WARLEN; AGE AND GROWTH OF LARVAL GULF MENHADEN 



Table 1. — Standard length (mm) and estimated age (num- 

 ber of otolith growth increments +5) of laboratory-reared 

 larval gulf menhaden. Values in parentheses are 95% inter- 

 val estimates. 



Known age Number of Mean estimated 

 (days) fish age 



Mean SL 



7 

 14 

 20 



13 

 16 

 22 



7.8 (±0.38) 



13.2 (±1.14) 



19.3 (±0.77) 



4.7 (±0.28) 

 6.4 (±0.38) 

 8.0 (±0.42) 



ages and the 957c confidence intervals included 

 the known age in each of the three groups. Some 

 of the variation in the number of gi'owth incre- 

 ments observed in known age larvae (Fig. 3, 

 Table 1) may have been due to 1) poor growing 

 conditions during rearing that could have re- 

 sulted in reduced growth in underfed larvae, 2) 

 variation in the inception of increment formation 

 as has been observed in other species (Laroche et 

 al. 1983; Fives et al. 1986), and 3) faintness of 

 growth increments in some larvae. In contrast, 

 increments on otoliths of field collected larvae 

 (Fig. 2) were usually very regular and distinct 

 and were more easily observed than those on 

 otoliths of laboratory-reared larvae. I assumed 

 that the gi'owth increment deposition rate was 

 also daily in wild larvae examined in this study. 



Age and Growth of Larvae 



Average growth of larval gulf menhaden dur- 

 ing their first two months of life was described by 

 the Gompertz growth model for pooled length at 

 age data for 2,003 fish representing collections 

 from all six RV Oregon II cruises (Table 2, Fig. 4). 

 Larvae ranged in age from 5 to 62 days (x=24A 

 days) and in SL from 3.4 to 28.0 mm (x = 12.6 

 mm). In the log-transformed model, age ac- 

 counted for 82% of the variation in length. Gulf 

 menhaden were predicted to have grown from 2.4 

 mm SL at hatching to 20.4 mm at age 62 days. 

 The size at hatching estimated from the Gom- 

 pertz equation was only slightly less than the 

 hatching size, 2.6-3.0 mm SL, observed in the 

 laboratory (Hettler 1984). Age-specific growth 

 rates declined from — 79f/day at age 10 days to 

 <0.4%/day at age 60 days. Maximum absolute 

 growth rate occurred when gulf menhaden larvae 

 were 7.9 mm SL and 13 days old. 



The asymptotic length of larvae (21.5 mm SL), 

 determined from the variables of the growth 

 equation, is approximately the size when larvae 

 begin to transform into juveniles. This transfor- 

 mation, described by Lewis et al. (1972) for the 



closely related Atlantic menhaden, B. tyrannus, 

 apparently ends when the fish reach 28-30 mm 

 SL (Suttkus 1956). 



In all instances except one transect (TX Decem- 

 ber 1981, where there was no convergence in the 

 parameter values in the computer fitting proce- 

 dure and the model would not fit the data), the 

 Gompertz growth model could be used to describe 

 the growth of gulf menhaden larvae from each 

 cruise and transect (Figs. 5-7, Table 2). The 

 growth model for the FL December 1980 larvae 

 approximates an exponential form because of the 

 exceptionally low value for a. This may be due to 

 the preponderance of small, young larvae. 



GROWTH COMPARISONS 



Louisiana - Seasons and Years 



There were statistically significant differences 

 (MANOVA, P < 0.001) in the growth curves for 

 larvae caught off Louisiana for two seasons (De- 

 cember, February) and three years (1979-80, 

 1980-81, 1981-82). To determine if differences ex- 

 isted between seasons in each year and among 

 any two years within each season, I selected 9 of 

 the possible 15 pairwise comparisons for testing. 

 The Bonferroni critical value in these tests was 

 0.0033 (0.05/15). The inability to fit a Gompertz 

 growth model to the TX December 1981 data pre- 

 cluded a comparison with the larvae collected off 

 Texas in February 1982. 



Pairwise comparisons for within years data for 

 Louisiana larvae showed significant differences 

 (P < 0.003) in growth curves between early sea- 

 son (December) and late season (February) for 

 each year. Faster growth of early season larvae is 

 evident if the respective curves (Figs. 5; 6a, c; 7a, 

 c) are compared. For any age, the predicted size is 

 greater for early season than for late season lar- 

 vae. Only for the third year did the length at age 

 40+ days of February-caught larvae exceed that 

 for December-caught larvae. 



In similar comparisons for larvae caught off 

 Louisiana in December of all three years, there 

 were significant differences (P< 0.003) in the 

 growth curves (Figs. 5a, 6a, 7a) for any two years. 

 As judged by the predicted size at any age, larvae 

 appeared to grow faster in 1979 than in either 



1980 or 1981. While the curves for the 1980 and 



1981 larvae overlapped, larvae from 1980 were 

 larger at 30+ days than were the 1981 larvae. 

 Significant differences were also found among the 

 curves (Figs. 5b, 6c, 7c) for larvae caught in 



81 



