METHODS OF INVESTIGATION. 



21 



(3) ANALYSIS OF TROUT FROM BERGST0VAND, OSVASDRAGET, IN THE Os DISTRICT 



(Of. Tables XXV., XXVI.). 



that we find absolute agreement, and even then it may be merely due to 

 coincidence. Whether the examples examined be fairly large or relatively 

 small in number our tests invariably show some discrepancy in the average 

 length of the different year-classes. For these discrepancies there may be 

 various reasons. They may be due to actual variation in the size of the 

 fish owing to the catches not having taken place at the same time or locality. 

 They may be due to the difference in the number of specimens in each 

 sample, or there may have been mistakes in measurement. These dis- 

 crepancies, however, are, after all, only small, and the probable error to 

 which one is exposed in basing growth determinations on samples of 50 

 to 150 individuals does not amount to more than 1 cm. in those year- 

 classes which are represented by a few individuals, and in those classes 

 which are better represented the error is less, and does not amount to more 

 than 5 to 6 per cent, at most in the total length of the fish. 



For exact determination of the rate of growth it is evident that a larger 

 amount of material is necessary, but the comparisons above clearly show 

 that the examination of even 50 individuals and upwards can furnish results 

 which are quite useful in arriving at general conclusions as to the rate of 

 growth in various localities. 



I have, however, for various reasons, considered it advisable to ascer- 

 tain whether the examination of even a smaller number of individuals might 

 not provide equally serviceable results for the purposes of comparison. To 

 a certain extent this question is of theoretic interest in considering generally 

 variations in growth. It would be of immense practical utility if we could 

 take a few fish from any lake and form a reliable opinion as to their growth 

 and as to the usual course of growth in that particular lake. If this were 

 possible it would naturally make it far easier to obtain a definite comparison 

 of the various conditions of growth in different waters. 



I must again insist on the fact that the scales are permanent structures, 



