Physiological Changes with Age in Fish 201 



weight regression. There was a significant difference due to 

 age on the ovary weight/length relationship in the March 

 sample, but on good grounds Bagenal considers this to be an 

 anomalous result. Weight increases at a power 3-11 of length 

 in maturing females while ovary weight increases at about the 

 3 • 5 power of length. The latter figure is an estimate because 

 the entire data would have to be recalculated in order to 

 obtain an overall coefficient for the two samples. Since the 

 ovary increases at a more rapid rate than body weight, atten- 

 tion is again directed to Orton's theory of" over-reproduction". 

 Whether or not there is a statistical difference between these 

 coefficients would require a separate analysis. 



There was no effect of age on fecundity /length or fecundity/ 

 weight relationships according to a covariance analysis. Thus 

 differences in these relations could not be attributed to age 

 but to the individual variations that occur in fecundity at 

 any given length or weight. This result makes us more cautious 

 in accepting the rather small decline in fecundity of six- and 

 seven-year-old haddock which Raitt found. He presented no 

 measure of deviations from the regressions, and since the 

 individual measurements were not given the computation 

 cannot be performed. 



Even though Bagenal was satisfied with statistical proof 

 that age played no significant role in determining ovary weight 

 or fecundity, he was disturbed about two features in his data 

 which did not coincide with this interpretation. Milinsky's 

 (1944) very large dabs from the Barents Sea did not produce 

 the egg numbers expected of their size, based on the Scottish 

 population. The difference might be explained by geographical 

 or by racial variation, which is often very great. Also, the 

 number of eggs did not increase in proportion to the weight of 

 the gonad. The regression coefficients between these two 

 variables are 0-6907 and 0-8117 in his two samples computed 

 from a regression of the logarithm of fecundity on the loga- 

 rithm of ovary weight. The latter result was, as Bagenal 

 says, "... unexpected since the larger gonads will have a 



