NATURE OF THE GENETIC EFFECTS 377 



preceded the reproduction of the chromosome to form chromatids. How- 

 ever, in the section on the consecjuences of a single break it was pointed 

 out that union of the broken ends sometimes fails to occur until after the 

 fragments have reproduced. It was seen that in some of these cases one 

 of the chromatids may undergo restitution while the other, failing to do 

 so, results in acentric and eventually also in dicentric chromosomes. 

 Similarly, when two or more breaks have occurred the union of some or 

 all of the broken ends may be delayed until after the pieces have repro- 

 duced to form chromatids. Here, too, the sister fragments may there- 

 after follow different courses so that, for example, a chromosome broken 

 at two places may give rise to one chromatid with an inversion and one 

 with a deletion having its breaks at the same points as those of the inver- 

 sion. Moreover, the chromatid fragments may become transferred in 

 such a way as to result in one daughter nucleus having a deficiency while 

 the other receives a duplication, which may be attached either to the 

 homologous chromosome that already contains (often in an adjoining 

 position) another representative of the same chromosome region, or to a 

 nonhomologous chromosome. The production, in this manner, of 

 daughter cells of different but often more or less complementary genetic 

 types, derived from one treated cell, is seen especially strikingly when 

 spermatozoa have been irradiated, for then a visibly mosaic individual 

 may be formed, approximately half of which is descended from each of 

 the two genetically different daughter cells of the "first cleavage" stage. 



8. NONRANDOM INCIDENCE OF THE CHANGES 

 PRODUCED BY CHROMOSOME BREAKAGE 



The frequencies with which structural changes of different types are 

 found following irradiation do not follow a purely chance distribution, 

 even when due allowance is made for the fact that some types are much 

 more subject than others to elimination before being found. The most 

 marked irregularity in distribution is to be noted in the great excess of 

 structural changes involving one or more breaks in a heterochromatic 

 (see last part of Sect. 3) region of a chromosome, when we take into con- 

 sideration how very short the heterochromatic portions of the chromo- 

 some threads are in comparison with the euchromatic parts. Most of 

 the heterochromatin of a normal chromosome lies in a short region on 

 either side of its centromere, and a very little adjacent to each of its free 

 ends or telomeres (Prokofyeva-Belgovskaya, 1937, 1938; Muller, 1938); 

 yet a high proportion of all the radiation-induced translocations, inver- 

 sions, and deletions formed in Drosophila have involved one or more 

 breaks in a heterochromatic region, usually in a region near the centro- 

 mere. So, for instance, in the X chromosome of Drosophila the hetero- 

 chromatin near the centromere occupies only about a twentieth the 



