A eee 
however, showed that this supposition only held good for the three first groups. Fig. 7 
shows the composition with regard to age according to the results of the scale measure- 
ments. It will be seen that the three first groups consisted, roughly speaking, of fish 
of ®/,, 12/,, and 2°/, years, whereas the fourth group exhibited no fewer than three differ- 
ent year classes, viz; 3°/3, 4?/; and 5?/, year fish. 
From the table of sizes of the small herring taken at Lofoten, and from Figs. 6 and 
7, it will be seen that the growth of the herring varies greatly for different individuals 
of the same year class. The method of age determination by measurement fails, there- 
6 7 8 9 10 1 1314 15 16 17 18 19 2021 22 23 2425 26 27 Bem 
6 7 8 9 10 11 12 15 14 15 16 17 18 1920 21 22 23 24 25 RG 27 cm 
Fig. 6. Sample of small herring and fat herring from North Coast of Norway, 
autumn, 1909. Arranged in cm. groups. (LEA). 
Fig. 7. Same sample as in Fig. 6; individuals of different year classes arranged in cm. groups 
according to age determination. 
1?/; denotes the curve for fish of 1°/; years old (Lea). 
fore, when dealing with older fish, the variation in growth in the older year classes being 
so great as to exhibit a difference in age of one, two, or three years between fish’ of the 
same size*). 
With so great a variation in size between the different individuals it is evident that 
the task of determining the average growth or rate of growth of the herring demands 
a great deal of work, and presents very considerable difficulties. These last are still 
further increased if it is desired to ascertain, not merely the average growth for a single 
*) Previous investigations, based solely upon measurement, have thus exhibited considerable 
errors in determining the age of the fish. This applies also, unfortunately, to my own first preli- 
minary investigations in this respect. 
9 
> 
