37 2 Gates and Rees . — A Cy to logical Study of 
nucleus of this cell has probably been formed by the close approach to each 
other of the inner ends of the two spindles present during the second mitosis, 
leading to the coalescence of the two inner daughter nuclei into one at the 
time of their formation. This frequently happens in tapetal cells owing to 
the crowding of the spindles, so that a number of the tapetal cells are trinu- 
cleate after the second mitosis. In Fig. 82 this central fusion nucleus is in 
a somewhat later synaptic condition than the two end nuclei. The fact that 
synizesis occurs in these cells cannot, we think, be attributed to the fixation, 
but must rather be an indication of a peculiar physiological condition of 
these nuclei. 
Fig. 11 shows a mother-cell in the typical synizesis stage. Even with 
the most careful treatment and differentiation, the synaptic knot is invariably 
so dense that it is impossible to trace the spireme thread for any distance. 
Soon, however, threads begin to appear at the margin, and the whole mass 
is seen to be composed of looser and thicker threads, until a stage is reached j 
such as shown in optical section in Fig. 12. The spireme emerges as 
a thick uniform thread which is apparently more or less continuous through- 
out its length, though occasionally free ends maybe observed. No trace of 
any parallelism of the threads has been found in these early stages, the 
thread appearing uniform and unsplit, even when weakly stained. As the j 
spireme spreads out,, through the nuclear cavity after synizesis, definite loops : 
very soon make their appearance, as seen in Figs. 18-23. At first it is 
impossible to make an accurate count of these loops. Figs. 18-20 show 
early stages in the process of looping. The sides of the loops come to 
be more or less parallel, and it will be seen that the two sides of a loop are j 
also beginning to twist round each other. This latter feature becomes more 
and more evident in the later stages. As the process of looping continues, ( 
it is possible to make a rough count of the number of these loops, and it is 
found to be approximately nine, i. e. the number corresponding to the 
gametic number of chromosomes (Figs. 23-29). Finally the loops gradually 
become detached from each other. There is at this time no definite 
polarization of the spireme loops. They do not radiate regularly from the 
centre of the nucleus, nor are they polarized in any other way corresponding 
to the ‘ bouquet ’ stage which occurs in the presence of a centrosome ; also no 
definite second contraction stage has been observed, such as occurs in 
various other plants. The loops continue to be well distributed through the 
nucleus. 
While the spireme, now a pachynema, is for the most part uniform in 
appearance during this period, it is by no means universally so. As the 
figures show, thinner threads are occasionally found connecting thicker 
portions of uniform thickness. Also some portions may have a beaded 
appearance owing to the alternation of lighter and darker areas along the 
threads. But none of these appearances when critically studied permit of 
