NATURAL SELECTION 227 



constitute the basis of all new evolutionary steps. We are 

 inclined to counter the argument that, because they are found 

 in certain forms to be very rare, they must depend on Natural 

 Selection for their survival and spread, by suggesting that we 

 do not as yet know enough about the mutation-rate at large, 

 especially under natural conditions. But, however that may 

 be, we have still to discover what is the part played by factorial 

 recombination. We have mentioned above (p. 25) that this 

 is capable of producing novel forms {e.g. the numerous cases 

 of ' novelties produced [immediately] by recombination ' ; 

 Castle's production of the hooded pattern in rats). Further- 

 more, the species within a genus tend to comprise very many 

 that represent permutation and combination of a common 

 stock of characters, and may very well (though we do not 

 know of any specific instances) exhibit distinctive and peculiar 

 characters which arise from factorial recombination. There 

 are, we admit, limitations to the possibilities involved in 

 ' evolution by hybridisation,' but, given a reasonable amount 

 of isolation, it seems to us likely that a considerable part of 

 the early stages of evolutionary divergence may be of this 

 nature. 



The Evolution of Dominance. — Before closing this section we 

 propose to discuss very briefly Fisher's theory of the evolu- 

 tion of dominance. His case is put forward in his book (1930, 

 chapter iii) and in a review (1931). Ford (1930, 1931) has 

 also summarised the evidence. Wright (1929) and Haldane 

 (1932) have not accepted Fisher's hypothesis. 



Fisher realises that the genetic conception of ' wild type ' 

 is in need of some explanation. The wild type exists because 

 the majority of genes in animals in nature are dominant to 

 their allelomorphs which have been detected in the laboratory. 

 Fisher endeavours to explain the dominance characteristic 

 of the wild form as the result of selection of the gene-complex 

 in such a direction that any given mutant will produce the 

 minimum possible visible effect in the heterozygote. It is 

 assumed from the data on Drosophila that most mutants, 

 especially the easily visible ones, will be harmful, and therefore 

 it will be to the advantage of the species to suppress their 

 effects as far as possible, i.e. in the heterozygote. The argu- 

 ments in favour of the theory may be considered under three 

 headings. 



