A mortality study was carried out among the workers of a plant that had produced ferrochromium and stainless steel, and was still producing stainless steel, in order to determine whether exposure to chromium compounds, to nickel compounds, and to polycyclic aromatic hydrocarbons (PAH) could result in a risk of lung cancer for the exposed workers. The cohort comprised 2269 men whose vital status were recorded between 1 January 1952 and 31 December 1982. The smoking habits of 67% of the cohort members were known from medical records. The observed numbers of deaths were compared with the expected ones based on national rates with adjustment for age, sex, and calendar time. A low mortality, achieving statistical significance, was found from all causes (observed = 137, standardised mortality ratio (SMR) = 0.82) and from benign respiratory diseases (observed = one, SMR = 0.15). With regard to mortality from lung cancer, a non-significant excess appeared in the whole cohort (observed = 12, SMR = 1.40). Among the exposed workers, however, a significant lung cancer excess was found (observed = 11, SMR = 2.04) that contrasted with a low SMR (0.32) in the non-exposed group. This excess is unlikely to be explained by smoking, as the tobacco consumption of these two groups was similar. No trend was observed for mortality from lung cancer either according to time since first exposure, or according to duration of exposure. A nested case-control study clearly suggested that this excess of deaths from lung cancer was attributable to former PAH exposures in the ferrochromium production workshops rather than to exposures in the stainless steel manufacturing areas.