Radiotherapy is associated with a board spectrum of early and late normal tissue injury. It is a basic clinical observation that even within a group of identically treated patients large variability exists in the incidence and severity of radiation sequelae. Although this is partly a result of the random nature of radiation-induced cell killing, there are at least two additional phenomena involved. One is that certain cofactors influence the expression of radiation damage. These include for example, age of the patient, hemoglobin level, and smoking habits. Another type of predisposing factors are related to coexisting morbidity, like collagen vascular disease, diabetes mellitus, hypertension, and infections. The influence of these factors on the expression of normal tissue injury is critically reviewed and some of the methodological problems involved in this field are discussed. The other phenomenon, that probably contributes to the patient-to-patient variability in the expression of normal tissue reactions, is the variability in intrinsic cellular radiosensitivity that has been shown among individuals. Clinical studies have shown that patients who express a given type of normal tissue injury in one treated area are more likely also to express this injury in another treated area. On the other hand, different normal tissue reactions seem to have a very limited, if any, intrapatient correlation. A number of genetic syndromes are associated with hypersensitivity to radiation, both clinically and in vitro. Also, studies have shown that highly selected patients who express an unusually strong response to radiotherapy, are likely to have in vitro radiosensitivities in the lower normal range. Recently, two studies on otherwise unselected patients support the hypothesis that in vitro radiosensitivity of normal human skin fibroblasts correlated with clinical normal tissue reactions.