124 TRANSACTIONS LIVERPOOL BIOLOGICAL SOCIETY. 



the case of ten herring (Table 141), from Bornholm, the 

 average formula being 2aII. For 25 herring from Korsdr 

 (Table 73), the so-called autumn herring formula also 

 holds good. If we, however, take the first ten of these Eor 

 comparison with the ten Bornholm fish, we find that the 

 formula no longer holds good. If Heincke had only 

 received 10 herring from Korsor instead of 25, then he 

 would have referred this " local race " to the spring 

 herring- with the average formula 2bII. (Indices D 2'25, 

 Y. 2057, A. 1*53). When one takes a certain number of 

 individuals from Heincke's tables for investigation, it 

 becomes possible to change autumn into spring herring, 

 and vice versa, at any rate occasionally. 



For instance take Table 23. Here we have 21 herring 

 which give the average formula 2all, that is the autumn 

 herring formula. If, however, the first ten only are 

 taken we get the average formula for spring herring 2bJ I. 

 (Indices D. 219, V. 205, A. 154). 



One is, therefore, from a consideration of Heincke's 

 own tables, forced to the conclusion that neither measure- 

 ments carried out on single specimens nor the average 

 formula obtained by measurements of an indefinite 

 number of individuals lead one to a certain and definite 

 separation of herring into the two groups of autumn and 

 spring herring. 



Without going so far as to actually say that no such 

 difference exists, it seems sufficiently obvious that more 

 measurements are necessary to establish Heincke's 

 arguments, and it is also obvious that the results of these 

 new measurements must be differently expressed. 



