340 ON THE FORMATION OF CORAL ISLANDS. 



out apparent order ; though it may he suspected that 

 there has here been a deficiency of observation. 

 Lastly, they are found extending in long lines, more 

 or less straight, or in rows. 



There is some reason to distrust the assertions that 

 have been made respecting the rapidity of their pro- 

 duction. It is easy to mistake one reef for another, 

 amid such crowds, and m seas so little known, where 

 also there is no guide but the ships' reckonings and 

 observations. And it is obviously an excuse for the 

 error of a navigator or an accident to a vessel, for a 

 bad reckoning or an incorrect chart, to find a new 

 rock where the old one had been forgotten or mis- 

 placed. If there is one evidence of an unexpected 

 rapidity of formation, given by a recent navigator, 

 who, after three years' absence, found some parts of 

 a reef which had scarcely reached the surface at his 

 former visit, clothed with vegetation, it will be im- 

 mediately seen that the last part of the growth of a 

 reef does not depend on the animals themselves ; so 

 that it proves nothing. 



There is reason to believe, from a considerable 

 examination of the soundings in this ocean, that both 

 the places of the coral reefs and their peculiar dispo- 

 sitions, are, very often, if not always, determined by 

 the forms of the bottom of the ocean where they lie, 

 and that they are placed on the submarine hills of 

 these seas. When disposed in straight or curved 

 lines, the windward side of the structure, which is 

 exposed to the breach of the sea, rises almost ver- 

 tically, like a wall, reaching the surface ; while, to 

 leeward, they shelve away, deepening the water in this 

 direction. It is this windward abruptness, in particular^ 

 which renders these rocks so dangerous : since there 

 is no warning through alterations of soundings. But 



