071 Some Stages in the Life History of Gnetum. 257 
striking, and any explanation of the origin of the Angiosperm-sac and of 
the significance of the endosperm-fusion must take them into account. At 
least the following conclusions are clearly indicated, viz. : (1) The number 
of the sac-nuclei is not constant ; the functions of the sac may be ade- 
quately performed by 4, 8, or 16 sac-nuclei ; (2) the degree of differentiation 
exhibited by the sac-nuclei varies within wide limits ; (3) the gamete char- 
acters are not always resident in the oosphere alone ; (4) the production of 
the primary endosperm-nucleus by the fusion of two polar-nuclei is not in 
any way dependent upon any specific characters present in the latter and 
absent from other sac-nuclei. 
The first three of these conclusions, which are completely established by 
the facts, constitute evidence in favour of the view that the Angiosperm-sac 
is descended from one in which the nuclei were not differentiated at all. 
Such a sac is known in Gnetum and perhaps also in Juglans *— a sac in 
which all the free nuclei are believed to be potential gametes, i. e. actually 
capable of being fertilised. It is therefore suggested that in exactly the 
same sense as the oosphere-nucleus of the Angiosperm is homologous with 
the female gamete of G-netum, so the other sac-nuclei of the Angiosperm are 
homologous with those of Gnetum, and therefore the endosperm of the 
Angiosperm is homologous with that of Gnetum. With regard to the last, 
Coulter t lays stress on the fact that the endosperm of the Angiosperm is 
undifferentiated — " growth, not organisation." Its undifferentiated character 
compared with Gnetum (p. 243) is explicable if it be a reduced form. The 
fact that five of the Angiosperm-nuclei have usually acquired new functions 
which prevent them from being fertilised and from participating in the 
endosperm-fusion does not affect this homology, though it places difiiculties 
in the way of its recognition. 
A marked feature of the sac in probably the majority of the Angiosperms 
is its polarity, which has usually exercised a considerable influence upon the 
theories advanced to explain the origin of the sac in this group. It is con- 
ceivable that this character has favoured the development of that degree of 
specialisation which we find among the sac-nuclei. The egg-apparatus, in 
which the retention of sexual characters is most marked, generally occupies 
the position which is most favourable to fertilisation ; the antipodal group, 
in which vegetative characters are most strongly developed, is placed where 
it can most efl&ciently perform a nutritive function. With regard to the 
polarity itself, it may be noted that it could not be strongly marked in a sac 
crowded with nuclei like that of Gnetum ; the fact that there are few nuclei 
in the sacs of most Angiosperms may of course be quite unconnected with 
their strong polarity, but when there are many (Euphorbia virgata, Feperomia 
Mspidula, Juglans, etc.) they are not usually segregated into groups at the 
* Nawaschin, 1895 ; Karsten, 1902. 
t Coulter, 1911. 
