A study to characterize particulate matter emissions from 195 in-use gasoline and diesel passenger vehicles was conducted during the summer of 1996 and the winter of 1997 in the Denver, Colorado region. Vehicles were tested as received on chassis dynamometers using the Federal Test Procedure (FTP) Urban Dynamometer Driving Schedule (UDDS). Both PM-10 and regulated emissions were measured for each phase of the UDDS. Approximately 88% of the PM-10 collected was carbonaceous material, of which the average organic fraction was 0.7 for gasoline vehicles and 0.4 for diesel vehicles. This suggests that the organic carbon (OC) to elemental carbon (EC) split may be useful in separating light-duty gasoline from diesel PM emissions. Sulfate emission rates averaged 0.45 and 3.51 mg/mi for gasoline and diesel vehicles, indicating that the EPA's mobile emissions model overpredicts sulfate emission rates. Elements identified by X-ray fluorescence averaged between 3 and 9% of the PM-10 mass. Polynuclear aromatic hydrocarbon (PAH) profiles developed may help distinguish between gasoline and diesel vehicles in source apportion ment studies. Total PAH emissions, however, were not a good candidate as a tracer of gasoline PM emissions. Hopane and sterane emissions were very similar across the fleet and may be useful tracers for mobile source PM emissions. Overall, emission rates varied significantly with vehicle classification and driving condition, suggesting that a single profile representing the entire fleet will need to carefully reflect the local fleet composition and the local weighting of cold, hot, and hot-stabilized emissions.