FISHERY BULLETIN: VOL. 87, NO. 3, 1989 



Table 3. — Comparison of growth rates of northiern anchovy larvae 

 between Site 1 and Site 2 for post-yolk-sac larvae less than 18 days 

 old, using analysis of covariance. 



Linear regression of grovrth. L is length, f is time since hatching, and 

 n = total number of larvae sampled. 



Sitel L, = 7.33 + 0.46 (f - 9.99) n=197 



Site 2 L, = 7.82 -i- 0.46 (f - 11 .54) n = 79 



were used to estimate the larval production at 

 hatching (Pq) and mortality coefficient ((3). 



Mortality Estimates 



Data in three arrays were used to estimate 

 mortality: bongo larvae alone, bongo larvae plus 

 2.5-4.0 mm larvae from CalVETs, and bongo 

 larvae plus 3.0-4.0 mm larvae from CalVETs 

 (Tables 4, 5). The nonhnear regression gave esti- 

 mates of p ranging from 1.15 to 2.18 (equivalent 

 to IMR of 0.115-0.218 for larvae of age 10 days) 

 for Site 1 and 1.06-3.03 for Site 2. The lower 

 values of (3 were caused by low numbers of 2.5 

 mm larvae in CalVET samples. The log-linear 

 regression produced more consistent estimates 

 of p than those from nonlinear regression: Site 1 

 values ranged from 3.32 to 3.94 and Site 2 values 

 ranged from 2.87 to 3.47. Estimates of Pq at 

 hatching age ranged at Site 1 from 12.03 to 

 28.38, and for Site 2 from 1.37 to 4.23. For all 

 data arrays, site differences in larval production 

 rates were apparent (Tables 4, 5), but site differ- 

 ences in mortality rates were not (Table 6). 



Comparison of Larval Production and 

 Mortality Coefficients Between Sites 



Larval samples from bongo tows were used to 

 compare the mortahty between sites. Analysis of 

 covariance was performed on the logarithmic 

 transformation of age and larval production. The 

 results of the analysis indicated that the site 



difference between mortality coefficients was 

 not statistically detectable whereas site differ- 

 ence between larval production rates was (Table 

 6). 



We also compared the larval production and 

 mortality at the two sites (Table 5) with those 

 over the northern anchovy spawning domain as 

 sampled by the CalCOFI 8502 survey (P„ = 4.81 

 with SE = 0.30, and p = 2.21 with SE = 0.11 for 

 37 positive CalVET tows out of 45). Larval pro- 

 duction at Site 1 was much higher than the aver- 

 age over the CalCOFI 8502 survey region, but 

 production at Site 2 and larval mortality at both 

 sites were similar to the survey region averages. 



Figure 5 shows the larval production-at-age 

 curves at the two sites and, for comparison, 

 larval production curves for anchovy spawning 

 domains as sampled on CalCOFI Cruise 8502, 

 just prior to our site-intensive program. Site 1 

 displayed high production of all larval ages, 

 which was higher than its subdomain (CalCOFI 

 Region 7), whereas Site 2 production was some- 

 what above that of its subdomain (CalCOFI 

 Region 8) and somewhat below that of the entire 

 spawning domain as sampled ("all regions"). 



Starvation Incidence 



Midgut cell height is the histological criterion 

 of larval feeding history. The range of cell height 

 and the cell height change with larval size were 

 similar at both sites (Fig. 6). The similarity 



682 



