496 BOTANICAL GAZETTE [JUNE 
because I can think of nothing more likely. Saxton (5) has since 
met the same difficulty and has hazarded the same guess. It may 
be that the mitoses are both siniultaneous and passed through with 
extreme rapidity, and that it is merely chance that they have been 
missed. No indications of amitosis have been observed. Above 
the 64-nucleate stage the numbers are not regular, being usually 
somewhat less than the exact power of 2. Enlargement of the sac 
and multiplication of the nuclei continue up to the latter part of 
January. At this time there are more than 2000 free nuclei present. 
The cytoplasm still remains extremely scanty. Figs. 9-16 show 
the progress of development at intervals of about a month. 
In January a change in the method of development occurs. 
Without any considerable increase of nuclei, the cytoplasm increases 
rapidly. As soon as it has become somewhat thicker, vacuoles 
make their appearance. The result (fig. 17) is that the central 
cavity is surrounded by a rapidly thickening sac of vacuolated pro- 
toplasm, with the nuclei largely confined to the inner border (fig. 18), 
In many cases the walls between the vacuoles break through. 
leaving the inner plasma membrane connected to the outer one 
merely by tenuous strands. The nuclei usually lie at the points 
where these strands join the inner plasma membrane. A few are 
found at the outer membrane, especially at the micropylar end of 
the sac. The protoplasm is thicker at this end also. This process 
goes forward very rapidly. The inner border advances on the 
central vacuole and the nuclei multiply somewhat. They now pass 
out along the cytoplasmic strands (figs. 19-21). By the time the 
inner border has closed up on the vacuole completely (fig. 20), most 
of the nuclei have migrated outward. 
With the continued increase of the cytoplasm, most of it remains 
in the peripheral regions, especially near the micropylar end. It 
collects along the strands and plates until distinct uninucleate 
vacuolated sacs are formed. The nuclei are now generally sus- 
pended by still more delicate strands in the central portion of the 
sac. Fig. 22 shows how these sacs behave under the action of the 
killing reagents. Each sac appears to have its own wall of inclosing 
protoplasm capable of being separated from that of its neighbor. 
Mitoses now occur plentifully in preparations of the peripheral 
