a new Natural Order. 407 
would be the posterior one which should alternate with the two posterior 
petals, and of course be intermediate between the two posterior and connate 
stamens. Now the position thus theoretically assigned to the wanting sta- 
men is almost practically ascertained by the fact of a small tooth or short 
filament being found in the place just mentioned. To sum up then, the struc- 
ture, apparently so anomalous, of the flowers of Melianthus, we may use the 
following formula :—* Calyx deeply cut into five unequal segments. Poste- 
rior segment cowl-shaped, short; anterior ones large, ovate-oblong ; lateral 
ones broadly-linear, shorter than the anterior. Petals five, comparatively 
small, alternate with the calycine segments; inferior one more or less abor- 
tive, distant from the others; lateral and posterior ones converging towards 
the cowl-like sepal and connected together by the woolly margins of the 
apex of their claws. Disc incomplete, horseshoe-shaped, and at the same 
time concave, with its aperture turned towards the ovarium and the inferior 
{apparently superior) side of the flower, interposed between the two posterior 
stamens and the two posterior petals. Stamens (normally 5) hypogynous, 
alternate with the petals: the posterior one always abortive, occasionally 
represented by a short sterile filament; the inferior ones free, the interme- 
diate or lateral ones more or less connected at their base. Ovarium four- 
celled, its cells alternating with the stamens.” If to these floral characters 
we add an awl-shaped style, with four small stigmatic toothlets; four ovules 
inserted in two rows above the middle of the internal angle of each cell; a 
membranaceous, inflated, almost tetrapterous, four-lobed capsule, the carpels 
of which open along the obliquely-internal suture of their superior lobe; and 
lastly several seeds (generally two in each carpel), horizontal or ascending, 
anatropous, with a hard shining testa, a large somewhat horny albumen, a 
narrow, straight embryo with a claviform radicula longer than the plano- 
convex cotyledons ;—in recapitulating, I say, all these characters and com- 
bining them with those of general: habit, vegetation, foliation and properties 
already mentioned, we shall have a comprehensive idea of the leading features 
of Melianthus,—a sketch sufficient at least to guide us in the search of the affi- 
nities of that singular genus. 
Thus, if we find shrubby plants with pinnate leaves, winged, or at least 
marginated petioles, intrapetiolar and connate stipules, simple racemes, and 
VOL. XX. 3H 
