124 



Fishery Bulletin 90(1). 1992 



We combined data from Oregon and California in this 

 analysis because sample sizes (before spawning) were 

 inadequate for application of the model separately. 

 However, in Table 17 we provide estimates and fitting 

 parameters based on definition IV for each region as 

 well as those based on the model using the combined 

 data. The ML50 estimated for definition IV was lower 

 in California than in Oregon. However, analysis of co- 

 variance of the log transformation of fraction mature, 



Ln + 1 , on length and locality indicated that 



the difference between states was not significant (P 

 0.625, F 0.26). Thus our ML50 estimate for Dover sole 

 along the California and Oregon coasts is 332 mm with 

 95% CI of 315-349 mm (Carter et al. 1986; Fig. 11). 

 The ML50 we estimated from data in Hagerman (1952) 

 for Dover sole from the Eureka California fishery is 

 high (363 mm) compared with our final estimate for 

 Oregon and California coasts (Fig. 11). However, 

 Hagerman collected his specimens during the spawn- 

 ing season, and his estimate is similar to the ML50 for 

 females taken during the spawning season (Definition 

 IV, Table 16). 



Discussion 



Validation of fecundity assumptions 



In the Introduction, we specified four assumptions re- 

 quired for an unbiased estimate of annual fecundity in 

 Dover sole. These assumptions were that (1) fecundity 

 was determinate; (2) potential annual fecundity was 

 equivalent to actual fecundity; (3) females used to 

 estimate annual fecundity had not spawned; and (4) 

 recruitment of oocytes into the advanced stock of 

 yolked oocytes had ceased for the season. The follow- 

 ing is a review of the evidence for the four assumptions. 

 Five lines of evidence support the assumption of 

 determinate fecundity for Dover sole: (1) in mature 



ovaries (mean diameter of advanced oocytes >0.85 

 mm), a hiatus existed between the advanced stock 

 of mature oocytes and smaller, less mature oocytes; 



(2) total fecundity declined over the spawning season; 



(3) total fecundity was lower in females having 

 postovulatory follicles; (4) the mean diameter of the 

 advanced oocytes increased over the spawning sea- 

 son; and (5) our analysis of the order of spawning 

 batches was consistent with the determinate fecundity 

 assumption. 



The second assumption, lack of significant atresia, 

 also was supported by our analysis. Overall, atretic 

 losses of advanced oocytes were negligible during the 

 years of our study. Multiple regression analysis in- 

 dicated that atresia had a small but significant effect 

 on total fecundity of the California females that had 



