FISHERY BULLETIN: VOL. 84, NO. 3 



on the combined 1981 and 1982 observer data, the 

 proportion of males was 0.51 (N = 1,375) in the cen- 

 tral area and 0.43 (N = 8,271) in the northern area. 

 Since the proportion of males still differed signifi- 

 cantly between areas (2x2 contingency table, x 2 

 = 30.7, df = 1, P < 0.001), it is unlikely that the 

 change in the proportion of males was due to size 

 selectivity. Furthermore, since the difference in the 

 proportion of males appears to be established before 

 maturity, biological explanations such as sexual dif- 

 ferences in migratory behavior or natural mortal- 

 ity are also unlikely. 



Although we cannot explain the latitudinal varia- 

 tion in the proportion of males, we believe that the 

 depth variation, especially the abrupt increase in the 

 proportion of males in the shallowest depth zone, 

 is due to sexual segregation. Sexual segregation by 

 depth has been observed for another slope-dwelling 

 crab, C. tanneri (Pereyra 1966). Adult female C. tan- 

 neri occur within a rather narrow depth zone 

 throughout the year while adult males undergo a 

 seasonal migration from relatively shallow water in 

 summer to the deeper water occupied by females 

 during the winter mating period. To determine if 

 golden king crab have a similar seasonal migration, 

 we examined the proportion of males from the 

 northern area at depths <400 m (the northern area 

 had nearly equal sampling in all four seasons). Using 

 pooled 1981 and 1982 data, analysis of variance 

 showed that the proportion of males did not vary 

 significantly between seasons (F = 0.13, df = 3, 179, 

 P > 0.05). Although adult males of golden king crab 

 probably congregate in somewhat shallower water 

 than adult females, unlike C. tanneri this segre- 

 gation appears to be maintained throughout the 

 year. 



REPRODUCTIVE BIOLOGY 



significance. Since the coefficient was not signifi- 

 cantly different from zero (F = 3.85, df = 1, 57, 

 P = 0.06), we chose a linear relationship to describe 

 the data. 



Fecundity-size relationships for females with 

 uneyed embryos (N = 46) and eyed embryos (N = 

 19) from the central area were compared to deter- 

 mine whether the relationships changed with stage 

 of embryo development. Analysis of covariance 

 showed that the slopes did not differ (F = 0.77, df 

 = 1, 61, P = 0.38) but that the intercept for eyed 

 embryos was significantly less (F = 4.89, df = 1, 

 62, P = 0.03) than that for uneyed embryos. At 114 

 mm, the median size of adult females in all areas 

 combined, uneyed clutches were 18% greater than 

 eyed clutches. Similar to other crab species (Wear 

 1974), golden king crab lose a significant number 

 of embryos between egg extrusion and the appear- 

 ance of embryonic eyes. 



Fecundity-size relationships were then compared 

 between the northern (N = 59), central (N = 46), 

 and southern (N = 24) areas considering only those 

 clutches with uneyed eggs. Analysis of covariance 

 showed that the slopes did not differ (F = 0.74, df 

 = 2, 123, P = 0.48), but the intercepts differed 

 significantly between areas (F = 4.38, df = 2, 125, 

 P = 0.01). Pairwise i-tests indicated that southern 

 and central intercepts did not differ (P = 0.99) from 

 each other, but that both differed significantly (P 

 = 0.01, P = 0.04) from the northern intercept. Data 

 from the southern and central areas were therefore 

 pooled and compared with those from the northern 

 area. Again, the slopes did not differ (F = 1.25, df 

 = 1, 125, P = 0.27), but the northern intercept was 

 significantly greater (F = 8.83, df = 1, 126, P = 

 0.004) than the pooled central and southern inter- 

 cept. Assuming equal slopes, the resulting fecundity- 

 size relationships are 



Fecundity 



Fecundity-size relationships for golden king crab 

 were estimated stagewise by examining 1) the form 

 of the relationship, 2) whether the relationships 

 varied with stage of embryo development, and 3) 

 whether the relationships varied between areas. 



The fecundity of king crabs has been reported to 

 increase as either a linear (Haynes 1968) or a curvi- 

 linear (Somerton 1981b) function of carapace length. 

 To determine which form was more appropriate for 

 golden king crab, a second degree polynomial was 

 fitted to the fecundity and size data from the north- 

 ern area (all clutches contained uneyed embryos) and 

 the coefficient of the quadratic term was tested for 



Northern 



Central-southern 



E = -24815 + 323 CL 

 (N = 59, R 2 = 0.79) 



E = -26145 + 323 CL 

 (N = 68, R 2 = 0.74) 



where E is number of uneyed eggs and CL is cara- 

 pace length in millimeters. Females from the north- 

 ern area carry, on average, 1,330 more eggs than 

 equal-sized females from the central and southern 

 areas. For 114 mm females, this represents a 12.6% 

 difference in fecundity. 



Northern females may be more fecund than equal- 

 sized central and southern females because they are 

 older and size-specific fecundity often increases with 

 age (Pianka and Parker 1975). But, it is also likely 



580 



