260 Transactions of the Royal Society of South Africa. 
P. hispidula is of particular interest ; in each case the fusing-nuclei are 
all those which are free. In Gunner a macrophylla there are seven free 
nuclei which all fuse.* In the 8-nucleate sac of Alchemilla sericata, the 
antipodal nuclei remain free and all four fuse with the upper polar, f These 
facts certainly afford strong support to the opinion already expressed by 
Coulter, that the " conditions in the embryo-sac favour fusion of any free 
nuclei, in any number and of any origin." % But if the whole explanation 
of the act of fusion is found in the fact that the participants are free to 
respond to conditions favouring fusion, these conditions being inherent in 
the structure of the sac, the results of the fusion itself may nevertheless 
be of profound importance. By hypothesis, the fusing-nuclei in their 
primitive form were potential gametes ; in any particular case they may 
or may not have undergone a measure of reduction. If they have so far 
reverted to the conditions of vegetative nuclei that they can directly produce 
a nutritive tissue, then an endosperm may be formed without fusion. 
Otherwise the formation of an endosperm appears to be dependent upon 
nuclear fusion which is comparable with the act of normal fertilisation 
in so far as it produces a stimulus to vegetative growth. § We should, 
therefore, expect to meet cases in which the endosperm is eliminated if 
no fusion occurs, as in Went's Podostemaceae ; || and others in which the 
vegetative characters of some of the sac-nuclei are sufficiently developed 
to enable them to produce a vegetative tissue without fusion, as in Helosis 
and a few other forms. ^ A vegetative tissue so formed by the polar or 
any other of the sac-nuclei is, on the view adopted here, the morphological 
equivalent of that which frequently results from the division of the 
antipodal-cells. The sac-nuclei from which it arises have reverted to the 
vegetative condition from which they have emerged in Gnetum. The 
tissue is gametophytic though it lacks one of the essential characters of 
the prothallus in that its organisation is necessarily later than that of the 
gamete, which therefore it does not bear. 
In the Angiosperm usually only one fusion- nucleus is constituted, but 
there may be two**; in G-netum there are many, and the sac becomes 
septate before fusion. In the Angiosperm no more than 14 sac-nuclei are 
known to enter into this fusion; in G-netum and Welwitschia some hundreds 
are concerned. In view of the later history of the fusion-nucleus there 
is probably a limit to the number of gamete-nuclei which can enter into 
* Samuels, loc. ext., fig-. 18. 
t Vide Coulter and Chamberlain, 1903, fig. 42. 
X Coulter, 1911. 
§ Cf. Sargant, 1900, p. 705 ; Pearson, 1909, p. 384. 
II Went, 1908. 
^ Vide Coulter and Chamberlain, 1903, pp. 166, 167. 
** Campbell, 1911, p. 786. 
