Bees may drive the reproduction of four sympatric cacti in a vanishing coastal mediterranean-type ecosystem

BACKGROUND: Sympatric congeneric plants might share pollinators, or each species might avoid competition by evolving specialized traits that generate partitions in pollinator assemblages. In both cases, pollen limitation (a decrease in the quality and quantity of compatible reproductive pollen) can occur, driving the plant mating system to autogamy as a mechanism of reproductive assurance. We assessed the relationships between pollinator assemblages and mating systems in a group of sympatric congeneric plants. We attempted to answer the following questions: (i) How similar are pollinator assemblages among sympatric cactus species? (ii) Which mating systems do sympatric cactus species use? METHODS: We studied sympatric Eriosyce taxa that inhabit a threatened coastal strip in a mediterranean-type ecosystem in central Chile. We performed field observations on four taxa and characterized pollinators during the years 2016 and 2017. We estimated differences in the pollinator assemblages using the Bray–Curtis index. To elucidate the mating systems, we conducted hand-pollination experiments using three treatments: manual cross-pollination, automatic self-pollination, and control (unmanipulated individuals). We tested differences in seed production for statistical significance using Kruskal–Wallis analysis. RESULTS: Eriosyce subgibbosa showed a distinctive pollinator assemblage among the sympatric species that we studied (similarity ranged from 0% to 8%); it was visited by small bees and was the only species that was visited by the giant hummingbird Patagona gigas. Pollinator assemblages were similar between E. chilensis (year 2016 = 4 species; 2017 = 8) and E. chilensis var. albidiflora (2016 = 7; 2017 = 4); however, those of E. curvispina var. mutabilis (2016 = 7; 2017 = 6) were less similar to those of the aforementioned species. E. curvispina var. mutabilis showed the highest interannual variation in its pollinator assemblage (18% similarity). Reproduction in E. subgibbosa largely depends on pollinators, although it showed some degree of autogamy. Autonomous pollination was unfeasible in E. chilensis, which depended on flower visitors for its reproductive success. Both E. chilensis var. albidiflora and E. curvispina var. mutabilis showed some degree of autogamy. DISCUSSION: We observed differences in pollinator assemblages between E. subgibbosa and the remaining Eriosyce taxa, which depend on hymenopterans for pollen transfer. Pollinator assemblages showed considerable interannual variation, especially those of E. subgibbosa (ornithophilous syndrome) and E. curvispina var. mutabilis (melitophilous syndrome). Autogamous reproduction in these taxa may act as a reproductive assurance mechanism when pollinator availability is unpredictable. Our study contributes to improving our understanding of the reproductive systems of ecological interactions between threatened species in a Chilean mediterranean-type ecosystem.