?5 



way. This depends very much on the character of the early stages, 

 and also on the manner in which it becomes attached to the spindle. 



1) In those cases in which the chromosome is an ellipse opened 

 at one end (Fig. 3b), the long sides sometimes lie parallel, but 

 sometimes are twisted over each other. When they are parallel, the 

 chromosome may bend at its middle part, so that the two ends are 

 brought into close juxta-position, and then there is every reason to 

 suppose that the same series of changes is passed through as those 

 we have already described for the closed ring- or elliptical-chromo- 

 somes. But often the long axis of the chromosome lies in the equa- 

 torial plane of the spindle, and is directed radially outward from it 

 (Fig. 10), In this case the spindle fibres become attached to each of 

 the two longitudinal halves at some point near the end of the chromo- 

 some. The greater portion of the entire body projects out from the 

 spindle. When this is the case curious forms arise which will be 

 easily understood by reference to Figs. 10, 11. 



It should be pointed out that sometimes the closed elliptical 

 chromosomes also may be attached to the spindle in this manner 

 (Fig. 9), and we thus explain the appearances which are sometimes 

 to be met with in an obviously young chromosome with incurved ends 

 which are still attached to one another (Fig. 9). Sooner or later, 

 however, they will split asunder, and follow the same course as in 

 those which begin as open bodies. Not infrequently also the chromo- 

 somes in which the ellipse is opened at both ends goes through the 

 same stages. In any case the further fate of the two diverging halves 

 is simple in that they are pulled apart to the poles, either straighten- 

 ing out along the spindle (a rare occurrence) or increasing their in- 

 ward curvature as the primitive longitudinal split opens out. But 

 they appear never to give rise to the figure shewn in our sketch 

 (Fig. 8) and it is to the existence of this type of fission just described, 

 that the omission of the form shewn in Fig. 8 (a by no means un- 

 frequent occurrence) is due. 



2) In those cases in which the chromosome-halves are twisted 

 round each other, the figures produced are often complex. The chromo- 

 some is open at one end only (Fig. 12), and the twist is a slight one, 

 so that the ends project over each other but a little way (Fig. 3 ft), 

 it commonly becomes attached to the spindle at the point of the 

 crossing of the limbs. Sometimes here also, the limbs straighten 

 themselves out along the spindle, but oftener they curve outwards, 

 and are attached to the fibres by their inner convex surfaces (Fig. 15)^ 

 Then as they diverge farther their free ends usually become curved 



