CA^'ADIAN FISHERIES EXPEDITIOy, WL'rlo 107 



Kristiansand. The first nineteen samples of large herring, in the winter of 1015-lt}, 

 presented the same appearance as noted so many times l>efore, the year-class 1904 being 

 still predominant, and that of 1910 represented by only a few specimens. The twen- 

 tieth sample, however, the first of the true spring herring, was a great surprise, as it 

 contained one single specimen only of the 1904 year-class, and was otherwise composed 

 mainly of young fish. 6. 5 and 4 years old. i.e.. the year classes 1910 (with 26 per 

 cent), 1911 (18 per cent), and 1912 (43 per cent). Here, then, was the 1910 year- 

 class, but associated with two younger ones, of which the class 1912 especially was 

 present in force. The following spring herring samples (from February, 1916) 

 reverted once to the old style of composition, with 40 to 50 per cent 1904, but in 

 March this is gradually changed, and we have first the 1910 year-class, and later the 

 1912 class, accompanied by the less numerously represented intermediate year-class 

 1911. Fig. 26 shows the entire series of age-analyses for seine-caught spring 

 herring, the previous series of drift-net-caught large herring samples is here omitted 

 for want of space, but may be supplied by imagining the nineteen earlier samples of 

 fairly the same character as Nos. 2 to 4 in the figure. 



A glance at the figure will show that the group of mature fish must have under- 

 gone a thorough change during the course of the season. At the commencement,, 

 there were evidently two very different sub-groups (represented by samples 1 and 2-4), 

 As the season goes on, however, these intermingle, so that the curves for age-distribu- 

 tion exhibit a highly peculiar appearance. Altogether, the different samples agree 

 very well one with another, and this despite the fact that the situation this year was 

 highly variable, and not, as hitherto, practically stationary. 



The entire material of mature Norwegian herring, it will doubtless be admitted 

 distinctly indicates as possible the statistical recording of age-distribution and its 

 variations in this group of the stock. The results arrived at in methodical respects 

 from analysis of this material is, that the year-classes which have passed over into the 

 group in question become so thoroughly mixed that it is possible, even with relatively 

 few and small samples, to keep trace of the condition, when the same is, as during 

 the years 1910-13, mainly stationary. During periods where marked alterations take 

 place, as in 1908, 1914, 1915i, and 1916, the number of samples will need to be greater, 

 in order to furnish a view of the actual changes occurring. Thus, in 1916, no single 

 sample can be taken as representative of the conditions, and erroneous conclusions 

 would certainly have resulted had not several samples been available from various 

 parts of the season, and different fishing grounds. 



It is interesting, from a methodical point of view, to consider somewhat more 

 closely samples from those periods when new year-classes began to make their appear- 

 ance. We have here, first of all, the immigration of the 1904 year-class, in 1908 ; then 

 that of the 1908 class in 1914; the 1910 class in 1915; and finally, the year-classes 

 1910-12 in 1916. 



The 1904 year-class was first observed among the mature fish in 1907,. but only 

 in very small numbers; in 1908, it is found in the first sample from February, although 

 not numerously represented ; in the next and last sample from this season, however, 

 it makes up 65 per cent. In the sample from 1909, it amounts to about 40 per cent, 

 i. e. a decrease in the proportion. This rule, a rise of the percentage to a maximum, 

 followed by a fall, applies to all cases where investigation has been made. The 1908 

 year-class appeared on the south coast at first in great numbers, later on in the same 

 year the percentage was lower. The 1910 year-class reaches a maximum in the penul- 

 timate sample from 1915, with 48 per cent, in the last sample from that year percen- 

 tage is only 33 per cent. The immigration of a new year-class, and the intermingling 

 of the same with the older components of the stock, evidently takes some time, and 

 wovild appear to commence with the entrance of the immigrants in a body, which 

 wedges itself into the stock already on the grounds within a restricted jwrtion of the 

 same. The absence of the 1910 year-class during a great part of the fishing season 



6551— Hi 



