BROWN: EMBRYO-SAC AND EMBRYO IN PHASEOLUS 537 
among them (PLATE 25, FIG. 1). This cell is typically five-sided 
in section; it contains a very large nucleus and stains more deeply 
than the adjacent cells. At a later stage the young macrospore 
mother cell is separated from the epidermal layer by another layer 
of cells (Fic. 4). Still later, when the macrospore mother cell is 
fully grown, there are often two layers of cells between it and the 
epidermis (FIG. 5); this is not always the case, however, for not 
infrequently the elongated macrospore mother cell is in the third 
instead of the fourth layer of cells. No division figures were seen 
either in the first differentiated hypodermal cell or in the cells of 
the epidermal layer. For this reason, I have been unable to 
determine whether the hypodermal cell that is early distinguished 
by its size itself functions as the macrospore mother cell, or whether, 
on the other hand, this hypodermal cell divides, one of its daughter 
cells becoming the macrospore mother cell. On the whole, the 
arrangement of the subepidermal layers at the later stages (Fic. 
5) supports rather more strongly the former hypothesis. Prepara- 
tions were obtained in which there were what seemed to be two 
young macrospore mother cells in an axial row (Fic. 3); another 
preparation showed two young mother cells lying side by side, 
and in one case two fully grown mother cells lay side by side. In 
no case, however, was the further development of more than one: 
macrospore mother cell observed. 
The fully grown macrospore mother cell is about three times _ 
as long as wide (FIG. 6); its nucleus is near the micropylar end of 
the cell; the chalazal end is usually pointed (Fic. 6), but in some 
cases quite rounded (Fic. 7). The nucleus remains in the micro- 
pylar end of the cell during the prophases of the ensuing division. 
Fic. 7 shows the nucleus in synapsis. One preparation was ob- 
tained showing the heterotypic division; the spindle. lies approxi- 
mately in the center of the cell (Fic. 8). One of the two daughter 
cells formed by this division fails to undergo a second division, 
since, so far as my preparations show, a row of but three macro- 
spores is formed (Fic.9). Guignard (1881) reported the formation 
of but three macrospores in Phaseolus multiflorus but did not deter- 
mine which of the two daughter cells, formed from the division of 
the mother cell, fails to divide; he also found a case in which a longi- 
tudinal division took place in one of the functionless macrospores; 
