666 



Fishery Bulletin 98(4) 



areas were not statistically distinguishable. Males from 

 the Columbia area grew more slowly and reached larger 

 sizes, on average, than males from the Vancouver, Eureka, 

 and Monterey areas (Fig. 2A). Estimates of L^ and K for 

 the Columbia area differed from estimates for the com- 

 bined Vancouver, Eureka, and Monterey areas by S'/t and 

 -26%, respectively. Parameters of the alternative gi'owth 

 curve were more precisely determined with CVs of 1-3% 

 for the Columbia and combined areas in comparison to the 

 standard von Bertalanffy model where CVs were 2-14% . 

 Residual variance for the Columbia area was lower than 

 that for the three areas combined. However, residuals 

 from the Columbia area curve did not satisfy the normal- 

 ity assumption (P=0.01), whereas residuals from the com- 

 bined areas did (P=0.83). Overall, there was no latitudinal 

 cline in growth parameters. 



Similarly, we found geographic variation in growth of 

 female Dover sole by area (Table 3). The hypothesis of 

 identical female growth parameters across areas was 

 rejected (P<0.001, X"=91.70). Female samples from adja- 

 cent INPFC areas were grouped and growth cun'es were 

 estimated. Homogeneous growth was rejected for Van- 

 couver and Columbia samples (P=0.002, X"=14.92) and 

 for Columbia and Eureka samples (P<0.001, c = 56.90) 

 but was accepted for Eureka and Monterey samples 

 (P=0.09, x^=6.45). Homogeneous growth was rejected for 

 combined samples from Eureka, Monterey, and Columbia 

 (P<0.001, X"=89.15) and weakly supported for combined 

 samples from Eureka, Monterey, and Vancouver {P=0.025, 

 X-=14.47). 



Growth of female Dover sole exhibited geographic varia- 

 tion (Fig. 2B). In particular, females from the Columbia 



