The General Motors sulfate dispersion experiment was conducted in October 1975 at the General Motors Milford Proving Ground. The experiment simulated a four-lane freeway; 352 catalyst-equipped automobiles were driven at 80 km/h, resulting in a traffic volume of 5462 vehicles/h. The runs were conducted in the morning to obtain the most adverse conditions for pollutant dispersion. The maximum catalyst sulfate exposure near the roadway averaged 8 ug/cu m for sixty-six o. 5-h runs. The average sulfate emission rate for each vehicle was 0.023 g/km. Near the roadway, mechanical mixing due to the traffic dominated the mixing caused by the ambient turbulence. At low cross-road winds, plume rise becomes important. The US Environmental Protection Agency's HIWAY model was found to overestimate the concentrations at the pedestriat level under stable conditions. These overestimates become worse as the wind speed decreases, as the wind direction approaches parallel to the road, and as the distance from the road increases. A simple line-source model was constructed to remedy many of the limitations low cross-road winds. It also avoids a cumbersome numerical integration required in the HIWAY model. An advection-diffusion model was also constructed in which the eddy diffusivity was determined from dynamic considerations. The influence of traffic was approximated by an additive component in the diffusivity tensor. Good agreements with observations were found, even when the off-diagonal terms of the diffusivity tensor were neglected. It is also expected that when the vehicle velocity is reduced, the extent of pollutant dispersion would also be reduced./Author/