110 R- H. MOLE 



however, a direct relevance to the interpretation of experiments on factors 

 which alter the incidence of leukaemia in irradiated mice. Such experiments 

 are commonly carried out by choosing some particular course of radiation 

 exposure which reproducibly induces leukaemia in high frequency and then 

 scoring the effects of experimental intervention in terms of leul^aemia 

 incidence. This is the way in which the miportance of hormones and of various 

 organs like the spleen and thymus have been demonstrated. However, with 

 dose-response curves which are markedly curvilinear or dependent on 

 fractionation, scoring in terms of leukaemia incidence may not, in fact, mean 

 very much. The problem of interpretation seems to be closely analogous to the 

 problem of interpreting the activity of substances which reduce acute 

 mortality from radiation when administered before exposure. It is easy with 

 a lot of substances to show reduction in mortality from say 80 to 10% but due 

 to the curvilinearity of the dose-response curve this is equivalent to altering 

 the lethal effect of a roentgen by only a few per cent. Experiments on 

 physiological factors affecting radiation leukaemogenesis do not ever seem 

 to have been carried out with more than one schedule of radiation exposure so 

 it is impossible to know whether the change in the leukaemogenic efficiency 

 of a roentgen is large and important or small and trivial. Thus it would seem 

 that judgment should be reserved on the quantitative importance of the 

 different physiological factors which have been shown experimentally to 

 change the incidence of radiation-induced leukaemia. 



One logical problem is always going to face the experimenter. Experiments 

 are concerned with tens, sometimes with hundreds, rarely, if ever, thousands 

 of animals. Whatever mechanism can be demonstrated at the incidence levels 

 which are measurable m groups of these sizes, say 100 down to 10 or even 1%, 

 it is always possible that some other mechanism will be the important one at 

 the lower levels of incidence which are of practical concern to human popula- 

 tions. The only way round this seems to be to produce leukaemic cells by in 

 vitro treatment, for by varying the number of cells inoculated, it should then 

 be possible to quantitate rare cellular events with probabilities of do^\ai to 10 ° 

 or so. 



EEFERENCES 



Kaplan, H. S. (1948). J. nat. Cancer Inst. 9, 55. 



Kaplan, H. S., and Brown, M. B. (1952). J. nat. Cancer Inst. 13, 185. 



Mole, R. H. (1959a). United Nations Peaceful Uses of Atomic Energy Proc. Second 



International Conference, Geneva, 1958, 22, 145. 

 Mole, R. H. (1959b). Brit. J. Radiol. 32, 497. 



Upton, A. C, Wolff, F. F., Furtil J., and Isjumball, A. W. (1958). Cancer Res. 18, 842. 

 Upton, A. C, Klviball, A. W., Furth, J., Christenberby, K. W., and Benedict. 



W. H. (1960). Cancer Res. 20, part 2, 1. 



8 



