Olacineae.  Certainly in habit, corolla, etc.  Olacineae are allied to Aurantiaceae, but they are nearer akin to Santalaceae.  The processes are indubitably modified stamina, with a great tendency to irregularity; in one species from Tagoung only three fertile, and five sterile stamina were observed:  the three fertile generally, but not invariably, alternate with the petals.

To Santalaceae they approach in processes, valvate corolla, and placentation, also to Loranthaceae.

Eight stamina thus accounted for; when two opposed to petals, belong to outer series—­also single one.

In Punica, the structure of the ovaria is highly curious.  We find the bottom of the tube is occupied by two cells, partially filled with ovula, which are attached both to the axis and to the base, as well as to the lower part of the outer paries of each cell; so far, it does not depart from the order, for in Aplexus the placentation is tolerably similar.

Above these two, are 4-5 cells, filled with ovula, which are attached entirely to the outer wall of each cell, but the placentae however would seem to have an obvious connexion with the axis, although this is very doubtful.

The formation of the stigma decidedly indicates a binary formation of carpella.

If these 4 upper cells are 4 constantly, and the base of the ovary is as constantly two celled, then the explanation is sufficiently obvious, though different from that given by Lindley. {113}

First, we have in the bottom from which the mere structure of an ovary is deduced, the normal dicarpellary structure, and there is in addition a tendency in excess toward a parietal placentation.

The anomalous formation arises first from parietal placentae being produced to the axis, and from spurious growth from the sides of the ovary also meeting in the axis, by which the ovula are divided into four bundles.

Lindley’s view seems to be questionable, because as in all cases the styles and stigmata are more permanent than ovaries, there should be as many styles, etc. as ovaries. 2nd, because according to this view the placental suture of the carpella would be turned from the axis, (look at Pomaceae,) although his view of Pomaceae being right would indicate an additional affinity with Mespilus, etc. which it does in habit and abbreviated lateral branches.

Are all Myrtaceae dicarpellar?

The true nature of the case is pointed out in the instance cited by Lindley of a permanent variety of apple, which has 14 cells and 14 styles!  With regard to Nicotiana and Nolana; have these one or two rows of carpella?

TECTONA.

Arbores, trunco crasso, cito ramoso, cortice albido, laevi, tenui.

Folia siliceo-aspera, inflorescent dichotoma.

Calyx aestiv. valvat. cor infundibul, subregularis laciniis, 5 rotundatis, demum reflexis aestivat. laciniis super 2, omnino exterior, facies barbato-villosa.