490 RADIATION BIOLOGY 



on the probability of double breakage would in itself explain not only 

 the relative frequencies of the observed "isochromatid" breaks but also 

 the strangely identical pattern of relative freciuencies of the observed 

 single chromatid breaks, without invoking a multiple action of ions on 

 any given chromatid. This is because, in many cases in which breaks of 

 both chromatids had occurred, one of them would finally manage to resti- 

 tute before the stage at which the observations were made (or, alterna- 

 tively, one fragment of one thread would unite with the nonhomologous 

 fragment of the other thread), and the observer would get the impression 

 of a single chromatid break. In such a case, simultaneous occurrence of 

 the two breaks would have assisted in the original separation of fragments 

 after breakage, even though that separation became permanent for only 

 one of the threads. It seems likely that a considerable majority of the 

 apparently single breaks are really relics of these original double breaks 

 and that few are derived from the breaks which were originally single 

 since the latter, although doubtless far more numerous, w^ould have had 

 a much stronger tendency to restitution. 



If this analysis is correct, the data under discussion give no ground for 

 following Catcheside and Lea's inference that the ionizations cooperate 

 with one another in producing individual breaks in the first place. The 

 greater frequency of apparently single breaks with a given type of radi- 

 ation would really be referable only to the greater frequency of the origi- 

 nally double breaks from which most of them had been derived. That is, 

 there is nothing to show that that particular spatial distribution of ions 

 which is more conducive to double breaks in relation to single breaks is 

 actually more efficient in the primary production of breaks. Hence it 

 must be left an open question whether clustering of ions favors primary 

 breakage. It seems at least equally tenable (and more so, by analogy 

 with what is known of gene mutations) that the individual ionizations 

 and/or activations are the effective agents in breakage, although any 

 single one, even if located within a chromosome, would give only a small 

 probability, P, of breaking the chromosome. A cluster of n of these 

 ionizations or activations would according to this mechanism afford 

 approximately n times this probability or, more accurately, a probability 

 of 1 - (1 - P)". 



Among the experiments on the efTects of radiation of different specific 

 ionization in producing chromosome changes in Tradescantia was a series 

 with a rays (Kotval and Gray, 1947). This radiation proved twice as 

 efficient as neutrons in producing breaks in tw^o sister chromatids simul- 

 taneously. This must have been caused in part by a greater likelihood 

 of the a rays being within simultaneous reach of both chromatids and in 

 part by a lower restitution frequency (Lea, 1946). In support of the 

 latter inference was the finding that, for a given frequency of observable 



