Cancer risk as a result of air pollution may be quantified by different approaches. We compared the sum of unit risk based effects of single pollutants with an epidemiology-based method by using PM(10) as a surrogate of the total air pollution. The excess rate for lung cancer cases attributable to an increase of 10 microg/m3 in average PM(10) exposure was estimated from available cohort studies. Applying the epidemiology-based risk method to the air pollution situation in the Basel area (Switzerland) resulted in 13.3 (95% CI = 6.9-19.8) excess lung cancer cases per 100,000 person years. This estimate was considerably higher than the unit risk-based estimate yielding 1.1 (range, 0.45-2.8) cancer cases per 100,000 person years. We discuss these discrepancies in light of inherent differences between approaches in toxicology and epidemiology.