584 



Fishery Bulletin 100(3) 



weight (TW ±0.01 g; measured after each individual was 

 blotted dry) were recorded for all thawed, undamaged 

 individuals. Gonads were excised from a subsample of in- 

 dividuals (-80%) of each collection, identified macroscopi- 

 cally (and microscopically if needed) as ovaries or testes, 

 blotted dry on a paper towel, and weighed (±0.01 g). Sea- 

 sonal conversion equations for length measurements were 

 generated by least-squares regression (SAS, 1990a). 



Age and growth 



Sagittal otoliths for aging studies were excised, cleaned of 

 extraneous tissue, and stored dry in vials. Otolith length 

 was measured under a binocular dissecting microscope at 

 lOx with a calibrated ocular micrometer, and an otolith 

 length (mm) versus fish standard-length relationship 

 was derived by using least squares regression. For age 

 determination, one whole otolith (either left or right) was 

 placed on black velvet to enhance the visibility of opaque 

 zones, submerged in a 60% glycerin solution, and viewed 

 at 32x with reflected light. Otolith edges were classified as 

 either opaque or hyaline to determine time of opaque ring 

 formation. The number of opaque rings was considered to 

 equal the age in years of the pinfish. Age determinations 

 were repeated three times on all whole otoliths. If two of 

 the three age readings were in agreement, the value was 

 accepted. If age varied between all readings, the structure 

 was not included in the analysis. The accuracy with which 

 whole otoliths can be used to age pinfish was confirmed 

 by comparing whole otolith ages from a subsample of the 

 largest pinfish to age readings made on thin sections of 

 the same otoliths. Methods of Hood and Johnson (1999) 

 were used to section otoliths. To verify the consistency of 

 the opaque ring counts, a random subsample ( 13% ) of all 

 whole otoliths was aged once by an experienced second 

 reader and the results were compared to the final age 

 determination. Pinfish larvae are present off the west 

 coast of Florida between November and February (Darcy, 

 198.5); therefore a 1 January birthdate was assumed for all 

 pinfish. Age in years was designated with a decimal exten- 

 sion that represented the date of capture in days from the 

 1 January birthdate. 



Growth was modeled by fitting length-at-age for all 

 years combined to the von Bertalanffy equation: 



L„,=L„.,*(l-exp<-*''"^'-'"") + f,^, 



where L, , = the standard length of the / th individual at 



aget: 

 L^ = the asymptotic maximum length; 

 k = a growth constant; 

 Iq = the hypothetical age at which length is zero; 



and 

 £ s = independent, identically distributed, normal 



random errors. 



Parameters L, k, and /„ were estimated by using SAS non- 

 linear regression (Proc NLIN) with the Marquardt method 

 (SAS Institute, 1990b). Lack of fit was assessed by using 

 residuals plots (Bates and Watts, 1988). 



Growth-rate differences between sexes were investi- 

 gated by using an approximate randomization test to 

 compare growth curves (Helser, 1996). Essentially, differ- 

 ences were investigated by comparing an observed test 

 statistic to an empirical probability density function of a 

 test statistic under the null hypothesis of no difference. 

 The observed test statistic for pinfish was developed by fit- 

 ting the von Bertalanffy growth functions to length-at-age 

 data for sexes combined and to length-at-age for each sex 

 separately. The sums of the residual sum of squares from 

 the two sex-specific models were then used to calculate the 

 observed test statistic 



t(x, 



)=E"''-''''-X'''' 



-/ r 

 't.p' 



i.i.p 



where tiXg) = the test statistic; 



Z, = the predicted length-at-age for all sexes 

 combined; and 

 Itp = the predicted length-at-age for each sex 

 (p=l,2) (Helser, 1996). 



To generate the empirical probability density function 

 (pdf), length-at-age data for both sexes were pooled and 

 then assigned randomly without replacement to two 

 groups with sample sizes equal to the original number 

 of observations per sex. Growth curves were then fitted 

 separately to the length-at-age data of the randomized 

 groups, and the test statistic under the null hypothesis of 

 no difference was calculated as 



tix}- 



(/,,-/,)- 



't.'.P 



■l.'.P^ 



>.t,P 



where tix) = the difference in residual sum of squares 

 between the von Bertalanffy fits to the entire 

 pooled data set and von Bertalanffy fits to 

 the randomized groups (denoted by *). 



The randomization procedure was repeated 1000 times to 

 obtain the pdf of tix). The null hypothesis of no difference 

 was rejected if /(.V|,)>Z(.v) at a=0.05. 



Reproduction, sex ratios, and maturity 



Gonads excised from selected individuals were classified 

 macroscopically as either immature or mature by using 

 the maturation criteria of Nikolsky (1963) and Cody and 

 Bortone (1992). Seasonality of reproduction was deter- 

 mined by noting when changes in the gonad condition 

 took place. A gonadosomatic index (GSI) was calculated 

 to show changes in gonad weight in relation to gonad-free 

 total body weight. The index was computed as 



GSI = gonad weight/itotal body weight - 

 gonad weight) x 100. 



Length at 50% maturity was estimated for pinfish cap- 

 tured during the season of gonadal maturation. The prob- 



