A 19 year follow up study was conducted to explore the association between occupations expected to be exposed to electromagnetic fields and the occurrence of leukaemia and brain tumours. Incidence of cancer between 1961-79 was calculated and the standardised morbidity ratio (SMR) with a 95% confidence interval (95% CI) was related to that of all Swedish working men. For all the selected "electrical occupations" the SMRs for total leukaemia and brain tumours were near unity. Increased risks were noted for all leukaemia among electrical/electronic engineers and technicians, (SMR 1.3; 95% CI 1.0-1.7) as well as in the sub-groups of telegraph/telephone (2.1; 1.1-3.6) and machine (2.6; 1.0-5.8) industries. Risk for chronic lymphoid leukaemia was increased in the same occupational category (1.7; 1.1-2.5) and in the sub-group of machine industry (4.8; 1.0-14.0), as well as for all linesmen (2.0; 1.0-3.5) and power linesmen (2.8; 1.1-5.7). Risk for acute myeloid leukaemia was increased among all miners (2.2; 1.0-4.1) and miners working in iron/ore mines (5.7; 2.1-12.4). Increased risk for all brain tumours (2.9; 1.2-5.9) and glioblastomas (3.4; 1.1-8.0) appeared among assemblers and repairmen in radio and TV industry. Raised risk for all brain tumours was seen for all welders (1.3; 1.0-1.7) and welders in iron/steel works (3.2; 1.0-7.4) and risk for glioblastomas was also increased for all welders (1.5; 1.1-2.1). No major changes in relative risk estimates were noted after the exclusion of persons who were over 65 at the time of diagnosis. Although a homogeneous pattern of increased risks of leukaemia or brain tumour was not noted, the hypothesis that magnetic fields might play a part in the origin of cancer cannot be rejected.