AFFINITIES OF DAVIDIA INVOLUCRATA. 315 
6. STRUCTURE OF THE OVULE AND SEED. 
The ovule consists of the nucellus and an extra-nucellar tissue, which may be regarded 
as the congenitally fused integument and raphe. Just prior to megaspore germination 
it is more or less ovoid, with a curved, obeuneate nucellus at its apex, pointing upwards 
and inclined a little towards the axis, the free portion of the integument appearing as an 
irregular neck around the nucellar bulge, more strongly developed on the abaxial side. 
At the time of megaspore germination this neck develops over the nucellar surface, 
leaving a circular aperture on either the right or left hand side—the micropyle. At this 
time the “hinge” area of the ovule is about halfway between its own apex and base. 
The raphal tissues lie on the placental side merging at the base into the chalaza. At 
maturity the point of attachment is towards the apex of the ovule owing to extensive 
chalazal development—the free part of the integument being the portion between the 
point of attachment and the apex. The integument arches over a little towards the 
abaxial side, its concave under surface roofing the micropyle, which is more or less 
triangular in section. The ovule takes up finally a radial, vertical position in the 
loculus—the vascular bundle of the raphe being between the placenta and the nucellus. 
The micropyle is directed upwards and laterally (either to right or left). A similar 
position of the ovule at maturity is characteristic of the Araliaceæ, U mbellifere, certain 
genera of Cornacese, Amygdalus, and the Spireas. The mode of development of the 
integument and the position of the micropyle in early stages suggest an oblique insertion 
of the nucellus as in Avalia (masked at maturity by the radial position taken up by the 
ovule). The nucellus is of medium bulk and consists of about three concentric vertieal 
or axial rows within the epidermis. The material proved not young enough to show the 
formation of these rows, the earliest stages being just prior to megaspore germination. 
They doubtless originate as a subepidermal plate of cells in the usual way. The cells of 
the rows resemble one another very closely in appearance and in staining reactions. 
Sooner or later certain cells, either of the same or of different rows, became active at the 
expense of the surrounding sporogenous tissue (sterile megaspores) in the way described 
by Benson * and others (Pl. 31. fig. 14). There is considerable variation in behaviour of 
the sporogenous tissue. The axial rows are rarely completely sterile." From this 
condition there are all gradations up to the germination of four or five megaspores, and 
perhaps more (Pl. 31. figs. 18, 19, 21). The germinating megaspores of any sagem ovule 
rarely give rise to more than one complete embryo-sac. Of the six, seven, or eight ovules 
in any one ovary very few were found with typical embryo-sacs. There are, however, 
all gradations between sterile ovaries and ovaries with all their ovules fertile— 
from the evidence of germinating seeds. The abnormalities in connection with the 
embryo-sac are numerous, relating to the number, disposition, and form of the nuclei. 
Saes may attain the two- and four-celled stage without nuclear differentiation, and then 
disintegrate (Pl. 31. fig. 18). Frequently the nuclei of sacs in the four-celled stage 
resemble synergids in the micropylar end and antipodals in the chalazal end. Again, 
* Benson in Trans. Linn. Soc., 2nd ser. Bot. iii. (1894) p. 412 &c. 
