EVOLUTIONARY MECHANISMS 297 



effective against even quite small selective advantages. It is quite 

 impossible for evolution to be directed by mutation unless the char- 

 acters in question have very little adaptive significance indeed, or 

 unless we postulate mutation rates of a totally different order from 

 anything we know at present. 



Probably the main connection in which mutation pressure has any 

 influence on the course of evolution is in causing the reduction of 

 organs which are no longer of any adaptive significance; in total absence 

 of selection, random mutations can accumulate. Now it is usual to find 

 that the highest mutation rate at any locus is to a hypomorphic allelo- 

 morph determining a reduction from the normal size. Thus in time the 

 population is filled with a hypomorph of low efficiency and the organ 

 is reduced or even disappears. A similar argument applies to parts of 

 the Y chromosome in which crossing-over is suppressed. Genes in this 

 region being always masked by their wild allelomorphs in the X will 

 tend to become "inert" by the pressure of mutation towards hypo- 

 morph and eventually amorph allelomorphs. 



7. Selection of the Genotypic Milieu 



There is considerable evidence that the genotypic milieu of any 

 species is adjusted to the particular genes which occur; this pheno- 

 menon has been discussed in relation to dosage compensation and 

 dominance (Chap. 7). It was in connection with the latter problem that 

 Fisher^ first drew attention to the evolutionary aspects of the matter. 

 He argued that there was no reason, on a priori grounds, why new 

 mutations should not be partially dominant, with some expression of 

 the mutant character in the heterozygote. It is therefore necessary to 

 find some explanation of the fact that most rare genes are recessive to 

 the wild type. Now these genes will usually be present as heterozygotes, 

 and the faa that they have remained rare indicates that they are dele- 

 terious. The main influence of selection on a partially dominant gene 

 will therefore be on the heterozygote, and these will be the more 

 stringendy selected against the more markedly they show the mutant 

 charaaer. If there are any modifiers which reduce the expression of the 

 mutant gene in the heterozygote, these modifiers will confer some 

 protection on heterozygotes containing them and will therefore have a 

 positive selective value. Fisher supposes that this is sufficient to cause 

 an accumulation of such modifiers, which will eventually suppress the 

 mutant charaaer entirely in heterozygotes, so that the mutant gene will 

 ^ Fisher 1928, 193 1, Ford 193 1. 



