428 BREAKDOWN OF GENETIC SYSTEMS 



accounted for in terms of the conclusion already arrived at from 

 general observations, that the chromosomes divide in the resting 

 stage, while the centromeres do not divide until metaphase. The 

 argument is as follows. We may suppose that mitosis is frustrated 

 in these cases during prophase so that the nucleus relapses into the 

 resting stage, the effect being merely an extreme example of the 

 delay already noticed. Then before a new division can occur the 

 chromosomes will divide again, while their centromeres, which 

 divide only at metaphase, will be unaffected. Each centromere will 

 then have four chromatids instead of two. Whether or not this 

 explanation is sound we see that not only the independence of the 

 two processes of division, but also the particular time-relationship 

 assumed on general grounds is corroborated by the structure of the 

 diplochromosomes. 



The diplochromosomes show relational coiling, like the 

 chromatids at an ordinary mitosis, but clearly visible owing to 

 repulsion, like the chromosome coiling at diplotene. The special 

 delay responsible for their occurrence might be due to direct damage 

 to the centrosome, or to a secondary effect brought on by treatment 

 at a particular stage of the centrosome-cycle. Since groups of cells 

 are sometimes affected, and the nuclear cycles of groups are syn- 

 chronised in the spermatogonia, we may infer the latter. 



The immediate effects of irradiation on meiosis differ from those 

 on mitosis in one respect, a disproportionately high production of 

 acentric fragments (Stone, 1933 ; Mather, 1934). With a low 

 dosage the number of bodies may be unaltered at mitosis owing to 

 each breakage being followed by a reunion. With a similar dosage 

 the number of bodies at meiosis is more than doubled in Tradescantia 

 and Vicia. Further, Tradescantia is exceptional inasmuch as these 

 fragments appear at metaphase ; in Vicia, Tulipa, Primula and 

 elsewhere, with and without terminalisation, the fragments do not 

 appear until anaphase. Probably the explanation of this difference 

 is that the fragments are held to their unbroken partners at pachy- 

 tene in most organisms which have complete pairing, while in 

 Tradescantia with incomplete pairing, they lie free. The explana- 

 tion of lack of reunion of breakage-ends of chromosomes at meiosis 

 is probably the pachytene pairing, which prevents the necessary 



