were few and sporadic (3% of the total tagged) , the data were used to 

 calculate population estimates using the Jolly (1965) method for open 

 populations. However, results were judged meaningless because of the 

 large negative estimates of recruitment, survival estimates much greater 

 than 1, and large calculated standard errors (NUSCo 1975). For these 

 reasons the 1973 and 1974 estimates have not been considered usuable for 

 impact analysis. 



In 1975, a more intensive sampling program was designed and undertaken 

 to estimate the number of winter flounder found in the Niantic River 

 during the spawning and immediate post-spawning seasons (NUSCo 1976). 

 In general, samples for population estimates from 1975 to the present 

 were taken over a 5 to 9 week period in late winter through early spring. 

 Most sampling was conducted from two or three research vessels each 

 using a 9.1-m otter trawl on Monday and Tuesday of every week. Most tows 

 were taken at the station or stations with greatest catch-per -unit- 

 effort (CPUE) of that week. Most tows were approximately 5 min in 

 duration, although some were as long as 30 min. Tow duration varied 

 according to size of trawlable ground, expected catch, and amount of 

 eelgrass and algae which clogged the nets. 



The design of the 1975 study was based on the requirements for the 

 triple-trellis method of population estimation (Ricker 1958) . Eight 

 stations throughout the Niantic River were sampled by trawl to capture 

 winter flounder for tagging (Fig. 1) . Fin clipping was used to mark 

 individuals and 10, 633 were marked during a 7-week period beginning on 

 March 31. All fish larger then 150 mm were marked, a practice which has 

 continued to the present. Three population estimates were calculated, 

 one based on the triple-trellis technique and two using the Jolly technique. 



