A critical assumption underlying terrestrial ecosystem models is that soil microbial communities, when placed in a common environment, will function in an identical manner regardless of the composition of that community. Given high species diversity in microbial communities and the ability of microbes to adapt rapidly to new conditions, this assumption of functional redundancy seems plausible. We test the assumption by comparing litter decomposition rates in experimental microcosms inoculated with distinct microbial communities. We find that rates of carbon dioxide production from litter decomposition were dependent upon the microbial inoculum, with differences in the microbial community alone accounting for substantial (approximately 20%) variation in total carbon mineralized. Communities that shared a common history with a given foliar litter exhibited higher decomposition rates when compared to communities foreign to that habitat. Our results suggest that the implicit assumption in ecosystem models (i.e., microbial communities in the same environment are functionally equivalent) is incorrect. To predict accurately how biogeochemical processes will respond to global change may require consideration of the community composition and/or adaptation of microbial communities to past resource environments.