owing probably to some inadequacy of the survey. Out of these deaths, 6 were due to skin cancer and 2 to aplastic anemia and these numbers were significantly higher than would be expected to occur among the population at large. Leukaemia was found in 5 cases, in no significant difference with the number expected. Concerning the average age at death, 52.7 years was the value found among radiology technicians, while the expected value would have been 48.6 years. 353. The mortality rates of United States radiologists in comparison with other medical specialists were re-examined by the Johns Hopkins University group [M28, M29] in two reports published in 1975 up to a total follow-up of 50 years. The comparison regarded male members of the Radiological Society of North America who were contrasted with fellows of the American College of Physicians and members of the American Academy of Ophtalmology and Otolaringology. The information through 1954 available from the previous study by Seltser and Sartwell updated for new members and decedents up to 1969. were traced for 99.5 per cent of the decedents. [838] was Deaths and causes of death The persons under comparison were about 30000 among all Societies for a total number of deaths of about 6500. 354. In the first paper [M28] the mortality rates from all causes were calcu- lated by the life-table method of analysis with age- and time-adjustments of the death rates in such a way that cumulated rates could be compared within any 10 year cohort and across societies. decade of entry. Mortality from all causes depended on the During 1920-1939 death rates of radiologists were higher than those of any other specialty group for both cancer and non-cancer causes. The differential between the rates for radiologists and other specialists was lower in the 1930-1939 cohort and it disappeared in the 1940-1949 cohort. So did the graduation of death rate radiologists > internists > other specialists which was noticeable in earlier periods. Removing the deaths from tumours in the 1940- 1949 cohort led to a disappearance of the difference between radiologists and non-radiologists noticed in earlier cohorts. The all-cancer mortality rates for radiologists were higher than those of other specialists up to the decade ending in 1949. The next decade had not aged sufficiently to show the expected peak of cancer mortality in the 60-64 years age group. It was pointed out that self- selection of the persons entering any one group and the life style after entering the specialty would have little influence on the data: thus, the presumed radiation exposure of the specialists under comparison would appear as the only rea- sonable way to explain the mortality differences and their trend over time.