SOMERTON and OTTO: DISTRIBUTION AND REPRODUCTION OF GOLDEN KING CRAB 



Table 2. — Number of trawl hauls sampled (ex- 

 cluding hauls without crabs) and number of 

 crabs sexed and measured by U.S. observers 

 aboard Japanese small trawlers within the 

 study area during 1981 and 1982. Data are 

 summarized by depth, latitude and month. 



data for each of the two depth strata within the 

 northern and central subareas to help illustrate 

 depth and latitude trends in the size distributions 

 (see Figure 2). Potential bias because of the varia- 

 tion of fishing effort with depth was minimized by 

 first partitioning the data into 100 m depth inter- 

 vals. Within each depth interval, a size-frequency 

 distribution and an average catch per hour (CPH) 

 were calculated. Size-frequency distributions, 

 weighted by the appropriate mean CPH, were then 

 summed over all 100 m depth intervals within each 

 of the two depth strata. 



Variations in mean size, CPH, and proportion 

 male with latitude and depth were also examined 

 using multiple regression. Two normalizing trans- 

 formations were used: 1) CPH was transformed to 

 the natural log scale and 2) proportion male was 

 transformed to the arcsine-square root scale after 

 replacing with 0.25/N and 1 with (N - 0.25)/JV, 



where N is the number of crabs within each trawl 

 haul (Bartlett 1947). 



Egg size was estimated by randomly selecting 10 

 eggs from each preserved egg mass and measuring 

 their maximum lengths (eggs are oval) to the nearest 

 0.1 mm with an ocular micrometer. The remainder 

 of each egg mass was air dried and, after separating 

 the eggs from the pleopods and setae, weighed to 

 the nearest 0.1 mg. Two subsamples of about 200 

 eggs each were randomly selected from each dried 

 egg mass and then weighed and counted. Fecundity 

 was then estimated by dividing the total weight of 

 an egg mass by the average of the two estimates 

 of individual egg weight that were obtained from 

 that egg mass. 



Male size at maturity was estimated from the 

 allometric growth of the right chela. When king crab 

 chela measurements are plotted against carapace 

 measurements on log-log axes, the data conform to 

 two straight lines that intersect at the average 

 carapace length at maturity (see Figure 3) (Somer- 

 ton 1980; Somerton and Macintosh 1983). To 

 estimate this size, we used the computer method 

 described in Somerton and Macintosh (1983) which 

 fits a pair of intersecting straight lines by iteratively 

 varying the carapace length at the intersection point 

 until the residual sum of squares about the lines is 

 minimized. Variance of the male size at maturity 

 was estimated using a computer technique known 

 as bootstrapping (Efron and Gong 1983). In our ap- 

 plication, the method consisted of randomly choos- 

 ing, with replacement, 50 subsamples equal in size 

 to the original data set. For each subsample, the size 

 at maturity was estimated by fitting the two line 

 model. Variance of the estimated size at maturity 

 was then computed as the variance among the 50 

 independent estimates. 



Although we attempted to detect and exclude par- 

 tially regenerated chelae in the field, we were not 

 always successful. Measurements from partially 

 regenerated chelae can increase the variance of 

 estimates of male size at maturity; therefore, these 

 measurements were removed from the data set 

 before analysis using a sequential outlier elimina- 

 tion technique described in Somerton and Macin- 

 tosh (1983). 



Golden king crab females were considered to be 

 mature, if they had eggs or empty egg cases at- 

 tached to the pleopod setae. Although we are not 

 certain that this is always true, for red and blue king 

 crabs, adult females extrude eggs soon after every 

 molt and the empty egg cases remain attached to 

 the pleopod setae until the next molt (Somerton and 

 Macintosh 1985). 



573 



