76 
Transactions of the Boyal Society of South Africa. 
with some emphasis. He says, indeed, that its structure does not differ from 
that of an ovary, but he proceeds : " Est-ce pour cela un ovaire ? Personne 
je crois, ne le soutiendra serieusement."* 
We may now consider the first whorl (Table 3), which is absent from 
the male flowers of Ephedra and G-netum, but elsewhere appears as an 
ovary. This identification is based upon the results of the authors' very 
detailed study of Welwitschia. f The evidence is mainly anatomical; the 
form of this envelope terminating in " a long style and a stigma " also 
has weight. + The authors even go so far as to describe this envelope as 
" un ovaire clos,"+ a description which is not justified by the structure. 
The anatomical evidence is furnished by four small vascular bundles,, 
which, in the male flower, pass into the base of the ovular envelope. Of 
these the two larger lie in the median, the two smaller in the tangential 
plane. The envelope is therefore regarded as an ovary formed by two con- 
crescent pairs of carpels. § With regard to these bundles on which such 
important conclusions are based, it may be noted that (1) at the level at 
which the envelope becomes free from the nucellus the two stronger are 
already extremely reduced 1| ; ( 2 ) their number do.^s not appear to be con- 
stant^[; (3) they do not always penetrate the base of the envelope.** 
For the inner envelope of the female flower of Welwitschia the evidence 
is more slender. This envelope admittedly arises as an equal ring from the 
base of the nucellus,tt already well advanced. Later the limb becomes 
lobed, the lobing being regarded as of no morphological significance.]: J The 
interpretation of the anatomy of this flower is " horriblement difficile." §§ In 
one case the anterior bundles of the floral axis " persistent meme suffisam- 
ment pour arriver a penetrer j usque dans la base du tegument interne 
(ovaire)." II II In others no bundles enter the free region of the envelope.^^ 
The authors regard this envelope as homologous with the ovular envelope 
of the male flower ; as an ovary of four carpels of which the lower pair are 
* Bertrand. loc. ext., p. 82. See also p. 65. 
t Lignier and Tison, 1912. 
X Loc. ext., pp. 100, 171. 
§ Loc. ci<.,p. 112. 
II Lignier and Tison, 1912, fig. 10. 
IF Sykes, 1910, p. 189. 
** Bertrand, 1878, p. 64 ; Sykes, loc, ext., p. 187. 
tt Lignier and Tison, 1912, p. 130 : Pearson, 1906, fig. 20 A. 
XX Lignier and Tison, loc. ext. The authors here state that this inner envelope is 
lightly lobed in the tangential plane. This statement is supported by a reference to 
their figure (C, fig. 24), in which the outer envelope (iii) is thus lobed. The inner 
envelope which is under discussion is not shown in this figure. 
§§ Lignier and Tison, 1912, p. 138. 
(Ill Loc. ext., p. 144. 
Loc. cit., p. 141 ; Sykes, 1910, pp. 198, 199. 
