POLYCENTRIC CHROMOSOMES 431 



centromere by irradiation in two ways. First, areas of endosperm, 

 after irradiation in Zea, may show the apparent results of 

 " deficiency." This sometimes consists in the loss of a whole 

 chromosome, since the recessive types of a group of linked factors 

 may appear. Islands occur in which the dominants reappear. 

 This may be explained as due to the failure of the chromosome 

 carrying the dominants to divide, and its later recovery of this 

 capacity (Stadler, 1930). Such an effect is most easily understood 

 as the result of a single change, and it may be supposed that this 

 change consists in injury to the centromere which temporarily 

 prevents its division and consequently the separation and redivision 

 of the two chromatids. 



Secondly, McClintock found (1933) that on two occasions a ring 

 and rod, each with a centromere, arose from one chromosome by 

 X-raying. Whether this is due to breakage of the centromere or 

 to the preservation of a ring from a multiple breakage and non- 

 disjunction, it is difficult to say on account of the special conditions 

 of survival of rings in diploids. Unless they are supernumerary the 

 complement will be defective. 



Evidence of a transverse doubleness of the centromere is uncertain 

 in most organisms. Its appearance at pachytene and at first 

 anaphase sometimes suggests transverse doubleness, but the 

 interpretation is open to doubt (D., 1933 a). In Ascaris, however, 

 the movements of the large multiple chromosomes indicate that 

 there is a row of centromeres distributed along the chromosomes at 

 intervals, for instead of the chromosome moving to the pole first 

 at one point, the whole of the middle part moves together (Schrader, 

 1935 ; White, 1936). The existence of several centromeres would 

 account for the chromosome breaking up into several independent 

 bodies when " diminution " takes place. It is merely necessary to 

 suppose that in the cells where no such breakage takes place the 

 chromatids have lain parallel between the centromeres, while in 

 those where breakage occurs there has been a persistence of chromatid 

 coiling. The irregular numbers of chromosomes found after 

 breakage may then be due to irregular coiling. In this way the 

 structural change of diminution will be physiologically controlled. 

 White has demonstrated the correctness of this view by irradiation. 



