72 
Transactions of the Moyal Society of So2ith Africa. 
terminates the axis of the female, as of the male spike.* The barren tip of 
the spike axis sometimes shows clear indications that it consists of a cupule 
fused with the true axis,t just as in the lower nodes the cupule and the 
axis sometimes remain concrescent.;}: The subapical cupule may therefore 
surround, even be concrescent with, a barren axis-tip; it does very frequently 
surround the base of a truly terminal flower. Therefore the simplest 
explanation of the cupule (envelope IV) described by MM. Lignier and Tison 
is that it is situated on the axis which bears the upper envelopes, i. e. the 
axis of the flower is the direct prolongation of the axis of the spike. The 
marked peculiarity of this cupule (envelope lY) is that it has assumed the 
form of a supplementary envelope, as stated in conclusion 2. 
The presence of a group of male flowers and hairs above the insertion of 
envelope II and of a group of hairs only in a similar position with regard to 
envelope III, seems to establish the first part of conclusion 4. Now for the 
first time we appear to be on quite safe ground in regarding the two outer 
envelopes of the complete (the single outer envelope of the incomplete) 
female flower of Gnetum as leaf-pairs homologous with the cupules of the- 
spike. 
Consideration of the second part of conclusion 4 may be deferred for th& 
moment. 
Conclusions 5 and 6 are theoretically of great interest. For the sake 
of comparison with Table 3 the former may be expressed as follows : 
Table 1. 
Gnetum 9 (complete!, 
nucellus. 
envelope I ( = " plurilocular " ovary) . 
not present. 
envelope " 1 ^ (cupules of Gnetum spike), 
envelope III J 
There is undoubtedly much to be said for this comparison ; but however 
great the probability of this or any other comparison between the flowers of 
G-netum and Welwitschia, the facts needed to establish it are not at present 
available. Apart from the identification of the innermost envelope as an 
ovary, which is considered below, perhaps the point which most invitee 
discussion is the treatment of the audroecium. MM. Lignier and Tison 
were obviously forced to consider the possibility that the group of male 
flowers standing above envelope II in their abnormal flower of Gnetum, was 
phylogenetically derived from the same primitive stru&ture which has given 
rise to the androecial whorl of Welwitschia. They have decided against this, 
* Strasburger, 1872, p. 158 ; Pearson, 1915 C, p. 159. 
t Caporn, 1916. 
X Pearson, loc. cit., fig. 8. 
Welwitschia <J . 
Nucellus = 
Envelope I = 
Andrcecium 
Inner perianth = 
Outer perianth = 
