On Some Stages in the Life History of Gnetum. 
247 
nuclei may escape enclosure in a compartment and so play no further part 
in the history of the sac. If fertilisation is not effected, endosperm-formation 
proceeds without interruption until it fills the sac. 
Occasionally, as in Welwitschia, an endosperm compartment encloses an 
unusually small number of free nuclei. In such a case fusion is certainly 
delayed and probably does not occur at all ; the cell undergoes a great 
increase in size and probably in the end collapses under the pressure of the 
surrounding tissue. Apart from such cases, which in Gnetum africanum 
appear to be rare, the primary endosperm [compartments always enclose 
many free sac-nuclei, so far as is known, the primary endosperm-cell being 
later constituted by nuclear fusion. In the seed the endosperm-cells are 
very frequently multinucleate, the nuclei showing signs of deterioration. 
This condition is possibly attained by nuclear fragmentation. The endo- 
sperm organises two growing points, one at each end, and an axial column of 
small cells among which the growth of the proembryos mainly takes place. 
4. The Oospore and Proemhryo. 
In G. africanum the oospore almost certainly gives rise to a small mass 
of proembryos. Each proembryo becomes a long tortuous tubular cell 
containing a number of free nuclei, as indicated by Earsten, Lotsy, and 
Coulter for other species. 
When the outstanding facts of the life-history of Welwitschia were 
presented certain theoretical questions, arising from the peculiar characters 
of endosperm found in that genus, were discussed.* More particularly, 
attention was directed to the possibility of the endosperm of the Angiosperm 
being derived from a more primitive form whose main characters are pre- 
served in Welwitschia. The use of the characters of the sac of Welwitschia 
in attempting to establish a hypothesis of this nature was always open to 
the criticism that their value as evidence of phylogeny is problematic, owing 
to the aberrant characters of this highly-specialised genus. But the value 
of the evidence furnished by Welwitschia can no longer be doubted now 
that it is established that the same type of endosperm, in a less specialised 
form, occurs also in G-netum. While the case from the side of the G-netales 
has been thus strengthened, later researches have not justified certain con- 
clusions then adopted regarding the Angiosperm-sac. Also there seems 
now to be a possibility of relating the Grnetum and Welwitschia type of 
endosperm to the prothallus of the lower gymnosperm — a possibility which 
was not then clear. Consequently it seems desirable to restate the case in 
the light of such new facts as appear to be relevant. f 
* Pearson, 1909. 
t The endosperm-hypothesis as stated in 1909 has been reviewed by Lotsy (1911) 
and by Samuels (1912). It is rejected by the former (see Pearson, 1915 a, pp. 324- 
327), and favourably considered by the latter, 
18 
