714 



Fishery Bulletin 100(4) 



o 

 O 



15-, 



10 



5- 



58 



 Non-vitellogenic 

 n Vitellogenic 



August Samples 



58 118 178 238 298 358 418 478 538 



September Samples 



538 



118 178 238 298 358 418 478 

 Mean maximum oocyte diameter (nm) 

 Figure 2 

 Comparison of mean maximum oocyte diameters (pm, 

 see text) for petrale sole with vitellogenic and nonvitel- 

 logenic oocytes, for samples collected m August and 

 September 2000. 



around 180 \im suggested that by September, the micro- 

 scopic classification of ovaries provided an accurate identi- 

 fication of maturity status of individual fish. Immature fish 

 were well represented in the September collections, indi- 

 cating the fish were still well mixed on the feeding grounds 

 (Table 2). Therefore, the September maturity data gener- 

 ated from microscopic inspection were used to fit age- and 

 length-at-maturity relationships for female petrale sole 

 (Fig. 3). The resulting curves fitted the data well and chi- 

 square tests on the residuals indicated no problems with 

 lack of fit caused by overdispersion (Pearson chi-square of 

 5.714 Iwith 20 degrees of freedom] for length, and 0.643 

 [with 11 degrees of freedom] for age, P>0.999). The fitted 

 relationships indicated that female petrale sole off the cen- 

 tral Oregon coast were 50% mature at about 33 cm and at 

 about 5 years of age. Complete maturity was obtained at 

 about 40 cm and about 9 years of age (Table 3). 



Examination of the monthly samples that, as an ag- 

 gregate, were used to fit the 1986-91 maturity curve 

 showed that numerous samples were included from 

 the IVIarch through August period, especially from the 

 port of Charleston, Oregon (Table 4). The inclusion of 

 these summer samples explains some of the flatness in 

 the fitted maturity curve from Turnock et al.'^ (Fig. 4, 

 L5Q=30.6, slope=0.29). A maturity curve fitted to these da- 

 ta, but excluding spring and summer samples, was much 

 steeper (Fig. 4, Lr,Q=33.56, slope=0.50). 



Inspection of the 1986-91 and 1992-2000 maturity 

 data from the winter months, when errors in determin- 

 ing maturity should be minimized (Tables 4 and 5, Fig. 5), 

 showed some evidence of the type of bias that can result 

 from sampling spawning aggregations. First, very few im- 



