u 



whilst its diverging portions are more centrally placed with reference 

 to the spindle axis, when seen in profile, and at right angles to the 

 polar axis. This again is unaccountable on the assumption that the 

 fissure above referred to represents the diverging halves of the origi- 

 nal chromosome. Either a complete rotation through 180" in the 

 radial plane of the spindle would be required, or else a closing of 

 the originally divergent ends, followed by a correspondig opening out 

 of the opposite ones must be assumed. But there seems to be no 

 evidence for such a complex series of movements, and in those chromo- 

 somes which have a ring-like structure they are not required ; in the 

 case of those which do not form rings the process is accomplished in 

 another way, as will soon be apparent. 



The divergent limbs which travel polewards along the spindle are 

 in reality the original halves of a chromosome, which has become 

 doubled, or bent, upon itself. The manner in which this complicated 

 figure is produced may be easily understood by taking an elliptical 

 hoop of wire as in Fig. 4, and bending it over upon itself as in Fig. 5. 

 Then the approximated and original ends E. E. will represent the 

 hump on the chromosome, and by pulling apart the two sides, at the 

 point where the bending has been elfected, in a direction at right 

 angles to the original bend, a fairly correct representation will be 

 obtained of what happens during the divergence of the limbs of 

 the chromosome (compare Figs. 4 — 7). As the two limbs are drawn 

 apart from each other, the obliterated original cleft again becomes 

 manifest, and at the same time the humped end also splits across 

 (Fig. 8). Thus the chromosome as a whole splits longitudinally 

 and the apparent transverse division of the ring at this period 

 is the result of the complex rearrangement of the various parts in 

 the manner just described. The singular appearance presented by 

 the chromosomes of Lilies at the time when they are dividing has 

 already been described by one of us '), and the possibility was con- 

 templated that it might be due to the splitting asunder of a ring 

 such as we have here described. The difficulty of following out and 

 explaining the process in Lilies is due to the fact that the ring is 

 hardly ever widely opened as in Amphibia, but that the two sides 

 are so closely approximated as to reduce the space between them to 

 a mere cleft which is often by no means easy to trace. The final 

 longitudinal fission of the chromosome is sometimes efiected in another 



1) Fakmeh, Loc. cit. 



