Six hundred and ten coke oven personnel were studied retrospectively over a period of 11 years . During this time 82 died, 88 retired and 80 left, but all were traced. The Standardised Mortality Ratios (SMRs) have been calculated for four principal causes which have then been further analysed into 30 subgroups. The work was limited because the population was small and the records not detailed enough to analyse the mortality rates according to environmental exposures. Differences between SMRs obtained at the two works were quite marked (73 and 98) and the overall SMR of 85 % was better than expected. The observed number of deaths from respiratory disease, malignant neoplasms and all other causes excluding cardiovascular, was very near that expected, but the numbers of deaths from cardiovascular disease and in particular arteriosclerotic disease, showed a significant deficit.Studies of death rates according to occupation have indicated that mortality from cancer is higher than normal in the carbonising and associated industries. Kennaway and Kennaway (1947) studying the certified deaths during 1921-38 from cancer in gas workers found that the observed number was 29 more than that expected from a population of 16 000, an increase of 84%. Doll (1952) studied 840 London gas companies' pensioners over 60 years of age who died in the period 1939-48, and estimated an expected 10-4 deaths from lung cancer, using the Registrar General's data for England and Wales, and an expected 13-8 from the data for London. The observed number of 25 was significantly different at the 0-001 and the 0-01 probability levels respectively. Reid and Buck (1956) Class A, and it was deduced that this probably reflected an occupational hazard; in addition, cancer of the scrotum had not been eradicated entirely.Kreyburg (1959) equated 3:4 benzpyrene pollution in one of the coke ovens he studied in Norway to 5000 cigarettes a day for workers on a 40-hour week. Referring to the high correlation between mortality from lung cancer and cigarettes smoked, he commented on the lack of evidence of high mortality rates from lung cancer in the works he investigated.In a series of six papers, Lloyd and Ciocco (1969), Robinson (1969), Redmond et al. (1969, Lloyd et at. ( ), Lloyd (1971