THE VARIATION, RACES AND MIGRATIONS OF THE MACKEREL. 259 



the Irish and the Channel means is much less emphatic for the partial 

 series than for che entire series of bars, owing to the more extended 

 range of the mean values for both localities. 



The explanation is, I believe, the same as for the greater range of 

 variation of the partial series as compared with the entire series, 

 viz., the irregularity introduced by the variability of the lateral line, 

 which renders the number 100 insufficient as a unit-sample. If this 

 is so, a closer approach to the former result ought to be revealed by 

 comparing the means of the local groups of fish as a whole, neglect- 

 ing groups of less than 200 fish. These means in fact for the 

 Lowestoft and Plymouth groups vary from 18'21 to 1823, and for 

 the Kinsale and Kerry groups from 18'37 to 18-45, thus displaying a 

 close agreement between the Lowestoft and Plymouth fish, and also 

 between the Kinsale and Kerry fish, but a considerable difference 

 between the Irish fish and those from the North Sea or Channel. 



This relation is precisely the same as that revealed by the variation 

 of the entire series of bars, the means for the North Sea and Channel 

 being lower than those for the Irish groups in each case. 



It is interesting to notice also that the close approximation between 

 the Brest and Scilly means in regard to the entire series of bars is 

 again shown in regard to those which meet the lateral line ; but 

 whereas the combined mean for Brest and Scilly only slightly exceeded 

 that for the North Sea and Plymouth in the former case, here it is 

 actually higher than the Irish mean. The significance of this differ- 

 ence it is impossible to decide at present, as the total number of fish 

 from Brest and Scilly only amounts to 174, a number which cannot be 

 regarded as sufficient to neutralise the error due to the curvature of 

 the lateral line. If the mean, however, be provisionally accepted as 

 approximately correct, it points to the conclusion that in some respects 

 the mackerel which are found ofi' the mouth of the English Channel 

 and the neighbourhood of Ushant in June may form a connecting link 

 between the Irish fish and those of the North Sea and Channel. From 

 the geographical relations of the areas under discussion such a result 

 would certainly accord with a priori expectations. 



The percentages of frequency of high and low values of the bars 

 may best be understood from the following condensed table : — 



