364 
DUNCAN S. JOHNSON 
The tapetal or parietal cell early divides by i or 2 periclines and two 
or three series of anticlines to form the 2 or 3 layers of parietal tissue 
found in the young seed above the sac (figs. 74, 77, 81, 89; also John- 
son, 1907, fig. 2). As the seed ripens the walls of these cells become 
but slightly thickened. The cells of the layer next the integument* 
often divide by a few more anticlines and elongate radially to form a 
series of columnar cells just below the micropyle (fig. 104). In the 
ripe seed the layer of parietal cells next the embryo has been crushed 
by its growth, the remaining layers persist intact and plump, but 
contrast with the perisperm cells in having few starch grains in them. 
At germination these parietal cells are pushed off, along with the 
style of the fruit, by the swelling of the endosperm as it bursts the 
seedcoats (see p. 375; fig. iii). 
The definitive archesporial cell, or embryo sac mother-cell, is 
at the time of its formation nearly cubical in shape. It is about 
20 yi long and 16 /i wide. Its nucleus when first organized is about 
10 IX in diameter, has a distinct wall, a prominent nucleolus and a rather 
close chromatin net just within the wall (fig. 58). As development 
proceeds this embryo sac mother-cell grows, until, when ready for 
the next nuclear division, it may be 30-45 [jl long and 25-30 /x broad. 
The nucleus itself continues to grow, up to and even after the time of 
synapsis. During the early phases of synapsis the nucleolus may 
have a diameter of 3 or 4 /x. Just before synapsis sets in the chromatin 
of the nucleus lies near the periphery in the form of a net, the threads 
of which have a diameter of .3 or .4 ti and form meshes i to 3 across 
(figs. 58, 59). The cytoplasm up to this time retains a uniform 
structure, with rather evenly distributed small vacuoles. 
2. The megaspore and embryo sac. 
The development of the embryo sac from the definitive arche- 
sporial cell is initiated by the occurrence of a genuine synapsis. The 
first evidences of the beginning of this process are seen in the increase 
in thickness of the chromatin threads of the reticulum, in the reduction 
in the number of meshes and in the gradual shrinking away of the 
chromatin from most of the nuclear wall. Later the whole chromatin 
thread forms but i or 2 close tangles of contracted loops near one side 
of the nuclear cavity (figs. 59, 60, 61). The examples of this stage 
seen do not give evidence of a doubling or splitting of the chromatin 
thread, before the compact knot stage is reached (figs. 60, 61). On 
emerging from the characteristic tight knot the chromatin takes the 
