Search

MANY species exhibit colour polymorphisms associated with alternative male reproductive strategies, including territorial males and 'sneaker males' that behave and look like females1–3. The prevalence of multiple morphs is a challenge to evolutionary theory because a single strategy should prevail unless morphs have exactly equal fitness4,5 or a fitness advantage when rare6,7. We report here the application of an evolutionary stable strategy model to a three-morph mating system in the side-blotched lizard. Using parameter estimates from field data, the model predicted oscillations in morph frequency, and the frequencies of the three male morphs were found to oscillate over a six-year period in the field. The fitnesses of each morph relative to other morphs were non-transitive in that each morph could invade another morph when rare, but was itself invadable by another morph when common. Concordance between frequency-dependent selection and the among-year changes in morph fitnesses suggest that male interactions drive a dynamic 'rock–paper–scissors' game7.

According to the Red Queen hypothesis--which states that interactions among species (such as hosts and parasites) lead to constant natural selection for adaptation and counter-adaptation--the disproportionate evolutionary success of parasites on common host genotypes leads to correlated selection for sexual reproduction and local adaptation by the parasite population. Here we determined whether local adaptation is due to disproportionate infection of common host genotypes, and, if so, whether infection of common host genotypes is due to commonness per se, or some other aspect of these genotypes. In a reciprocal cross-inoculation experiment parasites occupying the same geographical area (sympatric) infected locally common host genotypes significantly more often than rare host genotypes, whereas parasites occupying separate geographical areas (allopatric) showed no such significant difference. A mixed source of parasites (containing F1 hybrids) also showed no difference in infection between rare and common host genotypes. These results show that local adaptation results from parasite tracking of locally common host genotypes, and, as such, a necessary condition of the Red Queen hypothesis is met.

THE Red Queen hypothesis for the maintenance of biparental sexual reproduction suggests that, for species locked in revolutionary struggles with biological enemies, the production of variable progeny compensates for the genetic or ecological disadvantages of sex1–7. The advantage of sex and recombination under this hypothesis stems from the production of rare phenotypes, which are expected to be more likely to escape infection or predation by coevolved biological enemies. Like many evolutionary hypotheses, the Red Queen hypothesis is difficult to test directly, but its assumptions and predictions can be evaluated7–18. The most critical assumption is that biological enemies will disproportionately attack the most common phenotype19, 20. In this study of parasite loads of coexisting sexual and clonal fish, we find empirical support for this assumption.

Two classes of models attempt to explain why obligate parthenogenesis only rarely replaces sexual reproduction in natural populations, in spite of the apparent reproductive advantage that parthenogens gain by producing only female offspring1. The mutation-accumulation models suggest that sex is adaptive because it purges the genome of harmful recurrent mutations2,3. The ecological genetic models postulate that sex is adaptive in variable environments, particularly when the relevant variation is generated by revolutionary interactions with parasites4–7. Both of these models have considerable merit, but would seem to have limitations. The mutation-accumulation models require high rates of mutation3,8; the coevolutionary models require that parasites have severe fitness effects on their hosts9. In addition, parasites could select for clonal diversity and thereby erode any advantage that sex gains by producing variable progeny10. Here we consider the interaction between mutation accumulation and host–parasite coevolution. The results suggest that even moderate effects by parasites combined with reasonable rates of mutation could render sex evolutionary stable against repeated invasion by clones.

Nature 367, 554-557 (1994) DURING the production process, the contrast in shading in Figs 1 and 2 of this letter was lost; in addition, parts a and b of Fig. 3 were transposed. These figures should have appeared as they are shown here. The first sentence of each legend is given for guidance. FIG. 1 Results from computer simulations in which sexual populations were challenged by asexual lineages in the presence of coevolving parasites.

The widespread occurrence of sexual reproduction is an important problem in evolutionary theory, because of the cost of producing males in dioecious organisms and the allocation of limited resources in cosexes1–4. Here I compare the predictions of the major ecological hypotheses on the maintenance of sex by examining male frequency in Potamopyrgus antipodarum, a dioecious freshwater snail with both sexual and parthenogenetic populations5. The results do not-support the view that sex is maintained by a variable physical environment6–9, but they are consistent with the idea that sex is favoured by selection resulting from host–parasite interactions10–14.