About 42 000 ha of clonal Eucalyptus plantations have been established since 1978 on coastal savannas of the Congo. A short rotation silviculture is carried out in these plantations with high inputs (fertilization, weeding) and harvests every 7 years. Although soils are Ferralic Arenosols with very low reserves of total N and exchangeable cations, the MAI is around 15 m3 ha-1 y-1 under-bark. Field trials point out that nitrogen is the main factor limiting the growth of eucalypt plantations and that inputs of N by fertilization have to increase over successive rotations. The in-situ mineralization of N was studied in 3 sites located in a 500 m radius, on a plateau: a native savanna, a planted crop of eucalypt at the end of the rotation (age 7-8 years), and a replantation of the same clone of eucalypt after 1 rotation. An experimental design was set up in the replanted site to quantify the impact of the slash management on the soil properties and the growth of the stand. The treatments were: R, all above-ground organic residues removed from the plot; SH, stemwood only removed and DS, double slash treatment. A classical methodology of incubation in soil cores gave an index of N mineralization, useful to compare situations. Soil cores were contained in situ in 12 PVC tubes with anionic or cationic resins at the bottom. Soil was incubated (ca 14 days), sequentially sampled and mineral N measured. The experiment lasted one year in the savanna and the planted crop, and 2 years in the replanted site. The annual production of mineral N in the top soil (0-13 cm layer) of the savanna and the planted crop was 21 and 33 kg ha-1 y-1 respectively. The rate of net nitrification was much lower in the savanna than in the eucalypt stands (5% vs >77%). This low rate of nitrification in the savanna might be a result of an inhibition by the species Loudetia arundinacea of the micro-organisms responsible for nitrification. Planting eucalypts might remove this inhibition. In the repla