It is essential to establish more accurate N balances for different soil-plant systems in order to improve N use efficiency. In this study the N balance was studied in a poorly drained clayey loam soil under natural grassland supplied with either calcium ammonium nitrate or cattle slurry at two application rates. The aim was to determine the efficiency of the N applied and the factors which affect this efficiency. Mineralization-immobilization of N was calculated by balance between the quantified inputs and outputs of N. As N inputs increased, output via herbage yield was accompanied by an increase in apparent immobilization of N in the soil and by larger losses of N by denitrification. The difference between cattle slurry and N fertilizer was that the slurry behaved as a slow release fertilizer, its supply of mineral N being greater in the periods of time when fertilizer was applied a long time ago. Denitrification losses (up to 17% of the N applied) are suggested to be the main factor to mitigate in order to increase N use efficiency. A decrease in net mineralization (up to 136 kg N ha-1 year-1) was observed which was related to the mineral N application rate. There was evidence to suggest that this decrease was due both to the immobilization of the N applied and to a decrease in the rate of gross mineralization when mineral N was applied. Microbial biomass determinations could not explain the changes in the mineralization-immobilization equilibrium of N because of the great coefficients of variation for this determination (mean value of 18%). Nevertheless, it contributed to verify and explain some of the changes observed in this equilibrium