The amounts and forms of dissolved phosphorus (P) leached from undisturbed and cultivated soil in lysimeters were measured from sites that had received N fertiliser at 0, 200 or 400 kg N ha -1 yr -1 for 4 years. Soil in mini-lysimeters, 18 cm i.d. by 25 cm depth, was leached at monthly intervals over autumn and winter (500 mm total drainage) to simulate a relatively wet autumnwinter for eastern Southland (typical drainage, c. 350 mm yr -1 ). The major form of P leached was dissolved organic P (DOP), which represented 67 and 80% of total dissolved P (TDP) losses in the undisturbed and cultivated soil treatments, respectively. Historical N inputs had little effect on the amounts of dissolved reactive P (DRP), DOP or TDP leached from the undisturbed soils. However, DRP, and to a lesser extent TDP, losses decreased as historical N fertiliser inputs increased in the cultivated soil treatment. Significantly greater losses of DRP were observed from undisturbed compared with cultivated soils. The lower DRP losses in cultivated soil were attributed to greater sorption of P during matrix flow. These results suggest that, although DRP losses are of little concern following cultivation of pastoral soils, the greater mobility and loss of DOP in drainage, in both undisturbed and cultivated soils, represents a potential risk to water quality and loss of P from the plant rooting zone. Further study to examine the bioavailability of DOP is warranted to determine its potential risk to surface water.