Exposure to ambient particulate matter (PM) is associated with respiratory and cardiovascular disease and lung cancer. In this study, we used size fractionated PM samples (3-7, 1.5-3, 0.95-1.5, 0.5-0.95, and <0.5 microm), collected at four contrasting locations (three urban sites, one remote background) in the UK with a Sierra-Andersen high volume cascade impactor. The H(2)O(2)-dependent oxidant generating capacity of the samples was determined by electron spin resonance with 5,5-dimethyl-1-pyrroline-N-oxide spin trapping. In A549 human lung epithelial cells, we determined the cytotoxicity of samples by LDH assay, and interleukin-8 (IL-8) release as an indicator of their inflammatory potency. Oxidative DNA damage was measured by the formamido-pyrimidine-glycosylase (fpg)-modified comet assay. Marked contrasts were observed for all endpoints. Remote background PM showed the lowest oxidant potential, was neither cytotoxic nor genotoxic and did not increase IL-8 release. For the other samples, effects were found to depend more on sampling location than on size fraction. PM collected at high-traffic locations generally showed the strongest oxidant capacity and toxicity. Significant correlations were observed between the oxidant generating potential and all toxicological endpoints investigated, which demonstrates that measurement of the oxidant generating potential by ESR represents a sensitive method to estimate the toxic potential of PM.