Associations between occupational magnetic field exposure and non-Hodgkin's lymphoma (NHL), Hodgkin's disease, and multiple myeloma mortality were evaluated in 138,905 electrical utility workers. A job-exposure matrix based on measured magnetic fields was used to derive individual exposure estimates. There was a small positive association between all NHL and low-grade NHL and duration of employment in any magnetic field-exposed job, but only up to 20 years. Cumulative magnetic field exposure was associated with a rising, then falling, risk of NHL. Rate ratios for intermediate/high-grade lymphoma were increased for the highest levels of lifetime cumulative exposure (RR = 3.7 and RR = 2.3), and were most pronounced for those exposures occurring 10-20 years previously. Hodgkin's disease and multiple myeloma mortality did not appear to be associated with exposure. Associations were stronger for NHL subgroups, suggesting that further analysis by subgroup may be warranted, but the absence of dose-response gradients diminishes the likelihood that associations represent causal relationships.