ATTRACTION AND TORSION 493 



between chromomeres which is qualitatively specific and quantita- 

 tively limited to pairs. There is no attraction between all four 

 threads (chromosomes) of a tetraploid at zygotene, and there is no 

 attraction between the four threads (chromatids) of a diploid at 

 diplotene after the chromosomes have divided. This intermolecular 

 force operates at a greater distance than intramolecular attrac- 

 tion, and although it constitutes one of the primary physical 

 properties of living matter we can find a very ready inorganic 

 analogy {cf. Hardy, 1919, on Static Friction cf. 1936). 



In homozygous organisms there is, as a rule, no reason to doubt 

 that the attraction which initiates pairing also carries it on through- 

 out the chromosome. But in structural hybrids another force comes 

 into play. Chromosomes are brought together in the parts where 

 they are homologous by their specific attractions. But in linear 

 sequence with these are non-homologous parts. These are under a 

 torsion, as all chromosomes are at this stage, the torsion that deter- 

 mines relational coiling. They therefore continue this homologous 

 pairing by a non-homologous torsion-pairing (McClintock, 1933, on 

 Zea ; Lammerts, 1934, on poly-haploid Nicotiana, D., 1936 c on 

 Chorthippus) . Sometimes a chromosome may "pair" with itself 

 as a piece of twisted string does when released from longitudinal 

 tension. The chromosomes that show this property most strongly 

 are the inert B chromosomes of Zea Mays ; they show a correlated 

 lack of secondary attraction at mitosis and therefore perhaps have 

 a weaker primary attraction [v, infra) . 



There are various other consequences of torsion pairing. First, in 

 the triploids it is found that one of the three chromosomes is left 

 out of association at zygotene, the three chromosomes changing 

 partners intermittently. At a late pachytene it is found, however, 

 that all three lie very close together and may appear to be equally 

 associated in some small parts of their length (D., 1929 b, Olmo, 

 1934). This is evidently due to a chromosome which is unpaired in 

 an intercalary segment being dragged round its partners and there- 

 fore closer to them as the relational coiling develops. 



Further, we may recall that in ordinary homozygous diploids true 

 pairing is often incomplete, especially where the chromosomes are 

 very long. It is interrupted by the division of the unpaired parts 



