ment proceeds still further, and the pericarpium itself is perforated by the embryo, which, while preserving its connection with the parent plant, often attains the length of from eighteen inches to two feet. This happens in Rhizophiora and Bruguiera, or the Mangroves properly so called. In some of the spurious Mangroves, as Avicennia and Ægiceras, a lesser degree of development takes place, and in general their pericarpia remain entire till they have dropped from the tree. In both cases the final cause of 149] the economy is sufficiently evident; a greater than ordinary evolution of the embryo being necessary to ensure its vegetation in the unfavorable circumstances in which it is unavoidably placed.
But an analogous structure exists in other plants, where the final cause is less apparent, as in certain species of Eugenia, in which the integument of the seed is completely absorbed before its separation from the parent plant, and while the pericarpium remains entire.
An economy no less remarkable than that of the Mangroves, but of a nature diametrically opposite, takes place in the bulb-like seeds of certain liliaceous plants, especially of Pancratium, Crinum and Amaryllis; in some of whose species the seed separates from the plant, and even from the pericarpium, before the embryo becomes visible. This observation respecting some of these seeds was, I believe, first made by Mr. Salisbury; and in such as I have myself examined, I have found the fact connected with one no less interesting, namely, an unusual vascularity in the fleshy substance.
I have in another place,[1] in speaking of this substance, which constitutes the mass of the seed, and in a central cavity of which the future embryo is formed, stated it to be destitute of vessels, and entirely composed of cellular texture. But on a more careful inspection, of those seeds at least in which the separation precedes the visible formation of the embryo, I now find very distinct spiral vessels: — these enter at the umbilicus, ramify in a regular manner in the substance of the fleshy mass, and appear to have a
- ↑ Prodr. Flor. Nov. Holland. p. 297.
