The magnetic fields from overhead power lines and other electromagnetic sources were determined at the birth and diagnosis dwellings of all tumor cases reported in the county of Stockholm during the years 1958-73 for individuals 0-18 years of age. The study was limited to 716 cases having a permanent address in the county both at time of birth and diagnosis. An equivalent number of controls was matched to the cases according to church district of birth, age, and sex. Outside each dwelling, the occurrence of visible electrical constructions (6-200-kV high-voltage wires, substations, transformers, electric railroads, and subways) within 150 m of the dwelling was noted. Also, the 50-Hz magnetic field was measured outside the main entrance of the dwelling. Visible 200-kv wires were noted at 45 of 2,098 dwellings and were found twice as frequently among cases as among controls (P less than .05). The magnetic field measured at the dwelling varied between 0.0004 to 1.9 microT (mean value 0.069 microT). The magnetic field was higher (0.22 microT) at dwellings with visible 200-kV wires than at those without such wires. Magnetic fields of 0.3 microT or more were measured at 48 dwellings, and were found twice as frequently among cases as among controls (P less than .05). The difference was most pronounced for dwellings of nervous system tumors and was less for leukemias.