Beijing is a city with some of the world's worst particulate air pollution. Although there have been various control strategies implemented since 1998, there are still episodes of PM2.5 concentrations of hundreds of micrograms per cubic meter. In this study, samples were collected over a year in Beijing, chemically characterized, and the resulting data analyzed for source apportionment. The new error analysis capabilities built into EPA PMF V5.0 have been employed to better evaluate the profiles and assign them to source types. Secondary sulfate, local coal combustion and secondary nitrate were the major contributors to the PM2.5 mass. However, in this study, traffic was found to be more important as a PM compared to prior studies. It was actually the largest PM2.5 source in autumn and winter although local coal combustion is also a large source of PM in the winter months. These results demonstrate the value of using the displacement method to assess the variability in source profiles to improve our interpretation of PMF results. They also suggest more attention needs to be paid to traffic emissions in Beijing. (C) 2016 Elsevier Ltd. All rights reserved.