CAXADiAX Flail i:nn:>i kxi'fditiox, loi't-io 



101 



T.\ni.i: 2. — Showing the decrease with increasing age of the corresponding calculated 

 average values for the yearly inoremciit in loiiiith. (Table reprinted from 

 T.ea (V) tab. 7.) 



* l>eiiote incompleted increments. 



In the table above, t, indicates increment of leii°^th during first summer of life, 

 t.. tlie corres])onding increment for second summer, etc. It will be noticed that the 

 figures in the vertical columns exhibit decreasing values, and that a growth curve 

 drawn on the basis of the figures for fish at 3* years would differ from one based on 

 the values for fish 6J years old. This feature, which is very freiiuently encountered, 

 may render immediate comparison between the growth values for young fish and those 

 for older specimens precarious ; therefore, where possible, samples of herrings of more 

 or less equal age should be compared. 



As regards the degree of accuracy obtainable in such individual measurements of 

 growth, this is fairly high, and it is always possible, where greater accuracy is needed, 

 to measure several scales from the same fish and take their average. This branch of 

 the question need not therefore be further dealt witb here. It should, however, be 

 noted that all the sources of error which make themselves apparent in age determin- 

 ations also apply to measurements of growth. If a ring be overlooked, then two years 

 growth will be taken together as that of one, and the growth of the succeeding years 

 will be erroneously reckoned, etc. If, on the other hand, a secondary ring be mistaken 

 as a true one, the year's growth will be divided into two, and that of the following 

 years again incorrectly recorded. 



Save in the case of particularly diffii-ult scales, however, errors of this nature 

 will rarel.v occur in the values for growth during the first few years. 



VI. AGE DETEmnXATIOX AND (JROWTM MEASUKEMEXT TX PR.VCTICE. 



Where the winter rings are distinctly marked, and the fish young, it is of little 

 importance from what part of the body the scales are selected; in the reverse case, 

 however, this may be of the utmost imi>ortance. In all the scales of a herring the 

 winter rings are by no means equally distinct. Where possible, therefore, the scales 

 where these are most pronounced should always be chosen, i. e., the large scales from 

 the middle forepart of the body. 



VII. PREPARATION OF SCALES FOR PRESERVATION. 



In the earlier years of the Norwegian investigations, the herring scales were 

 scraped from the body with a knife, and placed in small envelopes, where they gra- 

 duallv dried. Before being used, they were soaked in a soap solution to which gl.vce- 

 rine was added, which afforded a good enough means of cleansing them. This is a 

 practicable method, and even presents certain advantages, when working on the fishing 

 grounds, or on board a vessel under unfavourable conditions. Of late years, another 

 method has been employed; three scales are plucked out with a pair of tweezers, clean- 



