DETERMINATION OF COILING 489 



two chromatids with an opposite direction of internal coihng is 

 evidently lost at the next mitosis, since whole arms seem to give 

 consistent results at each meiosis apart from crossing-over. Thus, 

 although the internal spiral must be determined directly by the 

 uncoiling of the molecular spiral, its direction must be facultative 

 and determined by an action of, or at, the centromere in each 

 division. 



The consistent direction of molecular coiling of each arm is per- 

 haps imposed by the centromere, through the coiling beginning 

 near the centromere and being transmitted consistently from par- 

 ticle to particle along the chromosome. This view agrees with the 

 observed priority of proximal parts of chromosomes at zygotene and 

 diplotene in many plants and animals. 



The conditions determining the direction of relational coiling of 

 chromosomes in pachytene seem to be more complex than those 

 determining internal coiling. Relational coiling seems to be 

 universal both from direct evidence and from the general occurrence 

 of non-homologous torsion pairing. To develop relational coiling 

 the two chromosomes must suffer an internal torsion in the same 

 direction as one another. We know in Tradescantia that their 

 internal coiling may be in opposite directions. The direction of the 

 spiral responsible for the two systems of coiling must therefore be 

 separately controlled. Now the relational coiling, like the internal 

 coiling, seems to be consistent throughout chromosome arms in 

 Fritillaria and Chorthippus apart from any interruptions due to 

 structural hybridity. But where such interruptions occur different 

 chromosomes with opposite coiling properties must sometimes be 

 brought together. These chromosomes must nevertheless be sup- 

 posed to coil relationally, since crossing-over takes place between 

 relatively translocated segments. Thus the torsion responsible for 

 relational coiling must itself be not individually but relationally 

 determined. We are thus brought face to face with a contradiction 

 between the principles of internal and relational coiling and with a 

 seemingly inexplicable property of the partner chromosomes 

 influencing one another's coiling direction in relational coiling. 



Relational coiling of chromatids must always be present in meiosis, 

 since interlocking of chromatids may always occur at first anaphase 



