Sturm and Salter: Scomberomorus cavalts m Trinidad waters 



363 



Figure 2 



Bimonthly percentages of otoliths of Sromheromorus cavalla with 

 opaque edges. A'^ = number of fish sampled. 



annually, mainly from October through March. Time 

 of ring formation is similar to that recorded for Brazil- 

 ian S. cavalla (Ximenes et al. 1978). 



The relationship between fork length (FL) and oto- 

 lith radius (OR) was represented by the following linear 

 equations: 



Males: FL = - 1.73 -f 1.490R (r = 0.86) 

 Females: FL = - 11.64 -i- 1.760R (r = 0.90) 



The intercepts were significantly different at the 

 P<0.001 level {F = 14.76, df = 1, 336); therefore, the 

 data were treated separately. Second-degree quadratic 

 models were fitted to the separate data, but analysis 

 of variance did not show significant curvature for males 

 (F = 3.03, df = 1, 98, P>0.05) nor females (F= 1.65, 

 df = 1, 235, P>0.1), therefore the linear equations were 

 used to back-calculate length from age for males and 

 females separately. 



Observed and back-calculated lengths for 99 male and 

 233 female S. cavalla are shown in Table 1 . Five age-0 

 fish (38.5-57.5 cm FL) were also recorded. The oldest 

 male was age VII and the oldest female age X, but 

 these were not included in the back-calculations. 

 Length variation within age groups was large, as was 

 the case for southeastern United States populations 

 (Beaumariage 1973, Johnson et al. 1983). For exam- 

 ple, age-II females ranged from 43.0 to 85.5 cm FL. 

 From age II onwards, females grow faster than males. 

 Figure 3 compares back-calculated growth of S. cavalla 

 from Trinidad with that from southeastern United 



