The increased sampling effort at station 2, prompted by the need 

 for additional winter flounder larval data, produced a significant 

 regression in 1980 (Table 27) . The 1980 samples were analyzed further 

 by partitioning them into balanced day and night data sets. Separate 

 regressions were calculated and the two apparent mortality coefficients 

 were compared. The day coefficient (0.404) was significantly larger 

 then the one for night (0.337) probably because of under- samp ling of 

 larger larvae during the day. The day exponents for 1979 and 1980 were 

 similar, however. Even more troublesome was the fact that no significant 

 regression could be calculated from 1981 data for station 2 because of a 

 flattened catch curve. This occurred despite the relatively large 

 sampling effort. The 1982 data, with even more samples, was not yet 

 available for analysis at the time of this report. 



Because the number of samples, sampling frequency, and tow type all 

 appeared to influence the catch curves, the effects of day or night and 

 mesh size were investigated. Adjustment factors for each 1.0-mm length 

 interval were determined and applied to station 2 data. Even with 

 correction for undersampling of large larvae, no appreciable changes in 

 mortality coefficients were found. 



The distribution of winter flounder larvae at station 1 produced 

 good catch curves and apparent mortality coefficients were generated for 

 1974, 1975, and 1981 (Table 27). The large densities of small yolk-sac 

 larvae at station 1 indicated that this was an area of prime spawning, 

 retention, or both. Preliminary results from 1982 indicated that larvae 

 were again very abundant at station 1. 



78 



