These explanations, which I venture to propose, of rather obscure descriptions did not occur to myself until after I had formed a new theory of my own, the result of a very careful examination of the ovary in all its stages from the earliest, up to the period of impregnation. At these early stages when the whole flower has not yet attained half an inch in length probably a fortnight or more before expansion I invariably find two rows of carpels, one inferior, of 4 or 5 and one superior of 5-6 or more. In the lower series the placentas are ranged round the axis with their base in the centre and the apex, which is free, towards the circumference. In the
upper the attachment, or base of the placentas, is in the circumference and the apex, also at first free, directed towards the centre. Between the two rows a diaphragm is always interposed. The apex of the upper placentas is occasionally, afterwards, prolonged and contracts adhesions to the axis.
In the accompanying figures I have attempted to represent these views. As the fruit advances in size considerable derangement of this structure progressively occurs which is apt to mask and confuse the appearances now described.
Having previously ascertained the occasional existence of inversion in the position of carpels, my first idea was that such an inversion took place in the upper row. This view, which, equally with the preceding, accounts for the crossing of the placentas I feel inclined to adhere to, though I confess not without some hesitation, because it implies a complexity of arrangement rarely met with in tire inimitably simple and beautiful operations of nature, but I think it as difficult to imagine the nearly equally complex and inconceivable operation of the folding in of one set of carpels over the other, which the explanation of Drs. Lindley and Arnott demand, while my explanation has the advantage of at the same time accounting for the double chamber which the ovary presents from its earliest stages, and renders unnecessary the doctrine of an adventitious verticel of carpels which for the present is mere assumption.
With these explanations I leave the question of structure to consider the one pending on its determination, viz. whether or not Granteae ought to be preserved as a distinct order or be reunited to Myrtaceae?
On this point so far as the unvarying evidence derived from cultivated plants is entitled to carry weight on a disputed point — and which I presume it must do until we find that evidence invalidated by the examination of others growing in a truly wild state — we must unquestionably, I conceive, adopt the views of those who urge the separation, because, the complex structure, above described, being constant here and unknown among true Myrtaceae, we have no right, in the total absence of direct confirmatory evidence, to assume that a part is adventitious, merely because it is at variance with our ideas of what should be, especially while we have in addition, differences in habit, in the formation of the seed (the cotyledons are foliaceous and spirally convolute) and in their pulpy envelope still further to confirm the correctness of these views.
In coming to this conclusion I do so mainly on the evidence I have myself adduced, attaching no value to the opinion of Mr. Don, which, being founded, according to his own showing, on most erroneous views of the structure of the fruit does not merit much consideration.
To the views of DeCandolle more importance must necessarily be attached, as the reasons he assigns are more satisfactory, though I do not think he has awarded sufficient value to the very peculiar " economy of the fruit" while he has laid too much stress on others of much less note, such as the want of pellucid dots, the absence of the marginal nerve of the leaves and the pulpy covering of the seed ; thereby, throwing into the shade the true essential character of the order, which unquestionably is the double row of carpels, with the upper placentas parietal and crossing the lower axillary ones, which, if I have rightly accounted for, constitute this a truly curious and unique fruit.
Affinities.}} According to the explanation now given, the affinities of this order remain to be determined, no known order presenting a similar combination of structure. But adopting the arrangement of Jussieu and DeCandolle, the one followed with some slight modifications in this work, we can scarcely find a more appropriate station for it than the one it now occupies, though but remotely allied to the orders among which it is placed.
