We conducted a population-based case-control study to evaluate the relation between exposure to drinking water contaminants (total and specific trihalomethanes and certain metals and nitrates) and childhood acute lymphoblastic leukemia. We compared 491 cases 0-9 years of age with 491 controls. We developed a municipality-exposure matrix based on municipal and provincial historical data, a tapwater survey in 227 homes, and information about residential history. We used average level of exposure and cumulative average over the period as exposure indices, and we measured risk for the pregnancy period as well as for the postnatal period. We show that risks were generally not increased for the prenatal period nor with average levels of exposure. Postnatal cumulative exposure for total trihalomethanes at above the 95th percentile of the distribution for cases and controls was associated with an odds ratio of 1.54 (95% confidence interval = 0.78-3.03); for that same period, risk associated with exposure to chloroform was increased (odds ratio = 1.63; 95% confidence interval = 0.84-3.19) as well as that for exposure to zinc (odds ratio = 2.48; 95% confidence interval = 0.99-6.24). Risks were also increased for exposure to cadmium and arsenic, but not for other metals nor for nitrates.