NOTE Davis and West Growth and mortality of Lutjanus vittus 



399 



could not be identified (104 juveniles) presented a 

 problem, because excluding them created a bias since 

 fish that could be sexed in age-class 1 (40 males and 

 40 females) were larger animals. There were significant 

 differences (P< 0.001) in back-calculated and observed 

 lengths between age-class 1 males, females, and juve- 

 niles (one-way ANOVA, F 50.7, df 2, 181; F 89.0, df 2, 

 181, respectively). Multiple comparison by the Tukey 

 test indicated that age-class 1 fish that could be sexed 

 were significantly larger than juveniles by about 17 mm 

 for back-calculated lengths and 33 mm for observed 

 lengths. To eliminate this bias, juveniles were ranked 

 by size. The smallest was randomly assigned a sex, and 

 then each juvenile in order was assigned to alternate 

 sexes. The mean lengths back-calculated to the last 

 annulus at each age for male and female (including 

 assigned sexes in age-class 1) are presented in Table 

 2 and Figure 7. 



Back-calculated length-at-age data minimize the ef- 

 fects of seasonal growth but do not completely elim- 

 inate it, because the time of check formation ranges 

 over several months (Fig. 5). Assigning an absolute age 

 using an arbitrary birthdate will only compensate for 

 growth differences between fishes caught at different 

 times of the year when there is little seasonal varia- 

 tion in growth rate. However, assigning absolute ages 

 does enable age-class data to be used in determining 

 growth curves (Fig. 8). While back-calculated lengths 

 cannot use age-class data, they do enable a more 

 realistic time-scale parameter (tg) to be estimated. 



400 



300 - 



E 

 E 



^200 



c 



100 - 



52 22 



14 2 



Males 



-1 — I — I — I — I — I — I — I — I — I 



400 



300 - 



E 

 E 



o) 200 



c 



CD 



100 



45 



17 1 



Females 



— I 1 1 1 r 1 1 1 1 1 



01 23456789 10 

 Age (years) 



Figure 7 



Growth of male and female Lutjanus vittus. Mean back- 

 calculated lengths to the last annulus, 95% confidence limits 

 (vertical bars), range, and sample size have been plotted. 

 Von Bertalanffy growth curves were fitted to individual 

 observations. 



The least-squares estimates of the von Bertalanffy 

 growth curve parameters are quite different between 

 the sexes for both forms of length-at-age data (Table 

 3). Independent of any assumed growth curve, there 

 were significant differences in mean back-calculated 

 lengths between sexes for age-classes 4-6 years 

 (ANOVA, F 42.1, df 1, 311, P<0.001) but age-classes 

 1-3 were not significantly different (F0.32, df 1, 

 605, P 0.569). Only fish whose sex was determined 



