Larvae were least numerous in 1986 in both the 

 river and bay. The abundance of larvae in the 

 river was similar in 1984 and 1987. Abundance 

 in the river did not necessarily reflect that in the 

 bay, as larvae were most abundant in the river in 

 1985, but were only found in moderate numbers 

 in the bay that year. 



Comparisons of armual spatial abundance of 

 the first four developmental stages were based on 

 the cumulative weekly geometric means; because 

 few Stage 5 larvae were collected, their abundance 

 and distribution was not examined. At the three 

 river stations (A, B, and C), Stage 1 abundance 

 was lowest in 1986, except for 1983 (Fig. 19). 

 The low abundance in 1983 was attributed, in 

 part, to undersampling due to net extrusion 

 (NUSCO 1987) and this was rectified in 1984 

 when a smaller 202-|im mesh net was used during 

 the early portion of the larval season. Comparison 

 among years at each river station showed a similar 

 pattern in Stage 1 abundance with 1985 the high- 

 est, followed by 1987, 1984, 1986, and 1983. Stage 

 1 larvae were rarely collected in Niantic Bay at 

 station EN and NB, suggesting that little, if any, 

 spawning occurred in the bay. By developmental 

 Stage 2, larvae were more prevalent in the bay, 

 but a majority were still collected in the river. 

 Stage 2 larvae were least abundant in 1986 at all 

 five stations. The order of annual abundance for 

 Stage 2 larvae among years was generally the same 

 for the three river stations, although 1987 dropped 

 from second to fourth in rank. Most Stage 3 

 larvae were collected in the lower portion of the 

 Niantic River (station C) and in the the bay (sta- 

 tions EN and NB) with very few present in the 

 upper river (station A). Abundance in 1986 was 

 lowest at all stations, but the magnitude of the 

 difference compared to other years was not as 

 great. The pattern and relative abundance of 

 Stage 3 larvae were similar among years at stations 

 EN and NB. Stage 4 larvae were collected pri- 

 marily in the lower river and in Niantic Bay. The 

 low abundance ol other developmental stages in 

 1986 was not as apparent at Stage 4 of develop- 

 ment. The large decline in abundance from Stage 

 3 to 4 in each year was probably related to less 

 effective sampling for older larvae. By Stage 4 of 



development, the left eye has migrated to or past 

 the mid-line and the larvae have become mostly 

 demersal and thus were less susceptible to either 

 the bongo sampler or entrainment at MNPS. The 

 similar abundance of all stages at EN and NB, 

 which are approximately 1 km apart, suggested a 

 relatively uniform distribution of larvae through- 

 out Niantic Bay. 



The decline in abundance of larvae as they 

 passed through developmental stages was quite 

 variable between years. For example. Stage 1 and 



2 larvae were the most abundant in 1985 at all 

 river stations compared to other years, but Stage 



3 larvae were among the least abundant, implying 

 that high mortality occurred during Stage 2 of 

 development. In 1984, however. Stage 1 and 2 

 abundances were moderate and Stage 3 abundance 

 at station C was the highest of the 5-year period, 

 indicating low mortality during Stage 2. Although 

 this variability could be attributed to the impre- 

 cision of plankton sampling, the consistency in 

 the relative ranking of years for Stages 1 and 2 at 

 the three river stations and for Stage 3 at the two 

 bay stations suggested that the precision in quan- 

 tifying larval abundance was good. Also, the 

 similar abundance of Stage 3 larvae at EN and 

 NB each year implied that both techniques pro- 

 vided comparable results, even though the sam- 

 pling methods at these stations were different. 

 Variability in the relative ranking of abundance 

 from stage to stage among years was reported by 

 Bannister et al. (1974) for egg and larval stages 

 of the plaice, which they felt was consistent with 

 density-dependent mortality. The lack of a pattern 

 in the decline in abundance among years indicated 

 that the processes that regulated larval winter 

 flounder abundance were complex and operated 

 at different levels from year to year. 



A comparison of the temporal occurrence of 

 developmental stages was based on the date of 

 peak abundance in the river and bay, which was 

 estimated from the inflection point of the 

 Gompertz function (Table 12). Because Stage 1 

 larvae were rarely collected in the bay, the dates 

 of peak abundance could not be estimated for 

 this area. The dates of peak abundance of each 



Winter Flounder Studies 



179 



