Previous epidemiological studies on one cohort of rubber industry workers indicated that observed mortality excesses from lymphocytic leukemia were related to solvent exposures. The results of a detailed examination of the solvent exposure histories of 15 cases of lymphocytic leukemia from this cohort and 30 matched industry controls are presented. Extensive review of historical solvent use, raw materials specifications and job description records were used to reconstruct exposures to specific agents. Cases were 4.5 times as likely as controls to have had direct exposures to both benzene and other solvents. Analysis by manufacturing production source of solvent revealed that cases spent greater proportions of their work experience in jobs with potential exposures to coal tar-based benzene and xylene. No differences were seen for petroleum-derived solvents.