100 DEPARTMENT OF THE ^^ATAL SERTICE 



to the possibility of approximately certain estimates of age. In the Canadian mate- 

 rial, some herrings were found which must certainly have been older than any which 

 I have examined among those from European waters, their age being over 20 years. 

 Where a definite age is assigned to them in the tables, this must expressly be noted as 

 approximate; in cases of this sort accurate determination is out of the question. 



Finally, mention should be made of a particular source of error which needs to 

 be guarded against by special precautions. 



In a sample of herring collected just at the time when the fish are developing 

 the first stages of a new summer zone on the scales, some specimens of a certain year- 

 class may be found where this has already visibly commenced, whereas others of the 

 same group have not yet reached so far. If, then, the rings be counted quite schema- 

 tically, and the observations recorded accordingly, the results will be that fish which 

 have already commenced their summer growth will appear a year older than those 

 which, albeit somewhat behindhand in this respect, are in reality contemporary with 

 the former, and the greatest confusion will ensue. 



In many cases, e. g., in dealing with young herring, the change is more or less 

 conspicuous and the error may be avoided ; as in such specimens, a narrow summer 

 zone is visible outside a broader one, or several such. It is a different matter, however, 

 when the fish are older, and have already one or more narrow zones near the margin 

 of the scale. A new narrow zone on such a scale does not occasion any conspicuous and 

 easily disting-uishable alteration. In this latter case, it is difficult to say what precau- 

 tions should be taken unless it be to avoid collecting samples during this transition 

 period, or, possibly better, to supplement such samples by others taken before and after. 

 The grown Norwegian herring, it may be noted, are not regularly fished for during 

 the transition period. 



V. THE SCALE OF THE HERRING AS AN INDICATION OF GROWTH. 



SOURCES OF ERROR. 



If, in addition to counting the winter rings on the scales, we measure the dis- 

 tance between them, these measurements will enable us to calculate how each fish has 

 grown from year to year. This, of course, presupposes that the growth of the scale 

 takes place at a rate simply proportionate to the rate of growth of the fish. Judging 

 from the investigations already made (vide Lea IX), the scales from various parts of 

 the body differ somewhat in this respect. On the whole, however, we may say that such 

 proportion does exist. ♦ 



It has been found, however, that in dealing with the material obtained from such 

 measurements of growth, the values arrived at in the case of j'oung fish appear to be 

 higher than the corresponding figures for the older groups (vide Lea XI and Lee XII), 

 so that possibly there may be a systematic error in the growth measurements. 



This is, in my opinion, only true to a certain extent. Moreover, some part of the 

 apparently systematic error will, as a matter of fact, be found to arise from other 

 causes, to wit, as the results of a biological process. Nevertheless, we may, in prac- 

 tice, until these problems are fully solved, do well to handle our material as if there 

 were a systematic error, of the nature shown in the following taW?:^ — 



