190 



obviously be undertaken. Meanwhile we hope that the 

 records given in the frequency distributions may serve for 

 comparison with other sea-areas. 



The possibilities of a shifting with growth of any one of the 

 characters investigated should also be considered. That this 

 actually happens in the case at any rate of young herrings of 

 from 30 mm. to 60 mm. is very clearly shown in Table V. 

 This is a series of measurements of young herrings from North 

 Wales made by Mr. Andrew Scott and examined by Professor 

 Johnstone. The distance D is expressed as a percentage of 

 T (total length) and correlated with T as tabulated, giving a 

 coejSicient of correlation of — 0-9696. This is almost absolute 

 correlation. The form of the table shows at a glance how the 

 dorsal fin is gradually moving forward as the young herring 

 is growing. 



This investigation of the shifting of characters was 

 extended to the sprat. Measurements were made by Mr. Scott 

 and tabulated by Professor Johnstone. Tables VI and VII 

 show the results. Table VI is a correlation between the 

 position of the dorsal fin and the total length. The result 

 shows that this character does not vary in any regular way with 

 increasing length. Table VII is a correlation between the 

 position of the ventral fins and total length. This shows a 

 somewhat irregular tendency to move forward as the fish grows, 

 but not so definite as in the case of the dorsal fin of the young 

 herring. 



Unfortunately, in the case of the samples of larger herrings 

 exammed by us, time did not permit of more than four 

 correlations being made. Three of these were based on 

 character A^ and one on character D. The results were 

 inconclusive, as Tables VIII to XI show. There is little or no 

 evidence of correlation in any case investigated. 



