Gaseous nitrogen losses, by NH ₃ volatilisation and denitrification, are mainly responsible for the low recovery of N fertiliser applied to irrigated maize on the North China Plain. Two field experiments were conducted to measure NH ₃ volatilisation and nitrification-denitrification losses from urea applied to maize (Zea mays L.) grown on a calcareous fluvo-aquic soil (Aquic Inceptisol) in Fengqiu County, Henan Province. The first was carried out in June 1998 (urea applied at 75 kg N/ha 3 weeks after sowing), and the second in July 1998 (urea applied at 200 kg N/ha 6 weeks after sowing). Each experiment included 3 treatments - control, surface-broadcast (SB), and deep point placement (DP) or broadcast followed by irrigation (BI). NH ₃ loss was measured by a micrometeorological method (NH ₃ sampler). Denitrification (N ₂ + N ₂O) was measured by the acetylene inhibition-intact soil core technique, and N ₂O emission was also measured in the absence of acetylene. The recovery of applied N was measured by a ¹⁵N balance technique. When urea was surface broadcast (SB) 3 weeks (75 kg N/ha) and 6weeks (200 kg N/ha) after sowing, 44 and 48% of the applied N was lost by NH ₃ volatilisation, respectively. The corresponding losses from the BI and DP treatments were only 18% and 11%, respectively. Denitrification was a significant process in this well-drained sandy soil, with average loss rates of 0.26-0.43 kg N/ha.day in the controls (from resident soil N), compared with 0.52-0.63 kg N/ha.day in the surface fertiliser treatments. Deep placement of urea reduced the denitrification rate to an average of 0.3 kg N/ha.day. The net denitrification loss from the fertiliser was <2% of the applied N, except for the SB urea treatment in the second experiment. The application of N fertiliser as urea increased N ₂O emissions from c. 0.3 to c. 2.3 kg N/ha over 57 days in the second experiment, with average N ₂O emission rates in the control and SB treatment of 0.006 and 0.042 kg N/ha.day, respectively. The significantly lower ratio of N ₂/N ₂O in the urea treatments compared with the control suggested that nitrification of applied N may have contributed to N ₂O production. Alternatively, the ratio of N ₂/N ₂O during dinitrification may have changed with the greater supply of NO ₃ ^-.