Nov., 1912.] The Reduction Division in Fuchsia. 
9 
to be seen. As division advances the thread continues to elongate 
up to a certain stage when it is rapidly thrown into loops (Figs. 12, 
13, 14), and begins to shorten and thicken constantly until the 
chromosomes are fully formed (Figs. 13, 15, 17, 21, 22). In the 
very earliest stages of the looping (Figs. 13, 14, 15, 16) the spirem 
may be traced for a long distance and the loops are found on the 
upper and lower surface of the nuclear wall showing that the 
loops are formed along the periphery of the nucleus and not as 
loops sticking out from a synaptic knot into the nuclear cavity. 
However, in the earlier stages there is a considerable crossing 
of threads in the center. Figure 13 shows four or more well 
defined loops already formed. These loops and the thread of 
which they are formed are still rather delicate. The nucleolus is 
in the center of the nuclear cavity. In some cases the cell wall 
begins to become somewhat indistinct at about this stage, 
(Fig. 12), but in others the wall remains well defined until the 
tetrads are fully formed inside the original cell. In most cases the 
sporocytes have not separated from each other nor from the 
tapetal layer, and have in consequence, not yet rounded up. 
The loops of the thread are formed in just such a way as loops 
would be formed in a heavy string if two ends of the string were 
held between the fingers and then twisted; twisting both ends in 
opposite directions. Some of these loops showed more than one 
twist. As the loops become tighter the spirem often appears as 
though it contained prominent knots. The granules are still 
very evident on the spirem where much looping has taken place 
but at this stage no doubling was visible. This does not neces¬ 
sarily indicate that division has not taken place; the granules 
may be lying too close together to be separated with the magni¬ 
fication used, or the differentiation possble with the sarfanin- 
haemotoxylin stain. As the looping proceeds the granules 
become less and less prominent until on the fully formed loops no 
granules are to be seen (Figs. 15, 16, 17). The loops finally break 
apart to form the bivalent chromosomes (Fig. 17). While chro¬ 
mosome formation is going on the nuclear cavity is apparently 
still enlarging (Figs. 15-19), but later as the nuclear wall disap¬ 
pears, the cytoplasm encroaches rapidly and fills the area around 
the contracting group of chromosomes (Figs. 20-24). Just about 
this time the sporocytes begin to separate from each other and 
assume a more rounded shape and the nuclear wall becomes 
more delicate. When all the loops are formed they lie around the 
periphery of the nucleus and can readily be seen and counted by 
focusing up and down. In each case the drawings were made 
from cells whose complete nucleus showed and had been undis¬ 
turbed in the cutting. It was somewhat difficult to draw correctly 
those loops which were to the side of the nucleus where it was 
often impossible to see the actual shape. In some cases the ends 
