CORAL ISLANDS. 395 



are the only ones capable of covering immense extents of 

 surface, do not commence their operations at a greater 

 depth than twenty-five or thirty feet, in order to raise 

 their habitations to near the surface of the sea. Frag- 

 ments of these species are never obtained, either with the 

 sounding line, or upon the anchors ; nor do we ever see 

 them, unless in places where the water is shallow ; while 

 the branched madrepores, which do not form thick and 

 continuous beds, either on the elevated places which the 

 ocean has left, or on the shores where they still exist, live 

 at considerable depths. 



It is evident, then, that these corals have erected their 

 fabrics on the^summits of submarine hills and mountains ; 

 and that all those reefs of Taiti, the Dangerous Archi- 

 pelago, Navigators'* Islands, the Friendly Islands, &c. 

 are composed of madrepores only at the surface. 



We thus consider it demonstrated, that the rocks of the 

 solid zoophytes or coral, are not capable of forming the im- 

 mense bases on which the greater number of the islands 

 that occur in the Pacific Ocean rest. 



There now remains for us to state how these animals, 

 by their union, are capable of raising small islets. Fors- 

 ter, as already stated, has given a very good description 

 of the manner in which this is effected. In fact, when 

 these animalcules have raised their habitations to the sur- 

 face of the water, under the shelter of the land, and they 

 remain uncovered during the reflux of the tide, the hurri- 

 canes which sometimes supervene, by the agitation which 

 they produce in those shallow waters, throw up from the 

 bottom sand and mud. These substances are detained 

 in the sinuosities and cavities formed between the corals, 

 and thus serve to fix them together, and connect them in- 



