Female Strobilus in Gnetum Gnemon. 991 
of these vascular structures might be taken as evidence for the hypothesis 
that the whole family originally possessed bisexual flowers. There are indi- 
cations that the organs in the whorl were grouped in .four sets of approxi- 
mately three members, for the main bundles just above the insertion of the 
complexes are linked together by anastomosing strands in this manner. 
Hence the whorl may have consisted of four members of threefold nature 
similar to the microsporophylls of the Welwitschia flower. The structure 
of the male inflorescence in Gnetum , however, points to an association of 
male flowers rather than of microsporophylls with the ovule ; hence it seems 
best to regard the female flower as representing at the present day a partial 
inflorescence and not a proanthostrobilus. 
It follows from such a view that the apparently simple spike now 
constituting the whole female inflorescence of Gnetum is really of com- 
pound nature, consisting primitively of a central axis with a series of cupules 
from the axils of which sprang secondary axes bearing a ring of male 
flowers and a single terminal female flower. From such a compound 
inflorescence, moreover, could be derived an inflorescence closely resembling, 
both morphologically and anatomically, the male spikes of Gnetum Gnemon , 
G. scandens , and other Indo-Malayan species. Suppression of the first 
internode of the secondary axis, a modification which constantly occurs in 
many species of Ephedra , would bring the terminal female flower and the 
ring of male flowers into the axil of the cupule, and the outermost protective 
covering of the partial inflorescence, being now unnecessary, would disappear. 
The shortening of the first internode of the vegetative branches is accom- 
panied in Ephedra altissima and frequently in E. nebrodensis by a shifting 
of the axillary buds of the first node from their normal lateral position to 
the abaxial side of the branch. The suppression of the internodes in the 
branches of the primitive inflorescence of Gnetum may have been similarly 
accompanied by a shifting of the male flowers so that they have become 
crowded together in the axil of the cupule below the abortive ovule, and 
reduced in number. 
In the vegetative branches of Gnetum Gnemon suppression of the first 
internode is not clearly evident, but the presence of small buds in the axil 
between the branch and its subtending leaf, basipetally developed and 
deriving their vascular supply from the bundles of the branch alone, may 
indicate that the same thing occurs here also. 
It seems probable therefore that Gnetum originally bore compound 
bisexual inflorescences from which both male and female spikes have been 
derived. The former have preserved in certain Indo-Malayan species 
their bisexual character, but the latter have as a rule lost all trace of it, for 
the strobilus of Gnetum Gnemon alone, as far as is known, retains in its series 
of vascular complexes surrounding the base of the ovule some vestige of 
its original complicated structure. 
