A cohort study on mortality among all members of the Swedish Chimney Sweeps Union who were active in 1950 or later and have been members for at least 10 a has been performed; 2,071 chimney sweeps fulfilled the criteria and the loss in the follow-up was less than 1 %. The observed numbers of death before the age of 80 were compared with sex-, calendar year-, and age class specific expectancy values from the national statistics of 1951-1979. Two hundred and thirty deaths were observed versus 197.6 expected. This result was due to a significant excess of deaths from tumors, particularly lung and esophageal cancer, and from nonmalignant chronic respiratory diseases. The multifold increased risk from these diseases could hardly be explained by extreme smoking or alcohol habits but rather by exposure to polycyclic aromatic hydrocarbons, nitrogen compounds, arsenic, and asbestos in combination with exposure to sulfur dioxide. The excess mortality had occurred in spite of favorable selection factors as demonstrated in different subcohorts. Preventive actions to diminish inhalation of soot and combustion gases should be taken.