As part of a long-term program to investigate the impact of air pollution on the health of a population in a polluted region in Northern Bohemia, mutagenicity of extractable organic matter (EOM) from air particles PM10 was investigated by the means of Salmonella typhimurium indicator strains TA98 and YG1041 using the Ames plate incorporation assay. The air samples were collected in both the polluted and the control districts during the summers and winters of 1993-1994. In the polluted district, the collection was repeated during the winter of 1996-1997. The crude extracts from filters pooled according to the locality and the season were fractionated by acid-base partitioning into acid, base, and neutral fractions. The neutral fractions were further fractionated by silica gel column chromatography into five subfractions. The induction of revertants with the crude extracts was higher in winter samples than in summer samples. Both indirect-acting and direct-acting mutagenicity were observed. The indirect mutagenic potency of aromatic subfractions containing polycyclic aromatic hydrocarbons (PAHs) was generally low. The mutagenic potency detected with TA98 was more distinct only in the winter sample 1993-1994 from the polluted area, where the aromatic subfraction accounted for 23% of total mutagenicity. In both strains, the highest direct-acting mutagenicity was found in slightly polar fractions containing nitro-PAHs. The mutagenic potency detected with YG1041 was about two orders of magnitude higher than that detected with TA98. No substantial locational- or time-related variances in the mutagenic potencies of EOM, or in the spectrum of chemical components identified in individual fractions were found. The polluted district, in comparison to the control district, was found to have higher amounts of EOM, carcinogenic PAHs and mutagenicity of air particles (rev/m(3)). The fractionating process, combined with the bacterial mutagenicity test, confirmed that nitro-derivatives are the most important contributors to the bacterial mutagenicity of air particles. However, this study did not fulfill the expectancy to bring substantially new, clear-cut information on the composition and the biological activity of air pollution in both districts.