42* 



40» 



740 72" 70" 



Figure 1. — Chart of the northern part of the Middle Atlantic Bight. 



38* 



relative frequency in catches to improve their 

 representation in the samples. Examination 

 of the length- weight data by size category 

 indicated that no apparent bias resulted from 

 this selection. 



Opportunities for sampling the commercial 

 catch depended on the level of fishing effort 

 for sumnner flounder and on landings of this 

 species at certain piers in the fishing ports 

 where sampling could be done. Because 

 sampling was about proportional to landings, 

 we believe it fair to assume that the estimated 

 length- weight equations adequately represent 

 the length- weight relation of fish in the New 

 England catch in each season. 



Length, weight, and sex were determined for 

 346 summer flounder in a single catch by a 

 research vessel in June 1961 at Menemsha 

 Bight, Mass. Because the measurements were 

 made in the laboratory, it is likely that they 

 were more accurate than those from com- 

 mercial san^ples, which were made under less 

 ideal conditions on the fish-unloading piers. 



Total lengths of all fish were measured to 

 the nearest millimeter. Weights in grams 

 were obtained with adjustable spring balances 



of three capacities in order to read all weights 

 to the nearest unit less than 0.5 percent of 

 the fish weight. We checked the balances oc- 

 casionally with standard weights and adjusted 

 them as necessary. Their weighing error was 

 about 1 percent. 

 The equation 



W = c L'' 



in which W = weight, L = length, and c and b 

 are constants was assumed to express the 

 relation of length to weight. The least squares 

 regression of the logarithmic transformation 



Y = a + bX 



in which Y = log W, a = log c, and X = 



log L was used for estimating values of c 



and b. 



Logarithms of lengths and weights were 

 rounded to 3 decimal places. Each fish was 

 treated as 1 point. Regression statistics for 

 all of the data are given in appendix table 1 . 

 Notation for regression analysis follows 

 Snedecor (1956). 



