PROPOSALS FOR FUTURE STUDIES 353 



exerted through restitution (''recovery") rather than through the other 

 two types of competing processes. 



So far we have seen that a study of dose-effect relations and their 

 modifying factors can yield a limited amount of information about 

 mechanism. In a few favorable cases, the basic shapes of dose-effect 

 curves, and the influence of RET and other factors on both shape and 

 slope, give us some ideas about the number of decisive entities and proc- 

 esses and about the time and place (s) at which the decisive processes 

 occur. If water content affects the dose-effect relations, we suspect that 

 the action involves activated water molecules; if oxj^gen tension has an 

 influence, we have evidence that oxido-reduction reactions are included 

 among the relevant processes; and so on. Undoubtedly much more in- 

 formation of this sort could be and is being obtained by diligent in- 

 vestigation of dose-effect relations and their modification. However, no 

 amount of this type of investigation, without other methods of attack, 

 can give us the main parts of the story : the identification of the physical 

 and chemical natures of the relevant and competing processes and their 

 interrelationships. 



Proposals for Future Studies 



At present there appear to be two theoretically possible means of ac- 

 complishing these ends (3). The first is direct: we work out the detailed 

 history of all the chemical and morphological entities in the cell from 

 the moment of energy transfer to the time when the end effect is ob- 

 served. This obviously is the theoretically satisfactory way to do the job 

 because, not only would radiobiological mechanism be completely under- 

 stood, but so would many basic mechanisms of normal cell biology. 

 However, in practice such a method implies advances in techniques and 

 knowledge of cell chemistry and physics which almost certainly will 

 never be attained in our time. 



Accordingly I suspect that any advances in the reasonably near future 

 will still be rather haphazard and will be attained largely by indirect 

 methods. To establish by such methods the relevance of a process to 

 any given end effect, the following general procedure would seem to be 

 minimal. 



First, a given process must be suspected of possible relevance. Our 

 chief source of suspects naturally is radiation chemistry. Any change 

 observed in a cellular constituent irradiated in vitro in an aqueous system 

 may fairly be suspected of relevance to any radiobiological action. We 

 have heard of many interesting examples in this symposium. However, 

 a process which occurs in vitro does not necessarily occur in a cell ; from 



