59 

 cm, Size 2 = ==60 cm). 



'Main factors chosen (P < 0.05) are shown in order of entry into the model. 

 ^Determined from significance and sign of between-level contrasts. 



^Chosen model has poor fit due to one anomalous cell: Size 2 in 1981 had much lower frequency of 

 occurrence of C ba/rdi than other cells. (Year was not significant.) 



812 



