Quantifying the effect of plant-soil feedbacks on plant coexistence: A meta-analysis of the microbially mediated fitness difference

Xinyi Yan, Integrative Biology, University of Texas at Austin, Austin, TX, Jonathan M. Levine, Department of Ecology and Evolutionary Biology, Princeton University, Princeton, NJ and Gaurav Kandlikar, Division of Biological Sciences and Division of Plant Sciences, University of Missouri, Columbia, MO

Background/Question/Methods
Dynamic feedbacks between plant species and their soil microbial communities play an important role in shaping a variety of processes in plant communities, including the maintenance of plant species diversity. Recent theoretical advances have shown that evaluating the net effects of such plant-soil feedbacks (PSFs) on plant coexistence requires comparing the strength of their stabilizing or destabilizing effects (via driving negative or positive frequency dependence, respectively), to the frequency-independent fitness difference they simultaneously generate among plants. Most empirical studies of PSF have tested only for stabilizing or destabilizing effects of soil microbes, with previous meta-analyses showing that such feedbacks often generate negative frequency dependent dynamics that should promote plant diversity. However, very few empirical studies have compared the (de)stabilizing effects of PSFs to the fitness differences they mediate, and the net effects of these feedbacks on plant coexistence remain unclear. We conducted a meta-analysis to evaluate the magnitude of microbially mediated fitness difference, and if it tends to overwhelm any stabilizing effect. We screened 69 papers from a published meta-analysis dataset and an additional 330 papers published since, to select studies that factorially grew plants in soils conditioned by conspecifics, by a heterospecific, and in an unconditioned (reference) inoculum.
Results/Conclusions
We identified 40 qualified studies, which used either autoclaved soil (n=24) or live soil collected from the field (n = 16) as the unconditioned reference. Across studies with live field soil reference (72 species pairs), the effect of PSFs is on average stabilizing (mean = 0.26, CI = (0.07,0.46), p = 0.007). However, we also found that PSFs also generate significant fitness differences among species pairs (mean = 0.64, CI = (0.44, 0.84), p < 0.0001). In 54% of species pairs, the mean tendency was for microbes to favor species exclusion over coexistence (larger microbially mediated fitness differences than stabilization). Using a simulation approach, we found that microbes drive significantly stronger fitness differences than stabilization in a quarter (17/72) of the pairs in our dataset. On the other hand, microbes only generated significantly stronger stabilization than fitness differences in 2 (3%) of species pairs. These preliminary results reveal the tendency of PSFs to generate substantial fitness differences that often undermine their stabilizing effects. Our work also highlights the need for empirical studies to include carefully selected reference treatments for calculating microbially mediated fitness differences and thereby inferring the total effects of soil microbes on plant diversity.