A system of plots designed to conform to the land was in operation on the Wheatland Conservation Experiment Station, Cherokee, Oklahoma, from 1942 through 1951. The arrangement permitted the use of moderately sloping land for plot sizes approaching field conditions. The experiment was designed to compare the effectiveness of different seedbed preparations for the conservation of runoff water for plant growth. The tillage practices conducted on both terraced and unterraced land in a factorial arrangement of plots were plowing, listing, basin listing, and stubble mulch tillage. This report evaluates the experimental technique of such a system of plots and determines the variability experienced and differences assessable from measurements of runoff, soil moisture, and crop yields. Runoff was measured with type H flumes and Friez FW-1 waterstage recorders. Amounts and intensities of rainfall were measured by Ferguson recording rain gages and checked by standard gages at 10 points located systematically on the station. Annual yields of winter wheat and soil moisture to a depth of 3 feet at planting and in the spring were measured. The system for measuring runoff water has been satisfactory. For the 10 years of data, the variability in all quantities measured was high. Nevertheless, significant differences were obtained on a site typical of a large area where soils and slopes were highly variable. The amount of runoff water from the terraced and contour cultivated land was significantly less than that from plots cultivated with the slope. When the four methods of seedbed preparation were compared, stubble mulching significantly reduced the yield of wheat.