FISHERY BULLETIN: VOL. 87, NO. 1 



AUG 



N'E5 



-29°N AP.IMAY N=62 



... <... I UAD.J 



GULF OF MEXICO 



JUNJUL m^fif^ 



OCT AUG 

 N-34 N-40 



FEBOCT 

 N'413 



•1986 

 1987 



95°W 



_i 



90'W 



_i 



Figure 1.— Purse seine sampling locations in the northern Gulf of Mexico. Points represent individual purse seine sets. N's refer to 

 the total numbers of red drum sampled from sets indicated. Precise locations were not available for 6 sets {N = 161). 



unavailable for 182 individuals. Sagittae were re- 

 moved, cleaned, and stored dry for later process- 

 ing. 



Length-weight regressions were fit to the data 

 using the model: weight = a FL*, where weight = 

 body weight (g) and FL = fork length (mm). Re- 

 gressions for male and female red drum were com- 

 pared using analysis of covariance (Ott 1977). A 

 Komolgorov-Smirnov two-sample test (Tate and 

 Clelland 1957) was used to detect possible sampling 

 bias by comparison of length-frequency distributions 

 of fish caught by different sampling gears. 



Otoliths were processed for age analysis by em- 

 bedding them in an epoxy resin (Spurr 1969) and 

 sectioning transversely (0.7 mm thick) through the 

 core of the left sagitta (or the right when the left 

 sagitta was not available), using a Buehler Isomet^ 

 low-speed saw. Sections were mounted on glass 

 slides with thermoplastic cement (Crystalbond 509 

 adhesive), sanded on 600 grit wet sandpaper to 

 remove saw marks, polished with alumina micro- 

 polish (0.3 ixm), and then examined with a compound 

 microscope (transmitted light at 40 x magnification). 

 Opaque zones (annuli) were counted in sections from 

 the core to the margin in the medial direction. Ap- 

 pearance of the margin was recorded as either 

 opaque or translucent. If the left sagitta was un- 

 readable, the right sagitta, if available, was prepared 

 and examined. Validation of age estimates was ac- 

 complished and the timing of annulus formation 

 determined by plotting percent occurrence of oto- 

 liths with opaque margins by month. 



'Reference to trade names does not imply endorsement by the 

 National Marine Fisheries Service, NOAA. 



Each otolith was aged by two readers, and the 

 resulting age estimates were compared. The coef- 

 ficient of variation was calculated for age estimates 

 in order to test the reproducibility of age estimates 

 independent of magnitude (Sokal and Rohlf 1981; 

 Chang 1982). If readers' initial age estimates for an 

 otolith did not agree, the section was reread. If the 

 resulting age estimates did not agree, the fish's 

 other sagitta was prepared and read. If the readers 

 did not reach agreement on an age or sections from 

 both otoliths were unreadable, the data for that fish 

 were not used in analyses. All ageing was done 

 without knowledge of the sample source or any 

 previous age estimates. 



Year-of-birth was back-calculated from age esti- 

 mates by subtracting estimated age from the year 

 of capture and assuming that the first annulus 

 formed in winter of year 2 (Beckman et al. in press). 

 Age-frequency distributions were compared using 

 a Komolgorov-Smirnov two-sample test (Tate and 

 Clelland 1957). 



Von Bertalanffy (1938, 1957) growth curves were 

 fit separately for males and females by nonlinear 

 regression. The growth equation for length was 

 L, = L„ [1 - e-*^*'-'oi] and for weight was W, = 

 W^ [1 - e-K((-*o)]3, where L, and VF, are the esti- 

 mated length and weight, L„ and W^ are the 

 asymptotic length and weight, K is the growth coef- 

 ficient, t is the age (years), and tf, is the hypothe- 

 tical age when length or weight would be zero. A 

 full model, in which separate parameters were fit 

 for males and females, was compared with a reduced 

 model in which sex was not considered. An F-test 

 (Ott 1977) was used to test for differences in the 

 models. 



18 



