282 THE VAIJIATIOX, IJACES AND MIGRATIONS OF THE MACKEREL. 



Standpoint, but the existence of the correlation can readily be demon- 

 strated. 



As the variation of both characters is most marked in the American 

 sample, I have analysed the data provided by it in the following 

 manner : — 



In 2 fisli liaving 9 rays . . . both had 6 finlets. 



had 6 



These figures show (1st) that the normal or modal number of finrays 

 (12) is constantly associated with the normal or modal number of 

 finlets (5) ; (2nd) that when the number of finrays is below 12, the 

 variation in the number of finlets is confined to deviations above the 

 mode ; and (3rd) that when the number of finrays is above 12, the devia- 

 tions from the modal number of finlets are exclusively below the mode. 



This relation can be shown still more clearly as follows : — 



2 had 9 rays. 

 Of 7 fish with G finlets 



91 



These figures show (1st) that the normal number of finlets (5) is 

 associated with a wide range of variation in regard to the number 

 of rays in the second dorsal fin (from 10 to 14), (2nd) that when the 

 number of finlets is above 5, the number of rays is constantly below 

 12; and (3rd) that when the number of finlets is below 5, the number 

 of rays is constantly above 12. 



This correlation is also exhibited by the mean values of each character 

 for the American sample, the mean number of finrays (11'850) being 

 the lowest observed, and the mean number of finlets (5'120) being the 

 highest observed. 



If the mean values for the various British localities are correct, they 

 ought to exhibit a similar correlation in regard to these characters ; but, 

 as will be seen from the following table, such a correlation between the 

 local means does not exist. The localities are grouped in order of 



