To determine whether occupational exposure to magnetic fields of 50-60 Hz was associated with cancer among electric utility workers, the authors used a case-control design nested within three cohorts of workers at electric utilities: Electricité de France--Gaz de France, 170,000 men; Ontario Hydro, 31,543 men; and Hydro-Québec, 21,749 men. During the observation period, 1970-1989, 4,151 new cases of cancer occurred. Each participant's cumulative exposure to magnetic fields was estimated based on measurements of current exposure of 2,066 workers performing tasks similar to those in the cohorts using personal dosimetry. Estimates were also made of past exposure based on knowledge of current loading, work practices, and usage. Workers who had more than the median cumulative exposure to magnetic fields (3.1 microtesla (microT)-years) had a higher risk for acute nonlymphoid leukemia (odds ratio (OR) = 2.41, 95% confidence interval (CI) 1.07-5.44). The same observation holds for acute myeloid leukemia (OR = 3.15, 95% CI 1.20-8.27). There was also an elevated risk for mean exposure above 0.2 microT (acute nonlymphoid leukemia, OR = 2.36, 95% CI 1.00-5.58; acute myeloid leukemia, OR = 2.25, 95% CI 0.79-6.46). However, there were no clear dose-response trends with increasing exposure and no consistency among the three utilities. Men whose cumulative exposure to magnetic fields was above the 90th percentile (15.7 microT-years) had an elevated risk for brain cancer (OR = 1.95, 95% CI 0.76-5.00) that was not statistically significant. No association with magnetic fields was observed for any of the other 29 types of cancer studied, including skin melanoma, male breast cancer, and prostate cancer. Controlling for potential confounding factors did not change the results.