Agricultural practices are believed to be the major anthropogenic source of enhanced nitrous oxide (N 2 O) gas emissions in New Zealand. Studies conducted in New Zealand generally suggest low N 2 O emission from pasture; however, there is little information for arable farming systems. This paper evaluates tillage and land use effects on N 2 O emissions using a closed chamber technique at an Ohakea silt loam (Gleyic luvisol) where winter oats (Avena sativa L.)/fodder maize (Zea mays L.) was double-cropped for 5 years. The tillage types included conventional tillage (CT) and no-tillage (NT) systems, and a permanent pasture (PP) was used as a control.Spatial variability in all treatments showed large inherent variations in N 2 O fluxes (a mean CV = 119%), which reflected natural soil heterogeneity, and perhaps the measurement technique used rather than the real differences due to the tillage and cropping systems evaluated. On an annualised basis, N 2 O emissions measured from December 1998 to September 1999 from the PP (1.66 kg N 2 O-N/ha per year or 19 g N 2 O-N/(m 2 h)) were significantly lower than the CT and NT fields averaging at 9.20 (or 105) and 12.0 (or 137) kg N 2 O-N/ha per year (or g N 2 O-N/(m 2 h)), respectively. However, there were no differences in N 2 O emission rates between the CT and NT treatments. Seedbed preparation using a power harrow which followed within a few days of first ploughing the CT field reduced N 2 O emissions by 65% within the first hour after power harrowing. However, N 2 O emission rates returned to the pre-power harrowing levels at the next sampling period, which was 1 month later.There was a strong relationship between log-transformed data of soil water content (SWC) and N 2 O emissions in all treatments with r = 0.73, 0.75 and 0.86 for the PP, CT and NT treatments, respectively. Seasonal variations in N 2 O emission from the PP were in the order of winter = autumn > summer. Although fluxes in the CT were higher in winter than in the autumn season, there were no differences between the summer and autumn data. The seasonal variations in N 2 O emission in the NT treatment were in the order of winter > autumn = summer.