Liquefied petroleum gas (LPG) is increasingly being examined as an alternative to gasoline use in automobiles as interest grows in reducing air pollutant emissions. In this study, emissions of regulated (CO, THC, NO(x)) and unregulated air pollutants, including CO2, particulate matter (PM), polycyclic aromatic hydrocarbons (PAHs), and BTEX (acronym for benzene, toluene, ethylbenzene, xylene), were measured before and after conversion of nine gasoline-powered automobiles to LPG/ gasoline dual-fuel retrofits. The tests were conducted on a standard chassis dynamometer in accordance with the United States Environmental Protection Agency FTP-75 test procedure, with the exception that all tests were conducted under hot-start driving conditions. The influences of LPG on air pollutant emission levels and carcinogenic potency were investigated and compared with gasoline. The results showed average emission factors of 0.14 g/km, 0.33 mg/km, 0.09 g/km, 0.44 g/km, and 197 g/km for CO, THC, NO(x), PM, and CO2, respectively, for LPG/ gasoline dual-fuel retrofits. Paired-sample t-test results indicated that the emissions of CO (p = 0.03), THC (p = 0.04), and CO2 (p = 4.6 x 10(-8)) were significantly reduced with the retrofit in comparison with gasoline-powered automobiles. The reduction percentages were 71%, 89%, and 14% for CO, THC, and CO2, respectively. The average total PAH emission factor for LPG was 217 microg/km, which is significantly lower than gasoline (863 microg/km; p = 0.05). The PAH corresponding carcinogenicities (BaP(eq)) were calculated via toxic equivalencies based on benzo(a)pyrene (BaP). Paired-sample t-test results fortotal BaP(eq) emissions showed no significant difference between gasoline (30.0 microg/km) and LPG (24.8 microg/km) at a confidence level of 95%. The discrepancy between PAH and BaP(eq) emissions resulted from the higher emission percentages of high molecular weight PAHs for LPG, which might be from lubricant oil. The average emission factors of benzene, toluene, ethylbenzene, and xylene were 351, 4400, 324, and 1100 microg/ km, respectively, with LPG as fuel, which were all significantly lower than those for gasoline (95% confidence level). The average reduction percentages were 78%, 61%, 57%, and 58% for benzene, toluene, ethylbenzene, and xylene, respectively.