Powell: Early-life-history traits of Brevoortia tyrannus and 5. patronus 



121 



survivors, which varied markedly. Growth and survival 

 between species were compared using this index. 

 Gain in biomass (B) for each tank was calculated as 



B = (W t -W M )(S), 



where W = mean dry weight/fish, t = age in days, and 

 S = number of survivors. 



A second series of growth experiments was conducted 

 to examine instantaneous daily growth rates and size 

 of larvae 10 d past first-feeding. Eggs and larvae were 

 reared in 10 L tanks at 16°, 20°, and 24°C. Larvae 

 were maintained on high food densities (~ >25 roti- 

 fers/mL). For each experimental temperature, 10 lar- 

 vae were sampled 10 d past first-feeding. Length (SL), 

 dry weight, and mouth width (an estimate of gape 

 size) were obtained from preserved larvae. Daily in- 

 stantaneous rates of increase in length and weight 

 were calculated according to Ricker ( 1975:207). 



Length-weight relationships were used to estimate 

 sizes separating early-life-history periods (Balon 1984). 

 Larvae were measured live, and dry weights were ob- 

 tained from larvae that were preserved by freezing. 

 For each species, an iterative process was used to de- 

 termine the break point (SL) between two regressions 

 that described the length/weight relationship. Itera- 

 tions were done with break point values of 8-14 mmSL 

 at 1 mm intervals. The two regressions that resulted 

 in the minimal mean square error were chosen. An 

 ANCOVA was used to compare the biphasic regressions. 



Statistical testing was done by using the General 

 Linear Model procedure (SAS 1985). Unless otherwise 

 noted, ANOVA was used to test differences. Each treat- 

 ment consisted of a group of larvae in a common tank. 

 The mean measurement (e.g., mean length or mean 

 weight of the group) was used in the statistical analy- 

 sis. Each experiental replicate was treated as an inde- 

 pendent observation. Differences between main effects 

 were considered significant at a=0.05. For interaction 

 between factors, differences were considered signifi- 

 cant at a=0.10 to reduce the chance of a type-II error. 



Results 



Atlantic menhaden eggs were larger in diameter and 

 nearly two-fold heavier than gulf menhaden eggs 

 (Table 1, 2). The volume of yolk was greater for Atlan- 

 tic menhaden, but oil globule volumes were similar 

 (Table 1,2). 



There was no discernible change in egg size during 

 the laboratory spawning season (Fig. 1), and the level 

 of variability throughout the season was smaller for 

 gulf (SD=0.02-0.04) than Atlantic (SD=0.04-08) men- 

 haden. There was a trend for field-collected Atlantic 



Table 2 



Summary of Atlantic and gulf menhaden egg data. Values in 

 parentheses indicate the number of replicates. For each ex- 

 periment -50 eggs were measured or weighed. 



Eggs 



Yolk 



Oil Globule 



Species 



Mean Mean Mean Mean 



diameter dry weight volume volume 



(mm±SD) (ug) (mm 3 +SD) Imm'+SDl 



Atlantic 1.61+0.08 

 menhaden (9) 



Gulf 1.22+0.04 

 menhaden (7i 



menhaden eggs to be larger at higher latitudes 

 (P<0.01), but there was a considerable amount of vari- 

 ability in egg size for any given latitude (Fig. 2). 



Newly-hatched, laboratory-reared Atlantic menha- 

 den were longer and heavier than gulf menhaden (Table 

 1, 3). Temperature affected length-at-hatching, in that 

 larvae (especially gulf menhaden larvae) were longer 



