Controversy remains about the importance of nonlinear sorption isotherms, desorption rate limitations, and aging effects, collectively referred to as nonideal sorption processes, in controlling the fate and transport of organic contaminants. Herbicide runoff from highway soils represents a good test case for assessing the relative importance of nonideal sorption because runoff flow rates are often high, soil-water contacttimes are short, and significant time is available for contaminant aging after application. This study examines the sorption and desorption of five herbicides with a wide range of properties (isoxaben, oryzalin, diuron, clopyralid, and glyphosate) on soil samples from two roadsides in northern California and uses the results to examine field runoff data from multiple rainy seasons. Nonideal sorption processes do not appear to be significant in determining herbicide runoff at the field sites because (i) sorption isotherms were linear or slightly nonlinear for all compounds but glyphosate, (ii) field runoff concentration ratios between isoxaben and oryzalin were consistent with linear partitioning predictions, (iii) runoff leaving the site appeared to be in equilibrium with local soil concentrations, and (iv) desorption distribution coefficients for aged herbicides on soil samples collected from the field site did not differ substantially from those obtained in short-term laboratory adsorption experiments. Collectively, these findings indicate that linear equilibrium models are adequate for predicting the concentration of herbicides in runoff in these field settings and that more complicated nonideal models do not need to be invoked. Vegetated slopes effectively reduced the herbicide loads, with average removals of 35-80% occurring as runoff traversed a 3-m segment 1 m from the edge of the spray zone.