350 BOTANICAL GAZETTE [NOVEMBER 
are few in number, probably not exceeding the number of chro- 
mosomes. 
In many cases the point at the pole. toward which the spin- 
dle fibers converge was occupied by a granule both in the first 
and second division spindles. This granule in well-stained 
preparations was always dark, but its inconstant occurrence 
was decidedly against its being considered a permanent struc- 
ture. The cell plate forms before the spindles of the second 
division (jig. 36). 
The body, which looks so much like a nucleolus, disappears 
previous to the first division at almost the same time as does 
the nucleolar membrane. At this period it presents a more of 
less irregular and lobed appearance, but vanishes so quickly that 
it was impossible to determine whether the process was one of 
fragmentation or solution. 
SECOND NUCLEAR DIVISION OF THE MOTHER-CELL. 
Before the second division there is a distinct resting stage 
An indistinct membrane is formed, and even a nucleolar body 
may appear. This latter, however, never becomes so large as in 
the archesporial nucleus, and often seems to be entirely absent, 
or at least indistinguishable from the chromosomes (fig: 3): 
A thick linin thread is usually formed, but the chromosomes 
remain distinct. In this character Potomageton agres® ” 
with Convallaria. During this resting stage it is again possible 
to count the chromosomes, when the number is Still found to ae 
seven or eight. This resting nucleus can be distinguished ea 
by its much larger size from the one formed after the secon 
division. 
The origin of the spindle could not be trac 
preparations showed it in the mature condition 
These spindles are smaller and more slender 
described above. Like the latter, they have very P 
and in both cases the poles are almost if not quite in con 
the cell wall. 
The chromosomes are closely aggregated in t 
ees well 
ed, but many 
(fig. 37 ). 
than those 
ointed poles, 
tact w! 
he nucleat 
