This study was performed to investigate the concentration of PM(10) and PM(2.5) inside trains and platforms on subway lines 1, 2, 4 and 5 in Seoul, KOREA. PM(10), PM(2.5), carbon dioxide (CO(2)) and carbon monoxide (CO) were monitored using real-time monitoring instruments in the afternoons (between 13:00 and 16:00). The concentrations of PM(10) and PM(2.5) inside trains were significantly higher than those measured on platforms and in ambient air reported by the Korea Ministry of Environment (Korea MOE). This study found that PM(10) levels inside subway lines 1, 2 and 4 exceeded the Korea indoor air quality (Korea IAQ) standard of 150 microg/m(3). The average percentage that exceeded the PM(10) standard was 83.3% on line 1, 37.9% on line 2 and 63.1% on line 4, respectively. PM(2.5) concentration ranged from 77.7 microg/m(3) to 158.2 microg/m(3), which were found to be much higher than the ambient air PM(2.5) standard promulgated by United States Environmental Protection Agency (US-EPA) (24 h arithmetic mean: 65 microg/m(3)). The reason for interior PM(10) and PM(2.5) being higher than those on platforms is due to subway trains in Korea not having mechanical ventilation systems to supply fresh air inside the train. This assumption was supported by the CO(2) concentration results monitored in tube of subway that ranged from 1153 ppm to 3377 ppm. The percentage of PM(2.5) in PM(10) was 86.2% on platforms, 81.7% inside trains, 80.2% underground and 90.2% at ground track. These results indicated that fine particles (PM(2.5)) accounted for most of PM(10) and polluted subway air. GLM statistical analysis indicated that two factors related to monitoring locations (underground and ground or inside trains and on platforms) significantly influence PM(10) (p<0.001, R(2)=0.230) and PM(2.5) concentrations (p<0.001, R(2)=0.172). Correlation analysis indicated that PM(10), PM(2.5), CO(2) and CO were significantly correlated at p<0.01 although correlation coefficients were different. The highest coefficient was 0.884 for the relationship between PM(10) and PM(2.5).