Experimental species removal reveals species contributions to positive pollinator-mediated reproductive interactions

Pollinator-mediated reproductive interactions among co-flowering plant species provide a canonical example of how biotic factors may contribute to species coexistence, yet we lack understanding of the exact mechanisms. Flowering-dominant and unusually attractive "magnet species" with disproportionate contributions to pollination may play key roles in such reproductive interactions, but their relative roles within the same community have rarely been assessed. We experimentally removed either a flowering-dominant or a highly attractive magnet species and compared effects on visitation frequency, pollinator richness, and seed set of co-flowering plants. Removal of either the flowering-dominant species or the magnet species reduced community-level pollinator visitation. Removal of the magnet species had the most consistent effect, including reduced pollinator visitation and richness, and reduced seed set of most co-flowering plants. These results suggest that the magnet species, which interacts with a wider range of pollinator species than does the dominant species, promotes the visitation and reproductive fitness of most other species. Removal of the flowering-dominant species affected only certain species, perhaps because these plants had floral traits similar to the dominant species. Our results highlight the role of attractive magnet species within a community in structuring reproductive interactions and identify potential mechanisms involved in coexistence facilitated by reproductive interactions.