192 BOTANICAL GAZETTE [MARCH 
as many as six layers of cells, three of which might become tapetal, 
while the others formed the sporangium wall. On the opposite 
side of the sporangium the tapetum was usually only one layer of 
cells in thickness, these evidently derived from the sporogenous 
mass (fig. 1). In longitudinal section of the sporangium, the spore 
mother cells were seen to lie in plates somewhat separated from 
each other (fig. 1). 
By the middle of February the separation of the plates of 
sporogenous cells from each other was marked (fig. 2). By this 
time the tapetal cells, especially those on the inner side of the spo- 
rangium, were beginning to break down and appeared as much 
distorted cells, while the tabular cells which formed the inner layers 
of the sporangium wall became very much flattened and took stain 
more readily (fig. 2). 
In the mature anther the cell walls had all disappeared except 
the well-developed endothecium inside of the epidermis. There 
are two pollen sacs which dehisce longitudinally with no definite 
stomium. 
The megasporangium 
The earliest stage observed showed the megasporangium as 
a rather massive protuberance inward from the side wall of the 
ovary cavity. On January 21 some of the ovules showed the 
beginning of the inner integument. By February 16 the outer 
integument had begun its development, and the growth of the 
ovule had been such that a median longitudinal line through the 
ovule had been shifted from practically perpendicular to the longi- 
tudinal axis of the carpel to a position nearly parallel with it. 
The growth of the integuments from this time on was somewhat 
irregular, sometimes the outer and sometimes the inner growing 
more rapidly. There was in either case apparently an excessive 
development of integuments, so that when the embryo sac was 
ready for fertilization they not only closed over the nucellus, but 
lay in folds above it (fig. 3). At this stage the innermost layer 
of cells of the inner integument had much greater cytoplasmic 
contents than the other cells. 
In the nucellus, which was relatively massive, there was evidently 
one archesporial cell picked out. The earliest stage examined 
