Abstract
Aim The outcome of interaction between plant individuals is often expected to be determined by differences in their sizes with other traits playing minor roles. However, such findings typically come from manipulative experiments in experimental gardens.
Consequently, little is known about the relative roles of size vs other traits in the field, particularly in communities other than forests. Here we aimed; (a) to determine the relative role of size-related traits in pairwise interactions between common species in a species-rich grassland community; and (b) to determine which other traits predict outcomes of such interactions in the field, and which traits predict mean competitive effects and responses of individual species.
Location The Krkonose Mts., Czech Republic. Methods We fitted dynamic models with explicit terms for pairwise interactions on a 31-year long data series of spatially explicit ramet counts of 19 coexisting species in a mountain meadow.
We compared these fits with those having total neighbour biomass as the only competition term to determine whether accounting for pairwise interactions increases model fit. Further we linked estimates of the parameters of pairwise interactions with species traits, including size-related traits.
Results Comparison of models with and without pairwise predictors showed that pairwise interactions provided better fits only in less than half of the cases. The main determinant of competitive interactions was leaf area and, to a lesser extent, mean ramet biomass; there was an additional signal of difference in rooting depth.
Conclusion Plant size and overall biomass have overwhelming effects on species-to-species interactions even in unproductive grassland communities, despite low differentiation of plant species along the size axis there. However, the weak albeit detectable signal of predictors not related to size of pairwise interactions points to possible mechanisms of within-community niche differentiation, which can be captured by inter-specific difference in functional traits and tested experimentally.