314 BOTANICAL GAZETTE [ocropER 
space in the protoplasm, but was more distinct than the outlines 
of the spores. The eighth spore of this ascus was in such a posi- 
tion that only part of its outline could be seen. In another 
ascus ( fig. 27) there were found eight well defined spores, the walls 
being now distinct. Seven of these were divided by single cen- 
tral transverse walls into two cells each. Each cell contained a 
single nucleus. The eighth spore was bent upon itself in such a 
way as to conceal one end and only one nucleus was visible in it. 
There can be little doubt, however, that another was present 
in the concealed end, and it may be asserted safely that at this 
stage the ascus contains uniseptate spores having a nucleus in 
each compartment. The septa in this case had been formed soon 
after the first nuclear division in the spore. In preparations 
from material slightly more mature, spores were found in which 
the number of nuclei varied from two to ten, corresponding im 
number usually to the number of cells in the spore. An excep- 
tion to this occurred in the case of one entire perithecium ( fig. 16). 
Here each spore contained from four to ten nuclei but no divid- 
ing walls could be seen. Either the staining was at fault, which 
seems improbable, or the formation of the wall does not always 
follow immediately upon nuclear division, but takes place instead 
after all the successive bipartitions of the nucleus are complete. 
The foregoing evidence, together with the fact that the ne 
occur constantly in pairs throughout the mycelium as well as th 
fruit, seemed conclusive proof that nuclear division does ee 
place, but for a long time no details of the process of division 
could be discovered, Later, cases were found in which the hya- 
line court had become elliptical in outline and contained some 
times one eccentric dark spot, and sometimes one at each end 
the ellipse. These at first seemed very similar, and the —_ 
in this condition corresponds closely with Trow’s (14) figures” 
what he calls “fragmentation” in the nucleus of Saprolegn® 
If his theory be correct then the central stained spot is pee 
a 
a chromatin plate, and by simple fragmentation aie all : 
plate for each daughter nucleus, and this grows quickly t 
size. This must be based upon the assumption that th 
ees 
