On Some Stages in the Life History of Gnetum. 
261 
its composition. The septation of the sac in G-netum and Welwitschia 
and the formation of many fusion-nuclei may be consequences of this 
limitation. 
In the Angiosperm the fusion- nucleus usually undergoes a series of simul- 
taneous divisions and the protoplasm in which the resulting free nuclei lie 
becomes completely septate to uninucleate cells or incompletely to multi- 
nucleate compartments in which ultimately all the nuclei fuse.* But in 
some cases, as in Peperomia, cell-walls are formed in all the divisions of the 
fusion-nucleus and its descendants,t and the endosperm is cellular from 
the beginning and throughout its development. This is also charac- 
teristic both of G-netum and of Welwitschia. When, as is usual in the 
Angiosperms, endosperm begins with free nuclear division, a rudimentary 
plate often appears suggesting derivation from an endosperm in which 
nuclear division was followed by cell-formation," % as in Peperomia. 
Attention has frequently been called to the curious resemblance between 
the common behaviour of the fusion-nucleus and its products in the 
Angiosperms and the process by which the prothallus is formed in 
the macrospore of the Gymnosperm and the massive proembryo in the 
archegonium of the Cycads and Ginkgo. The endosperm of the Angio- 
sperm is clearly not phylogenetically related to both these latter structures, 
and this fact in itself makes it the more probable that the resemblance is 
no evidence for a phylogenetic relationship with either. It is much more 
likely that the whole explanation is to be sought in the similar conditions 
under which a single nucleus produces a multicellular tissue in a large cell. 
The " triple fusion," now recorded for an increasing number of Angio- 
sperms, is peculiar to this group. § Its occurrence is not confined to the 
8-nucleate sac ; it is observed in Cypripedium ; || a sperm-nucleus is one 
of the five constituents of the primary endosperm-nucleus in Euphorbia 
procera,% and it probably fuses with the seven sac-nuclei in Gunnera.** 
The participation of the second sperm -nucleus is therefore independent of the 
number of fusing-nuclei contributed by the sac itself. It is also clear that 
it is not of universal occurrence among the Angiosperms. It sometimes 
fails in species in which it normally occurs (as e.g. Naias tnajor),ff and 
* Strasburger, 1880, figs. 46-64 ; Tischler, 1900. 
t Johnson, 1914. 
X Coulter and Chamberlain, 1903, p. 172. 
§ The " double fertilisation " described for Juniperus virginiana (Ottley, 1909, 
p. 40) and for Thuja (Land, 1902) is probably not in the same category as the triple 
fusion of the Angiosperms (cif. Samuels, 1912, p. 90). It appears rather to be a case 
of the true fertilisation of a gamete which in most Archegoniatae is arrested, correctly 
described as " double fertilisation." Cf. Blackman, 1898. 
II Pace, 1907. 
If Modilewski, 1908, fig. 7. 
Samuels, 1912. 
tt Guignard, vide Coulter and Chamberlain, 1903, p. 216. 
