In a series of experiments, 15N-labelled nitrate fertilizer was used to study the nitrogen balance of a heavy clay soil. Two of the experiments employed 60-cm deep, 16-cm diameter columns of packed soil. In the first, no added 15N was lost until heavy rain saturated the soil in the undrained columns. Subsequently, fallow and planted columns lost 25% of the 15N added. While denitrification appeared primarily responsible for these losses, some of the 15N may have been lost by weathering of the plants during senescence. This experiment was repeated with precautions to prevent excessive rain saturating the soil. After 16 weeks, only 6 ± 4.5% of the added 15N was not accounted for by soil and plant analysis. In the third experiment, I5N was added to 16-cm diameter confined microplots in situ. The microplots were fallow and were confined by 60-cm deep plastic pipe. After 16 weeks, during which 190 mm of rain fell, 97.7 ± 2.4% of the added 15N was recovered.