Occupational exposure to styrene was studied in nine workers of a hand lamination plant in Bohemia. Personal dosimeters were used to monitor the styrene workplace exposure, and the levels of styrene in blood and mandelic acid in urine were measured. Blood samples were taken at four occasions during a 7 month period to determine styrene-specific O6-guanine DNA adducts in lymphocytes and granulocytes, DNA strand breaks and hypoxanthine guanine phosphoribosyltransferase (HPRT) mutant frequency in T-lymphocytes. Seven administrative employees in the same factory (factory controls) and eight persons in a research laboratory (laboratory controls) were used as referents. DNA adduct levels determined by the 32P-postlabelling method in lymphocytes of laminators were remarkably constant and significantly higher (P < 0.0001) than in factory controls at all four sampling times. HPRT mutant frequencies (MF) measured by the T-cell cloning assay were higher in the laminators (17.5 x 10(-6), group mean) than in the factory controls (15.7 x 10(-6), group mean) at three of the four sampling times, but the differences were not statistically significant. However, a statistically significant (P = 0.021) difference between MF in the laminators (18.0 x 10(-6), group mean) and laboratory controls (11.8 x 10(-6), group mean) was observed at sampling time 4 (the only sampling time when this latter group was studied). This result indicates that styrene exposure may induce gene mutation in T-cells in vivo. DNA strand breaks were studied by the 'Comet assay' at the fourth sampling time. The laminators were found to have significantly higher levels of DNA strand breaks than the factory controls (P = 0.032 for tail length, TL; P = 0.007 for percentage of DNA in tail, T%; and P = 0.020 for tail moment, TM). A statistically significant correlation was also found between the levels of lymphocyte DNA adducts and all three DNA strand break parameters (TL P = 0.046; T% P = 0.026 and TM P = 0.034). On the contrary, no significant correlations were found between DNA adduct levels and the HPRT mutant frequencies or between the mutant frequencies and DNA strand breaks. Taken together, these results add further support to the genotoxic and possibly mutagenic effects of styrene exposure in vivo. However, no simple quantitative relationship seems to exist between the levels of styrene-induced DNA damage and frequency of HPRT mutation in T-lymphocytes.