338 H. A. Blair 



are not sufficiently accurate to decide whether this assumption is more correct 

 than the one that the effects per roentgen are more similar in going from strain 

 to strain or species to species. 



It will be observed that according to Fig. 4 divided doses cause only about 

 one-third the life-shortening per unit dose as that produced by single doses. 

 Because some of the divided doses were given in increments as great as 120 r 

 and because existing data are not sufficiently accurate to define small effects, 

 it is probable that the curve for the smaller single doses coincides with that 

 for multiple doses. It is certain, however, that substantial single doses such 

 as 200 r (one-third LD50) or more, have considerably more effect than the 

 same dose in smaller increments. 



The reason for this difference that immediately suggests itself, is that the 

 irreversibility of the injury is some increasing function of its magnitude rather 

 than the linear function assumed here. That this is not the correct explanation 

 is indicated by the fact that repeated daily doses calculated to produce as much 

 injury of the type defined here as a single substantial dose do not have the 

 same effect on life span. There may be some unidentified dose dependent 

 concomitant of injury which affects its reversibility. At this time, however, 

 all that can be said is that a appears to be a constant independent of dose for 

 doses of daily increments up to about 100 r but that it increases with dose 

 with greater daily doses. That this larger effect of substantial single doses occurs 

 at the time of irradiation and is not due to a dose dependent subsequent develop- 

 ment is indicated by a single set of data (13). Such observations should be 

 extended. 



As predicted, the multiple dose curve of Fig. 4 is probably nearly linear. 

 The single dose curve increases more rapidly than linearly if carried to higher 

 doses than those depicted. This is to be expected because, according to the 

 hypotheses, life shortening will be linear with dose only to the extent that the 

 threshold curve of Fig. 3 is linear. As irreversible injury becomes substantial 

 it will have more effect on life span per unit magnitude according to this curve. 



YocKEY (15) has postulated the identity of radiation damage with reduction 

 of somatic genetic information, and has related the present formulation to 

 the consequences of such damage in tenns of information theory. 



CONCLUSIONS 



The hypotheses used appear to give a fairly accurate over-all description 

 of radiation injury. The only one which is definitely known to be inaccurate 

 is the last, which probably should be restated: Reparable and irreparable injury 

 add in all proportions and death occurs when their sum attains a level which 

 is some function, not fully detennined, of the remaining life-expectancy. 



Certain details, such as recovery rates, probably must be regarded as tissue- 

 or region-specific rather than whole-body specific. This may also be true 

 of irreversibility which has not been systematically studied in this regard. 

 This latter problem is of particular interest with respect to human exposure, 

 much of which, especially from internal emitters, is partial-body. However, 

 even if each tissue, for complete description, requires a different set of constants 

 A, a and ^, this adds only complexity of detail and not of concept. 



