A number of hypotheses and conceptual models, particularly those emphasizing nonlinear dynamics and self-organization, postulate increases or decreases in complexity in the evolution of drainage basins, topography, soils, ecosystems, and other earth surface systems. Accordingly, it is important to determine under what circumstances and at what scales either trend might occur. This paper is concerned with changes in soil landscape complexity due to redistribution of sediment by fluvial, aeolian, and tillage processes at historical time scales in an agricultural field system near Grifton, North Carolina. Soil mapping and soil stratigraphic investigations were used to identify and map soil changes associated with erosion and deposition by water, wind, and tillage; reconstruct the pre-agricultural soil pattern; and identify transformations between soil types. The Kolmogorov entropy of the pre- and post- agricultural landscapes was then compared. The soil transformations associated with erosion and deposition created four distinct new soils and made possible new transformations among soil series, increasing the number of soil types from seven to 11 and the number of possible transformations from 14 to 22. However, the entropy and complexity of the soil landscape decreased, with associated increases in information and redundancy. The mass redistributions created a lower-entropy landscape by concentrating particular soils and soil transformations in specific landscape settings. This result is contrary to studies showing a trend toward increasing pedological complexity at comparable spatial scales, but over much longer time scales. These results point to the importance of temporal scale, and to the fact that environmental complexity is influenced by factors other than the number of different landscape units present.