The outcome of animal-plant interactions, such as pollination and seed dispersal, is often context-dependent and varies across temporal and spatial scale. More specifically, traits that characterize a specific interaction at the center of its distribution may confer different payoffs at the edge of the distribution of one or more of the interacting species. We tested whether variation in seed traits, which are associated with dispersal by ants, and ant seed-dispersing behaviors vary in a predictable manner toward a plant range margin. Theoretical models predict that populations at species range margin evolve reduced dispersal comparing to core populations. We used a model system of the ant-dispersed plant, Sternbergia clusiana, to test whether the study plant a) exhibits traits that allow them to interact with beneficial seed dispersers, and b) whether ant-plant co-occurrence, seed traits and ant seed-carrying behavior vary along a geographic gradient that includes the plant geographic range margin. We mapped the plant distribution at various scales. In a set of cafeteria experiments, we quantified ants' activity pattern and seed-carrying behavior, and we ‘interviewed’ them regarding their preference for various seeds.
Results/Conclusions
We found that S. clusiana’s seeds are dispersed by two ant guilds differing significantly in their seed dispersal rates and distances. Those among-guild differences in seed-dispersal quality were positively correlated with plant investment in the reward (elaiosome size). These behavioral differences were consistent across the wide geographic range in which S. clusiana occurs. In addition, we found that seed dispersal by ants was much reduced at the geographic range margin. However, this reduction in seed dispersal towards the range margin was not associated with differences in plant or ant traits, but by a decline in co-occurrence of plants and effective seed-dispersing ants at the local scale. In fact, at the plant geographic range margin, the plant and the effective seed dispersers were locally segregated. Additionally, the plant range margin populations exhibit lower investment in seed production. Our results support the predictions of selection for reduced dispersal at range. margins. However, the mechanisms underlying this reduction in dispersal may be casual and not necessary the result of local selection pressures.