The effect of N-15 labelling duration (24 h, 7 and 31 months) on the incorporation and extractability of residual N-15 in the humus (H) horizon of a Humo-Ferric Podzol was investigated. Extractability of residual N-15 was studied using 2 M KCl, 0.5 M K2SO4, autoclaving with 10 mM CaCl2, and acidic permanganate and fumigation-extraction. Incorporation of N-15 into the classical fulvic acid (FA), humic acid (HA) and humin fractions was also studied. Greater amounts of total and applied (fertilizer) N were extracted from the 24 h than from the 7 and 31 months treatments (P < 0.05), with the difference between the last two non-significant. The KMnO4 strength had no effect on the amount of N extracted. The extracted fractions were always enriched with N-15 relative to the bulk soil regardless of treatment. Among the extraction methods used, the autoclaving method extracted the greatest amount of total and applied N, but had the lowest extractability ratio (ER) for each treatment. The fumigation- and KCl-extraction methods were more desirable in obtaining biologically meaningful N fractions. Nitrogen-15 enrichment in microbial biomass was found similar to that in the inorganic N fractions extracted by KCl and K2SO4 for each labelling treatment. For the FA fraction, a greater percentage of applied N was recovered in the 24 h (52%) than in the 7 and 31 months treatments (3 and 2%, respectively), while less applied N was recovered in the 24 h (47%) than in the other two treatments (96 and 96%, respectively) for the humin fraction. ER values increased with shorter labelling duration and were always greater than 1 for the FA fraction, and less than 1 for the humin fraction; These results show that the extractability of residual N-15 was quickly reduced due to its incorporation into stable soil humin fraction. However, residual N-15 was more extractable than the bulk soil N regardless of the labelling duration used in this study. (C) 1998 Elsevier Science Ltd. All rights reserved.