Trait variation across biological scales shapes community structure and ecosystem function.

Trait variation underlies our understanding of the patterns and importance of biodiversity, yet we have a poor understanding of how variation at different levels of biological organization structures communities and ecosystems. Here, we use a mesocosm experiment to test for the effects of a larval dragonfly functional trait on community and ecosystem dynamics by creating artificial populations to mirror within- and between-population trait variation observed in our study area. Specifically, we manipulate variation in activity rate, a key functional trait shaping food webs, across three levels of biological organization: within-populations (differences in trait variation in a population), among-populations (differences in population mean trait values), and among-species (species-level differences of co-occurring dragonflies). We show that differences in activity rate alter prey communities, trophic cascades, and multiple ecosystem processes. However, trait variation among populations had much larger effects than differences between co-occurring species or even the presence of a predator, whereas within-population variation had a relatively minor impact. Interestingly, combined with earlier work in the same system, our study suggests that the relative importance of species vs. individual level differences for ecosystem functioning will depend on the spatial scale considered. Ecological processes, including biodiversity-ecosystem-functioning relationships, cannot be understood without accounting for trait variation across biological scales of organization, including at fine scales.