386 ON THE GROWTH OF 



inclosed lake. This lake, however, is not entirely in- 

 closed ; (and it could not be, for without supply from 

 the sea it would soon be dried up by the rays of the sun,) 

 but the exterior wall consists of a great number of small- 

 er islands, which are separated from each other by some- 

 times larger, sometimes smaller spaces. The number of 

 these islets amounts, in the larger coral islands, to sixty ; 

 and between them it is not so deep but that it becomes 

 dry at the time of ebb. The interior sea has in the mid- 

 dle generally a depth of from thirty to five-and- thirty fa- 

 thoms ; but on all sides towards the land the depth gra- 

 dually increases. In those seas where the constant mon- 

 soonsprevail, where, consequently, the waves beat only on 

 one side of the reef or island, it is natural that this side of 

 the reef, exposed to the unremitting fury of the ocean, 

 should be formed chiefly by broken-off blocks of coral, 

 and fragments of shells, and first rise above the elements 

 that created it. It is only these islands respecting 

 the formation and nature of which we hitherto know 

 any thing with certainty ; we are almost entirely without 

 any observations on those in the Indian and Chinese Sea, 

 which lie in the regions of the six months' monsoons. 

 From the charts given of them, it is to be inferred that 

 every side is equally advanced in formation. The lee 

 side of such a coral reef in the Pacific Ocean, which is 

 governed by the constant monsoons, frequently does not 

 shew itself above the water, when the opposite side, from 

 time immemorial, has attained perfection in the atmos- 

 pheric region; the former reef is even interrupted in 

 many places by intervals tolerably broad, and of the same 

 depth as the inner sea, which have been left by nature, 

 like open gates, for the exploring mariner to enter the 

 internal calm and secure harbour. In their external 



