The study is based on the work histories and mortality data for 3832 male workers first employed before 1967 at a copper smelter in northern Sweden and followed up from 1950 to 1981. From the 3832 workers a lead cohort consisting of 437 workers employed for at least three years at sites with considerable lead exposure during 1950-74 was selected. These workers had regularly had blood lead measurements performed since 1950. Based on the cumulative blood lead dose 1950-74 and peak blood lead values, the cohort was subdivided into high mean, low mean, high peak, and low peak groups. Standardised mortality ratios (SMR) were calculated for the six groups using general and local reference populations. The original cohort of 3832 workers showed considerable excess of deaths for total mortality, malignant neoplasms especially lung and stomach cancer, ischaemic heart diseases, and cerebrovascular diseases when compared with the general population. In the lead cohort where the workers had been subjected to a considerable lead exposure only the raised SMR for lung cancer was sustained (SMR = 162; not significant). No significant differences were found between high lead and low lead exposed smelter workers.