Two levels of sewage sludge (10 and 100 tons DM ha-1) were mixed in the 0-20 cm layer of an oxisol in Guadeloupe (French West Indies) (table I). A field experiment was conducted with microplots (20 m2) during a 500-d test period: control without sewage sludge (T), (10 T) and (100 T). After soil sampling, sewage sludge clods over 5 mm diameter were hand-sorted from soil samples. The evolution of various soil phosphorus forms total (P(t)), organic (P(o)), mineral (P(m)) and extracted with bicarbonate-fluoride (P(NaHCO3-F) was studied. In the 10 t treatment sludge clods disappeared within 110 d, and the total soil phosphorus content significantly increased 3 months after buying up to 600-650-mu-g.g-1. In the 100 t treatment, sewage clods were still observed after 500 d, and total soil phosphorus showed a 3-fold increase (1 400-mu-g.g-1) (figs 1, 2). Kinetics of phosphorus forms showed that i) the application of 10 t.ha-1 increased mineral phosphorus and bicarbonate-fluoride extract, organic phosphorus content showed a lower increase (fig 3); ii) 100 t.ha-1 had the same qualitative effects, but the various phosphorus forms increased 3-fold with respect to the control, phosphorus in the bicarbonate-fluoride extract rapidly increased and remained over 300-mu-g.g-1 up to 500 d (fig 4). Application of 100 t.ha-1 of sewage sludge may be recommended to build up phosphorus reserves in highly degraded soils. An input of 10 t.ha-1 seems to be sufficient to fertilize short cycle crops as it significantly increased phosphorus of the bicarbonate-fluoride extract.