Soil carbon pools reflect a long-term balance between carbon inputs and outputs, driven primarily by plant production and microbial decomposition. It is important to understand plant-microbe feedbacks on carbon cycling in order to predict how the balance between these processes will change in a future atmosphere of elevated CO₂ (E). We examined the influence of E on microbial community composition, microbial activity, and soil carbon content in a Florida scrub oak ecosystem. The field experiment consisted of replicate (n = 8) open-top chambers exposed to ambient or twice-ambient atmospheric CO₂ concentrations for 6 years. E soils had higher fungi:bacteria ratios (p=0.04) and polyphenoloxidase activity (p=0.090) than ambient CO₂ (A) soils. Polyphenoloxidase degrades recalcitrant organic substrates. We also conducted a laboratory-based experiment in which soil organic matter (SOM) and leaf litter amendments had distinct isotopic signatures. Adding litter stimulated microbial respiration rates more in E soils than A soils (p = 0.055), suggesting a potential stimulation or ‘priming’ effect on SOM mineralization. We present two lines of evidence to argue that E has stimulated SOM mineralization in the field study. First, E soils consistently lost SOM relative to A soils between 1998 and 2002 (r²=0.98, p=0.009). Second, the size of the laboratory-derived priming effect was negatively related to the size of the in situ SOM pool (r²=-0.49, p=0.003). These data suggest that E enhanced soil organic matter degradation through changes in microbial community composition and activity; this mechanism may be an important constraint on how much of the ‘extra’ carbon assimilated by plants in an E atmosphere enters long-term storage in soil carbon pools.