of Pilularia globulifera , L. 247 
cells undergo division by horizontal walls, but in the central 
part all the walls are vertical, so that the central cell of the 
archegonium, and later the embryo, are only separated from 
the cavity of the spore by a single layer of cells. 
In the meantime the cells of the upper part of the prothal- 
lium have also been undergoing rapid divisions. In the peri- 
pheral cells are formed numerous radially placed vertical walls 
(PI. XIII, Fig. 34), so that the central cell, seen from above, 
appears surrounded by a single circle of small cells. The 
central cell next divides by a wall parallel to its outer surface 
(PI. XIII, Fig. 29), the outer cell being the mother-cell of the 
neck of the archegonium, the inner one giving rise to the 
oosphere and canal-cells. The contents of all the peripheral 
cells are less uniform than those in the central cell; and the 
nuclei are much smaller. 
The development of the archegonium proceeds as follows. 
The mother-cell of the neck becomes divided by two cross- 
walls into four equal cells, and soon after the central cell has 
a small discoid cell, the primary canal-cell, cut off at the top 
(PI. XIII, Fig. 31). 
Up to this time, about thirty hours from the beginning of 
germination, the prothallium has increased but little in size 
and is still completely enclosed in the spore, and all the cells, 
including the central one, are very much flattened. A rapid 
growth in height nowbegins. Thecells, which hitherto have been 
divided by vertical walls for the most part, now form horizontal 
walls, and at the same time increase in height, so that the 
young prothallium rapidly assumes its completed form. Each 
of the four primary neck-cells divides by a transverse septum 
into two, and the upper cells so formed project when full grown 
as a colorless papilla beyond the spore-membrane. As the neck 
increases in length the canal-cell elongates with it, and the 
ventral canal-cell is formed (PI. XIII, Fig. 32). It was impos- 
sible to get specimens where the nuclear division was taking 
place, which would of course determine the matter posi- 
tively, but from the relative position of the walls in the canal- 
cells one would certainly conclude that the ventral canal-cell 
