Anim. Behav., 1989, 38, 1067-1078 The good parent process of sexual selection GUY A. HOELZER* Department of Ecology and Evolutionary Biology, University of Arizona, Tucson, AZ 85721, U.S.A. Abstract. Two general processes have been proposed to explain the evolution of epigamic traits (i.e. secondary sexual characteristics that evolve in response to mate choice). The good genes process maintains that epigamic traits may be used to assess the genetic quality of potential mates. The Fisherian process maintains that these traits can evolve without any effect beyond conferring sexual attractiveness to the bearer. A third process.leading to the evolution of epigamic traits is presented here: the good parent process. Epigamic traits arise through this process by clarifying the differences in non-heritable parental quality among potential mates. The good parent process is validated by a haploid population genetic model, which further suggests that increases in the frequency of good fathers in the population and phenotypic plasticity of the trait enhance the evolution of a good parent trait. Species in which this process should be active can be identified, and data from natural populations can falsify or support the hypothesis that the good parent process dominated in the evolution of a particular epigamic trait. The three processes of epigamic trait evolution are not mutually exclusive. They may all occur in nature with different frequencies and intensities. It has long been recognized that traits associated with courtship can evolve as a result of mate choice (Darwin 1871). Males generally display these epi- gamic characteristics, whereas females are usually unadorned and choosy (although, in a broad sense, theories of epigamic trait evolution could apply equally well to males or females). The traits may be conspicuous morphological or chemical features, brilliant colour patterns, behavioural displays, or some combination of the above. In many cases it is evident that natural selection on viability or fer- tility (where the term fertility, as used in this paper, excludes any effects of attractiveness) could not have produced these traits. Thus the traits confer a mating advantage at the expense of viability or fertility. The numerous models that have been proposed to explain the evolution of epigamic traits fall under two general categories (see Borgia 1987). The Fisherian process (Fisher 1930) is still being modelled by contemporary population geneticists (e.g. O'Donald 1980; Lande 1981; Arnold 1983). This model allows for an initial variance in the heritable quality of potential mates with respect to viability or fertility, but the advantages of female mate choice are quickly reduced to the production of sexy sons and choosy daughters. Fisherian exaggeration of an epigamic trait may ultimately * Present address: Genetics Laboratory, Department of Anthropology, Columbia University, New York, NY 10027, U.S.A. 0003-3472/89/121067+ 12 $03.00/0 reduce the adaptive fit of the population to its environment. The second general process, the good genes process, grew out of Zahavi's handicap principle (1975) and has been elaborated by others as well (e.g. Hamilton & Zuk 1982; Kodric-Brown & Brown 1984; Nur & Hasson 1984; Andersson 1986; Pomiankowski 1987, 1988). These models suggest that, when there is additive genetic variance for fitness among males, a female can improve the viability or fertility of her sons and daughters by choosing a mate with 'good genes'. Epigamic traits evolve because they are correlated with the genetic quality of a potential mate. This mechanism in- creases the level of adaptation in the population. A third mechanism for the evolution of epigamic characters is proposed here: the good parent process. In this process epigamic traits evolve to advertise the non-heritable component of parental quality. That is, when paternal care influences the viability or fertility of offspring, males can be selected to communicate honestly to females the quality or quantity of care they are likely to provide her progeny (see Thornhill 1976). Thus, the pheno- types of her offspring will be directly improved. Motro (1982) modelled the situation where an epigamic trait evolves because it reveals a heritable component of male parental quality. This repre- sents a special case of the good genes process. Recent reviews of sexual selection theory make no mention of any models describing the evolution of epigamic traits that indicate non-heritable quality 9 1989 The Association for the Study of Animal Behaviour 1067