Evaporative and running loss emissions were measured in a fleet of ten (1 989) current and seven (1983-85) older vehicles with fuels whose compositions varied in aromatic, olefin, and MTBE content and 90 percent distillation temperature (T9O). Emission compositions from each test were analyzed for individual hydrocarbon species. The individual hydrocarbon profiles were used to calculate evaporative and running loss emission reactivities using Carter maximum incremental reactivity (MIR) and maximum ozone reactivity (MOR) scales. Ozone reactivity estimates were expressed as Ozone Forming Potential (gO3/test) and Specific Reactivity (gO3/gNMOG) for both reactivity scales. The data were analyzed by regression analysis to estimate changes in the mass and reactivity of evaporative emissions due to changes in fuel composition.Previous studies have focused on how fuel volatility affects evaporative emissions without regard for the chemical composition of the fuels. By contrast in this study, fuels differed principally in their chemical composition, while volatility was more tightly constrained. Consequently, observed effects are analyzed with respect to chemical composition.Evaporative emissions were found to be well controlled in the two fleets. Loss mechanisms other than conventional vapor-liquid equilibria, however, were found to be influencing the diurnal and hot soak emission compositions in both fleets by significant amounts.Emission speciation provided evidence that permeation and/or fuel leakage contribute significantly to the mass of diurnal and hot soak emissions. Results further suggest contributions to evaporative emissions by these mechanisms are influenced by changes in fuel composition. These mechanisms become a relatively larger source of emissions at the low emission levels of the vehicles tested.Diurnal and hot soak evaporative emissions were reduced by 25 to 30% by lowering aromatics in the current and older fleets, and were increased by 30 to 40% by lowering T90 in the two fleets. Analysis of the data relative to the small variations in actual test fuel RVP showed current fleet hot soak emissions were reduced by 4% for each 0.1 psi reduction in RVP.In the current fleet lowering T90 increased diurnal and hot soak Ozone Forming Potential about 47% and 27% respectively. Lower aromatics reduced hot soak Ozone Forming Potential about 23%. Lowering olefins reduced both diurnal and hot soak Ozone Forming Potential by about 17%.Similar effects were found in the older fleet. Increases in Ozone Forming Potential due to T90 reduction were around 32% and reductions due to lowering olefins and aromatics ranged from 23 to 45% depending on the emission phase and reactivity scale.In the current fleet, diurnal and hot soak Specific Reactivities were reduced about 25% by lowering olefin content. Lowering T90 or adding MTBE reduced hot soak Specific Reactivity by smaller amounts.In the older fleet, lowering olefins reduced diurnal and hot soak Specific Reactivities by up to 26%. Lowering aromatics increased diurnal and hot soak Specific Reactivity, and adding 15% MTBE or lowering T90 reduced hot soak Specific Reactivity by smaller amounts.