AND STERCULIACE.E. 119 
views I stated many years ago (Hook. Kew Journ. Bot. i. 358), 
and of which I have seen no refutation, these glands are repre- 
sented by the anther-cells, the petiole by the filament and con- 
nective, and the lamina either totally abortive or represented by 
petaloid appendages to the connective. Keeping this theory in 
view, we may well conceive that a dédoublement of the petal may 
produce the inner petaloid scale of some Sapindacex, Violacee, 
Bixacez, &c., or the fimbriate scales in the tube of Cuscuta and 
other gamopetalous flowers, or the corona of Passiflora, the 
cup-shaped nectary of Narcissus, &e.; or, again, that a dédouble- 
ment of the stamens may result in the staminal corona of Ascle- 
piadez when arising from the gynostegium. But that there is any- 
thing of the kind in Stereuliacez is, I think, fully disproved by 
the Buettneriez, where this supposed formation has been most 
relied upon; for here the petiolar gland of the petal-leaf forms 
the apex of the hood, eonnivent over or adhering to the staminal 
column, and is perfectly distinet in origin and position from the 
anther which it so curiously encloses*. 
How then are we to account for the disturbance occurring in 
Sterculiacew of the usual alternation in the staminal whorls? I 
find that on soaking the andreecium in several genera, and more 
especially in Glossostemon, it separates very readily into five bodies 
(adelphia), normally alternating with the petals, each one ending 
in a point or appendage (the teeth, barren lobes, or staminodia of 
the staminal column), and bearing the anthers on its margin on 
each side of the central point. . Might we not consider each such 
body or fascicle of stamens as one staminal leaf, with branched 
veins, the central vein bearing no anther (or altered gland), the 
lateral branches each terminating in an anther (or altered gland) 
corresponding tothe marginal glands on the stem-leaves of several 
species of Homalium, Ranara, Euphorbiacee, &c.? Where the 
number of anthers between each two staminodia is an even one 
(2, 4, 6 or 8), the staminal leaf has the same number of anthers on 
each side of the central nerve; where it is an odd one (1, 3, or 5), 
there is one more on one side than on the other,—a cireumstance 
readily explained by the great tendency to obliquity in the parts 
of the floral whorls, where the estivation is so strongly contorted. 
In confirmation of this view I may also observe, that in A/elhania 
Inever find the anther-bearing stamen exactly opposite the centre 
* The supposed lateral dédoublement in the staminal leaves of the inner 
whorl in Cruciferze is, to my mind, equally mythical ; and I hope, on an early 
occasion, to lay before the Society my reasons for coming to this conclusion. 
