226 THE VARIATION OF ANIMALS IN NATURE 



We have hitherto considered variation in terms of single 

 mutants. We will now turn to the question of recombinations 

 of the existing hereditary material. We believe that this must 

 be quite a secondary problem, since the very possibility of 

 recombination depends, in our opinion, mainly on the prior 

 spread of single mutants through large sections of the popula- 

 tion. But, though in this sense the problem is secondary, it 

 demands a brief consideration. 



The complex genetic basis of a combination puts it at a 

 disadvantage with changes in a single gene as regards rate 

 of establishment. This disadvantage might be compensated 

 for by a substantial measure of isolation. In some crosses 

 between plants where the parents are rather unlike, the hybrid 

 may be itself a new type which breeds true and cannot effec- 

 tively cross with the parents (polyploids) : but in animals 

 such a process is almost unknown. 



Fisher (1930, p. 96) points out that, while it is clear that 

 without mutation evolutionary change must come to a stand- 

 still, ' it has not often been realised how very far existing 

 species are from such a state of stagnation or how easily, with 

 no more than 100 factors, a species may be modified to a 

 condition considerably outside the range of its previous 

 variation.' We have already alluded to this subject (p. 192) 

 in discussing the experimental production of new races by 

 selection, and we saw that in practice, though entirely novel 

 forms may be produced, selection may come to an end very 

 soon. We hardly think Fisher is right in speaking of residual 

 heredity with such confidence as a source of evolutionary change. 

 Moreover, it seems hardly correct to picture a typical character 

 as determined by as many as a hundred factors, each subject 

 to selection. Such a rich source of variation as Fisher indicates 

 no doubt exists if all the segregating characters of a species 

 are reckoned together : but, if the character subject to selection 

 is mainly dependent (as is more likely) on a few factors, the 

 amount of residual variability will be low and Natural Selection 

 would not be capable of carrying out protracted improvement. 

 Fisher is right in saying that there are millions of different 

 ways in which a species may be modified : but this does not 

 mean that all these are available for a single selective step or 

 for continued development in any one direction. 



We do not deny that in the last resort gene-mutations 



