Rates of loss of trifluralin (α,α,α-trifluoro-2,6-dinitro-N,N-dipropyl-p-toluidine), fluchloralin [N-(2-chloroethyl)-2,6-dinitro-N-propyl-4-(trifluoromethyl)aniline], and pendimethalin [N-(1-ethylpropyl)-3,4-dimethyl-2,6-dinitrobenzenamine] from soil surfaces were evaluated in a series of field sampling studies. Herbicides were sprayed onto soil enclosed in rings with nylon mesh bottoms, exposed to field conditions, and analyzed periodically by gas chromatography. The herbicides were lost rapidly in the first 3 to 5 days from Bosket silt loam and Sharkey clay, although the initial rate of loss was less rapid in the clay. Trifluralin and fluchloralin were lost more rapidly than pendimethalin. Half-life values for the first 3 to 5 days for trifluralin, fluchloralin, and pendimethalin, respectively, were 2, 1.5, and 4 days on silt loam and 2, 4, and 6 days on clay. Rates of loss were much lower for the remainder of the 7- or 12-day sampling periods. Rainfall on the first day of the sampling period resulted in increased initial loss of all herbicides. Shading from direct sunlight greatly reduced the loss of fluchloralin and pendimethalin as compared to exposure to full sunlight, although trifluralin loss remained fairly rapid under shaded conditions. Simulated rainfall on the first day resulted in significantly increased dissipation of fluchloralin and pendimethalin.