Most studies of phosphorus (P) movement in soil have based their conclusions on patterns of extractable soil P as a function of depth, which has led to the assumption that no substantial leaching loss occurs because of high P-fixation capacity in mineral soils. Few studies have involved high-quality leachate samples collected below the root zone; rather, most have involved tile drainage systems. Equilibrium-tension lysimeters installed at a depth of 1.4 m were used to evaluate and compare P leaching from a restored tallgrass prairie and corn (Zea mays L.) agroecosystems on Plano silt loam soil (fine-silty, mixed, superactive, mesic Typic Argiudoll) in southcentral Wisconsin during a 5-yr period. The corn agroecosystem treatments included nitrogen (N)-fertilized (f) or N-unfertilized (nf) and no-tillage (NT) or chisel-plowed (CP). Mean volume-weighted molybdate-reactive phosphorus (MRP) and total dissolved phosphorus (TDP) concentrations were similar within replicate samples, but always higher in NTf corn than in the prairie or CPf corn systems, though drainage from the CPf corn was always higher than from the NTf corn system. Water-extractable soil P concentrations at any given depth were not positively correlated with leachate concentrations, suggesting that macropore flow causes infiltrating runoff to preferentially bypass the bulk of the soil matrix. Leachate-P concentrations from the natural and managed agroecosystems exceeded 0.01 mg P L(-1) and leaching losses were significantly higher from N-fertilized corn, regardless of tillage, than from the prairie or N-unfertilized corn systems, from which leachate-P concentrations and loads were similar. Increased root growth from N fertilization could cause more macropore formation, preferential flow, and P mineralization from decaying roots compared with N-unfertilized systems, which could contribute to a N-fertilization effect on P leaching.