Soil heterogeneity generated by plant–soil feedbacks has implications for species recruitment and coexistence

Summary Most studies of soil heterogeneity have focused on underlying abiotic factors such as soil nutrients. However, increasing recognition of plant–soil feedback ( PSF ) effects on plant growth, combined with the observation that PSF s operate at small spatial scales, suggests that heterogeneity due to PSF could affect plant population and community dynamics. The consequences of PSF ‐generated heterogeneity for coexistence depend on heterogeneity's effects on vital rates and how those vital rates influence population‐level recruitment dynamics. We measured vital rates and recruitment dynamics of three congeneric pairs of introduced perennial plants grown as monocultures in experimental PSF ‐generated soil environments. Field soils collected from conspecifics and congeners were alternated in patches or mixed together to produce heterogeneous and homogeneous soils, respectively. We quantified the effects of PSF ‐generated heterogeneity on germination and establishment and determined how these vital rates affected recruitment. We calculated net pairwise interaction coefficients to predict whether PSF s could mediate coexistence between congeners. Soil heterogeneity altered the relationship of vital rates to recruitment dynamics for some species. For example, S olanum dulcamara recruited later into heterogeneous than homogeneous soils, and germination was a stronger predictor of the timing of recruitment in heterogeneous soil, while mortality was a stronger predictor in homogeneous soil. Contrasts between soils of different origin suggest that mixing soils had non‐additive effects on vital rates (e.g. R umex crispus mortality was higher in homogeneous than in conspecific or congener soil). Interaction coefficients predicted that PSF s in heterogeneous soils might mediate stable coexistence only of R umex congeners. Synthesis . Heterogeneity generated by PSF s had species‐specific effects on vital rates, with consequences for recruitment dynamics. Mixing soils of different origin often resulted in non‐additive effects, which may indicate an interaction between soil abiotic and/or biotic properties and could predict non‐additive responses to soil disturbance. Finally, quantifying the reciprocal effects of PSF s on congeners suggests that PSF ‐generated heterogeneity may promote coexistence of certain species, which was not evident from individual PSF responses. Future studies should determine whether such mechanisms might operate for more distantly related species.