Management practices, designed to reduce ammonia (NH₃) volatilization and total N loss from flooded rice fields, after application of urea, were assessed in the dry season at Mabitac, Philippines. The assessment was made in replicated 4.8-m × 5.2-m plots by determining NH₃ and total fertilizer N losses with bulk aerodynamic and ¹⁵N balance methods, respectively. The bulk aerodynamic method was first evaluated by comparison of ammoniacal-N concentrations, pH values, and equilibrium NH₃ concentrations for a 50-m diam, circular area and small plots receiving the same fertilizer treatment. Good agreement was obtained so the bulk aerodynamic method was used to determine NH₃ loss from the small plots receiving the different management treatments. Ammonia loss was high when urea was broadcast into floodwater 10 d after transplanting—48 and 56% of the applied N was volatilized as NH₃ during the first 8 d after application of 53 and 80 kg N ha⁻¹, respectively. Total N losses were 60 and 59%, respectively. Ammonia loss was reduced slightly (to a mean of 43% of the applied N for the two application rates) by broadcasting the fertilizer into 0.05-m-deep floodwater and incorporating it into the soil by harrowing before transplanting. Removal of the floodwater before applying urea onto saturated soil and incorporation by harrowing reduced NH₃ loss to a mean of 9% and total fertilizer N loss to a mean of 33%. Fertilizer N recovery by rice plants was significantly increased by incorporation without surface water. Contribution from IRRI and CSIRO.