Measurements were made of nitrous oxide (N2O) emissions from N-fertilised ungrazed grassland and arable land at sites widely distributed across Great Britain during 1999–2001. The closed static chamber method was used throughout. Emissions varied widely throughout the year at each site, and between sites. Daily fluxes up to 1200 g N2O–N ha−1 d−1 were recorded. The highest annual flux was 27.6 kg N2O–N ha−1 at a grassland site in Wales, whereas the lowest, 1.7 kg N2O–N ha−1, occurred on a soil overlying chalk in southern England. The key factors affecting N2O emissions from agricultural soil were soil WFPS, temperature and soil NO3––N content. On grassland, rainfall (particularly around the time of N application), with its consequent effect on water-filled pore space (WFPS), was the main driving factor during the growing season. Annual emission factors (EFs), uncorrected for background emission, varied from 0.4 to 6.5% of the nitrogen (N) applied, covering a similar range for grassland to that found previously for sites restricted to Scotland. Continued monitoring at a grassland reference site near Edinburgh showed that annual EFs vary greatly from year to year, even with similar management, and that several years' data are required to produce a robust mean EF. The overall distribution of EFs in this and previous studies was log-normal. The EFs for small-grain cereals (and oilseed rape) peaked at a much lower value than those for grassland, whereas the values for leafy vegetables and potato crops fitted well into the grassland distribution. These differences in EF between various types of crop should be taken into account when compiling regional or national N2O emission inventories.