180 



BIOMETRIC INVESTIGATIONS ON THE HERRING. 

 By W. Birtwistle and H. Mabel Lewis. 



When this work was originally planned the intention was 

 to accumulate three series of data, on (1) the Manx summer 

 spawning herrings, (2) the Welsh winter spawning herrings, 

 and (3) trawled herrings. It was only possible to deal with 

 rather small samples, and it was intended to add together all 

 samples belonging to each of these three groups and so make 

 three large distributions. But it was found that samples of 50 

 fish taken from the same place but at different times, appeared 

 to be more different from each other than was expected on 

 the assumption that they were samples of the same race. 

 When these small sub-samples were added together the 

 frequency distributions of a single character became rather 

 irregular, suggesting that the material was heterogeneous in 

 respect of the character studied. This was not always the 

 case, but it was so often enough to cast doubts on the propriety 

 of adding together small sub-samples taken from the same 

 locality but in different months (to say nothing of different 

 years) to form one big sample. 



Scale readings suggested also that the hei rings invading 

 Manx waters in (say) June differed from those in (say) August. 

 This, too, is what the local fishermen appear to think. The idea 

 suggested itself that we have not to deal with a single " Manx 

 summer race " of herrings, but really with a small number of 

 sub-races or " genotypes." In some weeks the shoal may 

 consist predominantly of fish of one genotype, but at other 

 times predominantly of fish of another genotype, or perhaps 

 of a mixture of several. It is possible that seasonal conditions 

 may favour now one genotype and at another tune another. 



Therefore it would appear that it might be necessary to 

 measure big samples taken at the same time. But even then. 



