Frozen Evolution. Or, that’s not the way it is, Mr. Darwin. A Farewell to Selfish Gene.

The recessive gene hypothesis is another hypothesis explaining Haldane’s rule. It should be pointed out that the simple explanation offered by this hypothesis is not currently considered to be correct. In contrast to the dominance hypothesis, which assumes recessivity of genetic interactions, i.e. recessivity of phenotype manifestations of the joint products of at least pairs of genes, of which one is located, for example, on the sex chromosome and the other on the autosome, the recessive gene hypothesis directly considered the effect of the individual recessive genes without interactions. The X-chromosome is present in the hemizygous state in the cells of the heterogametic sex, i.e. is present in only a single copy, while it is present in the cells of the homogametic sex in two copies. The occurrence of any recessive mutation on the X-chromosome is thus necessarily manifested to a greater degree in the members of the heterogametic sex than in the members of the homogametic sex.
This model has two main inadequacies. The first problem consists in the fact that the effect of reduced fitness of the members of the heteogametic sex should be manifested in both hybrid and nonhybrid individuals. It is, of course, possible to argue that this effect is actually manifested in many species. For example in an ageing population, a gradual shift in the sex index in favor of members of the homogametic sex is often observed. It can be further objected that the negative effects need not be simply additive, but can grow at an faster rate (e.g. exponentially) with an increasing number of participating genes. Thus, the same number of recessive negative mutations can have a much greater effect in hybrids that have reduced fitness for a great many other reasons than in nonhybrid individuals. The existence of this nonlinearity, i.e. the snowball effect, is actually very probable. The individual genes can replace one another in their function. The probability that all mutually replaceable genes will be inactivated by mutations increases exponentially with the number of mutations in the genome (Orr & Turelli 2001).
A second, this time very substantial objection against the recessive gene hypothesis is that the number of recessive mutations on the X-chromosome should apparently be, on an average, much lower than in the genes on autosomes. While recessive negative mutations on autosomes can survive for a long time in the population, X-chromosomes with these mutations are constantly removed from the gene pool of the population by selection that acts on the members of the heterogametic sex. I am of the opinion that these objections were satisfied only by the hypothesis of somatic mutations (Gorshkov & Makar'eva 1999) (XXI.4.3.5), which can be considered to be a certain variation of the recessive gene hypothesis.