THE ORIGIN OF VARIATION 25 



heterozygotes are included. Thus, even if thousands of 

 millions of individuals are produced in any one generation 

 and no two individuals are genetically alike, only a small 

 fraction of the possible combinations would actually be 

 realised. The possibilities of recombination are much en- 

 hanced by the variation in the expression of genes when 

 combined with different gene-complexes. 



Outside domesticated forms it is not very easy to find good 

 examples of the effects of recombination. Permutations of 

 specific characters within a genus are, of course, very familiar, 

 but owing to the occurrence of sterility, etc., are rarely capable 

 of genetical investigation. Amongst domestic animals recom- 

 bination, leading to novel forms, was early recognised in 

 poultry, rabbits and pigeons. In a wild insect we may 

 mention the cases of Papilio polytes investigated by Fryer (19 13) 

 and of Aricia medon studied by Harrison and Carter (1924). 

 In the latter species two forms meet on the Durham coast and 

 a wide range of variants, many not known elsewhere, is 

 produced. More usually the meeting of two geographical 

 forms leads merely to the production of simple intermediates 

 (see p. 89). It is probable that recombinations of numerous 

 small gene-differences in wild populations are responsible for 

 a considerable part of the continuous range of variation in 

 size, colour, etc., often alleged to be fluctuational. 



It is very difficult to assess the actual evolutionary value 

 of the variation arising in this way. Some authors, such as 

 Lotsy, have supposed recombination, especially after crosses 

 between very different varieties or species, would supply all the 

 variation required for evolution. This theory is more plausible 

 in the case of plants — in which interspecific sterility is not so 

 much developed — than in that of animals. The problem is not 

 one of very easy direct approach, for genetical experiment on 

 a sufficient scale is lacking, but some indirect evidence may 

 be obtained. If the individuals of a species often differed from 

 one another in a large number of genes we should expect that 

 crosses of such individuals would give rise to a wide range 

 of variation, including some forms perhaps quite distinct from 

 either parent. Continued inbreeding of such a stock would give 

 rise to a large number of distinct lines. Apart from domesti- 

 cated forms, which are in quite a different category (cf. p. 188, 

 Chapter VII), it is not easy to find good examples. In some of 



