362. A third series was reported by Kohn, Bailar and Zippin [{K20, Z1] on about 500 cases of cervical carcinoma treated with x rays and/or radium obtained from two cancer registries in the U.S.A. These women were treated prior to the age of 55 and survived at least 5 years after treatment. By the time of the last report about 38 per cent of the women had died and this fact limited objectively the weight of the conclusions. The patients were grouped according to the stage of the disease at the time of treatment in stage I and stages IJ and ITI, respecively. There was also a grouping according to the regional dose delivered into low (2 - 22 Mem rad), middle (23 - 31 Mgm rad) or high doses (32 ~ 54 Mgm rad) which allowed survival to be evaluated as a function of dose. It was also re- ported that the incidence of leukaemia appeared in these patients to be lower than among patients treated for ankylosing spondilitis or for metropathia hemorragica. 363. Indirect information of a negative nature may be derived from a study of nearly 3000 children irradiated in infrancy to shrink their allegedly enlarged thymus. These infants were followed in time and compared with about 5000 non- irradiated siblings. In spite of a four-fold increase in tumours, particularly of the thyroid, among the irradiated subjects, the report [H15] shows good agreement between observed and expected numbers of death. This indicates the absence of excess non-specific mortality, which is not surprising in view of the small portion of the body irradiated. 364. In 1965 Court-Brown and Doll [C28] reported on a sample of 14.554 persons (12.161 mean and 2.393 women) treated for ankylosing spondylitis at various radiotherapy centers in Great Britain and Northern Ireland during the years 19351954. These patients had been followed for periods varying from 5 to 25 years up to the end of 1959 with reasonably good records of the treatment and edequate follow-up information in 98 per cent of the cases were available. The number of expected deaths in these patients if they had suffered only normal mortality rate were computed from the numbers of person-years at risk for each sex, agegroup and calendar period and they were multiplied by the sex- and age-specific mortality rates for each corresponding period. These calculations were perform- ed for all causes of death, all cancers, the principal types of cancer and the respiratory diseases. The total death rate among the patients was about 1.8 times as high as the corresponding national death rate. When the various causes of death were analysed separately, the following observations were made.