Evaluation of pollen transfer in wild plant communities revealing heterospecific pollen receipt is common, yet experimental hand pollinations have revealed high among-species variation in the magnitude of its effect on recipient fitness. The causes of this among-species variation are unknown, however, prompting the investigation of underlying factors. Here, we conducted a hand-pollination experiment with ten co-flowering species to determine whether the effects of heterospecific pollen receipt are mediated by the pollen donor or recipient species alone, or whether the effects are determined by the interaction between them. We further assessed species traits potentially mediating interactive effects in heterospecific pollen receipt by evaluating the relationship between heterospecific pollen effect size and three different predictors reflecting a unique combination of pollen donor and recipient characteristics. Our results show, for the first time, that the magnitude of the heterospecific pollen receipt effect is determined by the specific combination of donor and recipient species (i.e., interactive effects). However, we were unable to uncover the specific combination of traits mediating these effects. Overall, our study provides strong evidence that an understanding of heterospecific pollen receipt effects based on recipient or donor characteristics alone may be insufficient. This study is an important step toward an understanding of consequences of heterospecific pollen receipt in co-flowering communities.

The number of co-flowering species, floral density and floral trait diversity can be major determinants of pollinator-mediated plant–plant interactions in a community. However, evaluation of how each one of these co-flowering components affects the pollination success of a single focal plant species, and how these effects vary at different spatial scales, is lacking. Here, we evaluated the effects of functional diversity (flower morphology and colour), taxonomic diversity (reflecting potential sampling effects) and flower density (conspecific and heterospecific), on the pollinator environment (i.e. visitation rate and pollinator diversity) and pollination success (i.e. pollen load size and number of pollen tubes per style) of Cakile edentula (Brassicaceae). We applied structural equation models (SEMs) at the floral-neighbourhood (plot level) and community-wide scales to uncover the factors that mediate co-flowering community effects on C. edentula pollination success. We found that co-flowering community effects at the community level are more important than fine-scale floral-neighbourhood differences in mediating plant pollination success in our study species. Increasing plant functional diversity decreased pollinator visitation rate but increased the diversity of pollinator functional groups visiting C. edentula flowers. Taxonomic diversity negatively affected pollinator diversity suggesting that other unmeasured floral traits may be relevant or that single-species effects (sampling effects) may be important. Overall, our results suggest that functional floral trait diversity in a community may be the most important factor influencing pollination success of species in a community. We also found evidence for intra- and interspecific pollinator competition mediated by flower density, but none of these effects seemed to have a significant impact on pollination success. This study is an important step towards understanding the complexity of co-flowering community effects on the pollination success of individual plant species at multiple spatial scales. This study further reveals the potential importance of plant functional diversity in a community in helping predict competitive and facilitative interactions in co-flowering communities. Synthesis. Floral density and taxonomic and functional co-flowering diversity are important drivers of pollination success in flowering plants. The effects of the co-flowering diversity on the pollination success of plant species can largely depend on the spatial scale being studied. Only evaluating the outcomes of pollinator-mediated plant–plant interactions at multiple stages of the pollination process can lead to a complete understanding of their ecological consequences in nature.

Species' floral traits and flowering times are known to be the major drivers of pollinator-mediated plant–plant interactions in diverse co-flowering communities. However, their simultaneous role in mediating plant community assembly and plant–pollinator interactions is still poorly understood. Since not all species flower at the same time, inference of facilitative and competitive interactions based on floral trait distribution patterns should account for fine phenological structure (intensity of flowering overlap) within co-flowering communities. Such an approach may also help reveal the simultaneous action of competitive and facilitative interactions in structuring co-flowering communities. Here we used modularity within a co-flowering network context, as a novel approach to detect convergent and/or over-dispersed patterns in floral trait distribution and pollinator sharing. Specifically, we evaluate differences in floral trait and pollinator distribution patterns within (high temporal flowering overlap) and among co-flowering modules (low temporal flowering overlap). We further evaluate the consistency of observed floral trait and pollinator sharing distribution patterns across space (three geographical regions) and time (dry and rainy seasons). We found that floral trait similarity was significantly higher in plant species within co-flowering modules than in species among them. This suggests pollinator facilitation may lead to floral trait convergence, but only within co-flowering modules. However, our results also revealed seasonal and spatial shifts in the underlying interactions (facilitation or competition) driving co-flowering assembly, suggesting that the prevalent dominant interactions are not static. Synthesis. Overall, we provide strong evidence showing that the use of flowering time and floral trait distribution alone may be insufficient to fully uncover the role of pollinator-mediated interactions in community assembly. Integrating this information along with patterns of pollinator sharing will greatly help reveal the simultaneous action of facilitative and competitive pollinator-mediated interactions in co-flowering communities. The spatial and temporal variation in flowering and trait distribution patterns observed further emphasize the importance of adopting a more dynamic view of community assembly processes.