482 RADIATION BIOLOGY 



individual breaks, like individual gene mutations, are produced by single 

 ionizations or activations or by the individual natural clusters of them. 

 There must, of course, be a little falling off, at higher doses, from the 

 frequency of single-break losses expected on strict linearity; for at these 

 doses a larger proportion of the broken ends must become drawn off into 

 more complicated structural changes. However, the single-break cases 

 remain too numerous, relative to the others, to be greatly affected in 

 this way. That this is the case became more apparent later, when it was 

 found (Muller and Pontecorvo, 1941; Muller, 1941a; Pontecorvo, 1941, 

 1942) that those cases of X-chromosome loss which result in viable off- 

 spring represent only a small proportion of the total number of cases of 

 isochromatid formation caused by single breakage of the X chromosome, 

 inasmuch as the great majority of such cases involve the early death of 

 the zygote, presumably by bridge formation during mitosis. The fre- 

 quency of these lethal losses also follows the first power of the dose, as has 

 been determined by studies of the changes thereby induced in the sex 

 ratio (see Muller, 1928b, for the first evidence of this, before the basic 

 cause of this effect was realized to be chromosome breakage). 



Considering all the chromosomes, the frequency of their loss and of 

 bridge formation, caused by single breakage resulting in isochromosomes, 

 is enough to explain the great majority of the deaths of young zygotes 

 derived from spermatozoa that had been given lower and moderate doses 

 of radiation (Muller, 1940; Muller and Pontecorvo, 1941; Pontecorvo, 

 1942; Catcheside and Lea, 1945; Demerec and Fano, 1944). At such 

 doses the death rate of the young zygotes is therefore found to be linearly 

 proportional to dose, when allowance is made for the "saturation effect," 

 i.e., for overlapping of the lethal effects. At higher doses, however, 

 progressively more of the losses are caused by multiple-break aneucentric 

 structur9,l changes, which increase at a higher power of the dose than 1. 



Far more frequent than even the lethal losses are the cases of single 

 breakage followed by restitution. When a ring X chromosome is used, a 

 considerable proportion of the restitutions of the X chromosome also 

 become lethal or, less frequently, lead to chromosome loss in viable 

 zygotes (Bauer, 1939b, 1942; Muller and Pontecorvo, 1941; Muller, 

 1941b, Pontecorvo, 1941, 1942; Faberge, 1952). This effect must be 

 caused chiefly by twisting of the chromosome before its restitution, a 

 process resulting in the interlocking of the chromatids. It is found that 

 the frequency of the lethal effects and losses caused by these aborted 

 restitutions likewise varies with the first power of the dose. All in all, 

 evidence that the frequency of individual breaks in Drosophila is linearly 

 proportional to the dose is thoroughly established by varied genetic 

 evidence. On the other hand, the attempts which have been made 

 (e.g.. Lea, 1946) to find the absolute frequency of chromosome breaks 

 per nucleus produced by a given dose, including those breaks which 



