Here we report a combined analysis of carbon mass balance based on isotopic labeling and microbiological characterization during organic carbon stimulated bioreduction of a subsurface sediment in a large laboratory column experimental system. This combination of approaches allows quantification of both the cycling of carbon through multiple redox pathways and the associated spatial and temporal evolution of bacterial communities in response to this nutrient source. Carbon isotope mass balance facilitated by the use of 13C-labeled acetate as the electron donor showed evidence for a net loss of sediment organic carbon over the course of the amendment experiment. Furthermore, these data clearly demonstrated a source of isotopically labeled inorganic carbon that was not attributable to primary metabolism by acetate-oxidizing microorganisms. Fluid samples collected weekly over the duration of the 43-day amendment at <20cm intervals along the flow path were analyzed for microbial composition by pyrosequencing of ribosomal RNA genes. The microbial community composition was transient, with distinct occurrences of Azoarcus, Geobacter and multiple sulfate reducing species over the course of the experiment. In combination with DNA sequencing data, the anomalous carbon cycling process is shown to occur exclusively during the period of predominant Geobacter species growth. Pyrosequencing indicated, and targeted cloning and sequencing confirmed the presence of several bacteriovorous protozoa, including species of the Breviata, Planococcus and Euplotes genera. Cloning and qPCR analysis demonstrated that Euplotes species were most abundant and displayed a growth trajectory that closely followed that of the Geobacter population. These results suggest a previously undocumented secondary turnover of biomass carbon related to protozoan grazing that was not sufficiently prevalent to be observed in bulk concentrations of carbon species in the system, but was clearly identified in the partitioning of carbon isotopes. This study demonstrates evidence for predator-prey relationships that impact subsurface microbial community dynamics and provides a novel indication of the impact of this relationship on the flux of carbon through a system via the microbial biomass pool. Overall, our approach provides high temporal and spatial sampling resolution at field relevant flow rates, while minimizing effects of mixing and transverse dispersion. The result is a quantitative carbon budget accounting for a diversity of processes that should be considered for inclusion in reactive transport models that aim to predict carbon turnover, nutrient flux, and redox reactions in natural and stimulated subsurface systems.