CA^AD]A^' FISH ERIE f< EXI'EDITWX, lOl'rlo 



139 



The standard errors of the averages for dimensions t^ — t^ have been calculated for 

 each sample separately and for both taken together. With the aid of these, it has 

 been possible to estimate the importance to be attached to the differences. It was 

 found that all the differences were between once and twice as large as the correspond- 

 ing error, and they can thus hardly be considered as noteworthy. And as the remain- 

 ing growth dimensions have also been seen to exhibit no difference, the two samples 

 should be regarded as equal, so far as concerns the growth of the 1903 year-class. 

 Characteristic features common to both samples are: i, is large, distinctly greater 

 that t„; average values of less than 1 cm. are reached as early as t.; and finally, t^ and 

 t„ are less than the following dimensions, t^^ and i^^. 



If we now turn to the remaining year-classes in these samples, we find, as will 

 be seen from tabic 20, entirely analogous conditions, all year-classes exhibit an un- 

 mistakable likeness in respect of growth. There is the great ^i and the low values for 

 t; and onwards, and even so slight a detail as the fact that the last year's growth is 

 somewhat greater than that of the few years immediately preceding, is met with now 

 and again. As regards the 1904 year-class especially, which it would seem desirable 

 to compare with the same year-class in the samples from Xewfoundland, we find that 

 it forms no exception, save perhaps that the characteristic features are possibly even 

 more marked. 



Table 20. — Averages for year class 1903 compared with those for other year classes in 

 the samples from Magdalen Islands. 



We find then, for these two samples also, that the growth of the fish is alike in 

 both; and that 1903 year-class may be taken as representative for the remainder, at 

 any rate for the older fish. 



3. Samples from Nortlnimherland Strait. — In the five samples of grown fish from 

 this area, 1903 is the best common year-class among the older fish, and 1907 among 

 those of medium age. These two year-classes were therefore subjected to closer 

 examination. Tables 21 and 22 show the averages and standard errors for these year- 

 classes, in each of the five samples. 



6551—13* 



