236 THE BEHAVIOUR OF POLYPLOIDS 



their parts, since it leads to the corresponding parts lying parallel ; 

 and Sturtevant and Dobzhansky (1930) find that the four 

 chromosomes in an interchange heterozygote in Drosophila lie in a 

 ring at mitosis. Stern (1931) finds no evidence of attraction of the 

 distal end of a normal X for the corresponding part of its homologue 

 when this has been translocated to another chromosome, pre- 

 sumably owing to a conflict with the attractions of larger parts. 



A similar juxtaposition of different bivalent chromosomes at 

 meiosis has long been known in polyploid plants [cf. Kuwada, 1910 ; 

 Ishikawa, 1911), but its relation to the homology of the associated 

 bivalents and its independence of true meiotic pairing was not 

 recognised till recently (D., 1928). Like somatic pairing, this 

 secondary pairing is between chromosomes of similar size and shape 

 and it appears first at metaphase, i.e., it is not a continuance of a 

 prophase association by chiasmata ; it consists in approximation, 

 but never in " contact," i.e., it reaches an equilibrium with forces 

 of repulsion. It is too slight to determine or even modify the 

 anaphase separation. It may be continued between the pairs of 

 daughter bivalents during anaphase, and is often most marked at the 

 second metaphase. Thus it may be supposed that whatever attrac- 

 tion is responsible for it it has a cumulative effect in sorting out the 

 homologues during the two divisions. Again, like somatic pair- 

 ing, it is variable in its occurrence from division to division. 

 Cytologists have usually illustrated the nuclei that were freest of it, 

 to avoid the suspicion of bad fixation, and in consequence its 

 occurrence has been generally neglected, as Lawrence (1931) has 

 pointed out in a general review of the problem. Secondary pairing, 

 like many peculiarities of chromosome behaviour, is exaggerated in 

 appearance by bad fixation, but its essentially differential character 

 as between different chromosomes cannot be determined by an 

 external agent — it is not an artefact. 



Somatic pairing does not usually show at mitosis in plants 

 with secondary pairing at meiosis. The reason for this is that the 

 chromosomes in mitosis are always widely distributed, while at the 

 end of diakinesis (" pro-metaphase " of meiosis), when secondary 

 pairing begins, they are brought within close range of one another, 

 and therefore on analogy should attract one another more strongly. 



