Monitoring the air quality in subway passenger cabins is important because of the large number of passengers and potentially high levels of air pollution. This report characterized PM 10 levels in subway cabins in Seoul, Korea, and identified PM 10 sources using elemental analysis and receptor modeling. PM 10 levels in subway cabins were continuously measured using a light scattering monitor during rush and non-rush hours. A total of 41 measurements were taken during rush and non-rush hours, and the measurements were repeated in all four seasons. Filter samples were also collected for elemental composition analysis. Major PM 10 sources were identified using positive matrix factorization (PMF). The in-cabin PM 10 concentrations were the highest in the winter at 152.8 μg m −3 during rush hours and 90.2 μg m −3 during non-rush hours. While PM 10 levels were higher during rush hours than during non-rush hours in three seasons (excluding summer), these levels were not associated with number of passenger. Elemental analysis showed that the PM 10 was composed of 52.5% inorganic elements, 10.2% anions, and 37.3% other. Fe was the most abundant element and significantly correlated ( p < 0.01) with Mn ( r = 0.97), Ti ( r = 0.91), Cr ( r = 0.88), Ni ( r = 0.89), and Cu ( r = 0.88). Fe, Mn, Cr, and Cu are indicators of railroad-related PM 10 sources. The PM 10 sources characterized by PMF were soil and road dust sources (27.2%), railroad-related sources (47.6%), secondary nitrate sources (16.2%), and a chlorine factor mixed with a secondary sulfate source (9.1%). Overall, railroad-related sources contributed the most PM 10 to subway cabin air. ► PM 10 levels in subway cabins were highest during the winter. ► The in-cabin PM 10 levels were not associated with passenger number. ► Elemental analysis of PM 10 showed that Fe was the most abundant element. ► Four PM 10 sources in cabins were identified by positive matrix factorization. ► Railroad-related sources contributed the most PM 10 to subway cabin air.