918 MR. A. S. HORNE ON THE STRUCTURE AND 
The antipodals disappear with the elaboration of the vascular system of the seed—whep 
the conditions are different—and the endosperm is nourished uniformly just as the 
embryo in its turn js nourished within the endosperm. 
It would appear that the endosperm is a walled tissue ab initio. "This is so in an eight- 
celled endosperm of Sambucus and in slightly older stages of Cornus and Davidia. Ina 
four-celled pro-embryo stage of Davidia the sac is filled fairly uniformly with meri- 
stematie endosperm tissue— the pro-embryo being terminal. The pro-embryo is 
gradually displaced from its terminal position by the elongation of the sac and a corre- 
sponding encroachment of the endosperm, which soon, in this part of the sae, becomes 
more compact than in the main portion of more actively meristematic cells. The 
endosperm cells of Davidia, Cornus, and the Caprifoliaceze are very similar in appearance : 
the nucleus is fairly large and contains four or five prominent nucleoli, and it is embedded 
in a vacuolate eytoplasmie sac. "The cells are irregular in outline, owing to intermittent 
division during elongation of the sac, and the cell-walls are thin. As time goes on a 
compact tissue is formed which sooner or later becomes packed with reserve food. 
The youngest primary embryo observed consisted of four cells formed by a transverse 
and a longitudinal division (Pl. 31. fig. 25). A short suspensor, some four cells in section, 
was observed in a ten-celled pro-embryo (PI. 31.fig. 26). "The primary embryos are distin- 
guishable ez masse from endosperm tissue in being slightly more compact ; also the cells 
stain slightly darker and the nuclei are a little larger. The proximal endosperm cells, 
however, are frequently closely adpressed to those of the primary embryo and take a 
slightly darker stain than the subproximal endosperm. It is on this account difficult to 
distinguish between the two tissues. As the suspensor grows longer, the true embryo 
begins to develop as a small undifferentiated mass of cells. At the same time the 
basal cells of the suspensor disintegrate, leaving a bulbous cavity in the endosperm 
(Pl. 31. fig. 27). These basal cells clearly function as an absorbent organ as in the 
analogous case of the Alsinoidez described by Gibbs*. By the time they have dis- 
appeared the environment of the primary embryo has changed; the surrounding 
endosperm cells are relatively passive in so far as nuclear division is concerned and are 
in intimate contact with the suspensor and embryo. After a time the embryo begins 
to take upon itself the absorptive function of the suspensor. 
At the time when the primary embryo consists of some fifty or sixty cells the middle 
region of the extra-nucellar tissue shows signs of degeneration. The gap in my material 
between this stage and that of the fully developed fruit prevents me from giving a complete 
account of the developmental and nutritive processes. A much later stage in Benthamid, 
however, in which the fully stored endosperm does not appear to have been appreciably 
attacked, leads me to the opinion that the outer region with its peripheral vascular 
supply persists until the endosperm is organized as a storage tissue. 
The fruit at the time of germination consists of an exceedingly hard stony endocarp 
formed by the sclerosis of all the ovarian tissue within and including that of the inner 
ring of vascular bundles. It is very difficult to remove the seeds from the endocarp, since 
"* Gibbs in Ann. Bot, xxi. (1907) p. 40. 
| 
i 
| 
| 
: 
| 
