166 K. H. MOLE 



DISCUSSION 



lamerton: Dr. Mole had to put Dr. Brue's data tliroiigh a good many manipulations to 

 demonstrate linearity on a square law basis. I think I have a good reason for feeling 

 distrustful of this sort of analysis because what, in fact, has been done is to compare body 

 burden essentially with incidence. This is fine if you've got a uniform dose among the 

 cells at risk but not if there is a non-uniform dose. You can't prove a square law with a 

 non-uniform irradiation unless you weight the radiation in your irradiated individual 

 appropriately. This, I think, was also a snag m the curve that Dr. Mole showed the other 

 day of Dr. Burch's analysis of the spondylitic data using the mean dose to fit incidence 

 with the square law. But we do know for the spondylitics that there wasn't a uniform 

 exposure. Even if you only take the exposure to the spine there were separate fields and 

 non-uTiiform irradiation. Under those conditions you can't even guess what the weighted 

 dose is. This, in fact, is a criticism of the whole Court-Browoi and Doll analysis which I 

 think they recognize. They didn't show that there was linearity, they merely set out to 

 discover whether the results were consistent with hnearity, and to do this they made the 

 assumption that the relation was linear when they took a mean dose over the spme or 

 over the person examined. 



MOLE: I do not really think that that's a logical objection to the analysis of the bone- 

 tumour data, because, as long as the distribution of doses of different sizes is the same, 

 then you could apply a weightmg factor, whatever it is, to each dose and just go straight 

 ahead. If the distribution isn't the same with doses of different sizes than you are quite 

 right, but if it is the same I don't think it makes any difference. 



LAMERTON: In the case of spondyhtis the distribution wasn't the same because there ^^■ere 

 lots of different fields. 



mole: As far as I know the distribution of strontium is largely mdependent of the dose ; 

 this has been looked at. 



lamerton: I'm not convinced that if the distribution is the same you can still use mean 

 dose as your parameter. 



trpTON: i may prove to be mistaken on this point but I think if one adopts the thesis that 

 the carcinogenic transformation consists of two successive stages, then woidd it not 

 follow that, given an individual with one step completed, one would get a one hit type 

 of response curve. Looking at the data from Japan, you have a hodge-podge of different 

 kinds of dose response curves which, in the aggregate, simulate the effects of a certain 

 number of years of ageing. Burch has looked at the linear dose response curve mentioned 

 earher that Bond and Cronkite published on Sprague-Dawley female rats on the assump- 

 tion that tliis strain inherits already one of the mutations required and radiation simply 

 adds another. The fact that they have a liigh natural incidence means that they don't 

 have to live very long before they acquire the necessary additional mutation or whatever 

 event one postulates for the second genetic step. 



berenblum: I take it that the suggestion you made is based on the assumption that the 

 two events are both two single sudden events. Now would the same sort of reasoning 

 apply if the first event were a sudden one but the second depended on a long contmuing 

 action? 



mole: I don't thuik it would make any difference as long as the continuing action is 

 proportional to the number of radioactive events per unit of time, but it's much easier 

 with particles to think of each particle as doing something. 



CASARETT: I am concerned about the elimination of latent periods and this may have 

 something to do with the nature of the second event. I don't think that we need assume 

 the second event has to be an event in the cell of origin of the tumour, it can be a pro- 



