73 



Now as the process of mitosis advances, the relative proportion of those 

 chromosomes which are closed at both ends increases, and this is 

 especially true of the stage just before the advent of the achromatic 

 spindle. 



During the division of the spermatocytes of Triton and many 

 other animals (notably Amphibia) the formation of rings or elliptical 

 bodies is very marked in the earlier stages of the heterotype mitosis 

 (Fig. 2 a); in Salamander, and especially in Triton , they become 

 marvellously clear as the spindle is formed. At this stage, in Lilies, 

 each ellipse bends over on itself at the middle so as to bring its two 

 ends more or less into apposition (Fig. 4). It then assumes such a 

 position on the spindle that the two approximated ends point out- 

 wards (Figs. 5, 5i), whilst the spot where the bending occurred is 

 situated on the spindle and is directed inwards. Several modifications 

 may be introduced which depend on the extent to which the chromo- 

 some is doubled on itself; sometimes it is hardly bent at all, and 

 then it appears looked at from the pole, as a rod lying tangentially at 

 right angles to the spindle axis. If the doubling has been completely 

 effected it appears as a thick rod radiating from the spindle, and if, 

 as is more commonly the case, the approximation of the two ends is 

 incomplete, it resembles the letter Y, the vertical bar of which is directed 

 inwards (see Figs. 5, 6). The latter is a particularly instructive, as 

 well as a very frequent case, and has been figured by several obser- 

 vers. Indeed in Lilium Martagon, it is perhaps the commonest 

 form met with. Now it is obvious that it is extremely difficult to 

 harmonise such a figure with the view that the diverging limbs repre- 

 sent the original longitudinal halves of the chromosome and that they 

 will subsequently lie along the spindle and form the two daughter 

 chromosomes. Indeed such a view is negatived by the fact that the 

 appearance still persists after the formation of the spindle. For it 

 will be obvious that in order to fit in with any such hypothesis the 

 chromosome must first rotate through an angle of 90" before it can 

 place its limbs along the achromatic spindle. The difficulty will be 

 easily grasped by a reference to the figures 13, 14, in the memoir by 

 GuiGNARD already referred to. The longitudinal split (there shown as 

 extending throughout the chromosome) is seen in Fig. 13 from a 

 polar view, but in Fig. 14 a split is also seen in profile view. It is 

 obvious that these two fissions cannot be identical, since they occur 

 at the same stage, in planes at right angles to each othe But 

 further, everyone who is familiar with nucleus structures will admit 

 that the body of the chromosome lies peripherally on the spindle, 



