Premise of the study: It has been proposed that species of columnar cacti from dry tropical areas depend on bats for their reproduction, whereas species from dry subtropical areas are also pollinated by other species. To test this hypothesis, we examined the effects of pollinator guild and of variation in time and space on the reproductive success of a widespread species. •
Methods: Changes in fruit set, seed set, and pollinator activity through time were recorded in three widely separated populations of Stenocereus thurberi. Breeding system and sources of pollination limitation were determined by controlled pollinator exclusions in each population. •
Key results: Significant differences were found in the timing of activity and in the effectiveness of pollinators among sites. In the northern and central populations, reproductive success depends on bats, whereas in the southern population a combination of pollinators was more effective. No difference between open and hand cross-pollination treatments was found in the northern and central populations, which suggests no pollen limitation. However, significant differences were detected in the southern population, which indicates temporal differences in pollinator abundance or arrival time. •
Conclusions: Local variation in pollinator assemblages and reproductive success could greatly affect the evolution of pollination systems. The pattern of generalist pollination in the southernmost populations and specialized pollination in the central and northern populations contradicts the hypothesis of latitudinal variation. In the absence of nocturnal pollinators, the accumulated nectar can sustain visits by diurnal pollinators, a bet-hedging strategy that increases the chances of fruit set in some populations.