659 
Balls .— The Mechanism of Nticlear Division. 
seem that the black dots enlarge later, becoming the nodes of the reticulum 
in some cases. The wall of the nucleus seems to be entirely cytoplasmic in 
origin. Those spindle fibres which are divided by the wall fade away 
centripetally up to the wall, while the portion of each fibre inside persists, 
for a time at least, ending blindly in the wall at one end, and still connected 
at the other end with the chromosome from which it was formed. 
The method of formation of those spindle fibres which run from pole 
to pole without bearing chromosomes has not yet been mentioned. They 
do not appear to be common in the nuclei of the cotton-plant ; their mode 
of formation is the same as that of the other fibres, except that they are 
drawn out from the separating thread-rings at points other than those where 
the chromosomes are found. 
The nuclei of the vegetative cells of the cotton-plant are very minute. 1 
A few observations have been made on them, using the root-tip of seedlings, 
but the ordinary vegetative cells of the flower tissues are far too small 
for study. 
The scheme of division outlined above seems to apply to these cells 
also, with structural modifications due to the smaller circumference of the 
spireme relatively to the nucleus. The spireme has about the same 
diameter as the nucleus itself, and the achromatic threads are thicker; 
there are, consequently, no tangled loops to confuse the observer, but the 
larger number of chromosomes compensates for this. The chromosomes 
are not grouped to one side, but are symmetrically distributed in prophase 
around the periphery of the nucleus (Fig. 11). The two spireme halves 
separate directly in parallel planes, instead of rotating over one another 
as in the reduction division, being evenly tied together by the spindle fibres. 
The spindle thus produced (Fig. 12) is very fat and broad, while the rings 
at each end are frequently quite conspicuous. The actual separation of the 
spireme halves, and the isolation of the chromosomes, is thus more easily 
recognized (Fig. 11) than it is in the reduction division, although the nucleus 
is smaller. At the same time, the absence of loose loops of the thread-rings 
prevents the appearance of any striking structures. The c dots ' are often 
very distinct, and sometimes of appreciable size, being almost fit for descrip¬ 
tion as plasmosomes. 
The large plasmosome, or nucleolus, may either disappear or persist. 
In the latter case it may present appearances which suggest Wager’s 
figures of the nucleolus in Phaseolus , 2 excepting that his figures do not 
show the plasmosome to be in connexion with the thread-rings by means of 
thin filaments. These filaments give it the slightly stellate (Fig. 11) appear¬ 
ance. About equally common is another arrangement (Fig. 13), in which 
the plasmosome is merely a local dilatation of the thread-ring. 
1 About 10 /x in diameter. 2 Ann. Bot., 1903. 
