FLOWER OF GNETUM 
157 
The male inflorescence of G. scandens is probably correctly regarded 
as a stage in the reduction from the type of G. Gnemon to those of G. 
africanum, in which female flowers are very rarely present 1 , and G. Buch- 
holzianum in which female flowers are not known to occur 1 . Since 
complete and fertile female flowers frequently occur in the male in- 
florescence of G. Gnemon, in which they occupy positions otherwise filled 
by incomplete female flowers, the latter may be regarded as derived from 
the former by partial arrest ; in the absence of any direct evidence to the 
contrary, it is hardly possible to take any other view 2 . If this be so, 
the pure male inflorescence of G. Buchholzianum is the final result of the 
reduction of the female flowers of a functionally bisexual inflorescence; and 
intermediate stages in this reduction are preserved in G. Gnemon, G. scandens 
and G. africanum. 
In G. Gnemon (Fig. 2) and in G. funicular e s and probably in all species in 
which female flowers are present in the male inflorescence, the female flowers 
precede the male in ontogeny at each node. Assuming the existence of a 
functionally bisexual inflorescence in which all the female flowers were 
potentially fertile, an arrest of the nodal meristem immediately after the 
organisation of the female flowers would give the pure female inflorescence 
which apparently occurs in all species. If the male inflorescence has been 
derived by the partial, and finally the complete, arrest of the female flowers, 
it almost follows that the female inflorescence owes its origin to a similar 
arrest of the male flowers of a functionally bisexual inflorescence. It may be 
that stages shewing a preliminary reduction in the organisation of the male 
flowers will yet be described. 
The large number of male flowers formed in basipetal succession at each 
node of the male inflorescence is probably to be regarded as a derived 
character. Its biological result is comparable to that produced by a profuse 
branching of the inflorescence as a whole. The necessity for the production 
of male flowers in greater profusion may well have arisen as a consequence ol 
the introduction of the unisexual inflorescence. The basipetal development 
of lateral members appears to be confined to axes of limited growth . W ell 
known cases are those of the soral axis of terns of the “ Gradatae section 
and of many floral axes bearing numerous stamens 1 '. Iheie is veiy little 
information regarding the growth of the axis of the Gnetum infloiescence, 
but the fact that all the nodes are probably established before the develop- 
ment of the flowers commences (cf. Fig. 2) leads to the conclusion that me 
also the basipetal succession of the male flowers is associated with limited 
apical growth of the axis which bears them. 
1 Pearson, 1912, p. 608. 
3 Karsten, 1893, Taf. vm, fig. 11. 
5 Bower, 1908. 
2 But see Yon Wettstein, 1907. 
•* Goebel, 1900, i, 41. 
o Goebel, 1905, ii, p. 542, fig. 369. 
