362 



Fishery Bulletin 88(2), 1990 



one day from a beach, and catch per boat of S. cavalla 

 was usually small (<20 fish). Not more than three 

 catches were sampled up to a maximum of about 25 

 fish per sampling day, to enable accuracy of sample 

 treatment in the field. Entire catches were sampled, 

 and when this was not possible, large fish (>90 cm FL) 

 and/or small fish (<50 cm FL), when present, were 

 selected to direct attention to the ends of the distribu- 

 tion, and the rest of the sample was taken at random. 

 A total of 363 fish was obtained, 190 from hook-and- 

 line, 165 from gillnets (average stretched mesh size 113 

 mm), and eight from beach seines. In addition, 264 fish 

 were measured at the Port of Spain Fish Market. Fish 

 were measured to the nearest 0.5 cm fork length (FL) 

 and weighed to the nearest 20 g. Gonads were weighed 

 to the nearest 10 g. 



Otoliths (sagittae) were used to age 5. cavalla. When 

 viewed in a black dish containing water in reflected 

 light under a binocular microscope (x 10), the otoliths 

 revealed opaque and translucent rings. The non-mar- 

 ginal opaque rings were counted as annuli following 

 Beaumariage (1973), Ximenes et al. (1978), Johnson et 

 al. (1983), and Manooch et al. (1987), and distances 

 were measured along the longitudinal axis from the 

 focus to the distal edge of each opaque ring and to the 

 edge of the otolith. One person read the otoliths. 



The relationship between fork length and otolith 

 radius was determined by regression analysis using 

 both linear and quadratic models. When the relation- 

 ship was established, fork length at age was back- 

 calculated using methods of Bagenal and Tesch 

 (1978). 



The von Bertalanffy growth equation (1938)— L, = 

 L^ [l-exY)(-K{t + to))] where L, = fork length at 

 age t, L^ = asymptotic fork length, K = growth co- 

 efficient, and tfi = time when length is theoretically 

 zero— was fitted to weighted mean back-calculated 

 lengths using an MS-DOS/BASIC version of a compu- 

 ter program by Sparre (1987). 



Macroscopic stages of gonad development for both 

 sexes were identified as described for S. maculatus 

 (=S. brasiliensis) by Sturm (1978). They were imma- 

 ture-inactive, immature-developing, ripe, and ripe- 

 running. The gonadosomatic index (GSI) was computed 

 by dividing gonad weight by total fish weight and 

 multiplying by 100. 



Results 



Size composition of material examined 



Length frequencies oi Scomberomorus cavalla samples 

 from gillnets and hook-and-line were approximately 

 similar. Hook-and-line apparently selected slightly 



30 40 50 60 70 80 90 100 110 



Fork Length 



Figure I 



Length frequencies of gill net and hook-and-line samples of Scom- 

 beromontft cavalla from Trinidad waters. Lengths are grouped into 

 5-cm intervals. N = number of fish sampled. 



larger fish— 45-114.5 cm FL, with a modal length be- 

 tween 66 and 70 cm FL— than gillnets, which selected 

 fish between 38.5 and 105 cm FL, with a modal length 

 of between 61 and 65 cm FL (Fig. 1). The largest fish 

 sampled was a 127-cm FL female caught in a beach 

 seine. The size range of fish measured at the Port of 

 Spain Fish Market was 37-135 cm FL with a modal 

 length of 61-65 cm FL. 



Age and growth 



Only 2 of 341 otoliths were considered illegible. How- 

 ever, not all of the remaining 339 otoliths had clear 

 rings, though 85% were read with confidence. The 

 main difficulty was determining the extent of the first 

 opaque ring. The remaining 15% (52) were reread, with 

 79% (41) agreement between the two readings. The re- 

 maining 11 otoliths were then given new readings. For 

 tracing the frequency of opaque-edged otoliths through- 

 out the year, the data were grouped in bimonthly 

 intervals due to the scarcity of data in non-seasonal 

 months (Fig. 2). Otoliths with opaque edges were found 

 throughout the year. The minimum percentage (20%) 

 occurred in August-September after which there was 

 a large increase to maximum percentages (60%, 58%, 

 and 57%) from October through March, followed by 

 36% and 35% in April-May and June-July, respective- 

 ly. These results suggest that opaque rings are formed 



