Iv. THE HUMAN EXPERIENCE A. 329. INTRODUCTION In the present chapter the evidence about the existance of a non-speci- fic life-shortening effect in the human species is discussed. available comes from three different sources: groups of persons The evidence (radiologists, radiology technicians, physicians) exposed occupationally during the course of their professional life; patients who have undergone radiation treatments for different pathological conditions, but mostly for tumour therapy or for con- trol of ankylosing spondylitis; a large number of survivors of the A-bomb experience in Japan in 1945 and a few hundred people exposed in the Rongelap fall-out accident in 1954. The data will be discussed separately, since the modalities of the exposure are different in the three groups and the characteristics of the sample size and of the epidemiological observations are also quite different. 330. The studies performed on humans are subject to a number of limitations, mostly related to the lack of any control over the variables to be examined. In general, the sample size is small for effects which have often a marginal incidence over the whole population studied. The life-span study on the A- bomb survivors, numbering about 80.000 irradiated persons, is an exception in this respect. Often the time elapsed between irradiation and the epidemiolo- gical survey is insufficient to reveal effects which take a very long time to develop. This requires frequent updating to keep the time-course of the phe- nomena under control. In the case of radiotherapy patients there is the con- comitant presence of an important disease which causes a decrease of survival completely unrelated to the radiation exposure and induces a prevalence of associated disabling conditions altering the spectrum and the time of occurrence of the causes of death to be expected in a normal population. 331. Finding suitable control groups to match the irradiated group is always a problem: the distribution of ages, the geographical location, the differences in the socio-economic status and in the living and working conditions between the control and the test sample are often quite large. When the effects to be