Reclamation of Guadalquivir river marshes (SW Spain) constitutes a representative example of wetland reclamation in Southern Europe. Nowadays, this is an important area of tile-drained soils (40,000 ha) with an intensive irrigated agricultural production where high fertilizer rates are usually applied. In tile-drained soils, flow through macropores or cracks, which connect the nutrient rich topsoil with drain lines, can be an important pathway for nutrient transfer from soil. In order to study P loss in these soils and how it is affected by soil amendment usually applied in the zone (phosphogypsum and manure) an experiment was performed during two consecutive growing seasons on a reclaimed marsh soil from the Guadalquivir Valley. In the first season (1998)(1999), sugar beet (Beta vulgaris L.) was grown under sprinkler irrigation at a rate of 2.5 mm h À1 ; in the second (2000), cotton (Gossypium hirsutum L.) was grown under furrow irrigation at 8-10 mm h À1 . The amendments applied included manure (30 Mg ha À1 ), and phosphogypsum (13 and 26 Mg ha À1 ). Drainage events were recorded, and water samples collected and analyzed for total P (TP), dissolved total P (DTP), and dissolved reactive P (DRP). Total P in drainflow ranged from 0 to 0.818 mg l À1 in the 1998-1999 season and from 0 to 0.565 mg l À1 in the 2000 season. The major P form in drainflow was DRP, which accounted for about 50% of TP in the two growing seasons (the mean DRP concentration was 0.068 mg l À1 in 1998-1999 and 0.043 mg l À1 in 2000). Dissolved organic P accounted for a higher portion of DTP in the first season (37%) than in the second (13%). A larger load of phosphorus was observed on plots receiving manure. This treatment significantly increased (P < 0.05) the cumulative drainflow during the 1998-1999 growing season (sprinkler irrigation, low drainflow rates). This is consistent with the increased losses of TP, DTP, DAHP, and DRP resulting from this treatment in this growing season. In the following season, DTP loading were significantly increased by manure (P < 0.05). This seems to be related mainly to significantly increased DOP losses (P < 0.01), particularly during the first drainage event. The higher fraction of applied water was lost by drainage under furrow irrigation (high drainflow rates) is consistent with the high TP load during the 2000 growing season (199-285 g ha À1 ) relative to the 1998-1999 season (20-59 g ha À1 ). This difference in P losses was much greater than those resulting from amendment of the soil.Abbreviations: DAHP -dissolved acid hydrolyzable P; DOP -dissolved organic P; DRP -dissolved reactive P; DTP -dissolved total P; PG -phosphogypsum; PP -particulate P; TP -total P