232 Tra7isactions of the Royal Society of South A frica. 
for G. Ula* Pollination in G. africanum appears very frequently to fail. 
Lotsy had the same experience with G. Gnemon,-f due in this case to the 
local rarity of male trees of a species long cultivated in the east+ for 
the sake of the edible leaves and seeds. G. africanum, however, is not 
cultivated, and the failure of pollination here must be due to some degree 
of lack of adaptation to its environment. The fact that in this species the 
whole contents of the embryo- sac have frequently disappeared entirely is 
probably due to the exhaustion of available nutriment by the few ovules 
which do succeed in attaining their full size, with or without proembryos. 
The failure of fertilisation does not appear to affect adversely the develop- 
ment of endosperm and associated seed structures other than the embryo. 
Lotsy § records that in G. Gnemon the embryo- sac becomes filled with 
endosperm if no fertilisation occurs. In G. africanum, seeds whose external 
characters are indistinguishable from those in which many proembryos are 
found, frequently contain a normally developed endosperm in which no 
proembryos appear to be present. || The proembryos are not conspicuous, 
but with a little experience they are readily recognised, and their absence in 
these cases may safely be afiirmed. 
Another difficulty of constant occurrence in all our material is the 
almost synchronous development of all the ovules of the spike. In the 
early stages of the male spike the successive nodes do not develop in 
acropetal order,^ and presumably the female spikejhas the same character.** 
Later, the ovules at successive nodes are in practically the same stage of 
development at a given time ; still later one or more of them outpace the 
rest, which then develop no further.ff It is further probable that all the 
ovules on a particular plant at a given time represent a remarkably close 
developmental series. Certain it is that many ovules representing a single 
collection and presumably obtained from the same plant have yielded very 
few stages, and these close together. 
Finally, if we may judge from our material of G. africanum, fertilisation 
and the changes which immediately follow it are rapidly completed when 
once the pollen- tube has reached the sac. We have seen both pollen- tubes 
and free-nucleate proembryos in scores, if not in hundreds, but not a single 
case of fertilisation or of the first division of the oospore ; oospores themselves 
are recognised in only two of our preparations. 
* Lotsy, 1899, 1903. 
t Lotsy, 1899. 
X Karsten, 1893 a, p. 204. 
§ Lotsy, 1899, p. 107. 
II Cf. Thoday, 1911, p. 1101. 
% Pearson, 1915 b, fig. 2 a. 
** Cf. Lotsy, 1899, fig. 10. 
ft Cf. Lotsy, loc. cit., figs. 5, 6, 7. 
