Abookire: Biology, spawning season, and growth of Glyptocephalus zachirus 



353 



sification from Hunter et al. (1992). Females were first 

 classed as active or inactive by using the following 

 criteria. Through histological analysis, ovaries that 

 contained a sufficient number of AY/MN/HY oocytes 

 for one spawning were classed as active. Active females 

 were then classed as either spawning or nonspawning. 

 Ovaries without AY oocytes or with major atresia of AY 

 oocytes were classed as inactive. Inactive females were 

 first classed as nonspawning or postspawning and then 

 inactive nonspawning females were further classed as 

 mature or immature (Table 1). Active spawning females 

 showed evidence of past spawning (POP present) or 

 imminent spawning (MN or HY present), whereas active 

 nonspawning females had no evidence of recent or immi- 

 nent spawning but were presumed capable of spawning 

 in the near future. Females with ovaries containing 

 oocytes in early stages of vitellogenesis (EY oocytes 

 present) were considered mature but inactive. Immature 

 females had ovaries without vitellogenic oocytes. 



The fraction of active females and the fraction of 

 postspawning females was calculated for each month. 

 The start of the spawning season was defined by the 

 first observance of a hydrated oocyte or a POF and the 

 end of the spawning season was defined when the last 

 female with hydrated oocytes was observed. A one- 

 way ANOVA was used to test for differences in the 

 ovary wall thickness of mature females among months. 

 Assumptions of homogeneity of variances and normal 

 distribution of observations were met for the ANOVAs. 

 Bonferroni all-pairwise multiple comparison tests were 

 used after the ANOVAs to test for differences among 

 monthly mean values. Alpha was set at 0.05 for all tests 

 of significance. 



To estimate the length and age at which 50% of the 

 female rex sole were mature (ML^Qand MA^,,,), I used 

 the logistic regression model: 



Pmat = 1 / (1 -H 



+fcL 



where Pmat = the fraction of mature females per 15-mm 

 length-class; and 



L = total length in millimeters. 



Similarly, for estimating MA5o,age (in years) was substi- 

 tuted for length in the above equation. In each case, the 

 equation was solved for Pmat = 0.5 to obtain the length 

 and age at 50% maturity. The equation for the upper and 

 lower 95% confidence limits around Pmat when Pmat 

 = 0.5 was solved to yield 95% confidence limits around 

 ML,-|iand MA-,,. Results were compared with existing 

 length-at-maturity data for female rex sole off the Oregon 

 coast by reconstructing a length-at-maturity logistic 

 regression model from data presented by Hosie and 

 Horton (1977), taking the log transformation of both the 

 GOA and Oregon logistic regression curves to make them 

 linear, and then comparing the slopes of the two lines 

 according to Zar (1999). Direct statistical comparison 

 with existing age-at-maturity data (e.g., the age at 50% 

 maturity, and age at 100% maturity) for female rex sole 

 off Oregon (Hosie and Horton, 1977) was not conducted 

 because Hosie and Horton (1977) did not present stan- 

 dard error or logistic regression equations. Instead, the 

 probability that a specific-size (or specific-age) female rex 

 sole off the coast of Oregon would be considered mature 

 in the GOA was calculated as the fraction of mature 

 GOA females of the specified length divided by the GOA 

 sample size of the specified length (Zar, 1999). 



Size and growth 



The relationship between weight and length for female 

 rex sole in the GOA was estimated with the equation 



