Balls.—■The Mechanism of Nuclear Division. 657 
The mode of formation of these chromosomes rather suggests a trans¬ 
ference of chromatin from the nucleolus to the chromosome areas by pro¬ 
gressive chemical change in the original paired granules. 
The chromosomes are bunched together at one side of the nucleus. 
There is no regular peripheral distribution. In this respect cotton differs 
from most organisms, and from its own vegetative cells. These chromo¬ 
somes are minute, being only 0^6 micron in diameter, while the univalent 
chromosomes are correspondingly smaller still. In consequence of this 
fact the achromatic substance of the nucleus is quite conspicuous, in spite 
of its slight stain affinity. 
The nucleus of the microspore mother-cell thus consists at this stage 
of twenty chromosomes grouped to one side, and these chromosomes lie in 
a continuous split spireme thread which is composed of achromatic 
substance. The nuclear wall has faded away, and the periphery of the 
nucleus is occupied by a densely granular zone of cytoplasm, which has 
been visible outside the wall for some time. 
The next stage is the most important one, and also the most difficult 
one to observe. It is completed rapidly—to judge by its comparative 
rarity—and the achromatic structures form a most complex tangle. The 
sequence of events appears to be as follows :—The two halves of the split 
spireme thread move apart from one another at the side where the chromo¬ 
somes are lying, but this separation does not affect the chromosomes. The 
chromosomes are isolated by the removal of the spireme halves, but not 
entirely, for continuity with the latter is maintained by thin filaments on 
either side (Fig. 5). The insertion of these filaments (the young spindle 
fibres) in the spireme halves causes slight swellings which appear to the eye 
as black dots. 
From this point I propose to refer to the spireme halves as the 4 thread- 
rings retaining the term 4 fibre ’ for its usual subject. 
The two thread-rings continue to separate, the fibres becoming longer 
and longer, until the part of each thread-ring which bears the dots has 
moved to a pole of the nucleus. The dots are scattered round some ioo° 
of arc, and the remainder of each thread-ring forms an irregular tangle of 
loops, which lies between the centrally suspended chromosomes and the 
granular cytoplasm (Fig. 6). The looping of thread-rings seems to be due 
to the fact that the circumference of the close spireme is greater than the 
circumference of the nucleus. This stage constitutes the multipolar spindle. 
The next event is the contraction of the dotted portions of the thread- 
rings, bringing the dots nearer together, and forming the bipolar spindle 
of metaphase (Fig. 7). The looped thread-rings lying in the clear zone 
between spindle and granular cytoplasm are quite conspicuous ; this was 
the observation which initiated the present research. It sometimes happens 
that a stray dot is not drawn up into the cluster at the pole as soon as it 
