That dominance can be evolved has been demonstrated on several occasions by selection experiments. Ford obtained specimens of the Currant Moth Abraxas grossulariata (L.) carrying a gene which, when homozygous, changes the white ground colour of the wings to yellow. The heterozygote is intermediate in appearance. He divided his stock into two and, in one, selected from generation to generation for the palest heterozygotes and, in the other, for the yellowest ones. In a few generations the effect of the gene had become almost completely recessive in the pale line, and almost completely dominant in the other line, thereby demonstrating not only that dominance could be evolved, but that this could occur in either direction. The experiment also shows that modifiers are available for producing dominance, as have other similar experiments. Stocks used in the laboratory are usually derived from a few individuals, therefore it is evident that in the much larger populations found in the wild, many more suitable modifiers must be available. Following this work (in order to get a comparison from other material an the number of generations required for the selection) Fisher himself with S. B. Holt got an exactly comparable result (in approximately the same number of generations) working with a 'short-tail' gene in mice.
Although it has been shown that dominance can be evolved, this does not prove that it is normally evolved and that disadvantageous mutants usually become recessive. That they usually do so is, however, indicated by the fact that the common allelomorph found in natural populations is nearly always dominant for its major effects. That is to say, in Haldane's terminology, it has a safety factor of 2. This view is further strengthened by Ford's point that it is often only the more obvious characters controlled by a gene which are, in fact, dominant.
In theory, it is possible to test the hypothesis by putting mutants from one species into the gene-complex of another in which they have not occurred, and seeing if dominance vanishes. There are, however, several difficulties. To begin with, it is almost impossible to be certain that the mutant does not occur, or has not in the past in the test stock, for when two forms can be crossed and produce fertile hybrids they are likely to have arisen from a common ancestral stock in the not too distant past, and the mutant may well have occurred in this stock. Where the method has been used in cotton and apparently given evidence in favour of the evolution of dominance, the data are open to quite a different interpretation.
Some similar but perhaps more convincing evidence is provided by a study of the Swallowtail butterfly Papilio memnon L. On the island of Palawan males and females have a long projection (tail) on each hind wing. In other populations the allelomorph which controls the presence of these tails is not known, the individuals being homozygous for the alternative allelomorph. It has been shown, by making appropriate crosses, that the absence of tails is recessive in the Palawan gene-complex. However, the presence of tails is recessive in other populations. That is to say, the dominance is reversed and the allelomorph which is rare or absent, whichever it may be in a particular population, is recessive in effect, as would be expected an Fisher's theory of the evolution of dominance.
The final answer to the problem of how the recessiveness of deleterious mutants usually arises is only likely to be reached when we know more about how genes exert their effects. The data now available show that many genes are not active for much of the time but only when their products are required, and are switched off when there is an excess of them. The mechanism regulating the activity of the genes is clearly adaptive and evolved as the result of natural selection. It might be argued that a by-product of the evolution of such mechanisms would be the evolution of dominance if the rate of action of the gene is not critical. However, as pointed out previously this still does not explain why the homozygote is capable of manufacturing substance at twice the rate of production normally required. What is known of gene action at the present time does not help us explain this margin of safety.
The importance of the controversy over the method by which dominance is obtained is a direct result of the hypothesis that drift is important in evolution. For, if the recessiveness of deleterious mutants is evolved by the accumulation of genes modifying its effect, as Fisher maintained, very small selective values must usually be effective in controlling evolution, and the great importance of drift in evolution would be thereby disproved.
CLARKE, C. A., and SHEPPARD, P. M. (1960). 'The evolution of dominance under disruptive selection', Heredity, 14:73-87.