ON THE FORMATION OF CORAL ISLANDS. 343 



tional protection from the sea. Loose shells, breed- 

 ing among the coral, and washed up in the same 

 manner, add to the general height and solidity of the 

 fabric. The same process is also going on during 

 this time, in the interior parts, where the projecting 

 banks lie, so that all these at length extend and unite; 

 producing, according to the original numbers and 

 groupings, islands of various sizes, and often of great 

 magnitude. Occasionally, the lakes before men- 

 tioned are also filled up by the growth and decom- 

 position of vegetables, as in other similar circum- 

 stances. 



To clothe these islands with soil and vegetation is 

 all that remains. This is a more rapid process than 

 would be expected, as I have already remarked. The 

 first foundation of it is laid in the sand produced by 

 the destruction of the corals ; and as sea plants take 

 root and grow upon it, this becomes a vegetable soil. 

 Sea birds of various kinds then find a residence, con- 

 tinuing to add to this soil ; and at length the floating 

 seeds of various plants are arrested, when a terres- 

 trial vegetable covering succeeds. Hernandia, Mo- 

 rinda, Cerbera, Scsevola, Pandanus, are among these 

 plants, the natives of all the islands of this sea. 



Such are the uninhabited islands, for the process is 

 now completed. How man finds his way at length, I 

 need not say ; and how he has filled many of these 

 islands, is well known. 



But far more is yet wanting to explain the present 

 state of many of these islands ; and here commences 

 a new source of geological interest, of a very different 

 nature. It is plain that under this mode of construc- 

 tion, they can have no great elevation ; and accordingly, 

 the flat ones are scarcely raised more than five or six 



