Nonpoint source pollution associated with runoff from agricultural fields continues to be a problem in many areas of the Midwest U.S. due to excessive fertilizer application, runoff prone soils, and climatic conditions that produce large thunderstorms during and after spring planting and fertilizer application.  The objective of our research was to evaluate the influence of three cropping systems on dissolved nitrogen (N) and phosphorus (P) concentrations and losses in surface runoff at the field scale. The study was conducted from 1993 through 2001 on three watersheds ranging in size from 7.3 to 34.4 ha located in north-central Missouri.  Soils are primarily Vertic Epiaqualfs and are poorly drained because of a naturally occurring claypan soil horizon located 150 to 450 mm below the soil surface.  The cropping systems evaluated were a conservation tillage corn-soybean rotation with high chemical inputs (CS1), a no-till corn-soybean rotation with medium chemical inputs (CS2), and an adaptive cropping, tillage, and chemical input system that included a corn-soybean-wheat rotation (1993 through 1996) and a no-till corn-soybean rotation with variable rate N application (CS3).  The watersheds were instrumented to measure surface runoff and collect samples for water quality analysis.  Mean annual runoff losses were 275, 256, and 216 mm for CS1, CS2, and CS3, respectively.   Mean annual NO3-N, NH4-N, and PO4-P concentrations were < 5, 0.7, and 0.3 mg/L for the corn years of the rotation and < 2, 0.1, and 0.2 mg/L for the soybean years.  Annual NO3-N, NH4-N, and PO4-P losses ranged from 0.2 to 15.8, 0.02 to 3.2, and 0.08 to 1.2 kg/ha.   No-till did not decrease surface runoff relative to conservation tillage.  Fertilizer incorporation did not reduce dissolved N concentrations and losses relative to no-till.  Dissolved N concentrations and losses were much higher from CS3 than CS2, even for soybean cropping.