Natural populations are evolving in ecological communities. Ecological communities are structured by stochastic forces, introducing variability in where species are and at what abundances over short scales of space and time. It remains unknown if and how community processes constrain adaptive evolution, and the role that stochastic processes play in setting those constraints. In this talk, I address this knowledge gap in two ways. First, I in grassland communities in Northern California, I test the strength of local adaptation with and without the presence of intact communities using common garden experiments. Around each transplanted individual, I additionally characterized the interaction neighbourhood around each focal individual. This method allowed us to fit multi-species competition models to the fitness data, estimating interaction strengths between our focal B. hordeaceus populations and each of seven common neighbourhood species. Second, I performed metacommunity simulations to test when and how stochastic processes alter the strength and spatial scale of adaptation in ecological communities.
I found that interaction strengths with the local B. hordeaceus population evolved with two of seven neighbourhood species, increasing competition with one and decreasing it with the other. Higher-order interaction (HOIs) strengths did not differ among local and foreign populations on average, but instead, local populations showed a marked increase in variance of HOIs. Fitness of local populations decreased when neighbours were removed, causing local maladaptation, whereas all populations performed equally when neighbours were left intact. Our simulation results indicate that, as diversity increases in communities, dispersal among neighbourhoods increasingly constraints adaptive evolution. Together, our results demonstrate how species interactions evolve to determine fitness in complex communities, providing a richer view of adaptation in natural systems.