28 THE AGE AND GROWTH OF SALMON AND TROUT. 



The number of observations on which each average figure is based is given 

 in brackets. 



If we compare the average lengths for each of the groups into which 

 our material has been divided with the average lengths based on the whole 

 of the 108 fish, we find the following : Taking chiefly small fish for our 

 calculations, such as are contained in the first three groups, we find that 

 the average length very closely approximates to that based on the whole of 

 our material, at any rate as far as the younger year-classes are concerned. 

 It is only in the later years, and especially in the last year, that the figures 

 are less reliable. It will further be noted that as we include larger and 

 larger fish in our investigations the more the figures for the later years 

 coincide. If, however, we compare the figures based on the ten largest 

 fish, and also those of the ten selected at random, together with the average 

 of the whole of the 108 fish we find most remarkable agreement. The 

 differences only amount as a rule to a few millimetres, or exceptionally to 

 one centimetre. It is evident, therefore that the errors which might arise 

 from basing one's calculations on only a small amount of material are not 

 likely to exceed 5 or 6 per cent. 



It appears, therefore, that one would be justified in basing calculations 

 as to the rate of growth on only a small amount of material, provided, 

 however, that the individuals are fairly distributed amongst the various 

 sizes of fish occurring in a lake, or that they are selected from a few of 

 the largest fish to be found there. 



A careful analysis of other similar calculations shows on the whole the 

 same characteristics as the material from Chaigijok, and I give here a 

 similar classification of fish taken from the Sanna Lake, near Harstad.* 

 (See Table XXXIVu.) 



This analysis, in its general features and results, is absolutely similar 

 to what we found in our classification of the fish from Chaigijok. Again 



* The fish taken at random were selected in the same way as those from Chaigijok. 

 The numbers were as follows: 9, 22, 30, 38, 50, 62, 69, 81, 90, 102, 113. 



