Sudekum et al.: Life history and ecology of Caranx ignobilis and Caranx melampygus 499 



The 14 counts yielded a pooled mean of 51.6 marginal 

 increments (range 47-62, SD 4.31) and a 95% con- 

 fidence interval of 49.1-54.1 increments. One-way 

 analysis of variance, however, showed that significant 

 differences existed among the five fish in mean incre- 

 ment count (F 4.61, df 4, 9, P 0.027). Furthermore, 

 variation in marginal increment deposition within 

 specimens (s 2 8.80) was comparable to that between 

 specimens (s 2 10.59). When mean counts of the five 

 individual fish were compared, the 95% confidence in- 

 terval for the population mean was 47.0-56.1 marginal 

 increments. The number of days since injection (55) 

 falls within this interval, and in general the marginal 

 counts are reasonably close to this number. 



During the maximum feeding-rate experiment, the 

 six specimens of C. melampygus increased from an 

 average standard length of 174 mm (range 166-185 

 mm) to 239 mm (range 225-252 mm) over a period of 

 161 days. This represents an average growth rate of 

 0.40 mm/day. An instantaneous growth rate of 0.45 

 mm/day (1.02g/day) was calculated for "wild" fish 



Figure 4 



Ultraviolet-white light double-exposure micrograph of a 

 section of sagitta from Caranx melampygus, marked for 

 growth validation. The fish specimen had been injected with 

 tetracycline 55 days before sacrifice. The fluorescent tetra- 

 cycline mark (T) is visible as a light band inside the otolith 

 margin (M). 



within the same size range by using the von Bertalanffy 

 growth curve. 



Reproduction 



Sex was determined in 119 specimens of C. melampy- 

 gus and 110 specimens of C. ignobilis. In both species, 

 the sex ratio of the samples, M:F, was slightly skewed 

 toward females: 1:1.48 for C. melampygus (x 2 4.45, 

 df 1, P 0.04) and 1:1.39 for C. ignobilis ( x 2 2.95, df 1, 

 P 0.08). 



Using morphological criteria, gonads from 152 

 C. melampygus and 124 C. ignobilis were visually 

 classified into one of three reproductive stages (im- 

 mature, developing, or gravid). To test the accuracy 

 of this visual technique, gonadosomatic indices from 

 62 C. melampygus were compared with the reproduc- 

 tive stages assigned visually for each fish. A two-way 

 factorial analysis of variance, with stage (immature, 

 developing, or gravid) and sex (male or female) as 

 classification variables and log-transformed GSI as 

 the dependent variable, showed that stage (F 44.34, 

 df 2, 52, P<0.0001), but not sex (F 0.95, df 1, 52, 

 P 0.334) affected gonadosomatic index. The interaction 

 term was insignificant (F 1.69, df 2, 52, P 0.194). Mean 

 values of the untransformed GSI's for the three stages 

 of reproductive condition were 0.16%, 0.42%, and 

 4.54%, respectively. Sidak <-tests (SAS 1985) revealed 

 that all three means were significantly different from 

 one another. Thus, the two methods of establishing 

 reproductive condition show a high degree of concord- 



