184 



FISHERY BULLETIN OF THE FISH AND WILDLIFE SERVICE 



20- 



10 



BROWNS BANK 



/ V \ A 

 .s:;^^ ^::frws- :::» - — -^-^ "' --° '° 



1 30 

 cr 



UJ 

 Q. 



20 



10 



GEORGES BANK 



.^-O' 



?.^-- — --° ~6--'sr.m-?—-- — t. 



25 30 35 40 45 50 



LENGTH IN CENTIMETERS 



Figure 4. — Percentage age compositions of haddock catches from Georges and Browns Banks. 



55 



error. The variation within the age groups was 

 such that if sampHng continued, about 95 percent 

 of the mean lengths obtained might be expected 

 to fall within the limits indicated. 



The differences in average length between the 

 Georges and Browns Banks fish for each age were 

 found to be highly significant.^ The F-test (pro- 

 duced by the analysis of variance of the lengths 

 in tables 2 and 3) showed values far in excess of 

 the 1 percent level. The probability is much less 

 than 1 in 100 that such a large difference in the 

 average length at each age would be due to chance 

 sampling of a homogeneous population of fish. If 

 average lengths are plotted against age for the 

 Browns Bank data, it is seen that the curve is not 

 as smooth as that for the Georges Bank data and 

 that two points (2- and 3-year-olds) seem par- 

 ticularly at variance with (below) what we might 

 expect in a "normal" growth curve. We believe 

 this is due to the fact that, in general, there is 



^ .\ges 1 to .'» only. No tests of slcntflcance wpre con'puted for older ages. 



variability in growth rate between different year 

 classes and one cannot expect to obtain a smooth 

 or "normal" growth curve from a single year's 

 collection of data. Specifically, we contend that 

 the 1946 and 1947 year classes on Browns Bank 

 were subnormal in attained size. 



Table 6. — Mean length and length range of various ages 

 of haddock from Georges and Browns Banks 



[.\verages computed from tables 4 and 5: standard errors from tables 2 

 and 3. Figures In parentheses show number of fish for which ages were 



1 Mean±2 times Its standard error. 



