OBJECTIVE To provide quantitative estimates of exposure to benzene for cases and controls in an epidemiology study to investigate the risk of leukaemia in petroleum distribution workers. METHODS Work histories were obtained for cases and controls together with detailed information on the distribution sites. For each job in the work history, an estimate of exposure (parts per million (ppm)) was obtained by multiplying a measure derived from exposure data by modifying factors to reflect the differences between the conditions that existed at the time of measurement and those at the time of interest. The modifying factors used related to job activity, the number of road tankers loaded, the benzene content of the gasoline, the mixture of products handled, temperature, and loading technology. Cumulative exposures for each case and control were obtained by multiplying the exposure estimates for each job by the duration of time in the respective jobs, and summing these over all jobs in the work history. Peak exposure and exposure through dermal contact were quantitatively classified for each job. RESULTS Measured exposures were obtained for 30 job categories, and ranged from 0.003 to 8.20 ppm. 40% of work histories were assigned background exposures, with a further 34% assigned the exposure estimate for a driver carrying out top submerged loading of motor fuel into road tankers. Cumulative exposures ranged from < 1 to > 200 ppm-years, although 81% were < 5 ppm-years. Comparison of the exposure estimates for selected jobs with data from sources not used in the study showed similar results. CONCLUSIONS The estimates of exposure to benzene in this study provide a sound basis for the epidemiological analyses.