buminous seed and capsule dehiscing laterally, not vertically, is a very anomalous order, having,
apparently, no really near relations. As regards placentation it undoubtedly associates with,
and was at first referred to, my first group (where it is placed in plate 142), but the ovary
has only two carpels and the flowers are irregular. Owing to its “having the same calyx,
corolla, stamens and bivalve capsules,” as Scrofulariaceae, many eminent Botanists esteem it
more nearly allied to that order than to any other, and, in defference to their judgment, I was
at first disposed to place it in my last group, though I thought the relationship remote. I
now prefer placing it here because I feel disposed to view it as much more nearly related to
Cyrtandracece than to Scrofulariacece, and consider that this is its proper location in the
Vegetable system. The habit, the lateral not vertical dehiscence of the capsule, and the exal-
buminous seed, all combining to confirm this view and show that its affinities appertain
much more to Cyrtandracece than Scrofulariacece. Orobanchacece, which I have brought
here on account of the ovarial structure, is one of those complex families which claim relationship
with several others, but is, I believe, notwithstanding its albuminous seed to the full as
naturally stationed here as in either the Gentianal Alliance, where Lindley places it, or in -the
Solanal to which it is usually referred. It seems in fact to form the connecting link between
the Bignonal and Solanal groups, fluctuating between Cyrtandracece and Scrofulariacece, having
the ovarial structure of the former, the albuminous seed of the latter, and the flowers of both.
Acanthacese certainly belongs to this group.
The sixth group forms a very natural assemblage, Erycibece, perhaps, excepted, the true
affinities of which are still undetermined. Lindley refers the genus Erycibe to Ebenacese, but
I think erroneously. I have brought it here on account of its basal placentation, besides which
it has at different times been referred to both Convolvulacece and Borraginece and is finally
raised to the rank of an order by De Candolle and placed between them.
The orders of the seventh group seem all nearly related. Hydroleacece forms the connecting
link between them/having the flowers of Convolvulacece and the ovary of Solanacece.
Lindley, however, refers it to his Cortusal Alliance, which is nearly identical with my first
group, but evidently under the erroneous impression that Hydrophyllacece and Hydroleacece
form parts of one order which they do not. Choisy, with, I think, more justice, indicates the
vicinity of Solanacece as its probable station in the Vegetable system. Between Solanacece
and Scrofulariacece it is exceedingly difficult to find distinguishing marks; Lindley however
places them in different Alliances, viewing the first as the typical order of his Solanales, and
placing the second among his Bignoniales, principally, I believe, on account of the symmetrical
flowers of the one and the unsymmetrical ones of the other.
Botanists justly lay much stress on characters taken from the albumen, especially when
very copious, as being physiological and therefore, more constant and valuable, than structural
ones, which are liable to alter under the influence of variations in the external circumstances
under which they are produced. But this is liable to exceptions, especially in acquatic and
marshy plants. It is wanting, for example, in JEgiceracece and Lentibulariece, both of which,
by ovarial structure, associate with Primulacece and Myrsineacece, albuminous orders. In such
a case whether is the physiological or structural character to have the preference ? Lindley, in
constructing the diagnostic characters of his alliances, always assigns to albumen a high value,
but is not rigid in the application of his own rules in practice, by excluding orders greatly at
variance, in that respect, with the character of the alliance. As a generic or ordinal character,
the presence of a notable quantity of albumen furnishes a character of the greatest value, but
assuming Lindley’s alliances to be near approximations to natural groups, and bearing in mind
the often wide discrepancies in regard to its quantity, presented by different orders entering
into their composition, I feel disposed to look upon it as a difficult subject to deal with in
associations higher than orders. And even in the case of orders it requires to be used with
care, as far as possible in such a manner as to avoid superseding those taken from external
structure, unless, by the way, it is our wish to render them altogether subordinate, as for
example in the construction of classes or groups dependent on the presence or absence of
that organization, when, if rigidly adhered to, all others must of course give way to them.
But to an albuminous class, it may and will be objected, that it can only be completed
by assigning too high a value to this important physiological character and rendering it most
artificial in practice. Numerous genera and even whole sections, of what are now considered
very natural families, are deficient in albumen, while in the remainder it is copious. Oleacece
furnishes an example in point. In the section Oleinece albumen abounds, but is wanting in
Chionanthece; these two nearly allied sections, as regards external structure, of this natural
order would therefore be widely separated in a capological system, while the others, less
closely allied by external characters, would be kept together. This example shows both the
difficulty and importance of assigning a just value to characters, whether physiological or
structural, and that overvaluing either inevitably leads to disruption of affinities.
The position of the carpels, in dicarpillary ovaries, in relation to the axis, is another character
the true value of which has yet to be ascertained. I have occasionally used it in the
above synopsis, but always with hesitation, from finding it vary in different genera referable to
the same order, and from often finding it a difficult character to make out, and therefore one
requiring to be used with much caution.
In the following plate, No. 142, I have given figures explanatory of the characters of
my seven groups of Corollifloras. The drawings being taken from plants not intended for
representation in the subsequent portions of the work, I am thus enabled to introduce a
greater number of subjects in illustration of this division than I should otherwise have been
able to do.
EXPLANATION OF PLATE 142.
I . LENTIBUIiARIEE.
(Pinguicula Vulgaris.)
1. Flower-bud magnified.
2; Flower and spur seen from behind.
3. Expanded flower showing the stamens and pistil.
4. Pistil detached.
5. Ovary cut vertically»
6. Stigma, side view.
7. Mature capsule, natural size.
8. Same, anterior valve magnified.
9. Seed magnified.,
10. Same cut longitudinally.
11. Diagram (from Nees’ gen. plant.)
II. M yRSINEACEE. ’
(Masa indica.)
1. Detached corolla split open, stamens opposite the
lobes.
2. Ovary and calyx, the calyx split open and forcibly
separated from the base of the ovary.
3. Ovary cut vertically.
4. ------ transversely.
5. Placenta and ovules detached. •
6. Full-grown fruit.
7. A mature seed.
8. The same cut from above, showing the transverse
embryo.
9. Cut across, embryo in situ.
10. Embryo detached.
III. Sapotacee.
(Sapota Elingoides.)
1. Corolla split open, showing the alternate fertile
and sterile stamens.
2. Calyx and ovary.
3. Ovary cut vertically.
4. ------ cut transversely.
5. Mature fruit cut transversely, all the ovules but
one aborted.
6. Mature seed.
7- —— cut transversely, albuminous.
IV. Oleacee.
(Olea linocieroides.)
1. Expanded flower.
2. Calyx and ovary.
3. Ovary cut vertically, ovules pendulous from the
apex.
4 . ------ cut transversely.
5. Full-grown fruit.
' l i ' - 'cut transversely.
7. Detached seed.
8 . ------ cut longitudinally, albuminous.
9. Embryo detached.
V. Asclepiadee.
( Gymnema Decaisneanum.)
1. Detached flower.
2. Gynostigium removed from the flower.
3. Ovary and stigma with the pollen masses attached.
4. Detached pollen.
5. One limb of the ovary cut transversely.
6. Cut longitudinally.
7. A seed and coma.
8. Cut longitudinally, showing embryo in situ.
9. Cut transversely.
10. Detached embryo.
VI. CtRTANDRACEE.
(Didymocarpus oval folia.)
1. Corolla laid open to show the insertions of the
stamens.
2. Calyx and ovary.
3. Ovary cut transversely, showing the free revolute
placentiferous margins.
4. Cut longitudinally.
VII. ScROFULARIACEE.
(Limnophila hypericifolia.)
1. Calyx, ovary, and bract
2. Ovary cut transversely, placentas axile.
3. Ovary cut vertically.
R