THE NATURE OF THE GENE 367 



This coiling suggests one of two conclusions about the nature of the 

 chromosome thread. Either the attractive forces in zygotene are much 

 more intense than at later stages when each chromosome coils inde- 

 pendently of its partner; or each chromosome has a single "sticky face" 

 along which the adhesion takes place. The latter conclusion is perhaps 

 unlikely; it would certainly raise considerable difficulties about the 

 mechanism of reduplication of the chromosome (cf. p. 399). 



It should be pointed out that a similar relational coiling, due to the 

 assumption of a chromosome spiral, might occur between different 

 regions of an unpaired part of a chrom-osome; the mechanism can be 

 seen if a woollen thread is twisted tightly and the ends then brought 

 nearer together, when a twisted loop known in textile circles as a 

 "snarl" is produced. McClintock^ has described associations of non- 

 homologous regions in zygotene in maize which have been interpreted^ 

 as "snarl-pairings"; but until we have a physical picture of completely 

 specific pairing forces, it is impossible to exclude the possibility that 

 non-homologous chromosomes may sometimes pair in the normal way. 



3. The Connections of Chromomeres 



Chromomeres (by v/hich we shall in this section mean the region of 

 chromosome associated with one gene) are joined together in a linear 

 array and must therefore have two ends which can join up with neigh- 

 bouring chromomeres. There usually do not appear to be more than 

 two such ends; if there were, we should be able to get branched 

 threads in which one chromomere was joined to three others. Branched 

 chromosomes have indeed been described, both on genetical and 

 cytological evidence. But in some cases in Drosophila, Muller and 

 Offermann^ have shown by investigation of the salivary glands that the 

 genetical evidence was at fault, and the cytological evidence in other 

 forms seems to require further confirmation. At present it seems 

 justifiable to conclude that the chromomere has only two ends which are 

 capable of stable attachment. This may not be true of the centromere, 

 at which branching of the chromosome seems to be possible (p. 95). 



These two ends cannot be regarded as poles. That is to say, the 

 chromomere is not in any sense like a little magnet with different north 

 and south ends. This is clearly shown by the fact that they can join up 

 in reverse order in inversions. 



There is still considerable debate whether there are special one- 

 ended chromomeres occupying the ends of the chromosomes. If this 



^ McClintock 1933. ^ Darlington 1937. ^ Kossikov and Muller 1935. 



