6. Examination of available historical data 

 (commercial fishing CPUE for Rhode Island 

 and a URI trawl time-series) showed that 

 winter flounder abundance typically fluctu- 

 ated with sharp increases in catch most likely 

 related to the occurrence of one or more 

 particularly strong year-classes in succession. 



7. For the first time in the past 1 1 years, more 

 males than females were taken during the 

 Niantic River surveys in 1986 and 1987. 



8. The length of females at 50% sexual matu- 

 ration was 26.8 cm, when fish were 3 or 4 

 years old. Most spawning in the Niantic 

 River was completed by late March or early 

 April. Spawning appeared to have been re- 

 lated to water temperature as in relatively 

 cold years proportionately fewer females 

 spawned earlier during the season them in 

 warmer years. 



9. Based on the abundance indices of females 

 and their size distribution, a yearly index of 

 egg production was determined. This index 

 peaked in 1982 and has declined about 80% 

 since then. However, adult abundance and 

 absolute egg production were not the only 

 factors in determining year-class strength. 



1 0. The 1 2 years of Niantic River winter flounder 

 relative abundance data were used with a 

 Ricker stock-recruitment model. Parental 

 stock was defined as all winter flounder age 

 3 and older and recruits were those fish turn- 

 ing 3 years old each spawning season. The 

 two-parameter Ricker model did not explain 

 much of the variability (44%) seen in annual 

 recruitment. However, annual February wa- 

 ter temperatures were found to be significantly 

 and inversely correlated with recruitment in- 

 dices. The addition of a temperature param- 

 eter to the model resulted in a much improved 

 fit to the observed data (R = 0.78) and helped 

 to explain large differences seen in recruitment 

 for similar parental stock sizes. Although 

 the actual mechanisms affecting winter floun- 

 der recruitment were unknown, the February 



water temperature appeared to have been re- 

 lated to those factors. 



1 1 . Larval winter flounder studies have been con- 

 ducted in Niantic River and Bay since 1983 

 and entrainment collections have been made 

 since 1976. In 1986 and 1987, peak abun- 

 dance of larvae occurred first in the river and, 

 after a lag, in the bay. Comparison of the 

 dates of peak abundance showed that most 

 flushing of larvae from the river to the bay 

 occurred during Stage 2 of development. 



1 2. An apparently higher mortality of larvae was 

 found in 1987 compared to the previous year; 

 this occurred early in the season, when the 

 majority of the larvae were in the river. Ef- 

 fects of predation by jellyfish was not as ap- 

 parent in 1986 and 1987 as it had been in 

 previous years. Examination of length- 

 frequency distributions indicated that most 

 mortality occurred during the 3 to 4-mm 

 size-class, suggesting that this was a critical 

 period for mortality. Annual total mortality 

 rates were estimated from the difference be- 

 tween the abundances of larvae at 3 and 7 

 mm in length. For 1983-87, these estimates 

 ranged from 84.6 to 96.9%. 



13. In entrainment samples. Stage 3 larvae dom- 

 inated. As expected, total entrainment esti- 

 mates for 1987 following the startup of 

 MNPS Unit 3 were among the highest during 

 the last 12 years, even though the median 

 density (number per 500 m ) was among the 

 lowest. Entrainment estimates were depen- 

 dent upon plant operating conditions as well 

 as larval densities each year. The dates of 

 peak abundance for entrainment samples 

 were positively correlated with March and 

 April water temperatures. 



14. From the 12 years of entrainment data, the 

 shape of the abundance curve, as measured 

 by the k parameter of the Gompertz function, 

 was found to be a good predictor of subse- 

 quent recruitment of age 3 winter flounder. 

 The shape of the abundance curve was related 



Winter Flounder Studies 



217 



