THE MECHANICS OF THE CHROMOSOMES I3I 



sents the centromere. Clearly crossing-over between e and /would give 

 chromatids abc^defdycba and gefg. That is to say, one of the chromatids 

 would have two centromeres (dicentric) the other none (acentric). 

 Similar types of chromosomes appear if there are two cross-overs, and 

 a new type, a loop with one centromere, is also possible. The different 

 types of compound cross-overs give rise to characteristic sets of chro- 

 matids ; for instance, two complementary compound cross-overs in the 

 inverted segment give rise to two dicentric and two acentric chromatids. 

 The various types of compound cross-overs can be recognized, since 

 the different chromatids have characteristic behaviours.^ The dicentric 

 chromatids are pulled towards both poles at the first division and thus 

 remain as a bridge between the anaphase chromosomes; the acentric 

 chromatids are not moved from the metaphase plate and are lost, while 

 the loop chromosomes form a bridge in the second division, when their 

 centromere has divided into two daughter centromeres which pull in 

 opposite directions. Most of the sorts of cross-over chromatids which 

 arise are therefore eliminated in one way or another. Inversions, in 

 fact, are often first discovered as cross-over suppressors or C factors. ^ 

 The situation is peculiar in oogenesis in Drosophila, since the formation 

 of chromosome bridges orientates the spindles in such a way that the 

 abnormal chromosomes tend to be eliminated in the polar bodies. The 

 reduction in recombination is therefore not accompanied by an equi- 

 valent reduction in fertility.^ 



8. The Cause of Crossing-Over 



A complete elucidation of the mechanical forces which cause the 

 breakage and rejoining of chromatids and thus give rise to crossing- 

 over and chiasma formation cannot be provided until our knowledge of 

 the nature and physical properties of the chromosomes is in a much 

 more advanced state. One of the most plausible suggestions which has 

 been made* is that the breakage might be caused by a torsion applied 

 to the four threads at the moment when the chromosomes divide at 

 pachytene. This hypothesis enables one to understand how chiasma 

 interference may come about : the relief of the torsion by one breakage 

 and rejoining will reduce the force acting on the chromatids in that 

 immediate neighbourhood. The torsional force involved here may 

 possibly be related to the forces producing the spirahzation of the 

 chromosomes. Regional differences in crossing-over, and therefore the 



^ Cf, Upcott 1937. - Sturtevant 1926a. 



^ Sturtevant and Beadle 1936. * Darlington 1935a, 6, 1937- 



