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



molt during winter. However, our winter sam- 

 pling coverage has been limited and has not de- 

 tected this type of occurrence. 



With the exception of 1981, when stomach 

 analysts did not consistently record snow crab 

 sex, the proportions of female juvenile snow 

 crabs in stomach contents were close to 0.5 and 

 were not significantly different from a 1:1 ratio 

 of females to males in 1984. Adams (1979) re- 

 viewed the hterature on C. opilio and found that 

 the early life history of males and females are 

 similar with respect to size, growth, distribu- 

 tion, and habitat. Brethes et al. (1987) found 

 sex ratios of C. opilio <30 mm CW in the north- 

 west Atlantic to be 1:1 with no spatial segi'ega- 

 tion of sexes. The proportions of female snow 

 crabs <95 mm CW in the NMFS assessment 

 surveys were close to 0.5 in most years. Thus, 

 cod do not appear to be selecting snow crab on 

 the basis of sex and are probably preying ran- 

 domly on individuals on the basis of crab size. 



Pacific Cod Population Consumption of 

 Crab 



Daily rations that were derived using the 

 Ursin et al. (1985) model for incorporating prey 

 size effects on gastric evacuation rate appear 

 reasonable compared with estimates of daily ra- 

 tion for Atlantic cod from areas with bottom 

 temperatures higher than temperatures in the 

 eastern Bering Sea. My estimates ranged from 

 0.47 to 0.86% BWD; and estimates for Atlantic 

 cod of similar sizes range from 0.5 to 1.0% BWD 

 in the North Sea (Daan 1973), 1.6 to 2.0% BWD 

 in the Faroe plateau (Jones 1978), and 0.5 to 

 1.9% BWD on Georges Bank (Durbin et al. 

 1983). Livingston et al. (1986) calculated daily 

 ration for Pacific cod using the Elliott and 

 Persson (1978) model without correction for prey 

 size effects using a subset of the data presented 

 in this paper and obtained values of 0.31% BWD 

 for cod <55 cm FL and 1.30% BWD for cod >55 

 cm FL. When compared with cod gi-owth data, 

 however, the rations for small cod were too small 

 to account for gi'owth, and rations for large cod 

 were too large. The current approach seems to 

 coiTect for the deficiencies in the previous esti- 

 mates, and produces ration values which are not 

 so divergent for the two cod size groups. 



Other parameters involved in estimating pop- 

 ulation consumption are also subject to error: 

 predator biomass estimates, the percentage of 

 prey items in the predator's diet, and the num- 

 ber of days the crab species is vulnerable to 



predation. NMFS survey estimates of cod bio- 

 mass in recent years have 95% confidence inter- 

 vals of 12-18% of the mean biomass estimate 

 (Thompson and Shimada 1987), a minimum confi- 

 dence interval because of the assumptions of 

 complete vulnerabihty and catchability of cod to 

 bottom trawls. Cod also performed seasonal 

 onshore-offshore migrations (Wespestad and 

 Shimada 1984); a factor not taken into considera- 

 tion here that could change the biomass of cod in 

 a particular area over the time period in this 

 study. Errors in diet composition parameters 

 can arise from insufficient sample sizes, uneven 

 spatial distribution of samples, and possible diet 

 changes over space and time scales not con- 

 sidered in this study. Sampling effort for cod 

 stomachs was widely distributed over the whole 

 shelf area during 1981 and 1985, but some areas 

 were not sampled well during 1984 (Fig. 1). I 

 have attempted to reduce bias in diet composi- 

 tion estimates that may arise from uneven 

 stomach sample sizes within areas by averaging 

 diet percentages estimated for each 20 nmi wide 

 square where stomachs were sampled within a 

 crab consumption ai'ea. A similar practice has 

 been adopted in the North Sea stomach sampling 

 program, which provides diet composition esti- 

 mates for a multispecies virtual population anal- 

 ysis model (Mehl 1986). In this study, I have 

 estimated consumption that occurred only dur- 

 ing May through September, so the estimated 

 numbers of snow crab consumed by cod apply 

 only to that portion of the year. Livingston et al. 

 (1986) has shown that Pacific cod in the eastern 

 Bering Sea consume snow crab throughout the 

 whole year; therefore, the estimates presented 

 in this paper can be considered mainly as indexes 

 of the total numbers consumed by cod. There is 

 also great uncertainty about the size at age for 

 crabs and the allocation of crab size gi'oups to 

 age classes should be considered approximate. 



Impact on Crab Populations 



Predation mortality rate of a prey population 

 must increase with prey population size in order 

 to demonstrate that a predator population is reg- 

 ulating the size of the prey population (Helling 

 1959). NMFS resource assessment surveys pro- 

 vide annual estimates of female red king crab 

 abundance, which can be compared with the 

 total cod population removals of female red king 

 crab in the same year (Fig. 10). Both the female 

 red king crab population and the estimated re- 

 movals from the population by cod follow the 



822 



