FLOWER OF GNETUM 
165 
the lower (posterior) in origin and is in advance of it in all the later stages 
of its growth. Of two opposite primordia there is no sign whatever. When the 
lower (posterior) edge of the envelope and the anther-primordium are barely 
evidenced as slight convexities on the surface of the meristem, its upper edge 
(Fig. 9) is already quite distinct. Many young flowers in longitudinal 
section have been examined but no exception to this rule has been met 
with. The lateral parts of the collar are less advanced than the upper 
edge, more so than the lower (Fig. 10). Both the free edge of the envelope 
and its apparent insertion on the secondary axis are oblique from beginning 
to end. A comparison of the three stages of Fig. 10 shews that the short 
length of axis which in the adult flower (Fig. 5) stands below the origin 
of the envelope, is formed entirely by intercalary growth ; and that the 
deep fissure which separates the youngest flower from the undifferentiated 
meristem beneath it, is caused by this intercalary growth and the increase in 
bulk and surface area of the general meristem, which occur together. 
The collar grows more quickly than the anther primordia and eventually 
arches over them and completely encloses them save for a very narrow short 
slit (Fig. 11). This slit is displaced towards the lower (posterior) side of the 
flower by the more active growth of the upper part of the envelope ; the line 
of the slit is parallel to a line joining the centres of the two anthers (cf. 
Fig. 11 C and D). In consequence of the rapid growth of the antherophore, 
in its later stages, the edges of the envelope are forced apart and split. The 
splitting in G. Gnemon is usually irregular, but frequently into two approxi- 
mately equal lobes lying in the antero-posterior plane 1 (Fig. 3). This last 
method of opening is probably simply due to the fact that the main line of 
splitting is determined by the longer diameter of the two anthers — the back 
of each anther acting as a blunt cutting edge. 
The unequal development of the envelope is well known. Karsten 
figures a young male flower in which the upper edge is considerably in 
advance of the lower 2 ; he also describes similar relations in G. funiculare' 3 
and figures them in G. latifolium 4 . There is no indication that the incom- 
plete female flower exhibits the same tendency to a basipetal succession of 
its parts in development 5 , nor is it clear that it is formed by the same 
meristem which later produces the males (Fig. 2 B). But the study of 
earlier stages than are yet available might lead to a contrary conclusion. 
From the published figures of Ephedra it appears that the axial portion 
of the male flower is relatively slightly earlier developed than in Gnetum, 
and even in its very early stages the envelope either arises from the secondai \ 
axis or is concrescent with it. Nevertheless the upper (anterior) part of the 
1 Cf. Lotsy, 1911, fig. 
;i L.c. p. 344. 
213, 2, 4. 
5 L.c. Taf. viii, figs. 11, 
2 Karsten, 1893, Taf. viii, fig. 12. 
4 L.c. Taf. viii, figs. 8—10. 
16. 
