398 BOTANICAL GAZETTE [NOVEMBER 
mature sac is of the usual organization (fig. 46). The outer end 
_ enlarges greatly, crushing the parietal and adjacent nucellar cells, 
and comes to lie in contact with the epidermis of the nucellus. The 
extreme inner end remains small through this and subsequent 
development, and in it the 3 antipodal cells are cut off by walls. 
While they persist until the embryo is of considerable size, they 
never manifest any activity (fig. 47). The polar nuclei fuse early 
(figs. 44-47), and the fusion nucleus lies well toward the outer 
end of the sac. The egg apparatus is typical. 
DEVELOPMENT OF .STAMENS.—The stamen primordia are at 
first oval in cross-section (fig. 8), but early differentiation of the 
archesporial cells makes them somewhat rectangular, and sets off 
the anther region from the filament. The archespores are single 
rows of cells (fig. 48); each soon divides by a periclinal wall to form 
the primary parietal cell and the primary microsporogenous cell 
(fig. 49), then both divide anticlinally. Frequently the anticlinal 
division precedes the periclinal (fig. 50). There are 2 periclinal 
divisions in the primary parietal cell, and an appropriate number of 
transverse divisions to keep pace with the rapidly elongating 
anther. The first (fig. 54) sets off a layer that finally differentiates 
into a well marked endothecium with characteristic spiral thicken- 
ings; the second (fig. 57) forms the middle wall layer and the tape- 
tum, and takes place about the same time as the last division in the 
sporogenous tissue. The middle layer is soon crushed and obliter- 
ated (fig. 58). 
The tapetal cells appear to behave in all possible ways. Accord- 
ing to BonnETT (1) the nucleus usually divides twice by normal 
mitoses, after which there may occur a great variety of nuclear 
fusions and abnormal mitoses. In some cases there is but one 
division. Usually in Rumex crispus the tapetum becomes bi- 
nucleate; often it is multinucleate; and now and then there may 
appear large irregular nuclei with many nucleoli, as if formed by 
the fusion of a number of small nuclei (fig. 61). Apparently the 
normal condition is for the tapetum to persist as a functioning 
nutritive layer up to the liberation of the microspores from the 
tetrad (fig. 63); but in keeping with the widespread degenerations 
occurring throughout the flower, it may begin to degenerate while 
