CENTROMERE-SPINDLE 531 



changes in shape and structure that it undergoes after it comes into 

 relation with them. 



(e) Anaphase. The similarity of the centromeres and the centro- 

 somes or spindle poles in their action on the spindle and their 

 reaction in orientation suggests that in the metaphase of meiosis 

 we are dealing with an equilibrium in repulsions between four bodies, 

 two poles and two centromeres, instead of between three as at 

 mitosis. As we have seen, this equilibrium position is such that 

 bivalents with a single terminal chiasma in a short arm appear 

 already separated at metaphase. The change from metaphase to 

 anaphase should then be different in meiosis from that in mitosis. 



In mitosis it seems to depend directly on the division of the centro- 

 mere. That this division does indeed take place may be inferred, as 

 we have seen, from the behaviour of diplochromosomes and from 

 the lack of synchronisation in division of unorientated small 

 chromosomes in Tulipa galatica (Upcott, unpublished). The two 

 daughter centromeres at mitosis repel one another like the two 

 partner centromeres at meiosis during terminalisation. In meiosis, 

 on the other hand, no change such as division occurs at the beginning 

 of anaphase, and indeed the distinction between metaphase and 

 anaphase is more difficult to make when terminalisation is complete 

 and the chromosomes are short (as in (Enothera) . The centromeres 

 merely move nearer to their poles. The equilibrium position shifts, 

 but there is no sudden change of character. Where nmnerous inter- 

 stitial chiasmata remain to be disentangled, there is a lapse of attrac- 

 tion between chromatids, and it seems likely that this plays some 

 part in the change of equilibrirmi. The important agent in anaphase 

 movement both at meiosis (and one which no doubt plays a part 

 also at mitosis) seems to be the waning of the spindle pole repulsion, 

 for at late anaphase the chromosomes always approach close to a 

 pole from which they were repelled at metaphase. 



A secondary agent of movement depends on the centromeres 

 having already moved. The part of the spindle through which they 

 have passed changes its shape. It narrows and stretches, and in 

 doing so may be seen in living cells to push the chromosomes further 

 apart (Belar, 1929 a and h). The stretching of the spindle was first 

 clearly seen by Kuhn (1920) in Vahlkampfia, where T-shaped tri- 



