Modifier genes modify and control processes that produce genetic and phenotypic variation, such as mutation, recombination and migration. They are usually assumed to be selectively neutral. Hence, their evolution is a consequence of their effect on the production of variation in other genes that do affect fitness. Modifier genes are therefore under ‘indirect’ or ‘second-order’ selection.

The reduction principle, first introduced in the early 1980s, is an important result in the mathematical theory of modifier genes. It states that, under assumptions commonly used in evolutionary theory, selection is expected to favour modifier alleles that reduce the rate of variation production — for example, reduce the rate of mutation or recombination.