CANADIAN FIh;n FRIES EXPEDITIOW J9l.'rl5 



145 



Only the dimensions t^ — t^ will be of interest in this connection, as t^ is incomplete 

 in the samples from 1914. 



A test calculation of the fractioii -j- gave three ditferences of eighteen, which 

 were over three times as great as the corresponding standard errors. It was the 

 sample from Lockeport which differed in respect of ^ from the West Ardoise (August) 

 sample and from station 42, while station 42 again differed in respect of ^3 from the 

 West Ardoise (August) sample. In the remaining fifteen cases, the differences were 

 less than three times their standard error, 



5 differences being less than once, 

 11 " " " twice, 



15 " " '' three times, 



18 " " '' four times, 



the corresponding error. 



The similarity between the various samples is thus on the whole good, albeit less 

 so that in the cases previously dealt with. It should, however, be borne in mind that 

 the comparisons in this case are made with young fish, where the dissociation of year- 

 classes will be more likely to make itself felt. 



Of other year-classes which might be selected for comparison between one sample 

 and another, that of 1910 is the best; the number of specimens is here, howe\er, so 

 small, that it would not be worth while to make calculations of the standard errors. 

 Table 30 shows the averages for this year-class. 



Table 30. — Averages for the year class 1910 in the four samples mentioned in 



table 29. 



Locality and Date. 



\V. Ardoise, July, 1914 



August, 1!I14 

 Lockefwrt, Xovembei-, 1!»14 

 W. of Pt. Hood, July, 1«J15 



Ivo. of 



In- 

 divid. 



12 

 18 



28 



12 2G 



12-30 



9 82 



11 94 



8 Ofi 

 7-87 

 8-8S 

 785 



Save for the incomplete dimension t^ these samples agree, well enough, as will be 

 seen, always bearing in mind the small number of individuals dealt with. Only the 

 sample from Lockeport differs in respect of the small t^. The interesting sample from 

 station 42 (west of Port Hood) does not appear to differ in any essential degree from 

 the remaining samples. 



Finally, table 31 shows comparison for the 1908 year-class. 



Table 31. — Averages for the year class 1908 in the four samples mentioned in table 29. 



Taking into consideration the extremely small number of observations in each 

 sample, the resemblance here again is noteworthy. With regard to this year class also, 

 the sample from station 42 differs in no way from the remainder. 



