Grass abundance shapes trait distributions of forbs in an experimental grassland

Questions: During community assembly, plant functional traits are hypothesized to be filtered both by prevailing abiotic conditions and by competition from the surrounding community. While trait responses to abiotic conditions are well documented, whether and how traits respond to the biotic community is less studied. We investigated trait responses in specific leaf area (SLA) and plant height for eight forb species in response to increasing grass abundance. We asked: (1) do species' trait distributions respond predictably to increased grass presence; (2) do functionally similar species respond similarly in terms of trait means to increased grass presence; and (3) at the community level, does increasing grass presence drive change in functional diversity among forbs? Location: Glacial Heritage Preserve, Thurston County, Washington, USA. Methods: Eight forb species were grown in communities that differed in the proportion of grass seed sown into the plot (grass-rich, mixed and forb-rich) in an experimental grassland established under uniform abiotic conditions. SLA and height were measured for 29-33 individuals per species per treatment, and differences among species and treatments were tested using PERMANOVA and PERMDISP. To assess how Festuca abundance impacted trait distributions for the entire suite of forbs, functional richness (FRic), functional evenness (FEve) and functional dispersion (FDis) were calculated for each seeding treatment and compared using ANOVA. Results: Trait means and/or dispersion of most species responded to grass abundance, but the direction and magnitude differed among species. FEve was lowest in the grass-rich treatment for SLA, and FRic was significantly lower in the grass-rich treatment for both SLA and height. FDis for the combined traits was lowest in the grass-rich treatment. Conclusions: This study provides clear evidence that the biotic community can shape plant trait expression within and across species, and supports the idea that at the species level, trait values are not fixed, but rather can vary in response to biotic factors. This variation can impact functional diversity as well as community-level trait estimates and therefore also affect our understanding of ecosystem functions that are reliant on those estimates.