584 



Fishery Bulletin 94(3), 1996 



The observed catches by length were compared to 

 estimated catches by length from the model fit with 

 A = 1 and with constant M of 0.3, 0.4, and 0.5 (Fig. 

 81.A11 three fits of the model produced similar catches 

 by length and fitted the observed data very well, ex- 

 cept data for 1965 and 1982. The poor fit in 1965 

 was mainly caused by the observed catch-length com- 

 position in the initial year (19641. The rapid expan- 

 sion of fishing grounds during this period made the 

 catch-length compositions inconsistent between these 

 two years. The catch in 1982 was extremely low in 

 comparison to the past annual catches. 



The fishing effort was very high during the mid- 

 1960's and the early 1980s (Fig. 9). Like Bristol Bay 

 RKC, larger weighting factors produced estimates 

 of fishing effort closer to the observed effort. With 

 A = 1, a constant M of 0.3 fitted the fishing effort 

 much better than a constant M of 0.5 (Fig. 9). 



