MANNER OF PRODUCTION OF MUTATIONS 523 



of producing both types of effects, breaks and gene mutations, it seems 

 most likely that some of these cases involved one type of double effect 

 and some the other, although it would be hazardous at this time to 

 estimate the numbers of these two types relatively to each other. 



There is no reason to assume that the number of breaks or of gene 

 mutations in a cluster thus produced would never be higher than 2. If, 

 as a first approximation, the number of effects in different cases are 

 assumed to form a Poisson series of frequencies, it can readily be reckoned 

 from the frequencies of those having just one effect (the viable cases) and 

 those having more than one effect (the lethals) that there had been 

 approximately 50 per cent more actual mutational changes involved, 

 counting separately all those occurring at or between different loci, than 

 there had been cases of mutation found. That is, about a third of the 

 mutational changes are lost to view in a count that scores as one unit 

 each case of a point mutation involving a visible effect. Since the pro- 

 portion similarly lost to view is so small in the case of X and 7 radiation, 

 this means that the neutrons have not much more than two-thirds the 

 efficiency of the less densely ionizing radiation in so far as the production 

 of separately detectable mutational effects in cases of point mutation 

 involving genes with visible effects is concerned. It is evident, moreover, 

 that, if this is true in the cases studied in which the mutations had 

 originally been detected by means of their visible effects, it must also be 

 true of mutations detected by means of lethal effects. That is, when a 

 lethal occurs at a given locus (either as a result of gene mutation or 

 deficiency), there is, on the average, just as much chance that one or 

 more other lethals will have been simultaneously produced at some 

 neighboring locus or loci, as when a change with a visible effect occurs at 

 a given locus. It follows that, with neutrons, about a third of the lethals 

 actually arising in the detected cases of lethal point mutations fail to be 

 individually recorded because of their proximity to other lethals. 



These facts and considerations appear to explain why neutrons cannot 

 have much more than two-thirds the efficiency of less densely ionizing 

 radiation in producing lethal point mutations which are obviously 

 separate. These relations would not be much disturbed by the inclusion, 

 among lethal point mutations less accurately analyzed, of cases of dele- 

 tions of a size large enough to be seen cytologically, as must have hap- 

 pened in the earlier studies. For in the work by Muller and J. I. Valencia 

 on mutations at specific loci, demonstrable deletions were found to be 

 produced with about the same frequency, in relation to the cytologically 

 invisible point mutations, by neutrons as by X or 7 rays. 



Returning now to the finding by Timofeeff-Ressovsky and Zimmer that, 

 when doses of equal ionizing capacity are tested, neutrons appear to have 

 only about two-thirds as high an efficiency as X or 7 rays in producing 

 sex-linked lethals, it is evident that, within the limits of statistical error, 



