— 34 — 



Four points are very distinctly evident from the above table: 1) that the 

 younger fish (Group I) are found in greater numbers in the southern sections of 

 the fishery district, 2) that the reverse is the case with the older fish (Group III), 

 3) that the younger fish in the large herring district (a and b) are more numerous 

 at the beginning of the fishing season than later on, and 4) that the reverse is 

 the case in the true spring herring district (c). 



This remarkable grouping can certainly not be due to any of the sources 

 of error already mentioned, (there can here be no question of any save that of 

 net selection) ; practically all the samples from district b are net samples, and the 

 three seasonal observations from here are comparable; practically all the samples 

 from district c are seine samples, and comparable, while the seine samples lie 

 between two series of net samples b and d, so that any effect of selection would 

 have made itself apparent in the form of a break in the continuity from 

 north to south. 



Conclusion. 



The foregoing, and in particular the facts indicated by Tables II and VI, 

 will suffice to show what methods should be adopted in order to obtain as 

 representative material as possible. Table VI points the necessity of having the 

 greatest obtainable number of samples taken at different places and times, while 

 from Table II, we find that the samples themselves may be comparatively small. 

 For the purposes of an annual investigation of the Norwegian large and spring 

 herring fisheries, weekly samples of 200 specimens from each of the four districts 

 represented in Table VI would suffice. As the fishery goes, this would probably 

 mean some 40 samples in all. Tables IV and V, moreover, indicate that samples 

 should preferably be taken from seine hauls, but that net samples may also 

 be used.' 



Vardö, -"(j 1914. 



EiNAR Lea. 



