58 BOTANICAL GAZETIE 
its outline usually remains rather definite until late stages ( figs. 
46, 47, 48, 49,50). This same condition was found in Pistia Sira- 
a: 
Fic. 
antipodal nucleus is disintegrating. 
51. Two-celled sac in which 
Four - celled sac, >< 
Fic. 53. Sac in which egg ap- 
paratus is formed; probably polar 
nuclei below egg apparatus. X 1375. 
Fic. 54. Sac with egg apparatus 
and pollen tube. X 1375. 
- > which (hea 
In fig. 57 is shown a sac in the two-celled stage, in ni : 
micropylar nucleus is normal, while the antipodal nucleus 15 : 
[JANUARY 
totes; while in Allium ‘Canadense 
(19) the nucellus is said to dis- 
appear quite early. 
Up to this point the sequence — 
is quite regular, but later there 
appear many irregularities of such 
a nature as to indicate general 
unfavorable conditions for normal 
Although after 
prolonged search I succeeded in 
finding preparations showing the 
usual behavior of the angiosperm 
embryo sac, these preparations 
constituted a very small part 0 
development. 
those which should be considered — 
in studying the real conditions of 
the embryo sac of L. minor. a 
only exceptional cases in whic 
sacs beyond the one-celled — 
do not give some evidence of dis- 
organization. The sections 
selected for drawing do not fully 
represent these conditions, ee 
they were selected to illustrate 
normal as well as abnormal occur 
rences. 
: r- 
integrating. Both nuclei may again divide, forming the fou 
celled sac ( jig. 52), and each of these may divide in the 7 : 
It is quite common to have the micropylar 1u of 4 
develop normally, while the antipodal nuclei disintegrate Ns tt 
Manner, 
42, 40, 51, 53, 54). 
. rm 
he polar nuclei may fuse to form the primary endosper™” 
' 
boiled 
