130 DEPARTMENT OF THE NAVAL SERVICE 



which owing to the isohited locality of capture, situated so to speak, in the middle 

 of Cape Breton Island, must be considered apart, the samples may naturally be 

 divided, according to the age-analyses, into two groups, one embracing those from Main- 

 a-Dieu and North Sydney, the other those from West Ardoise and Nova Scotia. 

 The best sample in the northern group bears, as we have seen, a strong resemblance 

 to the single divergent sample from Newfoundland, while the southern samples again 

 exhibit features in common with a similarly exceptional sample from station 42, west 

 of Cape Breton island. In the northern group we find, presuming that summer growth 

 had not commenced in the beginning of June, the year-classes 1910, 1908, 1904 and 

 1903 as most numerously represented, in the southern, the 1911 year-class especially, 

 with, to a lesser degree, those of 1910 and 1908, predominating among the younger fish, 

 while in the case of the older ones, the uncertainty of the age-determination renders 

 it impossible to say definitely which year-classes are here the richest. 



e. Comparison of the different areas. 



In the foregoing, an attempt has been made to describe the age distribution in 

 the samples from the different areas, on the basis of a preliminary group arrangement. 

 And it was found that samples from dift'erent waters may dift'er altogether in point 

 of age composition, while those from one and the same locality exhibit a high degree 

 of similarity. Particularly striking is the likeness observable between most of the 

 samples from the Newfoundland coast, as also between those from Northumberland 

 strait. Ha\'ing now compared the separate samples and noted the points of 

 resemblance for the different areas, with due reference to such exceptions as occur, 

 it may finally be of interest to draw up as far as possible, a brief survey of the entire 

 area embraced, on the basis of the details furnished by analysis of the age distribution. 



Fig. 38 shows the age distribution for a number of samples from different waters, 

 chosen from among the 1914 and 1915 samples. The principles upon which such 

 selection has been based will be found justified by the foregoing. 



The figure shows, in its own way, how thoroughly unlike the samples from different 

 waters proved. Especially characteristic is the difference between the Newfoundland 

 samples, on the one hand, and those from the southern parts of the gulf on the other. 

 The samples from the Atlantic coast, however, also exhibit a marked dissimilarity to 

 the southern gulf samples and also to those from Newfoundland. 



On comparing the two parts of the figure, we find, feature for feature,, a consider- 

 able likeness between them. Particularly marked is the resemblance between the 

 Newfoundland samples in the two years and those from Northumberland strait. 



Having regard to the nature and quantity of the material, it must be admitted 

 that the analyses distinctly suggest the existence of several types of age composition in 

 the waters investigated, and it would also seem that in arranging the samples 

 according to age composition, we shall simultaneously have arranged them, roughly 

 speaking, according to locality of capture, i.e., that the samples from one and the 

 same locality exhibit similarity of age composition. That we can speak of such a 

 thing as a "type of age composition" is due to the fact that in a sample from a cer- 

 tain water, certain year-classes are found to be more niimerously represented than 

 the remainder, all year-classes present are not equally rich, and the curves for these 

 fluctuations thus acquire a typical appearance. 



These differences in the samples from the various waters give rise to the sup- 

 position that the area investigated may include several races or tribes of herring, 

 each for the present with its own characteristic age composition. Judging from the 

 age curves, it might possibly he advisable in the investigations by means of growth 

 measurements, to start from the hypothesis here indicated by the arrangement of 

 fig. 38. 



1. That the Newfoundland herring form a tribe apart, represented through- 

 out the present material in all samples, save the single exception. 



