[en] About 9.5 million hectares (ha) are cultivated to rice–wheat in several regions of China. The rice–wheat (RW) rotation is the predominant agricultural system in South-western China. However, the productivity and sustainability of the system is threatened by soil degradation and water shortage. Two field experiments were conducted on a purple earth soil in Wenjiang county, South-western China to investigate the effect of water regime and straw and fertilizer N management on nitrogen and water use efficiencies of the rice–wheat cropping system. Treatments included conventional flooded, no mulch treatment (FNL), non-flooded, no mulch treatment (ANL), non-flooded, plastic mulching treatment (AML) and non-flooded, straw mulching treatment (ASL). "1"5N enriched fertilizers (180 kg N ha–1 for rice and 120 kg N ha–1 for wheat) were applied in micro plots to measure the fertilizer N recovery by the rice and wheat crops and the balance of the applied fertilizer N in the crop–soil system. In experiment 1 "1"5N labelled urea (5 atom % "1"5N) was the N source whereas in experiment 2 two N sources, i.e. "1"5N labelled ammonium sulphate (9.8 atom % "1"5N) and "1"5N labelled potassium nitrate (10.0 atom % "1"5N) were applied. In experiment 1, soil water content monitored with a neutron probe during the rice season was high in the top 0–10 cm layer, and declined with depth. Generally, the water content in each soil layer was higher with FNL than ANL or ASL, indicating that the imposed water regime was very important for the maintenance of soil water content. For the mulched treatments AML and ASL, soil water content was higher in the top 10 cm of soil, compared with ANL, indicating that plastic and straw mulching have the potential to prevent evaporation losses from the soil. Non-flooded plastic mulching had a similar effect on rice yield, N uptake, fertilizer N recovery and water use efficiency as conventional flooded, no mulch rice. The yield of rice decreased by 10–15% in the non-flooded, no mulch cultivation treatment, but straw mulching did not significantly decrease rice yield in the 3"r"d and 4"t"h years. Wheat yields were not affected by mulching. Urea-N recovery by flooded rice was on the average 22–25% of applied N and estimated N losses were 39–45% of the applied N. Non-flooded cultivation did not significantly affect the urea-N recovery by rice, but significantly decreased the residual fertilizer N in the soil, resulting in an increased N loss from the soil–plant system. Non-flooded rice cultivation with straw mulching decreased urea-N recovery by rice but increased the soil residual fertilizer N, thus resulting in decreased N loss from the soil–plant system. Urea-N recovery by wheat was on the average 37–44% of the applied N and estimated N losses were 18–29% of the applied N. Non-flooded pre-treatment did influence the wheat fertilizer N recovery; less fertilizer N was recovered while soil N recovery was similar thus resulting in higher N losses from the system compared to the flooded treatment. Water use efficiency for rice was much lower than that of winter wheat. The wheat water use efficiency was related to the water regime from the previous cropping season. In experiment 2, regardless of the flooding conditions, rice/wheat crops could recover significantly more NH_4"+-N than NO_3"-N, and the soil immobilized more NH_4"+-N than NO3"--N, so the loss with NH_4"+-N was significantly less than that of NO_3"-N. In the non-flooded straw mulching treatment, rice/wheat recovered significantly less N from NH_4"+-N or NO_3"-N while the soil residual N was significantly more than that of the flooded only treatment, especially for NH_4"+-N. Conventional flooded rice took up more N from NH_4"+-N or NO_3"-N than non-flooded rice, while recovery of NH_4"+-N or NO_3"-N and soil residual N in the wheat season were not different between flooded and nonflooded treatments. (author)