Complex plant–soil interactions enhance plant species diversity by delaying community convergence

1. A plant that causes specific changes to soil biota may either positively or negatively affect the performance of the plant that subsequently grows in that location. These effects, known as plant–soil feedback, can affect plant species diversity at multiple spatial scales. 2. It has been hypothesized that positive plant–soil feedback reduces alpha (local) diversity by allowing dominance by early-arriving species, but increases gamma (regional) diversity by promoting community divergence (increased beta diversity) through the emergence of alternative stable states. In contrast, negative plant–soil feedback has been thought to increase alpha diversity by allowing local species coexistence, but to reduce gamma diversity by promoting community convergence (reduced beta diversity). Although widely accepted, these hypotheses do not consider the possibility that plant species differ in their effect on, and their response to, a given other species via soil biota. In reality, plant–soil interactions can be complex, with the strength of the interactions variable between plant species. Using a basic simulation model of plant community assembly, we investigated how complex plant–soil interactions might affect plant diversity during succession. 3. When we included only positive or negative intraspecific plant–soil feedback in the model, with no variation in the strength of interspecific plant–soil interactions, results were consistent with the conventional hypotheses. When we allowed the strength of plant–soil interactions to differ between species, plant–soil interactions enhanced alpha diversity initially and beta and gamma diversity subsequently. Diversity enhancement occurred not necessarily because alternative stable states emerged, but primarily because complex plant–soil interactions lengthened the time during which local species composition changed. Due to the longer time for changes in species composition, the high level of beta and gamma diversity at the early stage of succession was maintained for a long time despite eventual community convergence. Thus, diversity enhancement was often transient, though long-lasting, making the conventional concept of alternative stable states inadequate for explaining diversity. 4. Synthesis. Based on these findings, we propose the new hypothesis that complex plant–soil interactions enhance plant species diversity by delaying community convergence. This hypothesis highlights the role of plant–soil interactions as a driver of long-lasting transient dynamics of community assembly.