The nitrite reductase (nirS and nirK) and nitrous oxide reductase-encoding (nosZ) genes of denitrifying populations present in an agricultural grassland soil were quantified using real-time polymerase chain reaction (PCR) assays. Samples from three separate pedological depths at the chosen site were investigated: horizon A (0-10 cm), horizon B (45-55 cm), and horizon C (120-130 cm). The effect of carbon addition (treatment 1, control; treatment 2, glucose-C; treatment 3, dissolved organic carbon (DOC)) on denitrifier gene abundance and N2O and N2 fluxes was determined. In general, denitrifier abundance correlated well with flux measurements; nirS was positively correlated with N2O, and nosZ was positively correlated with N2 (P < 0.03). Denitrifier gene copy concentrations per gram of soil (GCC) varied in response to carbon type amendment (P < 0.01). Denitrifier GCCs were high (ca. 10(7)) and the bac:nirK, bac:nirS, bac:nir (T) , and bac:nosZ ratios were low (ca. 10(-1)/10) in horizon A in all three respective treatments. Glucose-C amendment favored partial denitrification, resulting in higher nir abundance and higher N2O fluxes compared to the control. DOC amendment, by contrast, resulted in relatively higher nosZ abundance and N2 emissions, thus favoring complete denitrification. We also noted soil depth directly affected bacterial, archaeal, and denitrifier abundance, possibly due to changes in soil carbon availability with depth.