Nitrous oxide (N20) contributes to global climate change and agricultural soils seem to be the major source. Lack of information led to this study on the influence of different amounts and sources of nitrogen on N2O emission from a maize (Zea mays L.)-groundnut (Arachis hypogae L.) crop rotation in an Ultisol of the humid tropics. The treatments were: inorganic N + crop residues (NC), inorganic N only (RN), and half of inorganic N + crop residues + chicken manure (NCM). The corresponding amount of N applied was 322, 180, and 400 kg ha(-1) yr(-1), respectively. The N2O emissions depended on the amounts and types of N. A maximum peak (9,889 +/- 2,106 microg N2O-N m(-2) d(-1)) was detected at 2 wk before maize sowing amended with chicken manure, showing a persistent influence on N transformations and N2O release. The mineral N from either applied source became low by 2 to 4 wk, coinciding with the small N2O fluxes or its consumption to a few isolated instances. The N2O flux significantly correlated with the mineral N and water-filled pore spaces. The direct annual N2O emission was 3.94 +/- 0.23, 1.90 +/- 0.08, and 1.41 +/- 0.07 kg N2O-N ha(-1) from the NCM, NC, and RN treatments, respectively. The corresponding N2O-N loss of the applied N plus N fixed by groundnut was 0.83, 0.49, and 0.59%. Overestimations of direct annual N2O emission using the Intergovernmental Panel on Climate Change (IPCC) methodology suggest a location-specific emission factor for variable N sources to be considered.