exposed in 1915-1924, in good agreement with the fact that the work regulations for the dial painting industry came into operation at about that time. Large- scale measurements of radium burden on these women were begun in 195) and an anlysis of the relationships of radium body burden to mortality was performed only on women alive in 1954 who had been measured at least once between 195 and 1975. Only 360 women in the group were available for an analysis as a function of dose and therefore the comparison with respect to cause-specific mortality was performed between two groups only: subjects with a body burden lower than 50 wCi or those with a burden of 50 or more pCi. Mortality ratios from all causes higher than 1 were observed only in the groups with the higher body burdens (1.91). Among these, all malignant tumours, bone tumours and other unspecified neoplasms were also significantly elevated. Among women with less than 50 pCi body burden, tumours of the large intestine were the only sig- nificantly increased cause of death. 358. Another paper by Stehney et al. [S39] is more specifically concerned with the possible presence in this group of women of a life-shortening effect ascribable to causes other than bone sarcoma and head carcinoma. The study was per- formed by the life table method using age- and time-specifiec mortality rates for United States white females for the comparisons. There were 1235 women exposed before 1930: they were on average 20 years old at employment and about 44 per cent of the persons in the group had died by the end of 1976. times thus covered a period of between 45 and 60 years. The observation Regarding death from all causes, 529 deaths before the age of 85 were observed versus 461 expected and the cumulative survival of the group was significantly less than expected, starting at 10 years after employment. When mortality rates for bone sarcoma and head carcinoma were subtracted from the mortality rate for all causes, there was no significant difference at the 5 per cent level in the total population (455 cases observed against 460 expected) or at any of the time intervals considered. A correction for the effect of competing risks was also made on the data cleaned from the radium-related tumours and the difference between observed and expected survival was similarly non-significant also under these conditions. When calculations on the expectation of life were performed at one year intervals from zero to 59 years after the first employment, differences between ex- pected and observed mortality were again not apparent. The conclusion from this study is that when radium-tumour deaths are removed from the exposed sample the average survival is indistinguishable from that of contemporary white females