FORMATION OF CORAL ISLANDS. 497 



these (sometimes measuring six feet by four) are washed up on 

 the windward side of the reef. Shells, coral-sand, and various 

 other debris, accumulate upon it in like manner, until it is at 

 last changed into an island, upon which there is a calcareous soil 

 capable of supporting various -kinds of vegetation. When these 

 have once established themselves, the elevation of the surface 

 continues with greater rapidity successive layers of vegetable 

 mould being deposited by the rapid and luxuriant vegetation of 

 these tropical islands, which are soon tenanted by various forms 

 of animals, and at some subsequent period afford a habitation to 

 Man. It is not usually, however, until after the windward 

 coast has attained considerable elevation, that the leeward side 

 is perfectly closed in. Two causes may be assigned for this. 

 There seems to be a natural instinct on the part of the animals, 

 which impels them to build with greater rapidity on the most 

 exposed side ; and the leeward side is therefore the part last 

 completed. Moreover the closure of tfiis passage is impeded by 

 the current almost constantly passing out of it, which is caused 

 by the dashing of the waves over the windward side into the 

 lagoon ; and until this ceases, there will always be a free exit in 

 the opposite direction. After the wall has been sufficiently 

 upraised, however, the lagoon is often completely enclosed. In 

 the centre of this lagoon, deep water is often found, and no living 

 animals can be seen upon its bottom ; but the shelving edges of 

 the shallower portion are clothed with luxuriant growths of the 

 more delicate species of coral, and its waters abound in marine 

 animals of all descriptions. 



1141. These lagoon-islands vary in diameter from one to fifty 

 miles ; the breadth of the elevated ring which encloses the lagoon, 

 is from 400 or 500 yards to about a mile, though seldom above 

 half a mile ; and its height above the water is not often more 

 than four or five feet in any part. The annular (ring like) cha- 

 racter of these islands naturally suggests the idea, that they may 

 have been built upon the edges of submarine volcanoes, or upon 

 circular elevated ridges, forming basins, resembling those whiofi 

 abound in the parts of the globe at present upraised, and whicn 

 may reasonably be supposed to exist in those still submrr^ert. 



