Does competition for pollinators contribute to structuring Erica communities?

Abstract Question Pollinators are often implicated in the origin of plant species, but their role during the assembly of diverse plant communities has only rarely been investigated. Intra‐specific competition for pollinators can theoretically facilitate plant species co‐occurrence by limiting the density of any one species, while inter‐specific competition can potentially structure communities through the exclusion of inferior competitors. We analyse plant community composition and use field experiments to test whether intra‐ and inter‐specific competition for pollinators structures communities of co‐occurring Erica species. Location Cape Floral Region, southwestern South Africa (ca. 34° S, 18° E). Methods and Results In the study area, the genus Erica has diversified into ca. 680 species that differ primarily in their floral morphologies, reflecting insect, bird and wind pollination syndromes. Vegetation plot data from nine regions throughout the Fynbos Biome were used to test whether the frequency of co‐occurrence of species with the same pollination syndrome differs from a null model of random community assembly. We found that Erica communities were assembled randomly with respect to pollination in six regions, but that pollination syndrome co‐occurrence was significantly lower than expected in three regions. This over‐dispersion of syndromes across communities is consistent with communities being structured by inter‐specific competition for pollinators. To better understand the processes generating these patterns, we conducted competition experiments with six Erica species that share the Orange‐breasted sunbird ( Anthobaphes violacea ) as pollinator. We transplanted inflorescences of a focal Erica species into communities containing another resident Orange‐breasted sunbird bird‐pollinated Erica species and recorded pollination and nectar robbing rates in both the resident and the transplanted species in relation to the floral density of the resident species. Consistent with the observed community structure, pollination rate in both the resident and the introduced species declined with increasing density of the resident species, but the effect was not consistent among all populations. Nectar robbing rates of the resident species were negatively related to its own density. Conclusion Our analyses concur that both intra‐ and inter‐specific competition for pollination can contribute to structuring plant communities, but it is not the dominant factor structuring Erica communities.