This paper reports a study on the distribution of dinitrogen between the atmosphere, floodwater and porewater of the soil in a flooded rice field after addition of15N-labelled urea into the floodwater. Microplots (0.086 m2) were established in a rice field near Griffith, N.S.W., and labelled urea (80 kg N ha−1 containing 79.25 atoms %15N) was added to the floodwater when the rice was at the panicle initiation stage. Emission of nitrous oxide and dinitrogen was measured directly during the day and overnight, using a cover collection method and gas chromatographic and mass spectrometric analytical methods. Ammonia volatilization was calculated with a bulk aerodynamic method from measurements of wind speed and floodwater pH, temperature and ammoniacal nitrogen concentration. Seven days after urea application the15N2 content of the floodwater and soil porewater was determined and total fertilizer nitrogen loss was calculated from an isotopic balance. Throughout the experimental period gas fluxes were low; nitrous oxide, ammonia and dinitrogen flux densities were less than 5, 170 and 720 g N ha−1 d−1, respectively. The greatest dinitrogen flux density was observed two days after urea addition and this declined to ~ 100 g ha−1 d−1 after seven days. The data indicate that, of the urea nitrogen added, 0.02% was lost to the atmosphere as nitrous oxide, 0.9% was lost by ammonia volatilization, and 3.6% was lost as dinitrogen gas during the 7 days of measurement. At the end of this period 0.028% and 0.002% of the added nitrogen was retained as dinitrogen gas in the floodwater and soil porewater respectively. Recovery of the15N applied as nitrogen gases, plant uptake, and soil and floodwater constituents totaled about 94% of the nitrogen added.