Polycyclic aromatic hydrocarbons, which are carcinogenic and mutagenic to humans, are primary compounds in coke oven emissions generated by the coking process. The authors examined the relationship between coke oven workers' urinary 1-hydroxypyrene levels and their exposure to polycyclic aromatic hydrocarbons, as determined on the basis of work category and pre- and postshift effects in a steel plant in Taiwan. Eighty-eight coke oven workers constituted the exposed group, and 61 office workers in a steel plant located 1.5 km from the coke plant constituted the control group. The benzene-soluble fraction in personal air samples, and 1-hydroxypyrene in urine samples, were measured for 3 consecutive days. The 3-day urinary 1-hydroxypyrene sampling results for topside workers (i.e., those most heavily exposed to emissions) in the coke oven group, and for the control group, as determined from postshift geometric means, were 23.8 microg/gm creatinine and 0.3 microg/gm creatinine, respectively. These values increased to 13.4 and 0.1 microg/gm creatinine, respectively, after an 8-hr work period. The major source of polycyclic aromatic hydrocarbons for the exposed group was occupational; therefore, the closer workers were to the coke oven, the greater their exposure, and, consequently, the greater their metabolite level. Urinary 1-hydroxypyrene levels of the exposed group were 80 times higher than those of the control group. Smoking had no significant effect on the excretion of urinary 1-hydroxypyrene. 1-hydroxypyrene levels in the workers' urine during an 8-day period was cumulative (half-life = 18.6 hr). The authors concluded that it would be desirable to switch highly exposed workers to a low-exposure work area, after a period of rest. In addition, urinary 1-hydroxypyrene was a confirmed, useful biological indicator for exposure to polycyclic aromatic hydrocarbons.