The fate of the added N on a sandy loam soil was determined in an improved fallow - maize sequence field experiment in Zimbabwe. Pre-season mineral N was determined in 20 cm sections to 120 cm depth by soil auguring in seven land use systems. Thereafter, sequential soil auguring was done at two-week intervals in plots that previously had 2-year fallows of Acacia angustissima, Sesbania sesban and unfertilized maize to determine mineral N dynamics. Using the static chamber technique, N2O fluxes were also determined in the same plots. Pre-season NH4-N concentrations were >12 kg N ha−1 in the 0–20 cm layer for treatments that had a pronounced litter layer. NO3-N concentrations below 60 cm depth were <3 kg N ha−1 layer−1 for Sesban, Acacia, Cajanus cajan and natural woodland compared with >10 kg N ha−1 layer−1 in the control plots where maize had been cultivated each year. There was a flush of NO3-N in the Sesbania and Acacia plots with the first rains. Topsoil NO3-N had increased to >29 kg N ha−1 by the time of establishing the maize crop. This increase in NO3-N in the topsoil was not sustained as concentrations decreased rapidly within three weeks of maize planting, to amounts of 8.6 kg N ha−1 and 11.2 kg N ha−1 for the Sesbania and Acacia plots, respectively. Total NO3-N leaching losses from the 0–40 cm layer ranged from 29–40 kg ha−1 for Sesbania and Acacia plots within two weeks when 104 mm rainfall was received to an already fully recharged soil profile. Nitrate then accumulated below the 40 cm depth during early season when the maize had not developed a sufficient root length density to effectively capture nutrients. At one week after planting maize, N2O fluxes of 12.3 g N2O-N ha−1 day−1 from Sesbania plots were about twice as high as those from Acacia, and about seven times the 1.6 g N2O-N ha−1 day−1 from maize monoculture. This was at the time when mineral N was at its peak in the topsoil. The unfertilized maize showed consistently low N2O emissions, which never exceeded 2 g N2O-N ha−1 day−1 for all the eight sampling dates. The decrease of mineral N concentration in the topsoil resulted in reduced N2O fluxes, despite very high soil moisture conditions. Total N2O-N emissions were greatest for Sesbania plots with 0.3 kg ha−1 lost in 56 days. We conclude that, under high rainfall conditions, there is an inherent problem in managing mineral N originating from mineralization of organic materials as it accumulates at the onset of rains, and is susceptible to leaching before the crop root system develops. We did not quantify nitric oxide and N2 gas emissions, but it is unlikely that total gaseous N losses would be significant and contribute to poor N recovery that has been widely reported.