Sex-dependent selection on flower size in a large-flowered Mediterranean species: an experimental approach with Cistus ladanifer

Larger flowers increase pollinator visit rates and reproductive success, so selection on flower size is usually mediated by pollinators. However, larger flowers involve costs imposed by resource limitation so environmental conditions may also modulate flower size. "Male function" hypothesis entails that the intensity of selection is sex-dependent, being greater through male fitness, whereas female fitness is more limited by resources. In this study we analyse pollinator-mediated phenotypic selection on flower size through both sexes in a large-flowered Mediterranean species, Cistus ladanifer. We experimentally manipulated flower size in two populations, measured its effect on male and female fitness and estimated the strength and direction of phenotypic selection through both sexes and populations. Unmanipulated control flowers received higher pollinator visit rates and dispersed a higher pollen amount than reduced flowers. This translated into selection towards larger flowers through male fitness in both populations. Nevertheless, flower size had little effect on female fitness. Fruit set was high but selection through this component of female function was not significant. Seed number increased in control flowers, especially in one population, where we detected positive selection on flower size. Our results suggest that pollinator-mediated phenotypic selection on flower size in this large-flowered Med-iterranean species is especially modulated by male fitness, but flower size adjustment may also be a result of a simultaneous selection through both sexes that, in turn, is dependent of ecological context.