heterogeneous growth of the scale and fish (NUSCo 1984). Therefore, length at each annulus was 

 calculated for each sex by the non-linear relationships: 



length = 3.557( scale sizef'^"^ for females, (n = 216, r^ = 0.93) (2) 



length = 3.777( scale size)"'^"'* for males, (« = 193, r^ = 0.94) (3) 



Annuli measurements for each fish were substituted into the appropriate regression equation for back- 

 calculation of growth. Mean lengths-at-age with 95% confidence intervals were then computed. 



Using the 1983 length-at-age data, the von Bertalanffy growth model (Ricker 1975; Gallucci and 

 Quinn 1979) was used to describe the growth of Niantic River winter flounder: 



L, = L«,(l-et-^{'-'»'!) (4) . 



where L/ = length in mm at time t 

 K = growth coefficient 

 Loo = asymptotic maximum length 

 ?o = hypothetical date at which a fish would 

 have zero length if it had always grown 

 in the manner described by the equation 



A nonlinear procedure using the modified Gauss-Newton iterative method (SAS Institute Inc. 1985) was 

 used to estimate the growth model parameters from the length-at-age data. The co parameter (the product 

 of Loo and K) of Gallucci and Quirm (1979) was calculated for comparisons of growth. 



Similarly, the growth model was applied to the entire 1977-83 age-length data set. As all age 1 and 

 2 and some age 3 fish were not sexed, these specimens were used with both females (through age 10) and 

 males (age 8). An independent assessment of growth was made using lengths at marking and recapture 

 of 129 females and 81 males tagged with Petersen discs (see Movements and Exploitation below). As 

 recommended by Sundberg (1984), the method of Fabens (1965) was used with these data to estimate Loo 

 and K. Most recaptures used in the analysis were from NUSCo sampling to ensure that length data were 



11 



