500 RADIATION BIOLOGY 



above the preceding straight line, so as to give a concavely ascending 

 curve. 



To explain why the data seem to show only a continuing straight line, 

 the view of the proponents of an association between gene mutation and 

 breakage assumes that the lethals accompanying gross structural changes 

 are not, in the main, expressions of position effects of the altered gene 

 arrangements but of permanent intragenic alterations which were pro- 

 duced in connection with the process of breakage itself. It is pointed out 

 by them that on this hypothesis most of those breaks which at a heavy 

 dose are involved in lethal structural changes, of types (2b) and (3), 

 would, if they had been produced at a light dose, have been followed by 

 restitutions which were, nevertheless, accompanied by lethal effects, in 

 that case classified as gene mutations. Thus there would not really be 

 a major additional class of lethals, namely, position-effect lethals, to 

 increase the total lethal frequency disproportionately at heavy doses. 

 Instead, the freciuency of lethals, following that of individual primary 

 breaks, would remain essentially proportional to dose. 



In this connection the evidence from rings should now be reexamined. 

 If the frequency of cases of perfect (nontwisted) restitution of rings is, as 

 Lea supposed, comparable with that of losses (including dominant lethal 

 losses) of rings, then the apparently undiminished frequency of induced 

 lethals in offspring containing irradiated rings would show that there is 

 very little association after all between primary breakage and the pro- 

 duction of lethal gene mutations. On the other hand, if, in order to 

 explain the results with small deletions, the alternative view is adopted, 

 that the cases of perfect restitution of rings inordinately outnumber those 

 of losses, numerous though these losses are, then it would follow that pri- 

 mary breakage is a far more frequent event than the production of a 

 lethal or sterile change attributable to gene mutation. That is, most 

 breaks would not be associated with such changes. Since, in contrast 

 to this, studies of gross structural changes show that much more than 

 half of them are accompanied by lethal or sterihzing effects (Muller and 

 Altenburg, 1930; Valencia and Muller, 1949), very few of these effects 

 could be explained by attributing them to a by-product of the process of 

 breakage. They must therefore have resulted from the gene rearrange- 

 ment which in these cases followed breakage. That is, they must have 

 been position effects, and, if restitution had occurred, they would have 

 been absent. As originally believed, then, the lethals associated with 

 structural changes would constitute a group which, in the main, would be 

 added to that of the apparent gene mutations found in chromosomes that 

 had not been structurally changed. Hence the continued linearity of the 

 lethal curve at high doses could not be explained on the supposition of 

 primary breaks being associated with lethal effects. 



There are additional facts which appear irreconcilable with the muta- 



