THE DISTRIBUTION OF VARIANTS IN NATURE 125 



categories, of a similar structural grade, have been termed 

 ' subspecies ' by some entomologists. These subspecies, unlike 

 geographical races, live side by side ; but they can be called 

 species only if we give up all attempts to indicate (in any 

 one group) the same degree of divergence by the latter term. 

 It is probable that these subspecies occur in some groups more 

 than in others owing to differences in the mode of reproduction, 

 particularly in the length of the breeding season, in the way 

 in which the sexes find one another and in the degree of 

 development of gregarious habits. Subspecies tend to occur 

 in any group in which non-geographical methods of isolation 

 are easily effective. The great possibilities of such isolation 

 have often not been sufficiently realised and undue weight 

 has been given to geographical effects. 



The following are the more important general results 

 which emerge from our survey : 



1 . We have discussed at some length the antithesis between 

 individual and regional and geographical variation. In 

 some cases the antithesis stressed by Rensch and others between 

 populations broken up into clearly defined regional or geo- 

 graphical groups and those in which the variants are more 

 universally distributed is clear and can be shown in some 

 instances to be due to differences in habits, size, etc. We 

 believe, however, that the distinction is more apparent than 

 real and that no particular significance is to be attached to it. 

 To begin with, there seems to be a likelihood that geographical 

 variation will be found to be less clearly cut when the relevant 

 forms are more exhaustively studied and knowledge of their 

 distribution is based on more material. Series of geographical 

 races are easy to demonstrate when the samples are not too 

 large. Secondly, while we admit that clearly-cut qualitative 

 divergences on a geographical basis are not so typical of 

 groups such as terrestrial molluscs and arthropods, it is quite 

 evident that the proportions of the variant types in these 

 groups define populations quite as definitely as average 

 dimensions, colour, etc., define those of vertebrates. It is of 

 secondary importance that the regional divergences among, 

 e.g., populations of land molluscs tend to be smoothed over as 

 a result of the size and habits of these animals and in certain 

 (but by no means all) of their characters by reason of their 

 plasticity. When many characters of vertebrate populations 



