346 E. L. Stephens. 
moves to the lower end of the sac to fuse with the lower one (Figs. 4 
and 5); in several cases the nuclei were found to have completely 
fused before fertilisation (Fig. 8). The three antipodal nuclei 
decrease in size and quickly lose their staining properties. Ill- 
defined cells are formed around them (Fig. 5), completely filling 
the narrow lower end of the embryo-sac, and whilst the rest of 
the sac grows rapidly, this end remains small. It consequently can 
be recognised, at a slightly later stage, only as a small pocket at 
the antipodal end of the sac ; while as the sac increases still more in 
size, at the expense of the antipodal nucellar tissue, it disappears 
along with this tissue (Fig. 6a). Even before this happens, the nuclei 
have degenerated, so that they are rarely recognisable at later 
stages than that of Fig. 5. The antipodal cells thus apparently 
take no part in the activities of the embryo-sac. The egg-apparatus 
is of the usual type. The nuclei of the synergids, with the 
surrounding protoplasm, take up stains very deeply, each synergid 
thus usually forming a deeply-staining mass with a small clear 
apical portion (Figs. 5, 6a). 
The nucellus is composed of a central strand of elongated 
cells surrounded by starch-filled parenchyma. The embryo-sac, 
elongating laterally downwards, digests this parenchyma, and 
simultaneously large grains of starch begin to appear in the proto¬ 
plasmic layer lining the sac (Figs. 5, 6a). These grains again 
begin to disappear as the embryo-sac attains its maximum 
development before fertilisation and the starch-bearing tissue 
becomes used up (Fig. 7). 
A copious endosperm is formed after fertilisation, filling up the 
embryo-sac. It is only partly resorbed by the developing embryo, 
which lies in a sap-containing cavity closely surrounded by it (Figs. 
9—11). 
As in the Penasaceae, there is no suspensor. The pro-embryo 
is at first pear-shaped (Fig. 9), but later becomes spherical (Fig. 
10), the whole of it entering into the composition of the embryo. 
The cotyledons, which are long and linear, become differentiated at 
a much earlier stage (Fig. 11), and are much better developed than 
is the case in the Penaeaceae. 
The chief point of interest in connection with the embryology 
of this plant lies in a comparison with the closely allied Penaeaceae, 
there having been some discussion 1 as to whether the peculiar type 
1 see Stephens, E. L. The Embryo-Sac and Embryo of certain 
Penaeaceae. Ann. Bot., 1909, XXIII., p. 363—378. 
