This paper addresses the impact of California phase 2 reformulated gasoline (RFG) on motor vehicle emissions. Phase 2 RFG was introduced in the San Francisco Bay Area in the first half of 1996, resulting in large changes to gasoline composition. Oxygen content increased from 0.2 to 2.0 wt%; and alkene, aromatic, benzene, and sulfur contents decreased. Gasoline density and T50 and T90 distillation temperatures also decreased. Light-duty vehicle emission rates were measured in a Bay Area roadway tunnel in summers 1994−1997. Vehicle speeds and driving conditions inside the tunnel were similar each year. The average model year of the vehicle fleet was about one year newer each successive summer. Large reductions in pollutant emissions were measured in the tunnel over the course of this study, due to a combination of RFG and fleet turnover effects. Between summers 1994 and 1997, emissions of carbon monoxide decreased by 31 ± 5%, non-methane volatile organic compounds (VOC) decreased by 43 ± 8%, and nitrogen oxides (NOx) decreased by 18 ± 4%. It was difficult to separate clearly the fleet turnover and RFG contributions to these changes. Nevertheless, it was clear that the effect of RFG was greater for VOC than for NOx. The RFG effect on vehicle emissions of benzene was estimated to be a 30−40% reduction. Use of RFG increased formaldehyde emissions by about 10%, while acetaldehyde emissions did not change significantly. RFG effects reported here may not be the same for other driving conditions or for other vehicle fleets. RFG effects on evaporative emissions are also important. The combined effect of phases 1 and 2 of California's RFG program was a 20% reduction in gasoline vapor pressure, about one-fifth of which occurred following the introduction of phase 2 RFG.