had done, and destroying the plants that may have begun to vegetate from cocoa-nuts, roots, and seed brought thither by birds, or thrown up by the sea.  This, doubtless, happens very frequently, for we found many cocoa-nuts, and some other things, just sprouting up, only a few inches beyond where the sea reaches at present, in places where it was evident they could not have had their origin from those farther in, already arrived at their full growth.  At the same time, the increase of vegetables will add fast to the height of this new-created land, as the fallen leaves and broken branches are, in such a climate, soon converted into a true black mould or soil.[157]

[Footnote 157:  Mr Anderson, in his journal, mentions the following particulars relative to Palmerston’s Island, which strongly confirm Captain Cook’s opinion about its formation.  “On the last of the two islets, where we landed, the trees, being in great numbers, had already formed, by their rotten parts, little risings or eminences, which in time, from the same cause, may become small hills.  Whereas, on the first islet, the trees being less numerous, no such thing had as yet happened.  Nevertheless, on that little spot the manner of formation was more plainly pointed out; for, adjoining to it was a small isle, which had doubtless been very lately formed, as it was not as yet covered with any trees, but had a great many shrubs, some of which were growing among pieces of coral that the sea had thrown up.  There was still a more sure proof of this method of formation a little farther on, where two patches of sand, about fifty yards long, and a foot or eighteen inches high, lay upon the reef, but not as yet furnished with a single bush or tree.”—­D.

In a former volume we quoted a passage from Dr Forster’s observations respecting the formation of coral islands.  Captain Flinders gives a similar account in vol. ii. p. 114, of his voyage, drawn up from his own observations on Half-way Island, on the north coast of Terra Australis.  It is too long for this place.  The reader will find it transcribed, together with Forster’s, in the notes to the translation of Cuvier’s work, already referred to.—­E.]

Perhaps there is another cause, which, if allowed, will accelerate the increase of these islands as much as any other, and will also account for the sea having receded from those elevated rocks before mentioned.  This is the spreading of the coral bank, or reef, into the sea, which, in my opinion, is continually, though imperceptibly, effected.  The waves receding, as the reef grows in breadth and height, leave a dry rock behind, ready for the reception of the broken coral and sand, and every other deposit necessary for the formation of land fit for the vegetation of plants.

In this manner, there is little doubt, that in time the whole reef will become one island; and, I think, it will extend gradually inward, either from the increase of the islets already formed, or from the formation of new ones upon the beds of coral within the inclosed lake, if once they increase so as to rise above the level of the sea.