The cancer incidence in a historical cohort of 10,059 metal workers employed during the period 1964-1984 was investigated. Standardized incidence ratios (SIR) were calculated based on registry extracts from the Danish Cancer registry. Lifetime exposure data (occupational and other) were obtained by a postal questionnaire in living cohort members and interviews by proxy for deceased and emigrated subjects. The incidence of lung cancer was increased among workers ever "employed as welders" (SIR = 1.38, 95% C.I. 1.03-1.81). There was a significant excess risk of lung cancer among "mild steel (MS) only welders" (SIR = 1.61, 95% C.I. 1.07-2.33) and "nonwelders" (SIR = 1.69, 95% C.I. 1.23-2.26) (indicating carcinogenic exposures other than welding), a borderline significant lung cancer excess among "MS ever welders" (SIR = 1.32, 95% C.I. 0.97-1.76), and a nonsignificant excess risk of lung cancer among "stainless steel (SS) only welders" (SIR = 2.38, 95% C.I. 0.77-5.55). In spite of signs of inconsistency in the risk estimation by duration and latency, we find the results support the conclusions of other studies: employment as a welder is associated with an increased lung cancer risk.