Narrative of the surveying voyages of his majesty's ships Adventure and Beagle (vol.3)

558 CORAL FORMATIONS. April, 1836. of ocean, of more than a thousand miles in one direction and several hundreds in another, scattered over with islands, none of which rise to a greater height than that to which waves can throw fragments, or the wind heap up sand. Now if we leave out of the question subsidence, the foundation on which these reefs are built, must in every case come to the surface within that small limit (we may say twenty fathoms) at which corals can live. This con- clusion is so extremely improbable that it may at once be rejected : for in what country can there be found a broad and grand range of mountains of the same height within a hun- dred and twenty feet ? But on the idea of subsidence, the case is at once clear: as each point, one after the other according to its altitude, was submerged, the coral grew upwards, and formed the many islets now standing at one level. Having endeavoured on general grounds not only to re- move any extreme degree of improbability in the belief of a general subsidence, but likewise to show that it is almost necessary to account for the existence of a vast number of reefs on one level, we will now see how far the same idea will apply to the peculiar configuration in the several classes. Let us imagine an island merely fringed by reefs extending to a short distance from the shore ; in which case, as we have before remarked, there is no difficulty in understanding their structure. Now let this island subside by a series of move- ments of extreme slowness, the coral at each interval grow- ing up to the surface. Without the aid of sections it is not very easy to follow out the result, but a little reflection will show that a reef encircHng the shore at a greater or less dis- tance, according to the amount of subsidence, would be produced. If we suppose the sinking to continue, the encircled island must, by the submergence of the central land but upward growth of the ring of coral, be converted into a lagoon island. If we take a section of some encircled island on a true scale, as for instance Gambler, which has been so well described by Captain Beechey, we shall not find