A 2-year water quality monitoring project was conducted to evaluate the removal of storm water contaminants by existing vegetated slopes adjacent to freeways. Objectives of the study were to generate design criteria and to determine whether standard roadway design requirements result in buffer strips with treatment equivalent to those specifically engineered for water quality performance. Variables such as width, slope, vegetation density, and hydraulic loading were evaluated by studying the runoff through existing vegetated slopes at four locations in northern California and four locations in southern California. At each location, concrete channels were constructed to capture freeway runoff after it passed through existing vegetated strips of varying widths. The quantity and quality of the runoff discharged from the buffer strip were compared with freeway runoff collected at the edge of pavement. Buffer strips consistently reduced the concentration of suspended solids and total metals in storm water runoff. The strips were also effective in removing dissolved metals when the edge of pavement concentrations were sufficiently high. Little or no change in concentration was observed for nitrogen and phosphorus. Concentrations of organic carbon, dissolved solids, and hardness increased. For the constituents exhibiting a decrease in concentration, steady-state levels were generally achieved within 5 m (16 ft) of the pavement edge for slopes commonly found on highway shoulders and where the vegetation coverage exceeded 80%. Slope, vegetation type and height, highway width, and hydraulic residence time had little or no impact on the discharge concentrations.