186 BOTANICAL GAZETTE : [MARCH 
an almost inconceivable regularity. The granules would have to be 
definite and fixed, of the same number in both the egg and sperm, and 
not be subject to increase or diminution (jigs. 14-16). In the early 
stages of the spirem there are, of course, numerous instances of threads 
lying parallel, side by side, but these appearances are equally common 
after the chromatin granules appear double (jigs. 15, 17). 
The ribbon now begins to show an arrangement into definite loops 
(fig. 10). It becomes much shorter and thicker (jigs. 18, 19) and finally 
shows a definite twisting together into twelve loops, which have their 
heads toward the nuclear wall (jig. 20). At this stage the chromatin 
- granules can still be distinguished lying side by side (fig. 21). But at 
this time the whole ribbon begins to undergo a change, so that it stains 
of a uniform, dense color throughout, and before the loops separate 
all evidence of chromatin granules is lost (figs. 22, 23). That the loops 
shown in figs. 18-23 are the incipient chromosomes is self-evident. 
By no manner of interpretation can such a conclusion be explained 
away. By the time the twelve loops have separated, the nucleoli 
have entirely disappeared from the nuclear cavity (fig. 24). The nucle- 
oli break up into micronucleoli and are thrown out into the cytoplasm 
(figs. 38, 51). The figures in which they do not appear were taken 
from material stained in such a way that the nucleoli were not evident 
or not very distinct. 
The chromosomes are exceedingly interesting on account of the | 
many fantastic shapes produced by the coiling of the ribbon. A series 
of distinct shapes is given in jigs. 25-37. Occasionally the loop shows 
evidence of its double nature (jig. 29), but usually it appears homo- 
geneous throughout. Much time was spent in a study of these chro- 
mosomes, and the variety of shape and coil could be extended indefi- 
nitely. In all cases the chromosome is continuous, the outer end 
of the loop always being closed. Occasionally the coiling takes place 
in such a way as to form a double loop (figs. 26, 27, 35). The chro- 
mosome is situated on the spindle with its head or closed end outward 
(figs. 28-30). Sometimes it is very difficult to unravel the nature of 
the coil, as in fig. 35. In other chromosomes it is an easy matter to 
follow out the details of the loop, as in fig. 28. 
The spindle threads are evidently attached some distance back of 
the free limbs of the loop (figs. 25-28). The limbs are gradually 
