those that may be derived from the analysis of the single experimental series discussed in the preceding paragraphs. 3. 2th, SPECIFICITY OF LIFE~SHORTENING After the overall quantitative analysis in the preceding paragraphs, the problem of the specificity or non-specificity of the effect should now be reviewed. Such a discussion presupposes the availability of experimental series with careful pathology of the animals at death or serial sacrifices to investigate the development of the pathology of aging. Reported experiments of this kind are indeed very few and even when pathology is reasonable, any direct com- parison with survival is made impossible by the presentation of the data. Therefore, the present section will be essentially qualitative. It will in fact be based on the conclusions of the authors themselves which are often unsatisfactory owing to inadequate pathology (mostly macroscopic) or to insuffi- cient statistical analysis. In other cases the conclusions of the experiments were based by the models of action assumed in the interpretation of the data. However, in the absence of the original data, no better treatment of the subject matter is possible. 215. It is commonly reported [G4, S19, C12, L9] that correction of the data for animals dying of leukaemia and of ovarian tumours, which are very common causes of death in the rodent, does lead to a reduction of the large variability of the effect between strains and sexes and to a reduction of the lifeshortening efficiency of the radiation treatment. This indicates that at least part of the reduction of life seen after irradiation must be attributed to tumour induction. 216. The first large series where pathology was of such a quality to allow analysis of specific death causes was that of Upton et al. [U5]. The authors could establish no clear-cut relationship between shortening of life and inci- dence of tumours since the dose relationships for tumour induction had variable form, some neoplasms being increased and some decreased with increasing dose. These data gave impulse to the idea that radiation might cause non-spe- cific aging by advancing in time all diseases by about the same degree for each given dose. However, 4 more recent reevaluation by Walburg [W1] with a statistical method allowing for competing probabilities of death justified the conclusion that life-shortening, which was clearly apparent when all death causes were considered together, disappeared when tumours were excluded from the analysis.