Positive assortative mating is the non-random pairing of individuals that are more closely alike than the average in one or more phenotypic traits (Lincoln et al. 1998). For quantitative traits, assortative mating inflates the variance of the trait in a population by increasing the coupling of alleles with similar effects (Lynch & Walsh 1998, p 154). The traits most likely to be affected in animals are those used for mate choice or species recognition, including color pattern, scent, shape, or size. In plants, the affected traits are those used by pollinators to find and recognize flowers.
When intermediate phenotypes have low relative fitness ("disruptive selection"), assortative mating is selectively favored over random mating. If such selection acts directly on a mating trait, assortment with respect to that trait can evolve quickly. The frequency of intermediate phenotypes drops for two reasons: lower production of intermediate offspring due to assortative mating, and selection against those that do arise. Ultimately, indirect selection favors the evolution of reproductive isolation between extreme phenotypes (Kondrashov & Kondrashov 1999). That speciation theoretically can occur in sympatry by this mechanism is not terribly controversial, but the condition that the mating trait itself be the trait under direct disruptive selection is quite restrictive. There are a few clear examples of disruptive selection by pollinators on floral traits, such as in Polemonium (Galen et al. 1987) and Ipomopsis (Campbell et al. 1997), but not as many as might be expected if this were an important mechanism of divergence in plants (Wilson & Thomson 1996; Goulson & Jerrim 1997).
A more likely general scenario is that disruptive selection acts on some other ecologically important trait - for example, one involved in resource use or acquisition. Phenotypic divergence may reduce intraspecific competition, with extreme phenotypes favored over intermediate ones. When mating is random with respect to an ecological trait, disruptive selection on the trait increases the phenotypic variance but does not produce the evolutionary branching that would allow further divergence (Dieckmann & Doebeli 1999). The key to divergence in such a scenario is a genetic correlation between the trait under disruptive selection and a trait that promotes assortative mating. In an animal-based illustration of their sympatric speciation model, with disruptive selection on body size and mate choice based on color, Kondrashov & Kondrashov (1999) suggest thinking of the process as "the recruitment of colour for providing reproductive isolation between individuals of the opposite size."
When there is genetic variance in a mating trait, correlations between the mating trait and other characters occur readily via drift or inbreeding, especially for polygenic traits. The correlations usually are temporary, soon broken up by recombination. However, selection against intermediate ecological phenotypes strengthens the genetic correlation, as recombinants would be more likely to mate with the opposite ecological type due to their mating phenotype, ifmating is at all assortative with respect to the mating trait. In addition to selection against recombination, assortative mating itself reduces heterozygosity at the mating loci, further constraining the ability of recombination to break up the developing gene complexes. A positive-feedback loop can form, with disruptive selection, assortative mating, and genetic correlation strengthening each other, culminating in genetically isolated groups that are distinct for both mating and ecological characters, according to two different recent models (Dieckmann & Doebeli 1999; Kondrashov & Kondrashov
In cases where floral traits are genetically correlated with locally adapted ecological races ("ecotypes"), assortative pollinator foraging improves pollen transfer between mates adapted to similar ecological circumstances. When the post-pollination barriers to hybridization are complete, such "pollen targeting" increases pollination efficiency by reducing gamete wastage. When the barriers are partial, so that hybridization occurs but hybrid offspring have low relative fitness or are ill-suited to either environment of parental ecotypes, pollen targeting increases average offspring quality by reducing the frequency of hybrid offspring. Local mating within plant populations contributes to population structuring, with some divergence of genetic neighborhoods (Turner etal. 1982; see Waser, this volume). When local adaptation results, assortative mating among neighborhoods helps to maintain co-adapted gene complexes. Overall, selection for assortative mating is likely to be very common in plants, with the strength of selection depending on the relative fitness ofhybrid or intermediate offspring.
In general, we might expect those lineages that can attain assortative mating relatively easily to be more diverse and speciose than others. In angiosperms, animal pollination provides several additional mechanisms (beyond floral phenology and pollen-pistil interactions, for example) for assortative mating via pollen targeting.
Was this article helpful?