in length, with “one of its sides covered nearly the whole way with high trees.”[26] The islets, also, on Keeling atoll, it has been shown, have increased in length, and since the construction of an old chart, several of them have become united into one long islet; but in this case, and in that of Matilda atoll, we have no proof, and can only infer as probable, that the reef, that is the foundation of the islets, has increased as well as the islets themselves. [25] Cook’s “Third Voyage,” book III., chapter x. [26] Beechey’s “Voyage to the Pacific,” chapter vii. and viii. After these considerations, I attach little importance, as indicating the ordinary and still less the possible rate of _outward_ growth of coral-reefs, to the fact that certain reefs in the Red Sea have not increased during a long interval of time; or to other such cases, as that of Ouluthy atoll in the Caroline group, where every islet, described a thousand years before by Cantova was found in the same state by Lutké,[27]—without it could be shown that, in these cases, the conditions were favourable to the vigorous and unopposed growth of the corals living in the different zones of depth, and that a proper basis for the extent of the reef was present. The former conditions must depend on many contingencies, and in the deep oceans where coral formations most abound, a basis within the requisite depth can rarely be present. [27] F. Lutké’s “Voyage autour du Monde.” In the group Elato, however, it appears that what is now the islet Falipi, is called in Cantova’s Chart, the Banc de Falipi. It is not stated whether this has been caused by the growth of coral, or by the accumulation of sand. Nor do I attach any importance to the fact of certain submerged reefs, as those off Tahiti, or those within Diego Garcia not now being nearer the surface than they were many years ago, as an indication of the rate under favourable circumstances of the _upward_ growth of reefs; after it has been shown, that all the reefs have grown to the surface in some of the Chagos atolls, but that in neighbouring atolls which appear to be of equal antiquity and to be exposed to the same external conditions, every reef remains submerged; for we are almost driven to attribute this to a difference, not in the rate of growth, but in the habits of the corals in the two cases.