CAXADI.iy FISHERIE^^ EXPEDITIOy, 191',-lo 



111 



example of this size-regulated dissociation, among immature fish, is furnished by two 

 samples from hauls made at an interval of eight days and 21 km. apart. Table 5 

 shows the age-distribution in these two samples, and the difference between the average 

 sizes of the year-class 1913. 



Table 5. — Showing difference in average size of herring belonging to the same year- 

 group (1913) in two samples. In one sample the herring considered were found 

 associated with younger herring, in the other with older herring. 



Finally, it should be mentioned, that mature, full specimens of the same year- 

 class were also found among the spring herring in the spring of 1916, the average 

 size of these was 25-8 cm., or considerably in excess of that found in any sample of 

 immature herring, where the maximal average size was 22-7 cm. Neither the differ- 

 ence in size nor in degree of maturity, can easily be explained as a result of growth in 

 the period between late autumn 1915 and the spring of 1916, the discrepancy being too 

 great, the time too short, and the season, as has been shown by experience, being a 

 period of stagnation. It is therefore most reasonable to suppose that the said year- 

 class was dissociated into at least three groups, one consisting of small fish, associated 

 with younger ones; another of large, but immature individuals associated with their 

 seniors; and finally one comprising those of large size and completed maturity, 

 associated with the adults. 



The remaining two important year-classes in the material, viz., those of 1914 

 and 1912, do not exhibit fluctuations so violent as in that of 1913; there is, however, 

 also here a suggestion of dissociation according to size, wherefore it would be well to 

 reckon with the possibility that a year-class may become dissociated and grouped 

 according to size of individuals throughout the entire period of growth until full 

 maturity is attained. 



What takes place after this time, when all surviving individuals of the year-class 

 have reached maturity, cannot be stai;ed with certainty as yet; there is, however, reason 

 to believe that the process of maturing should be regarded as a phase of development 

 which, when completed, leads the separate components of a divided year-class to 

 reassemble, i. e., that the sub-groups formed by the varying rate of development will, 

 after maturity and subsequent spawning, reconsolidate into a whole. 



At any rate, it would seem, from the results of the Norwegian growth investiga- 

 tions, that the components of the rich year-class 1904 — which, from its numerical 

 importance and pecularities of growth, furnishes excellent material for a study of this 

 question — did reassemble after their separation during the process of attaining matu- 

 rity, and were later found mingled and collectively in the samples of mature fish. 

 The observations made from year to year present, when seen together, the picture of 

 a process terminating in a fairly stable mixture of the heterogeneous elements which 

 compose the year-class in question. This is especially noticeable in the case of that 

 growth-dimension which exhibits the greatest and most peculiar variation, viz., the 

 increment for the third year of life (^,). Fig. 27 shows curves of frequency for this 

 dimension, the observation? of each year being noted separately, and subdivide'1 into 



