NATURE OF THE GENETIC EFFECTS 379 



chromosomes a considerable number of regions which are much more 

 susceptil)le than the rest to liecoming involved in structural changes, and 

 that, according to the former investigator, these very regions show some- 

 what of the same tendency to conjugation with other such regions, as 

 well as with centromeric heterochromatin, as is shown by the hetero- 

 chromatin proper. She regards these interstitial regions, therefore, as 

 having a composition somewhat like that of heterochromatin. Still 

 further evidence of the effect of movement in promoting structural 

 change is to be found in Sax's (1942) finding that already acentric pieces, 

 produced by a prior irradiation, are much less likely than centric ones to 

 undergo further structural change when radiation is again applied. 



Not only are certain chromosome regions more likely than others to 

 engage in structural change, but it is found, when the frequency of 

 different structural changes is tabulated, that they show some tendency 

 for structural changes invohdng a break in one position to have their 

 other break in the same chromosome arm rather than elsewhere. This is 

 true despite the finding of Sax, just mentioned, that a break distal to 

 another one is in some circumstances more likely to restitute. Thus, in 

 chromosomes of Drosophila with two arms (regions on each side of the 

 centromere) of approximately ecjual length, the ratio of inversions in 

 which both breaks are in the same arm (paracentrically) to those in 

 which they are in different arms of the same chromosome (pericentrically) 

 or in other chromosomes (translocations) is about twice as high as would 

 be expected on a chance distribution, as shown by Catcheside (1938), by 

 Bauer, Demerec, and Kaufmann (1938), and by Bauer (1939). No doubt 

 this is in part an expression of the fact, to be discussed in Sect. 10, that 

 there is a distance limitation on the position, at the time of breakage, of 

 broken ends which will be able later to undergo union with each other; 

 this at the same time indicates that the movements of ends relative to 

 one another, between the time of breakage and that of union, are rather 

 limited. In spite of this there is no pronounced tendency for the distribu- 

 tion of inversions of different lengths, yet long enough to be readily dis- 

 covered by cytological means, to depart from that expected on a random 

 basis (Bauer, Demerec, and Kaufmann, 1938; and Bauer, 1939). 



There does, how^ever, appear to be a distinct tendency for structural 

 changes in which the breaks were very near together, resulting in minute 

 deletions, insertions, and inversions, to be more numerous than expected 

 on a chance basis. In part their apparently higher frequency is due to 

 the selective factor that, in the case of deletions (the largest class among 

 those readily found), the more minute ones survive better, but in part the 

 discrepancy in numbers seems to be a real one. It is probably to be 

 explained not only by the spatial limitations on union of broken ends 

 already mentioned, but also by a tendency for breaks to occur close 

 together, occasioned by the manner of operation of the radiation in 



