298 An introduction to modern genetics 



behave as a recessive. He has named the phenomenon the evolution of 

 dominance. 



The mathematical basis of the theory has met with criticism from 

 Wright,^ who maintains that the process is essentially a second order 

 one, and could not proceed fast enough to have any practical conse- 

 quences, and points out, further, that the modifiers would probably 

 have first order effeas of their own on fitness, which would quite out- 

 weigh the slight advantages they would confer in the rare cases when 

 they were in heterozygotes of the original mutant. Haldane^ made 

 rather similar criticisms of Fisher's original theory and put forward the 

 suggestion that natural selection of heterozygous harmful genes picks 

 out, not a set of modifiers which suppresses the expression of the gene, 

 but a wild-type allelomorph which has a considerable margin of safety 

 and thus has the same effect. We may suppose that the wild-type 

 expression is often the upper limit of possible gene effect; if the original 

 wild-type gene AA reaches this limit, but the heterozygote Aa does 

 not, selection will pick out another allelomorph ^', in which A' A' 

 again reaches the maximum and so does A' a (p. 185). 



Even if these criticisms are accepted, there is considerable evidence 

 that the genotypic milieu is subject to some sort of evolutionary control. 

 Harland^ has described a mutant Crinkled Dwarf in cotton. This is 

 completely recessive in sea island cotton (Gossypium barbadense) in 

 which it was first found, but gave intermediate heterozygotes in the 

 American species G. hirsutum where the mutation was not known to 

 occur naturally. This looked Uke a clear case in which the modifiers 

 conferring dominance on the wild type Had not been accumulated in 

 hirsutum where the gene did not occur and selection therefore had had 

 no chance to act. But further work showed that the situation is more 

 complicated. Harland discovered three normal allelomorphs of Crinkled 

 and three different genotypic miUeus. One of the normal allelomorphs 

 in G. hirsutum shows complete dominance in its own milieu, although 

 Crinkled does not occur. On the other hand, the three normal allelo- 

 morphs have different dominance relationships when got into the same 

 milieu, so there is some evidence for the type of phenomenon postu- 

 lated by Haldane.4 (Cf. p. 167). 



Fisher^ has made a special investigation to test his hypothesis, based 



1 Wright 1929. 2 Haldane 1930. ^ Harland 1936. 



* For a gene whose normal effect, the production of black pigment, is exag- 

 gerated into the production of melanotic tumours when it is in the genotypic 

 milieu of a foreign species, cf. Kosswig 1929. ^ Fisher 1935. 



