LOCALISATION AND TERMINALISATION 113 



large chromosomes suggests that terminalisation depends on the 

 size of chromosomes. When we recall that the effectiveness of the 

 body repulsion depends on the size of the nucleus we see that this 

 also is not surprising. Although similar or even greater forces 

 may be engaged in moving these large chromosomes they will 

 have less effect. This conclusion is borne out by the comparison 

 of chromosomes of different sizes within the complement. In 

 Stenobothrus , for example, we see that, while in the larger chromo- 

 somes there is a low degree of terminalisation, in the smallest one 

 the process is always complete. The same is true in the small 

 fragments of Fritillaria imperialis (D., 1930 d, cf. Ch. V). 



The degree of terminalisation is probably correlated with the 

 degree of longitudinal contraction. Thus in male-sterile Lathyrus 

 as compared with the normal (Upcott, 1936), and in the male 

 Macronemurus as compared with the female (Naville and de Beau- 

 mont, 1933) contraction is greater and terminalisation is more com- 

 plete. It seems that the greater contraction may directly determine 

 a greater repulsion and hence a greater movement, but the con- 

 nection might be indirect in several other ways. 



When we compare behaviour in related species we find that other 

 variables are concerned : there is no direct relation between size 

 and terminalisation, for example, amongst species of Allium (Levan, 

 1935)- Species therefore differ in the degree of centromere repulsion 

 during prophase, independently of the size of their chromosomes. 

 Nevertheless those with the longest chromosomes never have a 

 high degree of terminalisation. 



The question then arises why all chiasmata should be terminal 

 in one group of organisms with very large chromosomes, the tetra- 

 ploid relatives of Tradescantia virginiana (D., 1929 c ; Richardson, 

 1935). This species passes through a diffuse stage at diplotene, 

 and the stages of movements of chiasmata have not been seen. 

 It seems probable that pairing and chiasma-formation are localised 

 near the ends, that this species in fact corresponds with Fritillaria 

 in having interrupted pairing, but merely begins its pairing at the 

 ends instead of near the centromere. 



(viii) Anaphase : the Separation of Chromatids. When the 

 bivalent chromosomes have been arranged on the metaphase plate 



