Streamflow, precipitation, and suspended-sediment data were collected from two small agricultural basins in the Piedmont province of North Carolina. The data were used to determine the effects of land-management practices on sediment yield. One basin of 7.4 acres represents best land-management practices with strip cropping, crop rotation, contour farming, and grassed waterways. The other basin of 4.8 acres represents standard land-management practices with down-slope row orientation, unmaintained grassed waterways, and without crop rotation. Data collected during the 1985-87 water years were used to develop regression equations to describe the relation between suspended-sediment discharge and water discharge. Data sets consisting of suspended-sediment concentrations and corresponding instantaneous water-discharge data were developed. There were two data sets from each basin, one representing data collected during the growing season, May through September, and the other representing data collected during the nongrowing season, October through April. Four regression equations were developed, one for each data set, and were tested for goodness-of-fit by use of graphical analysis, influence diagnostics, significance tests, and residuals analysis. Following acceptance of the four equations, the slope of each individual line was tested to determine if season was a significant variable. Seasonally, the average sediment yields (2.7 tons per acre) from the basin having best land-management practices were only about one-seventh of those (20 tons per acre) from the basin having standard land-management practices. Comparison of annual sediment yields in the agricultural basins against the yield from a nearby forested basin, which represents a nearly undisturbed basin, indicates a 10- to 100-fold increase in sediment yields from the study basins. The forested basin sediment yield was 0.1 ton per acre in the 1987 water year. Sheet erosion, which represents soil moving from high areas to low areas in a basin but not necessarily reaching a stream, is also influenced by land-management practices. This type of erosion was estimated to be 3.6 tons per acre per year for the basin with best land-management practices and 45 tons per acre per year for the basin with standard land-management practices.