Species-specific interactions between plants and their soil-borne microbial associates have the potential to maintain plant species diversity and determine species relative abundance patterns. Although many greenhouse experiments demonstrate the existence of plant-soil feedbacks, fewer have explicitly tested whether greenhouse-generated feedback measures provide inference to temporal dynamics of mixed-species communities. We conducted a fully reciprocal experiment to determine the strength and magnitude of feedbacks among 18 tallgrass prairie grasses and forbs, in the presence or absence of soil biota. We also established a large, outdoor mesocosm experiment, where we drew from the same 18 species pool to assemble 3 and 6 species communities, initially with and without soil microbes. Change in species-specific biomass in this experiment was measured over three years. Results from these experiments were used to ask whether variation in strengths of plant-soil feedbacks predict plant community dynamics.
Results/Conclusions
We found strong evidence that our tallgrass prairie system is dominated by strong stabilizing interactions caused by soil biota, and the variation in strengths of these interactions predict temporal change in biomass in mixed-species communities. Specifically, in the greenhouse experiment, 10 of the 18 plant species exhibited negative plant-soil feedbacks when grown with live soil biota. When soil microbes were eliminated, these negative feedbacks vanished. When linking results from the greenhouse to those from the mesocosm experiment, we found that annual increase in biomass was higher for plant species exhibiting weak negative feedbacks, while such increase was suppressed for those species exhibiting strong negative feedbacks. There was no relationship between feedback strength and annual change in biomass when soil microbes were eliminated. Our finds provide strong support that plant-microbial interactions can shape temporal dynamics in ways that may set the stage for differences in relative abundance patterns.