Nutrient pollution of terrestrial systems can reduce plant diversity and lead to non-linear transitions in ecosystems, including hysteresis and regime shifts. These low diversity states can persist for decades after cessation of nutrient addition; however, the mechanisms reinforcing these alternative states are still unclear. Here, we test the effect of plant-soil feedbacks as a reinforcing mechanism of alternative stable states in plant diversity in an experimental grassland in Minnesota. Nutrient enrichment in this grassland has reduced plant diversity and led to the establishment of an exotic species as dominant, even 20 years after cessation of nutrient addition. Replicate experimental plots have either received no nutrients (controls) or nutrient addition for either 30 years or 10 years followed by 20 years of no nutrient addition. First, we performed a greenhouse experiment to test the direct effect of soil inoculum from these plots on native and exotic plant biomass production. Then, we sequenced microbial DNA from these soils, before and after the greenhouse experiment, in order to test for shifts in microbial community composition in response to the nutrient addition treatments, the specific plant species that dominate either regime, or both.
Soil inoculum, regardless of nutrient treatment, increased plant biomass when compared to plants growing on sterile soil, however, the effect was only marginally significant (P = 0.095). Soil inoculum from plots that had undergone two decades of cessation of nutrient addition increased plant biomass when compared to those still undergoing nutrient enrichment (P = 0.013). These effects were stronger when plots had received high nutrient inputs (P = 0.032). Our results indicate that altered plant-soil feedbacks, particularly those between plants and soil microbes, might accelerate the loss of plant diversity during nutrient pollution. However, plant-soil interactions do not seem to be reinforcing the persistent low-diversity state decades after cessation of nutrient enrichment in this grassland. This could be due to loss of important mutualists for native species, and possibly also to rapid evolution of less cooperative microbial mutualists due to nutrient enrichment. This has important implications for conservation and ecosystem service management, as restoration of soil biota might accelerate the recovery of biodiversity in these systems. Further experimentation will be carried out in 2016 to clarify the direct and indirect effects of microbial mutualists on native and exotic grass performance.