ABSTRACT THE long term effects of sediment redistribution due to the erosion-deposition process on soil properties and crop yield in a small watershed (1.6 ha, Udertic Paleustoll soils, 3% to 4% slope) were evaluated by an analysis of two-dimensional spatial trends, correlation structure of variability, and the interrelationship of selected topsoil properties and grain sorghum dry weight. Measurements were made in a 10 m, two dimensional grid of 108 points, which were supplemented by measurements of both top- and sub-soil properties from a 21 point line transect down the center of the watershed. Topsoil properties measured on the grid were texture, bulk density, and soil water storage after a rainfall early in the growing season. Crop dry weight, soil bulk density, and soil water storage were measured for two growing seasons. Soil properties measured on the line transect were texture, bulk density, macroporosity, organic matter, volumetric soil water storage at different suctions, pH, and available phosphorus content. Among the soil properties measured on the grid, sand content had the highest coefficient of variation, followed by clay content and soil water storage. The coefficient of variation of grain sorghum dry weight was greater than those of the soil properties. There were significant two-dimensional, curvilinear, trends for all properties, which were represented by least-square parabolic surfaces. Although silt and clay contents increased, sand content decreased downslope. Soil bulk density and water storage also increased downslope in 1983, but in 1984 the trend for bulk density was reversed and the trend for water storage was more complex. Grain sorghum dry weight in both years increased downslope, and was higher on both sides than in the middle of the field. A semivariogram was calculated for each property after the data were detrended using the parabolic fits. All variograms had a large nugget effect. The variograms for texture components showed a weak correlation structure, with a range between 30 m and 40 m. In both years, the semivariograms for soil bulk density and water storage did not clearly indicate a spatial structure. Also, there was no apparent spatial structure for dry weight, plant stand, or dry weight per plant in the 1983 variograms; while a weak structure was present in the 1984 variograms, with the range nearly the same as that for the soil texture (30 m to 40 m). Texture and available water in the top- and sub-soil explained a large percentage of the trend (increase) and variation of grain sorghum dry weight downslope. In all cases, the dry weight was correlated positively with silt content and negatively with sand content. The topsoil properties were relatively more important for yield under suboptimal growing conditions, while the subsoil was a limiting factor for yield under normal conditions.