12 



Fishery Bulletin 101(1) 



lan'ae is not large, e.g. to 7% (Theilacker, 1980; Fowler 

 and Smith, 1983; Kruse and Dalley, 1990). It is unlikely 

 that size-related shrinkage effects would have biased 

 our estimates of growth rate because these estimates 

 were based on larvae in similar size classes. Addition- 

 ally, Theilacker (1980) found that preserving northern 

 anchovy larvae after they had died during net capture 

 caused additional shrinkage, but this shrinkage was at a 

 constant rate that was proportional to fish length. Catches 

 of larvae were standardized to account for sampling effort 

 and expressed as number of larvae under 10 m- of sea sur- 

 face. This method of expressing the abundance of larvae 

 more accurately reflects station differences in abundance 

 than a mean density (number/ni'^) when fish larvae are not 

 homogeneously distributed throughout the water column, 

 as has been shown with other species from this area (Lyc- 

 zkowski-Shultz and Steen, 1991), and when sampling (sta- 

 tion) depths are variable, as they were in our study. 



Dry weights of larvae were determined by rinsing speci- 

 mens with distilled water, drying for 24 h at 60°C, and 

 weighing to the nearest 0.1 pg. Both sagittal otoliths were 

 removed following rehydration for 12 h. Otoliths were 

 mounted convex side up on a glass microscope slide with a 

 drop of Pro-Texx mounting medium and a cover slip. Oto- 

 lith growth increments were counted in the sagittal plane 

 under oil immersion (12.50x). 



A total of 140 Atlantic bumper larvae and 119 vermilion 

 snapper larvae were selected for age analyses. Specimens 

 were selected from stations where a wide size range of 

 larvae were collected. Daily otolith increment formation 

 has been validated for larval Atlantic bumper (Leffler and 

 Shaw, 1992). Daily increment formation has not been vali- 

 dated for vermilion snapper; however, otolith increments 

 observed in larval vermilion snapper were very similar in 

 width and spacing to validated daily increments found 

 in red snapper from this region (Szedlmayer, 1998; Lycz- 

 kowski-Shultz and Comyns^). Slopes of age-length regres- 

 sions for larval vermilion snapper (n=ll) and red snapper 

 (n=25) collected during July 1992 in our study area were 

 not significantly different, further indicating that vermil- 

 ion snapper, like red snapper, form daily otolith growth 

 increments. 



Otolith growth increments were counted by using the 

 sagitta (right or left) that provided the most distinct incre- 

 mental zones. Paired /-test analyses showed no significant 

 difference (P<0.05) in diameters of left and right sagittae 

 in both vermilion snapper (/j = ll) and Atlantic bumper 

 (n=20). Daily increments were counted along the longest 

 axis of the otolith from the core to the outer edge. Otoliths 

 were read once by a single reader, and a random subsam- 

 ple of otoliths from vermilion snapper («=30) and Atlantic 

 bumper (n=30) was read a second time to examine within- 



Lyczkowski-Shultz, J., and B. H. ("omyns. 1992. Karly Hfi- 

 history of snappers in coastal and sholf waters of tho north- 

 central Gulf of Mexico late summer/fall months, 1983-1989, 

 12 p. + 9 tables. 17 figures. Technical Report submitted to the 

 National Marine Fisheries Service, Southeast Regional Office, 

 9721 Executive Center Drive North, St. Petersberg, FL. 33702. 



reader variability. Otolith increment counts differed by 

 one day for only two of the 30 otoliths during the second 

 reading for both species. 



Data analysis 



Age-length and age-weight relationships were described 

 by using the exponential equation 



L or IV = exp(a -i- bt), 



where, in its linearized form, L = notochord or standard 



length in mm; 



W = dry weight in mg; 



a = y-intercept; 



b = slopeof regression line 

 (instantaneous growth 

 rate); and 



t = age of larvae in days. 



Values of a and h were calculated from the linearized 

 form of the growth equation after the length or weight 

 data were transformed to their natural logarithms. The 

 instantaneous growth rate (b), i.e. the slope of the log- 

 transformed, age-length or age-weight relationship, is 

 also referred to as the growth coefficient. Caution must be 

 exercised when making dry-weight comparisons because 

 of preservation-induced weight loss. Kristoffersen and 

 Salvanes (1998) found that body weight loss was as high 

 as 37-39% in small ethanol-preserved mesopelagic fishes. 

 Dry weight data were used only to determine whether 

 relative changes in weight tracked trends found in age- 

 length relationships. Analysis of covariance (ANCOVA) 

 was used to determine if differences existed among station 

 estimates of instantaneous growth coefficients (Sokal and 

 Rohlf 1969; SigmaStat, 1995). If differences were found 

 ( a=0.05 ), the simultaneous test procedure ( STP; Sokal and 

 Rohlf 1969) was used as an a posteriori test to determine 

 station differences. 



Cruise estimates of total larval abundance for each size 

 class (catch curves) were developed for Atlantic bumper 

 and vermilion snapper by summing the abundance esti- 

 mates of each size class under 10 m^ of sea surface from 

 each station. Length-frequency distributions were con- 

 verted to age-frequency distributions by assigning ages to 

 mid-points of the 0.5-mm size classes with the age-length 

 relationship previously described. Age-class abundances 

 were corrected for stage duration by dividing the abun- 

 dance estimate of each age class by their respective dura- 

 tions (Houde, 1977). It is necessary to correct for stage du- 

 rations of age classes if growth rates are nonlinear. Stage 

 durations of age classes were determined by assigning 

 ages based on previously determined growth equations 

 to end-points of the 0.5 mm size classes. This customary 

 method for constructing catch curves relies on the rarely 

 examined assumption that larvae at different sampling 

 locations are growing at similar rates. The high r~ values 

 of the age-length relationships (0.92 for Atlantic bumper; 

 0.84 for vermilion snapper) that resulted when aged lar- 

 vae from all stations were combined indicated that growth 



