The accumulation of mineral-nitrogen (N) in the top 10 cm of soil during the summer fallow was measured in 2 replicated field experiments following a range of crops including wheat, oats, canola, peas and lupins. At the first site, mineral-N was measured following harvest and in autumn before sowing subsequent crops across 3 seasons (1994–96). Crop residues were retained on the surface with intermittent grazing by sheep throughout the summer fallow and burnt before the autumn measurements. The smallest increase in mineral-N accumulation occurred following the cereals in all 3 seasons (mean increase 31 kg/ha). The highest accumulation of mineral-N in all seasons occurred following canola (mean 94 kg/ha), 3 times as much as that following cereals, and significantly higher than that after the legumes in 2 of the 3 seasons (mean 50 kg/ha). Differences in the amount, N content, or C : N ratio of the surface-retained crop residues are unlikely explanations for the observed differences in mineral-N accumulation. At a second site, measurements of the accumulation of mineral-N following canola and wheat were accompanied by measurements of populations of selected microorganisms involved with N cycling in soil. More mineral-N accumulated after canola than after wheat, however, populations of free-living, N-fixing bacteria, potential Azospirillim species and NH4+ oxidising bacteria were significantly lower following canola than following wheat, and populations of total bacteria and NO2− oxidising bacteria did not differ. These results suggest that greater mineral-N accumulation following canola does not result from a shift in those microbial populations which favour mineral-N accumulation, however, more detailed studies are required to resolve the exact cause of the differences. A possible explanation is that biocidal compounds released by canola roots during decay may cause a general ‘biofumigation’ and thereby result in a flush of mineral-N similar to that which accompanies chemical fumigation.