THE VARIATION, RACES AND MIGRATIONS OF THE MACKEREL. 265 



in regard to the nuraber of spots, the numbers from 1 to 11 occurring 

 with a frequency which decreases fairly regularly as the number of 

 spots increases. In one fish actually 17 spots were enumerated. 



It is not to be expected that a character which is altogether absent in 

 75% of the fish would show any great uniformity of variation in 

 samples of only 100 fish. 



The range of variation in the samples, excluding those of less 

 than 70 fish,* is, however, as follows : — 



Locality. Percentage of Spotty Fish. No. of Spots per 100 fish. 



Lowestoft and Ilamsgate . . 17% to 28% ... 24 to 57 



Plymouth .... 12% to 25% ... 25 to 57 



Brest and Scilly . 18% to 26% .. 32 (each) 



Kinsale 16% to 23% ... 22 to 40 



Kerry . . . . . 7% to 12% ... 8 to 20 



As we have already seen for previous characters, so here in regard to 

 spottiness, the table brings out the close relationship between the Xorth 

 Sea and Plymouth fish with a fidelity which is really astonishing ; and 

 the close affinity of the Brest and Scilly fish with those of the former 

 regions is also clearly shown. The novel feature of the table is the 

 remarkable difference between Kinsale and Kerry fish in regard to this 

 character, those of the West coast of Ireland having a very low degree 

 of spottiness, while those of the South coast occupy an intermediate 

 position between the Kerry fish and those of the Channel, approximating, 

 however, more closely to the latter. The maximum percentage of spotty 

 fish in the Kerry samples scarcely attains to the minimum percentage 

 in the samples from every other region, and the maximum number of 

 spots per hundred fish is actually less than the minimum number for 

 other localities. 



The reliability of the data as a test of racial peculiarities is established 

 in my opinion by the close conformity of the results in regard to the 

 North Sea and Plymouth fish, since these are groups the racial identity 

 of which is rendered antecedently probable by the sequence of events 

 in the fishing seasons, and by geographical and physical considerations. 

 This antecedent probability has already been confirmed by the results 

 described in the section on the number of transverse bars. 



The sensitiveness of the character as a test of racial differences is also 

 confirmed by the great contrast which has been shown to exist between 

 American and European fish in regard to this character, a contrast which 

 is borne out by the variation of almost every other character, but which 

 is more striking for this character than for the remainder. 



* The small lirandou sample, dated April IGtli, has been combined with that dated 

 April 22iul, in order to obtain a representative percentage for this locality. 



