This study presents the source apportionment of PM_2.5 performed by positive matrix factorization (PMF) at urban and rural sites in Beijing. These factors are interpreted as traffic emissions, biomass burning, road and soil dust, coal and oil combustion, and secondary inorganics. PMF failed to resolve some sources identified by CMB and AMS and appears to overestimate the dust sources. Comparison with earlier PMF studies from the Beijing area highlights inconsistent findings using this method.