AbstractQuestionsCommunity assembly is envisaged as filters that preclude some species in the regional pool from invading local communities. We tested whether the large floristic differences between adjacent calcicole and calcifuge grasslands are explained by either of five filters: environment (soil) or positive or negative interactions with plants and with soil biota.LocationSouthern Mexico.MethodsWe conducted a reciprocal‐transplant experiment with 20 species that were introduced to both habitats under three conditions: intact local community, without plants, and in sterilized plots. Each of the five filters mentioned above predict unique patterns in the performance (survival and growth) of plants in the six treatments. Thus, we used multimodel inference to determine which filters (patterns) were consistent with the evidence.ResultsWe detected at least one filter operating on all but four species. Survival data showed frequent support for environmental filtering, with interactions (mostly positive) playing a secondary role; however negative interactions became as frequent as environmental filtering when growth was considered.ConclusionsThe large physicochemical differences between the soils of both grasslands explain the high frequency of environmental filtering. Soils differed in nutrient availability, but also had toxic concentrations of different elements. Survival was strongly influenced by early mortality, while size was measured at the end of the experiment. It is thus likely that the differences between analyses based on survival and size reflect an ontogenetic change from positive to negative interactions. Other plants frequently facilitate seedlings, but this interaction often turns competitive over time. Soil mutualists provide nutrients that seedlings cannot access, but antagonists build up in the rhizosphere as plants age. Unlike studies that infer filters from extant plants in communities or from successful invasions, our approach provides direct evidence on which filters cause species to be absent from communities.
Read full abstract