squares estimates of these parameters and their 

 asymptotic 95 "/o confidence intervals were ob- 

 tained by fitting the Gompertz function to the 

 cumulative abundance data (based on the weekly 

 geometric means) using nonlinear regression meth- 

 ods (SAS Institute Inc. 1985). The a parameter 

 was used as an index to compare annual abun- 

 dances and the time of peak abundance was es- 

 timated as the date t; corresponding to the inflec- 

 tion point of the function defined by its parameters 

 P and K as: 



(log,P) 



'/ = 



The presence of compensatory mortality during 

 the early life history stages would help mitigate 

 the loss of entrained eggs and larvae. When abun- 

 dance estimates of both eggs and larvae were 

 available, density-dependent mortality was inves- 

 tigated with the following relationship (Rickcr 

 1975): 



lo8.(|) 



a-H pE 



where L = larval abundance estimate 

 E = egg abundance estimate 

 a = intercept 



P = index of density-dependent 

 mortality 

 If the slope (P) is positive the density-dependent 

 mortality is depensatory and if negative it is com- 

 pensatory. 



Armual entrainment estimates were calculated 

 for dominant ichthyoplankton species entrained. 

 These estimates were obtained by multiplying the 

 median density at EN during the period when 

 95% of the annual cumulative abundance oc- 

 curred times the total volume of water passed 

 through MNPS during the same period. A 

 nonparametric method (Snedecor and Cochran 

 1967) was used to construct 95% confidence in- 

 tervals around each median density and corre- 

 sponding entrainment estimate. 



Monthly impingement estimates were based on 

 the extrapolation of actual counts using a volu- 

 metric ratio. The daily cooling water volume was 

 calculated based on 15-minute flow rates from 

 0800 to 0745, the time corresponding to the actual 

 impingement period. Within each month, an es- 

 timate for every day not sampled was calculated 

 by multiplying the average impingement density 

 (number of fish per m' of cooling water) based 

 on the days sampled in that month times the 

 volume of cooling water on each day not sampled. 

 All of these daily estimates were then added to 

 the sum of the actual sample counts to arrive at 

 the monthly totals for each species. Annual 

 impingement estimates were calculated by sum- 

 ming the monthly estimates. 



As stated previously, seine sampling effort was 

 stratified by season and impingement sampling 

 was stratified by month. Therefore, whenever 

 appropriate, the length-frequency data were 

 weighted to account for unequal effort during the 

 year. Because seine sampling effort from April 

 through October was twice that during the re- 

 mainder of the year, data collected from November 

 through March were weighted by a factor of two. 

 I'or impingement collections, monthly weight fac- 

 tors of 4 (January), 2 (February and March), 6 

 (April), 7 (May through November), and 3 (De- 

 cember) were used to standardize the effort. 



Data on the annual abundance of fishes in LIS 

 and adjacent areas were examined to determine if 

 observed changes in the Millstone area were lo- 

 calized or evident over a larger area. Trawl and 

 ichthyoplankton data were available from moni- 

 toring studies at the Shoreham Nuclear Power 

 Station (SNPS) and summarized for 1977-82 

 ((Jcomct Tech. 1983) and 1983-1986 (EA Eng., 

 Sci., and Tech. 1987). SNPS is located on the 

 southern shore of LIS almost directly south of 

 New Haven, CT. The available trawl data were 

 converted to annual catch-per-unit-effort for day 

 collections. The ichthyoplankton data were sum- 

 marized in the reports for 1977-82 as the annual 

 sum of the mean densities ( 1 000 m ) per sampling 

 trip and for 1983-86 as the annual sum of the 

 monthly mean densities. Because the 1977-82 



Fish Ecology Studies 259 



