A population-based case-control study was conducted in Ontario, Canada, to assess the relation between the risk of childhood leukemia and residential exposure to magnetic fields. Participating subjects consisted of 201 cases, diagnosed at 0 to 14 years of age during 1985-1993, ascertained from the records at the Hospital for Sick Children (Toronto), and 406 individually matched controls. Where possible, point-in-time measurements of magnetic fields were made in all residences occupied by subjects during the period of inquiry in the defined catchment area. Three different classification schemes of wire code were assigned to each residence. Detailed information was collected by interviewer-administered questionnaires, which enabled risk estimates to be adjusted for socio-economic characteristics, medical history of parent(s) and child and environmental exposures. Inconsistent elevations in risk were associated with time-weighted averages of magnetic fields both inside and outside the home for subjects having residential point-in-time measurements that represented at least 70% of their etiological period. These risks increased in magnitude when analysis was restricted to children under 6 years of age at diagnosis or to those with acute lymphoblastic leukemia. For children younger than 6 years at diagnosis, outside perimeter measurements of the residence, > or = 0.15 microT, were associated with increased leukemia risk (OR = 3.45, 95% CI = 1.14-10.45). Evaluation of different exposure times for point-in-time magnetic field measurements and wire configuration suggested that exposures earliest in the etiological period were associated with greater risks for children diagnosed at a younger age (OR = 2.50, 95% CI = 1.14-5.49). Our findings did not support an association between leukemia and proximity to power lines with high current configuration.