The importance of subsoil denitrification on the fate of agriculturally derived nitrate (NO3) leached to groundwater is crucial for budgeting N in an ecosystem and for identifying areas where the risk of excess NO3 is reduced. However, the high atmospheric background of di-nitrogen (N2) causes difficulties in assessing denitrification enzyme activity (DEA) and denitrification potential (DP) in soils directly. Here, we apply Membrane Inlet Mass Spectrometry (MIMS) technique to investigate indirectly DEA and DP in soils by measuring N2/Ar ratio changes in headspace water over soil. Soils were collected from 0–10, 15–25 and 60–70cm depths of a grazed ryegrass and grass–clover. The samples were amended with helium-flushed deionized water containing ranges of NO3 and carbon (glucose-C) and were incubated for six hours in the dark at 21°C. The peaks for N2/Ar ratio, declined with increasing soil depth, indicating a reduced substrate requirements to initiate DEA en-masse (15–30mg NO3-N alone or with 60–120mgglucose-C, kg−1soil). The dissolved N2O concentrations were very small (0.004–0.269μgNkg−1soil) but responded well to the added N and C, showing a reduction in DEA with soil depth. In three separate studies, only subsoils were incubated for 3 days at 12°C with 20–30mgNO3-N±40–60mgglucose-C, kg−1soil. Denitrification capacity (DC, NO3 only treatment) was not statistically different to the control (no amendment) within a land use (0.03–0.05 vs. 0.07–0.22mgNkg−1soild−1), the highest being in ryegrass subsoils receiving groundwater. The DP was significantly (P