130 BOTANICAL GAZETTE [FEBRUARY 
the ovules are first beginning to be evident, the microsporangial 
tissue is well defined. 
There are, as usual, four microsporangia in early stages (fig. 
23), which later fuse to form the two pollen sacs (fig. 28). The early 
stages show the stamens bent inward and downward, but not 
connected with each other. Later growth, particularly of the 
connectives, makes a continuous tube, which is, however, easily 
separated into the constituent stamens (fig. 28). This tube extends 
beyond the sporangia, which occur on the side toward the perianth 
wall. In mature stages the pollen grains are slightly oval. They 
are very pale, almost transparent, with a pale green cast, due to 
little bodies, probably fat, many of which also occur in the perianth . 
parts. The microspores are loose and free, not massed together 
as in pollinia of orchids. Before shedding, the generative nucleus 
divides (fig. 29). A granular mass in the spore suggests the presence 
of a prothallial cell, but failure to secure stages in the present 
investigation must leave this doubtful at present. 
The ovules, very many in number, develop on parietal placentae 
(fig. 24), which swing free from the ovary walls in the center (fig. 
18). At this level ovules project on all sides of the placental 
column. In early stages the numerous primordia appear as in 
fig. 18. The fully developed ovule shows two integuments (figs. 
26-27). It is anatropous, with a long funiculus. As JoHOW (26) 
and TREUB (30), and lately ERNst and BERNARD, reported, there 
is a conspicuous differentiation of a few of the nucellar cells at 
the base of the embryo sac. This seems to have no significance at 
present. 
The seed is minute, with a testa two cells in thickness. The 
outer layer is composed of large, almost transparent cells. The 
inner one is constructed of smaller cells, with more contents, often 
appearing oil-like. The seed has a very evident endosperm, with 
cells of relatively large diameter, and an inconspicuous embryo 
of a few cells (fig. 30). In all respects it seems to agree with the 
accounts of TrEuB (30) and JoHow (25, 26), the former of whom 
first correctly interpreted the endosperm. Mr:Ers, in 1866, declared 
that the seed of Myostoma contained no embryo. Later, GRIFFITH 
interpreted the entire content of the seed as embryo. TREUB, 
