Imagine now the polyp to bud, as does _Palythoa_, and each bud to secrete its own cup, while the connecting branches also throw down the same limestone, so that the cups are connected on to one mass, and we have at once the formation of ordinary reef coral, of which perhaps the simplest possible case is the _Dendrophyllia_, figured on Plate XXVII, where each polyp gives rise to but one bud, which gives out one other, so each branch is like a simple chain of polyps and their cups. The other corals figured are rather more complicated, since one polyp gives off many buds, and their branches are correspondingly more massive. In the hemispherical corals, the connecting branches are short, practically non-existent, and the polyps are crowded together, as a kind of skin, over the solid mass of limestone they have secreted. Certain species of coral form enormous colonies, containing hundreds of thousands of little polyps. I remember a certain part of the fringing reef of Zanzibar[42] over which the water was 6 feet or more deep. Being perfectly clear, and so favourable to coral growth, it was inhabited by a species of the genus _Porites_, which formed huge cylinders, the flat tops measuring 6 to 12 feet across, level with that of the lowest spring-tide, since it is impossible for the polyps to live above that level. So closely were these great cylinders planted in the water that it was easy, by striding and jumping from one to another, to cross the channel to the shallower part of the reef on the other side. From what has been said of the formation of the coral cups it is clear that the quantity of living matter going to the formation of these great cylinders is very small, a mere gelatinous film over the surface.