32 DR. J. D. HOOKER ON WELWITSCHIA. 
forty and more grains that I have found with their tubes all produced. Furthermore, 
as is stated above, I have found pollen-grains on the nucleus before the elongation of its 
integument, and this in ovules contained in the extremely immature uppermost scales 
of an otherwise half-mature cone—ovules which, as T have reason to suppose, are 
developed long after the usual period of fertilization. 
In connexion with this subject, 1 may mention that the nearly mature cones are often 
bored through and through after the manner of flower-buds attacked by the larvee of 
Curculionide, and that a pollen-feeding group of Coleoptera, the Cetonie, abound in the 
regions inhabited by Welwitschia. 
The pollen remains on the apex of the nucleus for some time before any change takes 
place in the embryo-sac 5 and its large solitary tubes are apparently slowly emitted, and 
slowly make their way down the dense tissue of the nucleus. I have never seen more 
than one tube emitted; but this, after emission, sometimes elongates in both directions, 
or even forms a crutch or fork, between the arms of which the empty wall of the pollen- 
grain persists (Plate IX. fig. 88). The tubes reach about one-third down the conical 
apex of the nucleus, apparently always keeping near its periphery ; they are terete and 
even, they never branch, and swell at the base, only when they touch the secondary 
embryo-sac. 
Changes after Fertilization These commence within the nucleus of the ovule after it 
has assumed the form, &c., represented in Plate VIII. fig. 7, Plate IX. figs. 8, 9, and 
are conspicuous after it arrives at the stage seen in Plate IX. figs. 11, 12, which represent 
ovules taken from a perianth such as is figured at fig. 10. From this time onwards the 
nucleus developes rapidly in all directions, but the parts above the embryo-sac grow less 
than those below it; and as the greatest increment of all takes place at the very base, 
below the level of the insertion of the integument, the latter is carried up, and, assuming 
a higher and higher relative position as the seed ripens, is found at last towards the apex 
of the seed. This mode of ovular development is common, in a greater or less degree, 
to all Gnetacee, Cycadee, and to many Conifere. 
At the period last mentioned, the embryo-sac (Plate VIII. fig. 18, & Plate X. fig. 2) is a 
delicate membrane, more or less loosely investing an ovoid, orbicular, or transversely 
oblong compressed mass of endosperm-cells, in which, however, there is no visible dif- 
ferentiation of parts, nor any defined sharp boundary to the individual cells. At the 
next stage (Plate X. fig. 1) the nucleus is elongated both upwards and downwards, is 
dilated opposite the sac, and its integument is carried up considerably above its base ; 
the contained sac is also elongated, and rather broader above than below, becoming ob- 
ovoid. Still later (fig. 4), the integument is carried up above the middle of the nucleus, 
and the yet more elongated sac has descended, as it were, below the level of the insertion 
of the integument. The nucleus is now divisible into two distinct portions—an upper 
one above the embryo-sac, which I shall call the cone, and a lower one, containing the 
embryo-sac, which T shall call the body of the nucleus: the insertion of the integument 
marks externally the positions of these parts. Sometime between the last two stages, the 
membrane of the embryo-sac is found to have disappeared over the summit of the endo- 
sperm, and blunt tubular processes (Plate X. fig. 3) appear rising above the endosperm- 
