12 K. H. MOLE 



caused by radiation and that there are no problems in accepting a leukae- 

 mogenic effect of doses of whole-body irradiation as small as 5 r or less. It is 

 sometimes thought that in its favour is the association of mongolism and 

 leukaemia and against it the lack of any firm correlation between leukaemia 

 in identical twins, but these may well be arguments of dubious force. If 

 somatic mutation is indeed the basic mechanism of leukaemia it is a qualita- 

 tively, different kind of mutation, as pointed out earlier, since the cellular 

 change is not to a stable genotype but to a progressively altering one. There 

 seems no necessary correlation, as there would be for ordinary genetic muta- 

 tion, with the stable genotype characteristic of the mongol or of identical 

 twins. Brues's (1958) arguments against the somatic mutation theory are, as 

 he emphasized, the arguments against one particular kind of mutation, point- 

 mutation, as a sole cause; he was concerned to demolish the idea of a Imear 

 relation between radiation dose and carcinogenesis. 



Those who do have a first hand acquaintance with human cancer and 

 leukaemia may develop a very strong emotional bias against the somatic 

 mutation hypothesis from the feeling that, if it is true, nothing can be done 

 against cancer (Rous, 1959). Although there is no direct evidence in favour of 

 the hypothesis this kmd of prejudice is no evidence against. There has been 

 much discussion as to whether there is a "threshold" for leukaemia induction: 

 the reasons why this question is a theoretical one and can never be answered 

 by observation are given elsewhere (Mole, 1958). The data themselves can 

 never prove any particular hypothesis; they can only be compatible or 

 incompatible with it. 



The other general hypothesis, which relates leukaemogenesis to radiation- 

 induced tissue damage, is obviously true in the sense that a high incidence of 

 leukaemia in experimental animals as well as in man occurs only when the 

 radiation exposure has been enough to cause severe tissue damage, though just 

 what is the nature of the critically important tissue damage is not at all clear. 

 Is it possible to show that leukaemia does not occur unless there is a sufficient 

 degree of tissue damage? This is very like trying to prove a general negative 

 and if the idea is accepted that leukaemia, whatever the nature of its initia- 

 tion, starts in a small focus somewhere, then even if leukaemia can be caused 

 to occur without obvious tissue damage the problem of sampling is going to 

 prevent the sceptical biologist from denying that there might well have been 

 sufficient tissue damage somewhere where he did not happen to look. 



The experimenter with irradiated animals is always going to be biased 

 towards theories of tissue damage because an experiment can involve only a 

 relatively small number of animals and to get enough leukaemia to be "signi- 

 ficant" we know already that tissue damage must be caused. The human 

 epidemiologist is always going to be biased towards theories of direct cellular 

 action just because he deals with a few cases of leukaemia in thousands or 



