S rsuviaceae.—Calyx 5 parted, petaloid within. Corolla none. Ovary one (by abortion ?)
or several celled, placenta in the axis of the cells, or when one-celled, cohering to the side:
ovules several, superposed, attached by longish podosperms. Capsule usually becoming one-celled
by the obliteration of the partitions, or, spuriously, transversely 2 celled, circumscissile (dehiscing
transversely) little above the base.—Succulent herbaceous plants. Triant hema, Se-
suvium.
P aronicheaceae.—Sepals 5, more or less cohering at the base, persistent: petals 5, shorter
than the calyx, inconspicuous, or somewhat imperfect, marcescent. stamens 5, or three, by
abortion, shorter than the calyx : capsule 1-celled, 3-valved : placenta central in the base of the
cell with numerous seed attached by distinct podosperms to the base. Polycarpea, Hapaloseo#
Drymaria.
M olldgineae.—Sepals persistent: corolla wanting or imperfect, the petals resembling sterile
stamens : stamens 5-lU sometimes 3 by abortion : ovary free 3-5-celled, placenta in the axis
of the cells, extending their whole length ; ovules numerous superposed : capsule 3-5-celled
dehiscence loculicidal. Mollugo, Glinus, £ Orygia.
To this last genus our Glinus trianthemoides belongs. I t is referred to this section by
Endlicher and Meisner, but having a perfect corolla may possibly yet find a more suitable station,
though for the present I can propose no better place.
In this distribution I have excluded the name Ficoideae, which has especial reference to
Mesembryanthemum—the Fig marigold—a. Cape genus of great extent and differing, in appearance,
most widely from all the rest of the tribe, not only in its flowers, which are curious, much
resembling a radiate composita, but in its fruit, which is very remarkable, but with the structure
of which, Botanists, though the plants are now cultivated in'every green house in Europe,
seem imperfectly acquainted. Professor Endlicher (Genera plantarum) is the only writer, so far
as I am aware, who has described its appearance, with sufficient accuracy to lead to the suspicion
that it presents any marked peculiarity. DeCandolle describes the ovary as from 4 to 20-
celled adnate within the substance of the calyx. ( Ovarium calyci adnatum intus pluriloculare
( 4-20 ) saepius 5 —loculare) the capsule as many celled adnate to the calyx and dehiscing in a
star-like manner at the summit. No one would suspect from this any peculiarity. Endlicher
however takes a different view. He describes the ovary as composed of from 4 to 20-carpels,
horizontally verticelled round a thick central axis, with the margins of the carpels, forming the
ventral suture free above, or valvately connate, or introflexed forming a more or less perfect
spurious partition. Placentas linear, adnate to the middle nerve of the camel, occuDvinff the
bottom of the cell.”* F 5
According to this description the usual position of the placenta and consequently of the
whole carpel is reversed. The normal structure of a carpel is, for the ventral secture to be
placed in the axis and there bear the placenta. Dehiscence, when loculicidal then takes place
along the dorsum, or middle nerve, on the side most remote from the axis. Here, according to
the above description, we find matters wholely reversed, the dehiscence is loculicidal, but in the
place of the dorsum splitting to give passage to the seed, the ventral suture, which in other plants
is central and bears the seed, is said to be free and to open to give them passage; while the placentas,
in place of being produced by the union of the margins of the carpellary leaves, is said to be
generated by the midribs or costae. Can such an inversion really happen ? I certainly think not,
and, therefore, suspect the learned author has fallen into error in supposing the ventral suture
of the carpels free and dehiscent, in place of, as they appear to me, attached to a broad gynobase,
which, by raising the dorsal face of the carpel from the horizontal to the vertical aspect, produces
an appearance of inversion that does not exist. According to this view, which however I offer
with considerable diffidence as my acquaintance with the fruit of this genus is altogether derived
from the examination of dried specimens of three or four species, the flowers are not epigynous,
* Ovarium cum calycis tubo connatum, e carpidiis 4-20 conflatum, carpidiis circa axim centralem crassam sub-
nonzontaliter veticiUatis 4-20-loculare, earpidiorum marginibus suturam ventralem superam constituentibus valva-
tim connatis v. in septum plus minus distincte introflexis Placentae lineares earpidiorum nervo medis adnate funduin
loculi occupantes. ’ Endl. Gen. PI. p. 945-6.
in the usual acceptation of the term, but perigynous and the ovary free, as in all the other
genera which I propose combining into this Cvrvembryose class, thus establishing the justice of
my objection to its introduction intoLindley’s epigynous group, while it shows that it might have
been correctly admitted as a 5th alliance, into the gynobaseose group between Coriales and
Florkeales. Surianeae placed by Lindley, in this group, next Coriarieae apparently belongs
to my proposed class, though the albumen is sparing and less distinctly mealy than in the other
orders of the class.
A ff in it ie s . After the preceding extended discussion of the various relationships of the
orders under examination, it seems unnecessary to dilate further on this head than merely to
enumerate the orders which I propose combining to form this class ; these I shall quote from the
list of orders above alluded to in Lindlev’s first edition, preserving the succession he has
adopted (not being as yet prepared to offer an improved distribution), merely observing, that
there may be some other orders to be added with which I am unacquainted. They are, Caryo-
phylleae—Portulaceae— Ficoideae— lllecebreae—Amarantaceae—S clerantheae—C henopodeae
—Phytolaceae—Petiveraceae—Polygoneae—JYyctageneae—and ? Surianeae. This list of
orders, in all of which mealy albumen with an exterior embryo is found, sufficiently attest the
value of this structural peculiarity as an ordinal character and it seems not improbable that,
when the idea is extensively followed out, other groups of equal value may be thus formed :
Cruciferae and Caparideae for example may form the basis of such another—Leguminosae and
Rosaceae of a third Euphorbiaceae, Urticeae, &c. of a fourth Rubiaceae and Myrsineae of a
fifth and perhaps many more. These suggestions are thrown out, nearly at random, as hints, to
which, at the present moment, I attach little importance, though others, having a more extensive
knowledge of plants, may be able to turn them to good account.
G eographical D istribution. On this head I have but little to say, every part of the
world seems to possess plants referable to this group, usually frequenting arid, sandy, exposed
soils. In India, neither the genera nor species are numerous, but several of those that are met
with are very common. Portulaca quadrijida is met with every where; the species of Trian-
thema are about equally abundant, especially T. obcordata which is a troublesome weed in
almost every garden. The other species are also common, especially T. decandra, which is nearly
as common as T. obcordata. Talinum is more rare, seeking a rich rather moist soil in subalpine
jungles. Polycarpeae are common on hard red soils ; Hapolosea like Mollugo and Glinus
prefer cultivated ground, Drimaria is usually met with climbing among bushes in alpine jungles
and much resembles Holostium and some others of the chick weed family : Sesuvium is confined
to saline soils near the coast, Orygia, our Glinus trianthimoides, is of more rare occurrence,
usually met with in black soils, it is common in the ceded districts. At the Cape of Good
Hope they abound, to that country Mesembryanthemum one of the largest known genera, is
almost entirely confined, and species of nearly every genus of all the orders named are also met
with in that promontory, in common with all other kinds of succulent plants. As representatives
of these orders are found every where it is unnecessary to be more particular.
P roperties and U ses. On this head I have nothing to offer, Portulaca oleracea, as its
name implies, is occasionally cultivated as a pot herb. The young leaves of Trianthema obcordata
are sometimes similarly employed and, being mostly mild succulent plants, nearly all might
be, they possess no medicinal properties.
R emarks on G enera and S becies. While examining recent specimens of Trianthema
obcordata and decandra, I detected two grave errors into which we fell in preparing our generic
character of that, genus. It is there said “ capsule one-celled ( or with a spurious longitudinal
dissepiment projected from the placenta when the style is 2-partedJ splitting transversely,
a little above the base.” The words printed in italics is an error, for when the style is 2 partite
there are really two carpels and 2 cells the partition between which can be split from the summit
showing the cells perfec', each of which is furnished with its own placenta. Again it is said
“ lid coriacious containing one or two seeds attached to an unilateral placenta, bottom membranaceous
with one or more seeds attached to a unilateral placenta on the opposite side from