1918] OTTLEY—IMPATIENS 207 
In De l’ovule, WARMING (45) notes a few exceptions to the usual 
order of development of the integuments and gives Viola, Ficus, 
Convallaria, and Orchis as having two integuments which appear 
to grow as a single organ, and Tropaeolum as having at first two 
integuments which later appear as one. WARMING quotes STRAS- 
BURGER as saying that in Delphinium the integuments originate 
as one and elevate themselves as a unit; later at the summit the 
two integuments become distinct. Judging from his figure the 
conditions in Delphinium are much the same as in J. Sultani. 
As given earlier, the megaspore mother cell when it arises is 
completely surrounded, except at its base, by the epidermis of the 
nucellus, and the developing embryo sac also continues to lie in 
direct contact with the epidermis (figs. 2-4, 6, 8, 9). During the 
2-nucleate stage of the embryo sac the epidermis begins to break 
down, The disintegration first appears as a flattening of the cells 
and nuclei just below the apex, and then extends gradually to the 
base of the embryo sac (figs. 9, 10, 14, 37). In Oxalis corniculata, 
according to HAMMOND (25), the epidermis, which in this case 
serves as a tapetum, begins to disintegrate before the 2-nucleate 
embryo sac is formed. 
The apical cells of the epidermis are often longer-lived than 
those just below them, for it is quite common to find two or three 
cells surmounting the embryo sac and connected by only a line 
with those still persisting about the center of the sac (fig. 12). 
A somewhat similar appearance has been described by SmiTH (40) 
for Eriocaulon septangulare, where “the nucellar tissue lateral to the 
megaspores breaks down and is absorbed by the growing embryo 
sac. A few of the apical cells of the nucellus persist for a long time 
and enlarging assume the appearance of a tapetum. These too 
are ultimately absorbed and the embryo sac abuts directly upon 
the inner integument and micropyle.”’ In J. Sultani, however, 
these apical sibaaeilasss cells do not enlarge and seem to have no 
special function. 
While the epidermis continues to disappear it leaves but a line 
around the upper half of the sac (fig. 10). As the disintegration 
progresses downward, the cells near the base of the sac possess 
their normal tabular shape, while those nearer the middle of the 
