FISHERY BULLETIN: VOL. 84, NO. 3 



1 40 1 



30 



80 



-• — MALE 

 -— FEMALES 



1 1 1 1 



SOUTH CENTRAL NORTH 



AREA 



Figure 4.— For both sexes of golden king crab, estimated sizes 

 at maturity, and their 95% confidence intervals, are plotted against 

 area. 



areas. To test whether this is true, chela height and 

 carapace length relationships for adult males were 

 compared between areas. Analysis of covariance 

 showed that the slopes did not differ (F = 0.14, df 

 = 2, 1,998, P = 0.87), but the intercepts differed 

 significantly (F = 146.7, df = 2, 2,000, P < 0.001). 

 Pairwise i-tests further showed that each intercept 

 differed significantly (Bonferroni critical values; 

 maximum P < 0.05) from the other two and, similar 

 to the weight-size relationships, that the intercepts 

 progressively increased with latitude. Thus north- 

 ern males, which are the heaviest, have the largest 

 chelae. 



By itself, chela size is unlikely to be responsible 

 for latitudinal differences in weight because chela 

 weight is only a small proportion of total body 

 weight. However, chela size may be correlated with 

 other body dimensions (for example, length of walk- 

 ing legs) that also increase relative to carapace 

 length at maturity. We therefore used chela height 

 as a proxy for these dimensions and examined 

 whether the difference in chela height could account 

 for the difference in weight-size relationships. This 

 was done by comparing the weight-size relationships 

 between areas including the logarithm of chela 

 height as a covariate. Two additional modifications 

 of the previous weight-size comparison were made. 

 First, since weights and chela measurements were 

 not obtained from the same crabs in the southern 

 area, the comparison was restricted to the northern 

 and central areas. Second, since chela height and 



carapace length are linearly related only over the 

 adult (or juvenile) size range, the comparison was 

 restricted to males greater than the size at matur- 

 ity in each area. When the northern (N = 129) and 

 central (N = 614) areas were compared consider- 

 ing only carapace length as a covariate, the slopes 

 were not significantly different (F = 0.06, df = 2, 

 739, P = 0.81), but the intercepts were significant- 

 ly different (F = 7.36, df = 1, 740, P = 0.007). When 

 chela height was included as a covariate, however, 

 neither the slopes (P = 0.316) nor the intercepts (P 

 = 0.430) differed significantly between areas. This 

 indicates that latitudinal changes in chela size, and 

 perhaps other body measurements that also increase 

 at maturity, account for the observed latitudinal in- 

 crease in body weight. 



Juvenile weight-size relationships were also com- 

 pared between the northern (N = 10) and central 

 (N = 207) areas and neither the slopes (F = 0.06, 

 df = 1, 213, P = 0.938) nor the intercepts (F = 0.19, 

 df = 1, 214, P = 0.664) were significantly different. 

 The weight-size relationship for male golden king 

 crabs is therefore described by one equation for 

 juveniles and three equations for adults. Trans- 

 formed back to a linear scale, these relationships 

 are 



Juveniles W = 0.000365 CL 3 - 099 (N = 217, R 2 = 0.88) 



Adults 



Northern W = 0.000225 CL 3206 (N = 139, R 2 



Central W = 0.000219 CL 3206 (N = 632, R 2 



Southern W = 0.000204 CL 3206 (N = 100, R 2 



0.93) 

 0.91) 

 0.91) 



where W is body weight in grams and CL is cara- 

 pace length in millimeters. Within the adult size 

 range, males from the northern area are 10.3% 

 heavier and males from the central area are 9.8% 

 heavier than equal-sized males from the southern 

 area. 



Relative Abundance and 

 Proportion Male 



Relative abundance, or catch per hour (CPH), 

 based on combined 1981 and 1982 observer data, 

 is shown by sex, latitude, and depth in Figure 5. 

 Linear trends in CPH with depth and latitude were 

 examined statistically using multiple regression 

 (depth and latitude were considered simultaneous- 

 ly; interaction was ignored). The latitude coefficient 

 for males was not significant (P > 0.05) in either 

 year, but the latitude coefficient for females was 

 positive and highly significant (P < 0.01) in both 

 years. The depth coefficient for males was negative 



578 



