THE DISTRIBUTION OF VARIANTS IN NATURE 127 



instance of this. It seems evident that all animals are not 

 equally prone to receive the impress of their environment nor 

 in the same state of mutational activity. 



5. The general impression that one gets from a survey 

 such as the foregoing is that groups are formed by the spread 

 of individual variants rather than by mass transformation. 

 What we find is a gradation from single variants, or variants 

 represented only by a low percentage in the population, to 

 larger and more distinctive assemblages and eventually to 

 distinct regional geographical groups. We do not know, of 

 course, how many of the smaller groups may not be on the 

 way to extinction ; but we may assume that at least half of 

 them are not and that this possibility does not vitiate the 

 general conclusion that there is a process at work in nature 

 which facilitates the multiplication of single variants. If the 

 latter were spreading from single loci the mosaic of poly- 

 morphism is exactly what one would expect to find. Rensch's 

 attempt to show that variants are distributed in ' Rassenkreise ' 

 under the influence of differentiated environments seems to 

 us to break down on three counts : 



(a) The very general occurrence of polymorphism is 

 a proof that the environment is not the direct trans- 

 forming agency. The only way in which those who favour 

 that view could explain the occurrence of differentiated forms 

 living side by side in the same habitat is to suggest that they 

 acquired their differences elsewhere and have subsequently 

 met. But, as Robson (1928, p. 174) has pointed out, this 

 involves explaining (1) the frequent lack of epharmonic con- 

 vergence and (2) the means of spreading. 



(b) In numerous cases variants are not arranged with 

 reference to environmental gradients and many races range 

 unmodified through a variety of environments {cf. Sumner, 

 1932, etc.). 



(c) To argue that many of the observed changes that are 

 correlated with environmental differences may only be somatic 

 is but a negative objection ; but it is a great weakness of 

 Rensch's case that there is so little experimental evidence 

 that local races, etc., are of a fixed heredity. We do not wish 

 to ignore the many and striking cases of structural and en- 

 vironmental trends. We would even admit that in such cases 

 mass transformation of populations may be possible. But 



