Although microorganisms largely drive many ecosystem processes, the relationship between microbial composition and their functioning remains unclear. To tease apart the effects of composition and the environment directly, microbial composition must be manipulated and maintained, ideally in a natural ecosystem. In this study, we aimed to test whether variability in microbial composition affects functional processes in a field setting, by reciprocally transplanting riverbed sediments between low-and high-salinity locations along the Nonesuch River (Maine, USA). We placed the sediments into microbial 'cages' to prevent the migration of microorganisms, while allowing the sediments to experience the abiotic conditions of the surroundings. We performed two experiments, short-(1 week) and long-term (7 weeks) reciprocal transplants, after which we assayed a variety of functional processes in the cages. In both experiments, we examined the composition of bacteria generally (targeting the 16S rDNA gene) and sulfate-reducing bacteria (SRB) specifically (targeting the dsrAB gene) using terminal restriction fragment length polymorphism (T-RFLP). In the short-term experiment, sediment processes (CO 2 production, CH 4 flux, nitrification and enzyme activities) depended on both the sediment's origin (reflecting differences in microbial composition between salt and freshwater sediments) and the surrounding environment. In the long-term experiment, general bacterial composition (but not SRB composition) shifted in response to their new environment, and this composition was significantly correlated with sediment functioning. Further, sediment origin had a diminished effect, relative to the short-term experiment, on sediment processes. Overall, this study provides direct evidence that microbial composition directly affects functional processes in these sediments.