333-}im mesh collected significantly greater densities of the 3- and 4-mm size-classes with no differences 

 detected for the other size classes. The effect of mesh size (202- and 333-jim) and tow duration (6- and 

 15-min) on net extrusion of early developmental stages (Stage 1 and 2) was examined in 1984 during 

 February and March (NUSCo 1985). In the comparison between the two mesh sizes, Stage 1 larvae were 

 collected in greater densities with 202-|im mesh in 21 of 28 paired comparisons, which represented a 

 significant difference. No difference was found for Stage 2 larvae, with their density in the 202-nm mesh 

 greater in only 1 3 of the 28 comparisons. No difference in the collection density of Stage 1 and 2 larvae 

 was found between the 6- and 15-min tow durations (16 comparisons). The results of laboratory com- 

 parisons of 202- 333-, and 505-nm mesh (NUSCo 1986a) verified the fmdings of the field studies with the 

 greatest retention in 202 \im (92%), followed by 333 jim (78%), and 505 ^m (63%). 



Prior to 1984, all ichthyoplankton sampling at MNPS was conducted with 333- or 505-|im mesh nets 

 and during this period the early developmental stages were probably undersampled. This undersampling 

 limited the use of these data to examine the abundance and distribution of early developmental stages. 

 Although only a 333-tim mesh net was used for all recent entrainment collections, it probably did not 

 result in underestimates because 202-^m mesh nets were used at station NB during 1985 and very few 

 Stage 1 larvae were found in Niantic Bay (NUSCo 1986a). 



Die! behavior 



Diel behavior patterns of larval winter flounder could affect sample densities and bias abundance 

 estimates. A comparison of day and night collections was made at three stations where balanced day and 

 night sampling was conducted during several years: EN from 1976-85, NB from 1979-85, and C from 

 1980-83. The percentages that each 1-mm size-class made up of day and night samples were examined 

 (Fig. 9). At stations EN and NB an increasing percentage of the 5- to 6-mm and larger size-classes 

 occurred in night collections. Although the 5-mm and larger size-classes were more prevalent in night 

 samples at station C in the Niantic River, the difference was less apparent than found at the other two 

 stations. Comparison of day and night collections by developmental stage was possible with 1983 data 

 (Fig. 10). Only in this year were all three Niantic River stations (A, B, and C) sampled during both day 

 and night throughout the season. Stage 4 larvae at station A and Stage 1 larvae at stations EN and NB 

 were excluded because of low collection densities. No consistent difference between day and night 



61 



