Abstract A challenge in ecology and biogeography is to understand the drivers of the composition and distribution of biological communities. Environmental factors (especially pH) and dispersal limitation are thought to exert the primary controls on the composition of soil bacterial communities. However, quantifying their relative importance remains difficult because of analytical uncertainties. For example, the relationship between bacterial community composition and soil pH, which is often nonlinear, is typically evaluated with a linear test and without accounting for variability in rates of turnover along environmental gradients. Furthermore, potential drivers of variation in soil pH, and therefore bacterial community composition, are not commonly analyzed during microbial biogeographical studies. To address these issues we collected 700 soil samples across multiple spatial scales from beneath four late-successional tree species within 12 forests in the eastern United States. We performed high-throughput sequencing of 16S rDNA amplicons and measured soil properties thought to influence soil bacterial composition. Generalized Dissimilarity Modeling, a non-linear form of matrix regression, indicated that geographic distance and soil properties explained 77.3% of the deviance in turnover in overall bacterial community composition. However, only 2.1% of the explained deviance was attributable to geographic distance, indicating little contribution of dispersal limitation to bacterial ß-diversity across scales of ∼1.7 m to >1000 km. Although 81.7% of the explained deviance in overall bacterial composition was attributable to soil properties, particularly soil pH, the magnitude and rate of compositional turnover varied among bacterial families across the pH gradient. The ß-diversity of three dominant families (Bradyrhizobiaceae, Hyphomicrobiaceae and Burkholderia) was explained by neither soil properties nor geographic distance. Differences in soil pH between certain tree species likely led to distinct bacterial communities at several sites. Thus, shifts in soil pH, potentially as the result of shifts in tree composition, will likely have important consequences for the composition of soil bacterial communities.