LIVINGSTON: PACIFIC COD PREDATION ON THREE CRAB SPECIES 



of the eastern Bering Sea shelf in cooler years. 



Diet Composition by Year and Cod Size 



Analysis of frequencies of occurrence of red 

 king crab in Pacific cod stomachs through logistic 

 regi'ession shows that cod size is the most im- 

 portant factor in determining consumption of 

 whole red king crab; the frequency of occurrence 

 was significantly greater in cod larger than 60 

 cm FL. Blau (1986) found that Pacific cod, which 

 presumably had consumed soft-shell female king 

 crabs in the Kodiak region of the Gulf of Alaska, 

 ranged in size from 45 to 79 cm FL. That study, 

 however, did not examine differences in fre- 

 quency of occurrence within that size range, and 

 none of the crab eaten were whole. Interannual 

 differences in red king crab consumption were 

 significant in our study, decreasing from 1981 to 

 1985. NWAFC abundance estimates for the 

 female portion of the red king crab population 

 show a corresponding trend with a decrease in 

 numbers from 103.6 million in 1981 to 13.7 mil- 

 lion in 1985. This suggests that individual cod 

 predators responded to decreases in crab density 

 by consuming less crab. 



Although logistic regression of C. bairdi fre- 

 quencies of occurrence showed cod size as an 

 important variable, the model fit was poor due to 

 the low frequency of crab occurrence in cod s:60 

 cm FL in 1981. If the percentages by weight are 

 examined (Fig. 4), the size relationship looks 

 clearer. Smaller cod consistently ate more than 

 large cod, and the interannual changes within 

 size groups showed consistent, small increases 

 across years. NMFS surveys are not able to pro- 

 vide precise estimates of crab <40 mm CW, not 

 only because of escapement through trawl 

 meshes, but also because snow crab <100 mm 

 CW have a greater tendency to bury themselves 

 in bottom sediments (Conah and Maynard 1987), 

 reducing their vulnerability to trawl capture. If 

 cod respond to changes in C. bairdi density as 

 they appear to do for red king crab density, then 

 our data suggest the possibility of stable or 

 slight increases in juvenile C. bairdi population 

 numbers for crab <40 mm CW over the three 

 study years. The similarity of C. bairdi size dis- 

 tributions in cod stomachs for the three years 

 within each cod size gi'oup further supports the 

 suggestion of stable juvenile (<40 mm CW) C. 

 bairdi population size distributions over time. 



The model which best explains C. opilio con- 

 sumption by cod shows year as the most import- 

 ant variable (with consumption increasing over 



years) and size as the next important variable 

 (with larger cod consuming C. opilio more fre- 

 quently than smaller cod). If the occurrence of 

 juvenile crab <40 mm CW in cod diets can be 

 used as a measure of juvenile crab abundance in 

 the survey area, then our data show a probable 

 increase in juvenile C. opilio abundance from 

 1981 to 1985. Waiwood and Elner (1982) similar- 

 ly suggested that the increase in Atlantic cod, 

 Gadus morhua, predation on snow crab ob- 

 served in 1980 and 1981 in the Gulf of St. Law- 

 rence in the northwest Atlantic was due to in- 

 creased availability of small crabs. The size-fre- 

 quency distributions of C. opilio in cod stomachs 

 also indicate more C. opilio <35 mm CW in 1984 

 and 1985 than in 1981 at least in the survey area. 

 Somerton (1981) showed that large numbers of 

 C. opilio juveniles exist north of 61°00'N, an 

 area which was not sampled in this study. It is 

 possible that in colder years such as 1984 and 

 1985, an influx of juvenile C. opilio from these 

 northern areas could have entered southeastern 

 Bering Sea shelf waters and have become more 

 available to cod. Of course, there is a limit to the 

 distance juvenile crab can migi-ate in one year. 

 The observed downward shift in mean size and 

 increased numbers of C. opilio in cod stomachs 

 may not be the result of actual increases in 

 juvenile C . opilio abundance but might be the 

 result of progressive southerly shifts in the 

 geogi'aphic distribution of small juveniles in pro- 

 gi'essively colder years. More years of data need 

 to be examined, however, to determine what 

 happens to juvenile distributions over time, par- 

 ticularly between two consecutive years with 

 very different climatic conditions. 



Because of the large carapace lengths of red 

 king crab consumed by cod (50-160 mm CD, the 

 well-digested nature of red king crab in stom- 

 achs, the usual occurrence of whole crab only 

 around May, and the fact that one red king crab's 

 sex was determined to be female, we have as- 

 sumed that Pacific cod consume soft-shell fe- 

 males, which molt in Bristol Bay around April to 

 May (Hayes 1983). It seems highly unhkely that 

 cod could consume whole crab of those carapace 

 lengths in a hard-shell condition. Blau (1986) also 

 found cod consuming red king crab during the 

 king crab molting period in the Gulf of Alaska. 

 Because male red king crab molt in winter while 

 migrating to the mating grounds (Powell and 

 Nickerson 1965), our assumption that only soft- 

 shell females are consumed in spring seems sup- 

 portable. This does not rule out the possibility 

 that cod mav consume soft-shell males, which 



821 



