DR. J. D. HOOKER ON WELWITSCHIA. 39 
gates further, and reaches the sac (Hoffmeister, /. c. p. 415). Perhaps the whole course 
of the pollen-tube in Welwitschia corresponds with its first growth in Conifere, and the 
necessity for a second elongation is obviated by the ascent of the secondary embryo-sac. 
In Conifere the lower part of the pollen-tube always dilates more or less as it approaches 
the embryo-sac; in Welwitschia it behaves as in angiospermous plants, only dilating, if 
at all, on reaching the secondary sac. 
‘My very imperfect knowledge of the embryogeny of Gnetwm does not enable me to 
compare it with that of Welwitschia; but in G. scandens, G. Brunonianum, and G. Gne- 
mon I find many remarkable points of coincidence with it. In Gnetum, the cone of 
the nucleus is comparatively very small indeed ; the embryo-sac is early filled with 
endosperm, of which the upper cells are in G. scandens so very large, lax, and oblong, that 
I at first supposed them to be young secondary sacs, but abandoned the idea in favour of 
numerous filiform tubular delicate cells that traverse the endosperm longitudinally, 
within the embryo-sac. These thread-like secondary sacs(?) closely resemble those 
of Welwitschia in many respects, and they appear to be developed from cells which 
originally are not distinguishable from the surrounding endosperm-cells ; but they want 
the bulbous base, are found at all heights in the axis of the endosperm, and even towards 
its base, they sometimes branch downwards, and very few of them ascend and project 
beyond the mouth of the sac. Of those that did so ascend, one or two appeared to me 
either to enter the cone or to adhere to its base—I could not make out which; and I 
thought I found a pollen-tube adherent to the apex of one. Asin Welwitschia, the apex 
of the sac in Gnetwm seems to disappear early, and the large endosperm-cells to protrude 
from its cavity. I have failed to detect any canals or structure of any sort in the cone; 
and I have never found an embryo or funiculus in the very young state of the seed. 
From angiospermous plants Welwitschia differs in the naked ovule, free embryo-sac 
full of endosperm previous to fertilization, in the presence of secondary sacs, in the 
position of the germinal vesicle at their base, and in the compound highly developed 
suspensor; it, however, agrees with them in the germinal vesicle giving rise to one 
embryo only. 
There is a remarkable analogy in one respect between the processes of fecundation 
in Welwitschia and Santalum*, and still more in Loranthust, as described by Griffith, in 
both which genera the embryo-sac is produced beyond the nucleus of the ovule, ascends 
in the cavity of the ovarium, sometimes penetrating up the style, where it meets the pol- 
len-tube descending an open cavity leading downwards from the stigma. The canaliform 
style of Loranthus thus representing the canal in the cone of the nucleus of Welwitschia, 
it follows that in this respect the latter plant presents an intermediate stage between 
most Angiosperms, in which the pollen-tube perforates style and nucleus to reach the 
embryo-sac, and most Gymnosperms, in which it perforates the nucleus and embryo-sac 
too, and Loranthacee, in which it perforates the style only. If, as I suspect may be 
* See Griffith in Linn. Trans. xviii. 59 & 71; xix. 171 & 487; also Henfrey, ibid. xxii. 69. 
+ Linn. Trans. xix. 178. The late Prof. Henfrey possibly alludes to this, when he observes (Linn. Trans. xxiii. 299) 
that his investigations on Gnetum lead him to regard favourably the opinion expressed by Prof. J. G. Agardh, that the 
Gnetacee are related even more closely to Loranthacee than to Conifere. 
‘i 
' 
t 
t 
