A similar approach was used in examining the length-fecundity 

 relationships. The 1977 Niantic River equation has been used with 

 length frequency data and population estimates to calculate annual total 

 egg production in the Niantic River. The 1982 recalculated estimate 

 differed only slightly for 25 and 30 cm fish (2-3%) but was 10% higher 

 for a 40 cm specimen (Table 22) . The Millstone regression equations 

 produced estimates 8-12% lower (1977) and 14-19% higher (1982) than the 

 1977 Niantic River base. The addition of 16 specimens to the Millstone 

 data base in 1982 caused increases of 24 to 35% in predicted fecundities 

 for fish of the same length from that predicted in 1977. This is somewhat 

 troublesome and illustrates how much variability can be produced in such 



studies. The combined Millstone and Niantic River data only improved 



2 

 the r slightly although the standard errors were reduced considerably. 



The simple regression determined with combined data had calculated 



fecundity estimates from 7 to 14% greater than the base. The geometric 



mean regression using the combined data produced lower estimates for a 



25 cm fish (-9%), little change for a 30 cm fish (1%), and a large 



increase for a 40 cm winter flounder (17%) . 



A comparison was made among the 1977 Niantic River, the 1982 combined, 

 and the 1982 geometric mean regressions using the 1981 winter flounder 

 population survey data to calculate total and mean fecundity (Table 23) . 

 The 1982 combined and the 1982 geometric mean regression produced 

 total and mean fecundities that were 12 and 8%, respectively, greater 

 than the 1977 standard. 



Finally, Tyler and Dunn (197 6) reported that the winter flounder 

 seemed to have a reproductive strategy that allowed it to sacrifice egg 

 production to maintain body weight when food was scarce. Thus some 



56 



