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



13 



rates over the study area were similar enough to justify 

 use of this technique. 



Cruise estimates of larval mortality rates for Atlantic 

 bumper and vermilion snapper larvae were estimated 

 from catch curve analyses (e.g. Houde, 1977; Essig and 

 Cole, 1986; Watanabe and Lo, 1988; Deegan. 1990; Co- 

 myns et al., 1991 ). The instantaneous mortality rate (Z) 

 was estimated by the slope of the exponential function 

 relating duration-corrected larval abundance and age 

 (Ricker, 1975): 



D, = D^ exp i-Zt), 



where D, = total abundance of larvae at time /; 



Dy= total abundance of individuals at time 0; 



Z = instantaneous mortality rate; and 



t = age of size class in days since spawning. 



Age and abundance of size classes were fitted to this 

 exponential function with a nonlinear least squares 

 routine, and only the descending limb of the regression 

 was used to estimate mortality rates. To reduce poten- 

 tial biases associated with 1) any trend of increasing 

 variance in the length-at-age distribution with increas- 

 ing age, and 2) net avoidance by larger larvae, only 

 Atlantic bumper and vermilion snapper larvae smaller 

 than 6.1 mm and 6.0 mm, respectively, were used to 

 estimate mortality rates. Kolmogorov-Smirnov two- 

 sample tests showed no significant differences (P<0.05) 

 between size-frequency distributions for day versus 

 night catches within this size range for vermilion snap- 

 per and Atlantic bumper larvae. 



Results 



221 



0.2 

 18 

 0.16 

 0.14 

 0.12 



g 0.1 



n=19 

 1=26 „.,, n=22 T 



f T T ~ 



_L n=23 



I I — I — I — I I — I — 



12 23 24 34 37 41 42 44 47 



12 23 24 



34 37 41 

 Station 



42 44 47 



Figure 2 



Growth coefficients (horizontal bars) for Atlantic bumper 

 larvae collected at nine stations in the northcentral Gulf of 

 Mexico, during 14-16 September 1991. Vertical lines repre- 

 sent 95% confidence intervals around the growth coefficients, 

 and numbers above bars depict sample sizes; (A) shows age 

 versus In length growth coefficients and (B) shows age versus 

 In dry-weight growth coefficients. 



Age and growth 



Atlantic bumper larvae, which were commonly found 

 throughout the study area, ranged from 2 to 14 days old, 

 1.4 mm to 8.1 mm in length, and 0.003 mg to 1.446 mg 

 in dry weight. Estimates of age versus length growth 

 coefficients were not similar for all stations (ANCOVA; 

 P=0.001). The STP revealed no overlap in 95% confidence 

 intervals around growth coefficients for larvae collected 

 at station 42 and larvae collected at stations 41, 23, and 

 47 (Figs. 2A and 3). According to their respective growth 

 equations, Atlantic bumper larvae at station 42 grew at 

 approximately 0.43 mm/d and reached a length of 6 mm in 

 approximately 13.3 days. Larvae collected at adjacent sta- 

 tion 41 grew faster, approximately 0.63 mm/d, and reached 

 a length of 6 mm in 10.4 days. 



Similarly significant differences in station estimates 

 (n=9) of age-dry-weight growth coefficients were also 

 found (ANCOVA; P=0.01), and growth coefficients for 

 larvae collected at station 42 were significantly different 

 from larvae collected at stations 41 and 23 (STP; Fig. 2B). 

 By 11 days, the estimated dry weight of an Atlantic bum- 

 per larva at station 42 was 0.38 mg, whereas at station 41 

 larvae gained weight faster and the estimated dry weight 



of an 11-d-old larva was 0.58 mg. Adjacent stations 41 and 

 42 were 10 km apart, and water temperatures at these two 

 locations were very similar. Surface temperatures varied 

 by only 0.1°C (28.7°-28.8°C), and surface and midwater 

 temperatures varied by only 0.5°C. Daily surface water 

 temperatures recorded at a weather bouy within the study 

 area showed that temperatures varied by less than 2°C 

 during the 31-d period prior to our study. 



Significant differences in station (n=7) growth rates of 

 vermilion snapper larvae were also found in our 14-16 

 September 1991 cruise (ANCOVA; P=0.03). Vermilion 

 snapper larvae ranged from 4 to 16 days old, 2.5 mm to 

 6.5 mm in length, and 0.014 mg to 0.696 mg in dry weight. 

 Growth coefficients for larvae collected at stations 15 and 

 25 were significantly different (STP; Figs. 4A and 5). Ac- 

 cording to their respective growth equations, vermilion 

 snapper larvae collected at station 15 reached a length of 

 5 mm in 10.7 days, whereas larvae collected at station 25 

 grew more slowly and did not reach a length of 5 mm until 

 12.6 days. Stations 15 and 25 were located 17 km apart on 

 the inner shelf at water depths of 29-30 m. Surface water 

 temperatures at these stations varied by 2.2°C, and both 

 surface and midwater station temperatures differed by 

 less than 2°C. 



