187. It is very difficult to see an overall trend for more complex fractiona- tion patterns. On different species, treatment schemes involving from 3 to 90 dose fractions administered from 3 days to about 23 weeks have been tested. Very seldom is the fractionation pattern comparable within each experimental series, since for the same total dose both the dose per fraction and the total treatment times are changing. In some cases [S2, A2, AT, H5, LY, R3] dividing the dose into smaller and smaller fraction does lead to an increased survival time following x- or gamma-irradiation. So does the increase in the total treatment time for the same number and size of fractions. But in other cases (M1, K6, M20, C18] a paradoxical effect is observed, i.e., and increase of the life-shortening upon dose fractionation. A changing spectrum of the various diseases contributing to life-shortening with more leukaemia induced at longer fractionation times has been invoked to explain the oberservations. The out- come of a given fractionation treatment depends critically on the characteristics of the biological system as they reflect on the pathology of the animals at death. It is difficult therefore to give a numerical value to the effect of frac- tionation except to say that the effect, when present, is not very large and the biological variables appear in general more important in determining it, than the actual change in the dose-time relationships induced by dose fractionation. 188, Four neutron dose fractionation experiments have also been reported on mice. In the first [V1] splitting a dose into 3, 4 or 10 fractions in 3, 4 and 12 days did not result in any significant change in the mean after survival. The second series [V5] involved changing the number of fractions/week (1,3,6) for a given weekly dose (60 rad) and a fixed time of treatment (13 weeks) and was similarly negative. In the third series [AT] fixed doses (80 and 20 rad of neutrons) were given singly or subdivided into 24 fractions administered over 23 weeks and such a scheme produced a significant increase of life-shor- tening. Similar observations were also made in the last series [S47]. It should again be emphasized that the physical parameters of fractionating the dose are not prevalent over the changes of the pathological expression of the damage which largely determine the life-shortening effect. 189, In spite of the absence of a component of wasted radiation that has been claimed to confound the analysis, the effects of chronic terminated exposures reviewed in paragraphs 173-180 are more difficult to evaluate than the experiments involving duration-of-life exposure. In principle, for any given type of radiation, the effect to be expected should lie between the dose effect re-