Despite much recent research on pollination, we have amassed relatively little hard information about how animal pollinators contribute to angiosperm speciation and species distinctiveness. We usually assume that pollinators make important contributions by specializing on plant species and providing differences upon which plants specialize. But evidence for generalization in many plant-pollinator interactions calls this assumption into question. Generalization makes strong 'floral isolation' unlikely during and after the speciation process, and is seen by some workers as problematical for a scenario of pollinator-mediated disruptive selection leading to species formation. The timeis ripe to initiate several lines of further research. It will be valuable to characterize selection on plants more carefully at various spatial scales to understand whether different suites of pollinators can affect consistent selection in different directions on floral characteristics. Similarly, it will be valuable to explore how reproductive isolation arises within and among populations, races, and higher taxa of plants. Even if they are not strong agents of floral isolation, pollinators may contribute to the evolution of reproductive isolation in other ways, which largely remain to be discovered. One possibility is that reproductive isolation evolves more or less independently of diversification in reproductive characteristics. If so, floral adaptation may not be tightly coupled with the formation of biologicalspecies in angiosperms, as is often supposed. [ABSTRACT FROM AUTHOR]