196 
Allen . — Nuclear Division in the 
transformation of the reticulum into the spirem, it is plain that the double 
nature of the latter is not due to a longitudinal splitting of a single thread ; 
but that the two threads are formed independently of each other out of the 
materials of the reticulum, and approach each other while they are in 
process of formation. From such appearances as that of Fig. 9, it seems 
that the approximation of the materials of the two threads may begin 
during the period of the reticulum, resulting, perhaps, in a fusion of some of the 
knots in pairs. But Figs. 10 and 11 make it plain that this approximation 
is brought about for the most part while the material of the knots is being 
distributed along the strands of the future spirem, and that the approxima- 
tion of these strands has some causal relation to the massing of the whole 
system in the synaptic figure. 
There seems to be no rule as to the side of the nuclear cavity which shall 
be occupied by the synaptic mass. In a longitudinal section of an anther 
sac, there are no more cases of nuclei in which the massing is on any 
particular side — e. g. toward one end of the sac — than there are in which 
it is toward any other side. As has been said, the greater part of the 
nucleolar material is flattened against the nuclear membrane, nearly always 
in the immediate neighbourhood of the aggregated spirem. Sometimes 
two or three flattened nucleoles appear in the same nucleus. All stages 
may be found in the transition from the rounded to the flattened form, so 
there can be no doubt that these bodies are the same as the nucleoles of 
the earlier stages. The vacuolated appearance is also often retained, even 
in cases of considerable flattening. The nucleoles shown in Figs. 11, 12, 
and 14, represent only moderate instances of the flattening, which often 
continues until the nucleole is an extremely thin plate, just inside the 
membrane, and extending, in section, around a quarter or even a third of 
the periphery of the nucleus. Seen in surface view, this plate has an 
irregular, lobed outline, and is perforated by occasional irregular holes. 
The nuclear membrane is still very lightly stained, though plainly present. 
Fig. 14 represents a median section of a cell whose nucleus is in the 
condition just described. Both cell and nucleus have reached approximately 
their final size. The nucleus has, in most cases, taken up an eccentric 
position within the cell. In case the cell is longer in one diameter, the 
nucleus is commonly nearer one end. The long axis of the nucleus generally, 
though not always, coincides with that of the cell. The cells are still 
angular, united into a tissue, with orange-staining walls of about the same 
thickness as in younger stages. The tapetal cells of the inner layer, 
bounding the mass of spore mother-cells, are very much vacuolated and 
often shrunken, the majority of them with their nuclei and the remaining 
cytoplasm at the end of the cell which is in contact with the spore 
mother-cells. 
Even at the stage represented in Figs. 13 and 14, the parallel threads 
