duetion could not entirely explain the life-shortening observed, although thymic and myeloid leukaemia could account for most of the increase in mor- tality in irradiated groups. 71. By the use of a radioprotective agent (WR-2721 or S-2(3-aminopropylamino) ethylphosphorothioic acid) which protects against acute mortality more efficiently than it does against the life-shortening effects of radiation, Yuhas [Y1] expanded the range of doses studied. The shape of the dose-response re- lationships were consistently different for the two strains studied. In the A/J strain the curve was linear non-threshold at low doses and came to a plateau in the high dose range. In the C57BL/6J life-shortening was curvilinear over the entire range of doses. It is impossible to assess whether the radio- protective treatment altered the actual shape of the dose-response relation-— ship in ways and amounts different for the two strains used. considers the dose relations obtained at doses below the LD Actually, if one 50/30 without the use of the WR-2721, a certain amount of curvature can be seen in both sets of data, perhaps more pronounced in the C57BL/6J. 72. In amore recent experiment Grahn, Fry and Lea [G5] gave LAF1 hybrid mice of both sexes single exposures of °C gamma rays in the range of 390 to 900 R. Mean after-survival showed a curvilinear trend with dose. The principal life- shortening effect was attributable to excess tumour mortality up to 390 R, while at higher exposures the loss of life expectancy was not paralleled by a further increase of tumour incidence. Walburg [W1] commented on these data and interpreted them to show that when the life-shortening effect is 15 per cent or less of normal the increased mortality is attributable entirely to induction or acceleration of tumours. 73. Clapp et al. [C12] reported on a large-scale experiment on life-shorten- ing and disease incidence in RF/Un mice irradiated with 300 kVp x rays (50-00 rad) and with 60 MeV protons (47-372 rad). The data indicated a flattening of the dose-response curve at doses in excess of 200 rad and a reasonable straight trend at the lower doses. When animals dying from thymic lymphoma and myeloid leukaemia (which were induced in up to about 40 per cent and 25 per cent, respectively, of the mice) were removed from the calculations, the mean survival time of the remaining animals still showed a decrease as a function of dose. Gompertz's analysis confirmed that removing the leukaemic animals did bring the death rate curve more near and more parallel to the control line: the curves,