A comparative analysis of pollen limitation in flowering plants

Abstract

We conducted a comparative analysis to determine lie history and ecological correlates of pollen limitation among 224 species of animal-pollinated flowering plants. To test predictions concerning the occurrence of pollen limitation, species were classified categorically for six liie-history traits (self-compatible vs. self-incompatible, autogamous vs. non-autogamous, specialized vs. unspecialized floral morphology, nectariferous vs. nectarless, monocarpic vs. polycarpic, herbaceous vs. woody) and two ecological conditions (presence in open vs. forested habitats and temperate vs. tropical biomes). Pollen limitation of species in contrasting categories was compared using non-parametric tests (TIPS analysis) and phylogeneticallyindependent contrasts (PICs). The results of TIPS and PICs analyses were generally congruent, although fewer significant contrasts were evident with PIGS, probably because of low statistical power. Overall the results suggest that the influence of phylogenetic history on the intensity of pollen limitation was not particularly strong. Nonetheless, significant variation in the degree of pollen limitation was demonstrated among seven angiosperm families suggesting some phylogenetic component to the phenomenon. With both TIPS and PICs, pollen limitation was less intense in self-compatible and autogamous species. TIPS analysis demonstrated that herbaceous, nectariferous, and temperate species were less likely to be pollen-limited, but using PICs this could only be corroborated for those that were self-incompatible. None of the traits were singularly unambiguous predictors of pollen limitation, possibly reflecting the stochastic nature of pollinator service.