Field studies of emissions of nitrous oxide (N2O) from a fallow soil treated with different forms and amounts of fertilizer N showed that N2O emissions from plots treated with N in the form of ammonium sulfate or urea markedly exceeded those from plots treated with the same amount of N in the form of calcium nitrate. This supports evidence from laboratory research that most of the N2O evolved from soils treated with ammonium and ammonium-producing fertilizers is generated by nitrifying microorganisms during oxidation of ammonium to nitrate and not, as previously assumed, through reduction of fertilizer-derived nitrate by denitrifying microorganisms. Emissions of N2O from plots treated with fertilizer N as ammonium sulfate or urea increased with the amount of N applied. Emissions of N2O in 25 days were increased 329--524% by application of 125 kg N ha-1 as ammonium sulfate or urea and increased 1024--1319% by the application of 250 kg N ha-1 in these forms, but these applications did not markedly increase N2O emissions after 25 days, and the fertilizer-induced emissions of N2O-N observed in 96 days from plots treated with ammonium sulfate or urea represented only 0.11--0.18% of the fertilizer N applied. Emissions of N2O from plots treated with different amounts of N as calcium nitrate did not increase with the amount of N applied and were not appreciably greater than the emissions observed when no fertilizer N was added.