to the fury of the waves, it will be near the outer edge of the reef that the development will be most rapid. If the subsidence of the island thus surrounded should still continue, as mountain after mountain and island after island slowly sink beneath the water, fresh bases would be successively afforded for the growth of the corals, and the outer edge elevated by their continual labour, thus transforming the space into a sort of circular lagune. The madreporic deposits would thus form an isolated girdle, and the lagune, which occupies the centre, would become deeper and deeper in proportion to the lowering of the soil. This is the second stage of the madreporic isle. The existence of the atolls are thus subordinated to two principal conditions: the progressive subsidence of the shore washed by the sea, and the existence of coral formed of stony cells, the growth and multiplication of which are extremely rapid. It follows from this that madreporic isles cannot exist in all seas; that they only have their birth in the Torrid zone, or at least near the Tropics, for it is only in these regions where the warmth exists, so necessary to their development, that the madrepores show themselves in greatest abundance. The great field of madreporic formations, in short, is found in the warm parts of the Pacific Ocean. It is from this point, as from a common centre, round which are ranged the series of madreporic isles and islets, that it will be useful, in concluding this chapter, to trace their geographical distribution. We borrow the materials for this from Milne Edwards's _tableaux_ of their distribution in the principal seas of the world.