25-4 THE VARIATION, RACES AND MIGRATIONS OF THE MACKEREL. 



IGth, (45 fish). Two of these samples, it will be noticed, are small: in 

 one the modal number is 26, in the other 28. 



The first impression made upon a survey of the figures in the table is 

 undoubtedly the slight nature of the differences in the range and 

 frequencies of variation in the different samples. But the next 

 impression is the remarkable constancy with which mean-values of less 

 than 27 are associated with samples from localities in the North Sea 

 and English Channel, while the means for the Irish and American 

 samples exceed that figure with only one exception. This indicates, as 

 can easily be seen by comparison, that in North Sea and Channel 

 samples the values below 27 always occur more frequently than those 

 above 27 ; while in the Irish and American samples, with only one 

 exception, the reverse is the case. 



For North Sea and Channel samples, the means vary from 26"50 to 

 26'94, and, for Irish samples, from 26'90 to 27"56. The American mean 

 is 27-38. 



If we exclude all samples of less than 90 fish, the means, taken in 

 the same order, vary from 26-62 to 26-91, and from 26-90 to 27-32. 



Analysing this contrast still further, by means of the table of 

 percentages for the local groups (neglecting groups of less than 200 

 fish), we see that the frequency of the value 26 varies from 30 7o to 

 31 7o in the case of Lowestoft and Plymouth, but from 21 7o to 23% in 

 the case of Kerry and Kinsale. Similarly the frequency of the value 

 28 varies from 14% to 15% for Plymouth and Lowestoft, but from 

 23% to 25% for Kinsale and Kerry. The frequency of the modal 

 value 27 is constant at 38 7o for the two Irish groups, but varies from 

 38 7„ to 42 7o for Lowestoft and Plymouth. 



The conclusion is irresistible that the Lowestoft and Plymouth fish 

 resemble one another very closely, and that the Kerry and Kinsale 

 fish do so likewise, but that there is a comparatively serious difference 

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



Hitherto, however, we have not taken into consideration the actual 

 frequencies of the extreme values {i.e., those below 26 and above 28). 

 These values occur very rarely, and their separate frequencies cannot 

 consequently be compared with exactitude in samples of ordinary 

 magnitude. To bring their frequencies into consideration, it will be 

 necessary to merge all the extreme values into two groups, one of low, 

 and the other of high value, and to compare the frequencies of the 

 combined values. This has been done in the following condensed 

 table, in which, also, the Eamsgate data have been combined with the 

 Lowestoft records to form a single group representative of the North 

 Sea, and Brest has been combined with Scilly. No one, I imagine, will 

 be prepared, on the evidences provided in this report, or on any other 



