The rotation effect on corn was evaluated between 1989 and 2004 in a large randomized block experiment located on a hillside near Lexington, Kentucky. Corn was grown after corn, full-season soybean, and wheat/double-crop soybean. The rotation effect was evident in most years. Corn after corn yields were generally inferior, while corn after wheat/double-crop soybean yields were generally superior. In absolute terms, the rotation effect was generally larger as the quality of the season, measured in terms of the average seasonal yield, increased. In relative terms, the yield ratios among the rotations were fairly constant. Spatial instability was often evident, as the rotation effect was not consistent up and down the hillside. Temporal instabilty, measured in terms of the annual coefficient of variation, increased as the average seasonal yield decreased, and was greater for the corn after corn sequence.