No. 613] EXPERIME? 



43 



reduced to the genetic factors which Lang worked out for 

 Helix. But, exactly as in the classic cases, there was no 

 possibility of stating a definite relation of these factors 

 to the grouping according to localities. In some locali- 

 ties certain factors or combinations did not occur, but the 

 attempt to classify the material along this line proved a 

 failure. However, every group from each locality ex- 

 hibited beside these factorial recombinations certain 

 quantitative characteristics of size, proportions, etc., of 

 the shell which were characteristic for definite localities. 

 These, however, are the characters which probably fall in 

 line with those caused by the quantity of the genes. 



The difficulties which the facts of geographic variation 

 create for the conception of species-formation by selec- 

 tion have often been discussed. Bateson in particular 

 (' 'Problems of Genetics") scrutinizes them from the 

 modern genetic point of view. They are indeed insu- 

 perable if all characters which show variations and recom- 

 binations are considered from this point of view. The 

 extreme irregularity, for example, of the local combina- 

 tions of types of shells in Helix, Partula and Achatinella 

 makes it impossible to regard them as local adaptations. 

 This is certainly true, but may be without any bearing on 

 the species question at all. The factors and recombina- 

 tions occurring in Helix, Achatinella and Prodromus are 

 more or less the same, just as are the recombinations of 

 coat colors in different rodents. They constitute a set of 

 mutations and their recombinations which are proper to 

 the type of germ-plasm of the group. They occur, re- 

 combine or fail to appear as chance wills, and seem to have 

 no special selective value. We do not think that these 

 are the characters which play a part in the evolution of 

 species; they are, in most cases, independent of adapta- 

 tion. 



There are, however, reasons for supposing that such 

 differences of characters as are based on the quantitative 

 differences of the gene are those which are influenced by 

 selection and are important for the formation of the first 



