The increased risk of mortality was distributed over a number of assigned causes of death. In the Societies with a postulated high exposure mortality due to can- eer, cardiovascular-renal diseases and all other causes combined was increased. Leukaemia showed the highest ratio of observed/expected deaths at all ages combined. In general, with the exception of leukaemia and other cancers, the mor- tality ratios were highest in the oldest groups of age. The excess of leukaemia and all cancers combined was greatest during the last working years and the excess in other causes during the post-retirement years. 346. From the above data Seltser and Sartwell [S38] inferred that occupational exposure to ionizing radiation on the part of radiologists had in the past produced a non-specific life-shortening effect. But the validity of this conclu- sion depends on the demonstration that the groups compared are similar in all respect, except for radiation exposure. The authors examined certain charac- teristics of the samples compared such as the geographic distribution of the groups, the region of residence, the size of the living communities, the birthplace: none of these comparisons revealed any difference among groups. Exclu- sion from the comparisons of the first five years after the members had joined the Societies, in an attempt to eliminate a possible selection due to persons with poor health not joining the profession [C27], did not modify the conclusions. And the same was true for another possible cause of selection due to the unfit persons not entering the profession with the highest radiation risk. Factors known to affect the survivorship of other populations (smoking, diet, alcohol consumption, family longevity) could not be tested, but were not deemed to have caused significant differences among the population groups tested. 3h7. Seltser and Sartwell commented on the fact that the reduced survivial (about 5 years) among radiologists for the years 1935-1944 was remarkably near to that obtained by Warren [W2]. This occurred in spite of the fact that the method used by this latter author (but not his conclusions) was criticized by Lewis [L18] and by the same Seltser and Sartwell [S37]. Whether Warren [W2] reached the right conclusion with the wrong method, or not, the results from his and from Seltser and Sartwell's study are in any case in good accordance. 348. The differences with respect to the negative findings of Court-Brown and Doll [C27] could first be explained, in the opinion of Seltser and Sartwell [S38], by the methods of analysis. The absence of a comparison between medical spe- clalists in the British series would be a weakness, since specialist physicians