To confirm the relationship between exposure to coal tar pitch volatiles and bladder cancer among primary aluminum production workers, we carried out a case-control study among blue-collar workers who had worked more than 1 year between 1950-1979 in a major plant using mostly the Soderberg process in the Province of Québec. Cases of bladder cancer (ICD code 188) diagnosed between 1970-1979 (n = 69) were mostly included in a previously reported study. To these were added cases diagnosed between 1980-1988 (n = 69). Each case was matched to three controls on date of birth, date of hiring, and length of service at the company. Smoking habits were assessed from the medical records at the company. Benzene-soluble matter (BSM) and benzo(a)pyrene (BaP) were used as indicators of environmental exposure to coal tar pitch volatiles in the workplace. The estimated risk for current smokers was 2.63 (95% C.I. 1.29-5.37). Estimates of risk by occupational exposure were adjusted for smoking. Men who had worked in the Soderberg potrooms were at higher risk of developing the disease, the risk increasing with the time spent in these departments. Similarly, a strong association between risk and cumulative exposure to BSM or to BaP was observed. The risks associated with cumulative exposure to BSM (mg/m3-years) and to BaP (microgram/m3-years) were described with mathematical models. Using a linear model (1 + bx) and lagging 10 years before the diagnosis, BaP cumulative exposure was a better indicator of risk than BSM cumulative exposure. The risk for each year of exposure to BaP at a concentration of 1 microgram/m3 increased by 1.7% (0.8%-3.2%). Using the same model for BSM, a worker exposed to the current threshold limit value of 0.2 mg/m3 for 40 years will sustain a risk of 2.22 (1.56-3.48). Comparison of risks according to different periods of diagnosis (1970-1979 vs. 1980-1988) did not reveal any significant temporal changes on risk estimates.