Rr ee ae ee ee ae ee ee 
THE STAMINATE FLOWER AND MICROSPORE 5 
spike. The flowers are contiguous before dehiscence, and some- 
times show a somewhat spiral arrangement. Each flower is 
composed of a nearly sessile group of one to six yellow, purple, or 
mottled anthers, whose filaments are entirely confluent. The 
anthers are crescent- or U-shaped. They are usually arranged in 
pairs with the concave sides together; but in case of odd-numbered 
groups, this arrangement is lost. In some cases the groups are 
borne on pedicels as much as 2 mm. long, a condition common 
among staminate flowers at the top of a pistillate spike (PLATE 4, 
FIG. 59). In his brief discussion Rowlee (25, p. 369) described the 
filaments as cohering, and the anthers as simple in structure. The 
meaning of this statement is not quite clear. The writer finds a 
tendency of the anthers to be two-celled. There are always at 
least two groups of primary archesporial cells, which in many 
anthers are confluent long before the pollen mother cells are formed. 
In other anthers the locules remain distinct almost up to the time 
of dehiscence (PLATE 2, FIG. 29). 
At the time of preparation for the heterotypic division of the 
pollen mother cells, the sporangium wall is composed of a single 
epidermal layer and three or four layers of sterile cells. The two 
outer layers of sterile cells form the wall proper, and the one or two 
inner layers are clearly differentiated as tapetum. The mature 
sporangium has a wall composed of an outer layer of epidermis, 
an inner layer of partly disorganized sterile cells, and between 
these two, a third layer composed of palisade-like cells with 
thickened walls (PLATE 2, FIG. 29). 
The two divisions of the pollen mother cell are as in the lily. 
The first division is followed by the formation of a wall before the 
second begins. All the cells of one locule show about the same 
stage of development; although the different flowers of one spike 
may at one time show all stages from resting nuclei to mature 
pollen spores. In 1899 Atkinson (2) studied the details of the re- 
duction division. The writer has nothing to add at this time, as 
questions of a purely cytological nature are out of the realm of the 
present work. The second division has not been studied in detail, 
but an examination of the preparations at hand has shown noth- 
. ing unusual. As reported by Atkinson (2, p. 5) the gametophytic 
nuclei show sixteen chromosomes. The tetrad is spherical with 
