fication of the shape of the dose-effect relationship in the range 50 - 400 rad from a very complex pattern at 45 rad/min to a nearly linear shape at 8.3 rad/day. The large difference in effectiveness was related to an upward dis- placement of the regression line in the O - 50 rad range in the high-dose rate groups. Balb/c females showed a similar trend, in that the 50 rad dose point was displaced upward in the high~intensity curve, suggesting a dose-rate dependent injury component which saturated at high dose rates at about 50 rad, similar to the one identified by Sacher [S2] in female mice given 200 rad or more. The response to neutron irradiation of high (25 rad/min) or intermediate (1 rad/day) dose-rate was somewhat different. In RFM mice the low-dose-rate was less effective at 2h rad but more effective at 188 rad than the high doserate. In Balb/c mice little dose-rate dependence was seen at low doses but at 188 rad the low intensity was more effective. The results are somewhat less clear at doses below. 2. 158, Dose fractionation The experiments reported in the next few paragraphs (159-172) were per- formed by splitting a given dose or a series of doses into two or more frac-— tions, irrespective of the time over which the total dose was administered. The dose per fraction, the fractionation interval and the total time to complete the course of irradiation are variables that interact together in producing the final effect. They cannot be disentangled from each other under most of the experimental conditions used. Often the ocmparison is therefore between a dose given in a single treatment and the same dose over a very protracted course of fractionation. Only seldom is the accuracy of the data such that numerical protraction factors can be derived with the necessary degree of precision. 159. In the mouse Sacher[S2] performed experiments on the life-shortening of 400, 800 and 1200 R when given in equal fractions 5 day/week over 2 or 8 weeks. The life-shortening of mice dying from causes other than lymphoma decreased with increasing number of fractions, even though the effect of leukaemia induction increased by fractionation, was reported for the C57BL mouse by Kaplan and Brown [Ki2}. The dose fractionation experiments performed by Curtis and Gebhard in 1958 [C17] with fission spectrum neutrons and 250 kVp x rays in CF1 female mice (see also paragraph 194) did not show any change in RBE upon fractionation but the authors themselves recognized the peculiarity of this finding and attributed it to the use of fairly large doses. In their opinion, the re- covery rate from such doses would be essentially different in the x-ray and neutron groups, as compared to the recovery from small fractional doses. a -