Existing long-term cotton experiments established in 1976 with minimum and intensive herbicide programs were subdivided in 1985 and 1986 to determine the longevity of fluometuron and trifluralin soil residues after discontinuing herbicide application. In monoculture cotton, seed cotton yield increased when herbicide use was discontinued in 1985 and 1986 after 9 and 10 yr of continuous use, respectively, on a Sharkey silty clay and a Dundee silt loam soil. Yield increases coincided with reductions of fluometuron and trifluralin soil residues. Fluometuron dissipated from the Dundee silt loam soil by 10 mo after the last application but was present in the Sharkey silty clay soil at 0.20 ppmw after 28 mo. Trifluralin did not totally dissipate from either soil, and low levels were present in the Dundee (0.05 ppmw) and Sharkey (0.13 ppmw) soils 30 mo after the last application. Visual injury to fall-seeded wheat and vetch decreased as herbicide residues dissipated. Fall tillage had no significant effect on the rate of fluometuron or trifluralin dissipation from either soil but reduced seed cotton yields.