peak was in the spawning area of the Niantic River. The date of peak abundance progressed to the mouth 

 of the Niantic River, mid-Niantic Bay, and then throughout the Millstone area. It took from 21 d in 

 1975 to 36 d in 1974 for the peak to progress from the Niantic River to mid-Niantic Bay, but once in 

 the bay dispersal was rapid to other areas. This progressive pattern of peak abundance as a relative 

 measure of dispersal rate agreed with the findings of the larval dispersal model (Saila 1976). The Niantic 

 River has high larval retention characteristics, but once larvae enter Niantic Bay they spread throughout 

 the area. 



A more detailed examination of the abundance and distribution of larval winter flounder in the Niantic 

 River and Bay was possible with the data collected from 1981 through 1985. Abundance curves (Equation 

 7) were constructed from the Gompertz function (all R^ values of the Gompertz function exceeded 0.95) 

 for each year in the Niantic River (stations 1 and 2 combined for 1981 and 1982; A, B, and C for 1983 

 through 1985) and Niantic Bay (stations NB and EN combined). If collected, daylight samples during 

 May and .lune were excluded because they would have underestimated larval abundance during this period. 

 Larvae were most abundant in the river during 1982 and 1985 and the time of peak abundance for all 

 years occurred during mid to the latter part of March (Fig. 13). In the bay, larvae were most abundant 

 in 1982 and 1983, with similar numbers during the remaining 3 yr. The timing of peak abundance varied 

 more than for the river, ranging from about the second week of April in 1985 to the fu'st week of May 

 in 1981. Except for 1982, there was no apparent relationship between annual abundance in the river and 

 bay. For example, in 1985 larvae were most abundant in the river but were not abundant in the bay, 

 and in 1983 larvae were most abundant in the bay but were not abundant in the river. 



The abundance of each developmental stage was compared in the Niantic River and Bay from 1 983 

 through 1985, the only years that larvae were classified in developmental stages. The a parmeter from the 

 Gompertz function (Equation 6) was used as an index of abundance. This function fitted the data well 

 with all R^ values exceeding 0.97 and the 95% asymptotic confidence intervals for the a parameter were 

 small (Table 22). However, less than 20 observations were used and the actual intervals may be larger 

 because asymptotic theory generally requires large data sets to apply. The most noticeable difference in 

 the river was the high abundance of Stage 1 and 2 larvae in 1985 compared to the previous 2 yr. This 

 occurred despite decreasing estimates of egg production from 1983 through 1985 (Table 11). The low 

 abundance of Stage 1 and possibly early Stage 2 larvae in 1983 was partly attributed to net extrusion 

 through the 333-nm mesh nets (NUSCo 1985). The previously discussed high densities of larvae in 1985 



69 



