withdrawn by MNPS during the larval period for each year. A 

 distribution-free (nonparametric) method (Snedecor and Cochran 1967) was 

 used to construct a 95% confidence interval around each median and 

 entrainment estimate. 



An entrainment mortality study was conducted from April 25 through 

 May 17. Samples of entrained winter flounder larvae were collected in a 

 slow current area of the Millstone quarry downstream from the MNPS 

 discharges. A 0.5-m diameter, 1-m long, 0.333-mm mesh bridled plankton 

 net was lowered from a boat to the bottom (ca. 15 m) and slowly hauled 

 to the surface to minimize net damage. Samples were transported to the 

 laboratory and gently poured into 2-liter white porcelain trays. Dead 

 specimens were counted and removed. Test groups of five live larvae 

 each were transferred using wide-bore pipettes to 0.333-mm mesh 1-liter 

 holding chambers in flow-through effluent water. Effluent holding time 

 was 2 or 4 h to simulate retention in the quarry with three or two units 

 of MNPS in operation, respectively. Following effluent holding, chambers 

 were moved to ambient flow-through water and observed daily for latent 

 mortality over a 96-h period. All dead larvae or larvae surviving the 

 latent mortality holding period were classified by developmental stage 

 and measured . 



Harmonic regression models 



Fluctuations in the log-transformed catches of winter flounder 

 taken in various monitoring programs were analyzed using harmonic 

 regression techniques described in the Fish Ecology section. Data from 

 the trawl monitoring program were used to calculate the catch of winter 

 flounder for a standard tow of 0.69 km at six stations (Fig. 2), 

 including Niantic River (NR) , Niantic Bay (NB) , Intake (IN) , Twotree 

 Island Channel (TT) , Jordan Cove (JC) , and Bartlett Reef (BR) . Three 

 replicate tows were taken every other week and the log-transformed 

 catches were averaged to produce a single value for every sampling 

 period. Weekly means of the log-transformed catches were calculated 

 from the impingement, entrainment (station EN), and ichthyoplankton (NB) 

 monitoring programs. These values were used to construct various models 

 describing catches from October 1976 through September 1982. The models 



13 



