On Some Stages in the Life History of Gnetum. 
255 
the Coniferae and that they are not confined to this group. An embryo-sac, 
in which all three have reached their utmost limits, would contain free 
nuclei, each nucleus being capable of functioning as a a^amete — the condition 
believed to be realised in G-netum and Welwitschia before septation. 
In the primitive condition of such a sac it is probable that fertilisation 
would occur while all the nuclei were still free — as in species of Gnetum. 
The free gamete-nuclei which remained unfertilised would constitute the 
only material from which a new nutritive tissue for the needs of the embryo 
could be organised. Of the factors which led to the septation of the sac- 
protoplasm and the fusion of the free nuclei enclosed in each compartment 
so formed, we have no knowledge. As a result of such a series of reductions 
as has been assumed, new physical and physiological conditions would pre- 
vail in the sac. It is possible that the causes inducing the new behaviour 
both of the protoplasm and of the free nuclei were inherent in the physical 
and physiological conditions prevailing m a large cell under uniform 
pressure with vacuolated protoplasm and many nuclei, none of which 
possessed a full complement of vegetative characters. 
There is not sufiicient evidence at present available to justify the direct 
derivation of the embryo-sac of Gnetum and Welwitschia from that of the 
Coniferales, The seed appeared independently in more than one group. 
The tendencies which seem to be present in the evolution of the prothallus 
in the Conifers and less certainly in the Cycads, probably existed in all 
primitive seed-bearing groups, and may have prevailed in more than one. 
It is true that Ephedra " is most reasonably connected with the Coni- 
ferales," * and that striking comparisons can be instituted between Gnetum 
and certain Conifers. f But the relationships of Gnetum and Welwitschia 
to Ephedra are by no means clear, and Ephedra possesses very marked 
characters which do not occur in the Coniferales. Land's conclusion that 
" there is no proof that the Gnetales have been derived from or are directly 
related to any living group " I is still fully justified. 
A morphological comparison of the embryo-sac and endosperm of 
Gnetum with those of the Angiosperm is inevitable, because (1) the primary 
endosperm-nucleus or nuclei are in both cases constituted by the fusion of 
nuclei related in the same manner and degree to the functional gamete ; (2) 
the physiological results of the fusion are closely similar; and (3) these, 
with Welwitschia, are the only known cases in which endosperm is formed 
in this manner. § 
* Coulter and Chamberlain, 1910, p. 402. 
t Cf. Saxton, 1913, p. 340. 
X Land, 1907, p. 288. 
§ Land (1907, pp. 278, 279) calls attention to the occurrence in Ephedra of an 
intra-archegonial " ephemeral nutritive mass " whose origin and structure are 
obscure. He sees in this " a suggestion of the origin of endosperm ... in the 
