Sturm and Salter: Scomberomorus cavalla in Trinidad waters 



367 



ments, feeding habits, exploitation rates, methods of 

 capture, sample sizes, etc. The fishery in North Ameri- 

 ca is better developed than in Brazil (Collette and Russo 

 1984) and Trinidad, which enabled larger samples for 

 age and growth studies to be collected from the former 

 region. There were considerable differences between 

 sample sizes of older fish in the various studies, which 

 could introduce bias in the comparison of von Berta- 

 lanffy parameters. Best representation of growth of 

 older fish came from Manooch et al. (1987) and the 

 anomalous group of Louisiana females (Johnson et al. 

 1983). Changes in populations between studies brought 

 about by exploitation and other factors could also in- 

 fluence the results. 



Different methodologies could also contribute to 

 these differences. Nomura and Rodrigues (1967) 

 counted translucent rings to age Brazilian fish, thus 

 omitting part of a year. This probably resulted in their 

 obtaining the slowest growth of all the studies. Back- 

 calculated lengths were fitted to the von Bertalanffy 

 equation by Ximenes et al. (1978), Johnson et al. (1983), 

 and in this report. Beaumariage (1973) fitted observed 

 standard lengths and advised caution in extrapolation 

 of the theoretical growth curve for size at older ages, 

 because it was derived from Walford plots that ex- 

 cluded older fish; this resulted in conservative esti- 

 mates. Nomura and Rodrigues (1967) used both back- 

 calculated and observed lengths, which probably 

 resulted in large values of L^. Other differences in 



methodology, e.g., the weighting of mean back-calcu- 

 lated and observed lengths and more exact fitting of 

 the von Bertalanffy curve using computer programs 

 in the later studies, and the non-use of an intercept 

 value for back-calculation in some of the studies, could 

 lead to minor differences in the results. Comparisons 

 of results between the three areas evidently cannot 

 be properly done due to differing sample sizes and 

 methodologies. 



Estimates of maturity of 58.5 cm FL or age I-II for 

 Trinidad female S. cavalla generally agree well with 

 those of other studies. In Brazil, first maturity was 

 found to take place at 58.6 cm FL (Alves and Tome 

 1967) and 63.5 cm FL or age III (Gesteria and Mes- 

 quita 1976). Another study showed that females first 

 matured between 43.5 and 56.5 cm FL or age III, with 

 50% maturity at 77 cm FL or age V-VI (Ivo 1972). Size 

 of maturity of Trinidad females compares well with 

 that in Brazil. However, greater age at maturity of 

 Brazilian fish could be due to overestimated age at 

 length based on the use of translucent rings for age- 

 ing by Nomura and Rodrigues (1967). Finucane et al. 

 (1986) reported size at maturity at 44.9 cm FL in the 

 southeastern United States. In Florida waters, Beau- 

 mariage (1973) round ripe eggs in age-I females (61.4 

 cm SL = 65.6 cm FL) similar to Trinidad results. How- 

 ever, he concluded that these eggs were aborted or 

 reabsorbed, and spawning did not really take place until 

 age IV (88.0 cm FL). As with females, ripe males 



