Abstract The source of N 2O in terrestrial ecosystems has long been debated. Both nitrification and denitrification produce N 2O but their relative importance remains uncertain. Here we apply site preference, SP (the difference in δ 15N between the central and outer N atom in N 2O), to estimate the relative importance of bacterial denitrification (including nitrifier denitrification) to total N 2O production from soil. We measured SP over a diurnal cycle following the third year of tillage of a previously uncultivated grassland soil at the Kellogg Biological Station (KBS) in southwestern Michigan. Fluxes of N 2O in our study ranged between 7.8 and 12.1 g N 2O–N ha −1 d −1 and were approximately 3 and 10 times greater than fluxes observed in managed agricultural and successional fields, respectively, at KBS. Consequently, our study captured a period of high flux resulting from the cultivation of a historically never-tilled soil. Concentration weighted SP values decreased from 12.9‰ in the morning to a minimum value of −0.1‰ in the afternoon. Based on SP values reported for bacterial denitrification (−5 to 0‰; Toyoda et al., 2005; Sutka et al., 2006), hydroxylamine oxidation (nitrification) and fungal denitrification (33–37‰; Sutka et al., 2006) we found that production attributable to bacterial denitrification increased from between 52.9 and 60.9% in the morning to between 87.5 and 100% in the afternoon. Further, we observed diurnal variation in flux and SP that is consistent with increased production from bacterial denitrification associated with temperature-driven increases in respiration.