106 DEPARTsMENT OF THE NATAL SERVICE 



The recognition of this important phenomenon leads us to the conclusion that 

 an investigation of the age-distribution in a herring stock should include a study of 

 the different groups, and that the interchange between them should be most attentively 

 observed. One of these groups, that comprising the mature fish, has been under 

 observation for ten years in Norway. The experience gained during the course of 

 the work distinctly points to the possibility of following the age-composition, and its 

 variations, within this group. These investigations have, in methodical respects 

 especially, furnished interesting results; a brief description and discussion will 

 therefore here be given of the material collected for the study of this age-composition 

 and renewal of the mature group. 



During the first years, from 1907 to 1913, only a small number of samisles, from 

 two to four, of the true full herring were collected annually, with, in addition, one to 

 three samples of the so-called large herring, i. e.i, mature fish with genital products 

 large but firm, which we now know to be very closely related to the actual spawning 

 herring. Despite the small number of samples, and of specimens investigated, the 

 samples were found to agree so closely one with another, and the features observable 

 with regard to age-composition so marked as to convince the investigators at work on 

 the material that even these few samples, and this small number of specimens, might 

 yet be taken as representative, so far related to a certain very important point in the 

 distribution of year-classes in the group of mature Norwegian herring. The main 

 point in question was the fact that the year-class 1904 exhibited a marked numerical 

 superiority over all others. 



This year-class made its first appearance in any great number in a sample from 

 April, 1908. It was thereafter found, in every single sample investigated during the 

 years in question, to be enormously superior in numbers to all other year-classes. The 

 striking contrast between the numbers of this one year-class and those of the many 

 others, and the fact that this contrast was maintained for several years, served largely 

 to confirm the conviction in the minds of the investigators. In order to give the 

 surest possible foimdation for the observations;, a larger number of samples was 

 collected during the folio-wing years, from 1914 to 1916. The analysis of this material 

 confirmed most emphatically the presumption already arrived at, the year-class 1904, 

 despite its continually increasing age, being still found to occupy a dominant position. 

 One sample from the southern verge of the spawning grounds (Ivristiansand, February, 

 1914) contained, besides the 1914 year-class, another numerically strong age-group, 

 that of 1908 (vide Hjort V), but as this peculiarity was not obsers-ed in subsequent 

 samples from either the same area or elsewhere it was presumed that the sample in 

 question only represented a slight local disturbance occasioned by the immigration of 

 a new year-class, which supposition was later confirmed. The winter season 1914-15 

 passed, and still the 1904 year-class, with its now 11-year-old fish, was seen to pre- 

 dominate. The first samples then investigated, however, already indicated that the 

 1910 year-class would now come to occupy a distinctive position, being more numerously 

 represented than the adjacent classes, albeit far from equal in this respect to the old 

 1904 class. At first, also, there was but a slight degree of uniformity between the 

 different samples as to the numerical value of this new year-class. Not xuitil the 

 commencement of March, 1915, was it seen to be evident beyond question that this 

 year-class was decidedly richer than its older neighbour 1909, and in the two last 

 samples of the season, it even rivalled 1904, the last but one containing 48 per cent 

 1910 and 25 per cent 1904, the figures for the last sample being 33 per cent and 38 per 

 cent, respectively. 



With these results in mind, the following season would appear to be doubly 

 interesting; in the first place there was the question, would the 1904 year-class, now 

 no less than 12 years old, continue to assert itself as heretofore; and, secondly, would 

 the new year-class, maintain the same distinctive position as in the two last samples 

 from 1915, or fail to maintain it as had been seen in the case of the 1908 sample from 



