Stylosanthes viscosa is a common 'herb' in the tropical sand dunes of the Brazilian coast with clumped distribution in urban sand dunes of Salvador, BA. Its tiny keel flower with a very small amount of nectar hidden inside the corolla tube, a single open flower/inflorescence/day and the low floral density altogether should not be attractive to large bees. This should be true mainly upon the high-demanded energy for thermoregulation within the tropical sand dunes. Contrary to this expectation, most of the bee species sampled on S. viscosa flowers were medium to large sized bees and harvested nectar in legitimate visits by activating the keel's pollination mechanism of flowers. However, the keel was always inactivated after being visited by the enormous and frequent (78%) Xylocopa bees. All the flowers visited by these bees (including those with inactivated keels), and those experimentally protected from bee visiting produced fruits (100%). The specialized embolus mechanism of the anthers assures self-pollination. Because of this autonomous self-pollination and self-compatibility it was not possible to discriminate the relative importance of both facilitated self-pollination and cross-pollination by bees. Nevertheless, the large Xylocopa bees didn't limit self-pollination. Xylocopa cearensis was the most abundant bee (63%) and it visited several flowers (13.0 ± 4.27) in each flight route in the same clump, in spite of the low floral density therein (5 to 20 flowers.m-2). Flight distances between successive flowers varied between one and two meters (49%), and so they were higher than would be possible (15 to 70 cm) given the floral density in the clumps. The flight routes were directional, and most of shifting direction between successive flowers was below 90° with high frequency of small angles (<30°). This general pattern is explained by foraging optimization with adjustments to critical resource level thresholds. Paradoxically, floral biology and flowering pattern of S. viscosa should discourage the largest dune dweller bees, but this strategy works just temporarily. In this melittophilous species, the autonomous self-pollination (assured by the embolus) can be better explained by the automatic selection hypothesis rather than by reproductive assurance hypothesis, considering the availability of large bee pollinators in the tropical dunes.
automatic selection hypothesis; flight pattern; reproductive assurance hypothesis
