Emissions of nitrous oxide from intensively managed agricultural fields were measured over 3 years. Exponential increases in flux occurred with increasing soil water-filled pore space (WFPS) and temperature; increases in soil mineral N content due to fertilizer application also stimulated emissions. Fluxes were low when any of these variables was below a critical value. The largest fluxes occurred when WFPS values were very high (70--90%), indicating that denitrification was the major process responsible. The relationships with the driving variables showed strong similarities to those reported for very different environments: irrigated sugar cane crops, pastures, and forest in the tropics. Annual emissions varied widely (0.3--18.4 kg N2O-N ha-1). These variations were principally due to the degree of coincidence of fertilizer application and major rainfall events. It is concluded therefore that several years' data are required from any agricultural ecosystem in a variable climate to obtain a robust estimate of mean N2O fluxes. The emissions from small-grain cereals (winter wheat and spring barley) were consistently lower (0.2--0.7 kg N2O-N per 100 kg N applied) than from cut grassland (0.3--5.8 kg N2O-N per 100 kg N). Crops such as broccoli and potatoes gave emissions of the same order as those from the grassland. Although these differences between crop types are not apparent in general data comparisons, there may well be distinct regional differences in the relative and absolute emissions from different crops, due to local factors relating to soil type, weather patterns, and agricultural management practices. This will only be determined by more detailed comparative studies. ¿ 1999 American Geophysical Union