Comyns et al,: Spatial and temporal variability in growth and mortality of fish larvae in the Gulf of Mexico 



15 



ally exhibited a similar decrease in abundance with suc- 

 cessively larger size classes; however, the smallest size 

 class (1.1-1.6 mm) was the most abundant in only three 

 of the six cruises (Fig. 6, A, B, and F). This under repre- 

 sentation of the smallest size class in several cruises was 

 likely influenced by several potential factors, including 

 a possible decrease in spawning prior to sampling and 

 patchiness of eggs and newly hatched larvae caused by 

 the aggregation of spawning adults. Cruise-estimates of 

 mortality coefficients, which were derived by pooling data 

 from all stations sampled during a cruise and omitting the 

 smallest size class, ranged from 0.20 to 0.37 (Fig. 6). It is 

 likely that mortality rates varied between stations, but as 

 previously mentioned, an average cruise-estimate of mor- 

 tality was determined to ascertain a realistic level about 

 which the effects of small variations in growth rates could 

 be assessed on the cumulative survival of larvae. Standard 

 errors of Z estimates were low, ranging from 0.02 to 0.05. 



Size-frequency distributions were derived for vermilion 

 snapper lar\'ae (n=2581) taken during two September 

 1991 cruises, and single late-September cruises in 1992 

 and 1993 (Fig. 7) when vermilion snapper larvae were 

 abundant. Mortality estimates could not be estimated for 

 five September cruises during the period 1990-93 because 

 relatively few larvae were collected. Larvae collected dur- 

 ing three of the four cruises when they were abundant 

 showed a steady decrease in abundance of successively 

 larger size classes (Fig. 7, A, C, and D). During the fourth 

 cruise (late September 1991; Fig. 7B), the size-frequency 

 distribution showed a distinct peak in abundance of in- 

 termediate-size larvae (4.0-mm size class). Mortality coef- 

 ficients (Z) from the four cruises ranged from 0.19 to 0.30 

 and standard errors for the mortality coefficients were 

 relatively low ranging from 0.02 to 0.05. 



Discussion 



Plankton collections taken in the northcentral GOM 

 during September showed that growth and mortality 

 rates did vary in time and space for Atlantic bumper and 

 vermilion snapper larvae, and that these differences were 

 great enough to significantly impact the cumulative sur- 

 vival of larvae in a subtropical climate where larval-stage 

 durations are short (i.e. two weeks). Growth and mortal- 

 ity estimates of vermilion snapper larvae were previously 

 unknown. Two previous studies of growth and mortality of 

 Atlantic bumper larvae (Leffler and Shaw, 1992; Sanchez- 

 Ramirez and Flores-Coto, 1998) provided no information 

 on variability in growth rates at small spatial scales and 

 no estimates of mortality during the period when our 

 study was conducted. 



Highly significant between-station differences in growth 

 rates were observed for both Atlantic bumper and vermilion 

 snapper larvae. The largest difference in age versus length 

 growth coefficients for Atlantic bumper larvae was found 

 at adjacent, inner-shelf stations located approximately 10 

 km apart. According to growth equations, the faster grow- 

 ing larvae grew to a length of 6 mm 2.9 days sooner than 

 larvae at the adjacent station, and differences in larval 

 weight gain as expressed by dry weight of 11-d-old larvae 

 varied by over 30^?^. Water temperatures at these two sta- 

 tions were extremely similar; surface temperatures varied 

 by only 0.1°C, midwater temperatures varied by 0.4°C, and 

 surface and midwater temperatures varied by 0.5°C. It is 

 likely that a similarly small temperature differential was 

 present during the two-week period prior to this cruise, i.e. 

 throughout the life of larvae used in our study because dai- 

 ly surface water temperatures recorded at a weather bouy 

 within the study area during the previous month showed 



