The selective coefficients that keep modifying genes in a population.

A gene’s effects on fitness can be determined only if it produces a distinct phenotype of a known polymorphism. It is then possible to measure the phenotype’s chances of survival against other phenotypes, as, for example, did CLARKE and SHEPPARD (1966) for the melanic and non-melanic forms of the moth Biston betularia. But if a character can take any value in a continuous range, there is often complete ignorance of how many genes are causing the variation, of how they are organized into chromosomes, of what pleiotropic effects they may have on other characters and of what their fitnesses may be. The character may be some easily measured part of an organism, like the length of a particular organ. Or it may be the degree of expression of a particular genotype in a polymorphism. For example, the degree of dominance of two alleles is determined by genetic variations in the expression of their heterozygote. The genes that produce quantitative variation of this sort may be called polygenes or modifying genes or modifiers. A particular allele at a particular locus modifies the phenotype by a particular amount on the average. In general the modifiers may interact with each other so that the effect of one modifier may vary according to the other modifiers in the genotype. To understand how selection may change a character, it is necessary to know how the modifiers are held in the population. Do the modifiers of one character also modify other characters? If so, do their effects on the several characters cause variations in fitness? And how large are the selective coefficients? KIMURA (1968) has shown that one nucleotide must have been substituted for another in the genotype every two years on average during the evolution of the mammals. This rate of replacement implies that many of the mutations can have no effect on fitness; for if they did, the genetic load would be intolerable. This means either that many mutations are simply without any phenotypic effect and make no contribution to the variance of any character, or that a lot of the variation has no effect on fitness. Suppose the modifiers do have a considerable pleiotropic effect on fitness. The simplest way of maintaining the modifiers in the population is by a superiority of the heterozygotes over the homozygotes. Frequency-dependent selection could also maintain them. YARBOROUGH and KOJIMA (1967) have shown that frequency-dependent selection maintains the polymorphism at the esterase-6 locus in Drosophila melanogaster. If selection is then applied to the modifiers in one