RESUMEN
KEY MESSAGE: A relationship between vertical acropetal inflorescences with protandrous flowers and bee pollination was hypothesized by Darwin back in 1877. Here we provide empirical evidence supporting this association across the angiosperms. Plant reproduction is not only determined by flower traits but also by the arrangement of flowers within inflorescences. Based on his observations of the orchid Spiranthes autumnalis, Darwin proposed in 1877 that bee-pollinated plants presenting protandrous flowers on vertical acropetal inflorescences, where proximal flowers open first, can exploit the stereotypical foraging behavior of their pollinators (i.e., upward movement through the inflorescence) to promote pollen exportation and reduce self-pollination. In these inflorescences, male-phase flowers lie spatially above female-phase flowers. To examine this untested hypothesis, we compiled literature information from 718 angiosperms species and evaluated the association between vertical acropetal inflorescences with protandrous flowers and bee pollination within a phylogenetic comparative framework. Results reveal that this type of inflorescence is indeed more common in species pollinated by bees. Moreover, this association does not seem to be weakened by the presence of alternative self-pollination avoidance mechanisms, like self-incompatibility, suggesting that this inflorescence type benefits mainly male rather than female fitness. Other inflorescence types placing male-phase flowers above female-phase flowers, e.g., vertical basipetal inflorescences with protogynous flowers, do not provide strong evidence of a differential association with pollination by bees. Female-biased nectar production in vertical acropetal inflorescences with protandrous flowers may reinforce the behavior of bees to fly upwards, rendering Darwin's configuration more adaptive than other inflorescence configurations.