Information on soil properties that influence herbicide sorption throughout the soil profile is needed to identify soil-herbicide combinations with varying degrees of risk for groundwater contamination. In this study we investigated the effects of soil properties on the sorption of five herbicides commonly used in grain production in the Atlantic Coastal Plain (atrazine, cyanazine, dicamba, 2,4-D, and metolachlor), using subsoil and surface horizons from six Delaware soils. Our objective was to develop simple screening models, based on soil properties available from routine soil tests or regional soil databases, to identify soils that may have an increased risk of groundwater pollution. Each soil was sampled by horizon to a depth of approximately 2 m to include major variations present in the soil profile above the surface water table. Herbicide distribution coefficients (Kdm) were determined for each horizon and herbicide by sorption studies using a batch equilibrium technique and14C-labeled herbicides. The ranking of herbicide sorption, over all soil horizons, was metolachlor > cyanazine > atrazine > 2,4-D > dicamba (Kdm= 1.03, 0.94, 0.85, 0.65, and 0.25 L kg-1, respectively). Multiple regression analysis between Kdmand soil properties identified two rapid soil tests (KCl exchangeable acidity and organic matter content) as important predictive variables for atrazine, cyanazine, dicamba, and 2,4-D sorption (R2= 0.81***, 0.56***, 0.26*, 0.68***); metolachlor retention was best described by oganic matter content and effective cation exchange capacity (R2= 0.78***), another easily measured or estimated soil property.