256 THE VAUIATION, RACES AND MIGRATIONS OF THE MACKEREL. 



existence of a peculiar American race of mackerel will be disputed. 

 But the differences between the Irish and the Channel fish are not 

 so sharply defined in all respects as those between the American and 

 European, so that the same degree of certainty cannot be attached 

 to my conclusion as to the existence of a separate Irish race. Never- 

 theless, I would point out that, if the racial distinctness of the 

 American fish be admitted, it is impossible to avoid a similar conclusion 

 in regard to the Irish fish also, if we confine our attention to the 

 evidence of the variation of the transverse bars, for the difference 

 between the Irish and Channel samples is greater than the differ- 

 ence between the Irish and American, and is almost as great as the 

 difference between the American and the European as a whole. 



IMoreover, the general difference between the Irish and English 

 fish is not confined to the gross samples from these localities as a 

 whole — in which case it might be attributed to errors of observation 

 or calculation — but it has been shown to be also characteristic of the 

 numerous individual samples of fish from these localities, almost 

 without exception. This fact, in my opinion, is one of very great 

 importance. 



It must also be stated that the differences between the samples 

 are not due to any increase in the number of bars with the growth 

 of the fish. This can be inferred merely by comparing the sizes of 

 the fish in the different samples. Thus the highest mean number of 

 bars occurs in the American sample, which consisted almost entirely 

 of relatively small fish, while the lowest mean number occurs in the 

 Lowestoft sample, dated June 28th, in which the distribution of sizes 

 was practically identical with that in the American. (See Table A.) 

 Again the values are almost equally high for the Kinsale samples 

 as for the Kerry samples, although the former consist exclusively of 

 small fish, and the latter exclusively of large fish. 



The following facts which I have ascertained point unequivocally 

 to the same conclusion. 



The transverse bars are formed at an early period, and the basis 

 of their formation is a deposit of pigment along the free surface of 

 the myotomes. It is this relation to the myotomes which gives the 

 bars in the majority of cases a marked > shaped curvature. Now 

 the number of myotomes corresponds to the number of vertebne, which 

 is almost invariable in most species of bony fishes, and particularly so 

 in the mackerel. From the relation of the bars to the myotomes it 

 may, accordingly, be assumed that no change in their number takes 

 place after formation. 



If the correspondence between transverse bars and myotomes were 

 perfect, tlie number of the bars would be the same as the number 



