110 



DEPARTMENT OF THE NATAL SERVICE 



the samples as a rule include more than ten age-groups, this result cannot but be said 

 to be surprising. It might easily be imagined, for instance, that the difference in 

 size between the younger and the older fish would be so considerable as to involve a 

 degree of net selection far from negligible, the smallest and largest specimens avoiding 

 capture, whereby the net samples would be seriously biassed. That this proved not 

 to be the case, with the Norwegian samples of mature herring, is due to an important 

 biological phenomenon, which we shall now have occasion to examine more closely, 

 in considering the question of how best to procure representative material for the 

 study of growth. 



As already mentioned, the individuals in a year-class may fall into several sub- 

 groups, appearing in company with older or younger fish. This subdivision of a year- 

 class does not appear as merely accidental, but seems on the contrary to be regulated 

 in a definite manner according to the stage of development at which the fish have 

 arrived. This is, of course, natural enough in cases where some individuals of a year- 

 class attain maturity and are ready to spawn, while others are still immature. Noth- 

 ing could be more reasonable than to suppose that the mature fish should part com- 

 pany with those of their contemporaries that have not reached that stage, and move 

 off by themselves to the spawning grounds. The year-class is thus divided up anrl 

 dissociated according to the degree of maturity of the genital products. We now find, 

 however, that the mature individuals in a year-class thus divided are of larger — often 

 considerably larger — size than those still immature. There exists a positive correla- 

 tion between degree of maturity and size. The year-class thus divided according to 

 degree of maturity will therefore also be found dissociated according to size, the 

 larger specimens of such a year-class will be found associated with older, mature fish, 

 the smaller with younger, immature companions. 



This being ascertained, the question then arises as to whether the attainment of 

 maturity should be regarded as the only phases of development accompanied by such 

 dissociation in point of size among the individuals of a year-class. 



Table 3. — Age distribution (showing percentage) in four samples of immature 



herring from northern Norway, autumn 191.5. Group 3 (2§ years old fish) are 



in two samples found in association with older herring, and in two samples 

 with younger herring. 



The answer here must be that the dissociation already mentioned (vide table 3) is 

 also regulated according to size; we find, for instance, that the average length of the 

 2J-year-old fish (group 3 in table 3) is considerably greater in the two first samples, 

 where this age-group was associated with older fish, than in the two last, where it was 

 found in company with younger fish (22-1, 22 -61, 17-0 and 18-3 cm., respectively). 

 The four samples in the table are from hauls made within a restricted area, the 

 greatest distance between places of capture being about 40 km. and within a short 

 space of lijne, vi?., from 15th October to 11th November. An even more striking 



