“…Initially, the optimal hypothesis was based on group selection (Berg, 1941(Berg, , 1942(Berg, , 1948 and related genetic load (Kimura, 1960(Kimura, , 1967Levins, 1967) arguments, but later the analysis based on individual selection, which is more appropriate for amphimictic populations, was developed (Leigh, 1970(Leigh, , 1973. Under invariant direct selection positive mutation rates can be favoured only when an equilibrium population is polymorphic and either the population size is finite (Gillespie, 1981a), mating is non-random (Holsinger & Feldman, 1983), or selection acts on fecundity (Holsinger et al 1986;Twomey & Feldman, 1990). Changing selection generally favours positive mutation rates (Gillespie, 19816;Semenov & Terkel, 1985;Ishii et al 1989), although zero mutation rates can also be established under some values of the parameters (Leigh, 1970(Leigh, , 1973Gillespie, 1981 b).…”