Abstract
Plant-pollinator interactions are essential to ecosystem functioning, yet the mechanisms that determine why some plant species become highly connected within interaction networks remain insufficiently understood, particularly in tropical coastal systems. Here, we examine how multiple plant traits predict the interactive role of species within a bee-plant network in a coastal ecosystem in the Gulf of Mexico. Using an existing dataset comprising 35 plant species and 47 bee species, we quantified each plant's interactive role through species degree, betweenness, and closeness centrality. We then evaluated how six traits (i.e., flower number, flower size, flower color, number of stamens, plant height, and life form) influence these network positions. Our results show that four traits significantly predicted plant interactive roles. Plants surrounded by more open flowers and those with larger flowers interacted with a greater diversity of bee species, indicating that resource detectability and accessibility strongly shape visitation patterns. Herbaceous species also exhibited higher interactive roles than woody plants, likely due to their rapid growth, abundant and synchronous flowering, and predictable resource availability in dynamic coastal environments. Additionally, yellow-flowered species received disproportionately more visits and achieved higher interactive roles, consistent with known sensory biases of bees toward yellow wavelengths. In contrast, plant height and stamen number showed no detectable influence on network position. Overall, our findings demonstrate that a combination of vegetative and floral traits (particularly those signaling abundant, accessible, and visually detectable resources) drives the emergence of key plant species within bee-plant networks. Integrating plant traits with network metrics provides a powerful framework for identifying species that sustain pollinator diversity and for predicting community responses to environmental change.