Although challenging to study, soil is clearly a critical ecological arena for plants.  It is where plants are rooted and draw many resources, it teems with microbial life, and it is the seat of key biogeochemical processes.  Three aspects of soil – fertility, resource heterogeneity, and microbes – are of particular importance to the structure and dynamics of plant communities.  We discuss the theoretical frameworks for understanding how each of these aspects of soil affect vegetation structure and dynamics, and evaluate these frameworks in the light of observational and manipulative studies from both natural and restored plant communities.  While the patterns of plant community diversity responses to changes in soil fertility are clear, the mechanisms underlying these patterns are not yet well understood and the success of manipulating soil fertility in restorations has been mixed.  Additionally, new work is showing that experimental increases in soil fertility can produce dramatically different endpoints in community composition.  Although studies are limited, theoretical predictions are at odds with the results of experimental manipulations of soil heterogeneity.  Responses of clonal plants are strongly implicated in this disconnect for both natural and restoration contexts.  Two microbial processes that have emerged as potentially key drivers of plant community structure and dynamics are microbially mediated niche partitioning and plant-microbe feedbacks.  Of these, feedbacks have the most empirical support to date, and there is increasing application of positive feedback dynamics in restoration.  A challenge for the future lies in understanding how soil fertility, resource heterogeneity and plant-microbe interactions interact in driving vegetation structure and dynamics, and ecological restorations provide an excellent context in which to further develop and test such theory.