Excessive nitrate leaching from the U.S. Corn Belt has created serious water quality problems and contributed to the expansion of the hypoxic zone in the Gulf of Mexico. We evaluated the effect of implementing the late spring nitrate test (LSNT) for corn (Zea mays L.) grown within a 400-ha, tile-drained subbasin in central Iowa. Surface water discharge and NO3 concentrations from the treated subbasin and two adjacent subbasins receiving primarily fall-applied, anhydrous ammonia were compared. In two of four years, the LSNT method significantly reduced N fertilizer applications compared with the farmers' standard practices. Average corn yield from LSNT fields and nonlimiting N fertilizer check strips was not significantly different. Autoregressive (AR) models using weekly time series in surface water NO3 concentration differences between the LSNT and control subbasins indicated no consistent significant differences during the pre-LSNT (1992-1996) period. However, by the second year (1998) of the treatment period (1997-2000), NO3 concentrations in surface water from the treated subbasin were significantly lower than the concentrations coming from both control basins. Annual average flow-weighted NO3 concentrations for the last two years (1999-2000) were 11.3 mg N L(-1) for the LSNT and subbasin and 16.0 mg N L(-1) for the control subbasins. Based on these values and the AR models, widespread adoption of the LSNT program for managing N fertilizer where fall N application is typically practiced could result in a > or = 30% decrease for NO3 concentrations in surface water.