ON THE FORMATION OF CORAL ISLANDS. 341 



the remarkable fact is the utility of this proceeding : 

 since the shelter produced to windward enables the 

 work to go on in the opposite direction, with com- 

 parative facility. If the instinct of working upwards 

 to the surface is remarkable, this is even more so : 

 while, in all we trace the evident intention, in that 

 Power which gave these instincts, to effect the very 

 ends that are attained. 



Some peculiarities occur in the circular groups, 

 which are particularly deserving of notice, from the 

 results by which they are followed, and to which I 

 have already alluded. The first appearance consists 

 in a number of detached rocks and islands, forming 

 a sort of chain ; while, in further progress, inter- 

 mediate ones spring up, so as gradually to unite the 

 whole into a continuous line or circular form. On 

 the outside of this line, or ring, the water is deep 

 and the walls vertical ; but, within, it shoals in dif- 

 ferent places, in a sort of general declivity from the 

 margin. Thus it produces a solid mass, of a shape 

 like that of a plate or elevated basin ; being a kind 

 of platform with a depressed centre within a vertical 

 wall. In the smaller circles, when this process is 

 completed, the reefs bear this resemblance to a cir- 

 cular basin very accurately. And this cavity is, at 

 first, a kind of salt lake ; becoming shallower as the 

 animals, still acting within it, prolong their works 

 upwards. And thus it gradually becomes so con- 

 tracted, that the fall of rain is sufficient to freshen 

 the water ; whence follow the death of the animals 

 and the cessation of their operations. Thus it re- 

 mains a cavity, and becomes a lake ; forming a sup- 

 ply of water for future inhabitants, under an ar- 

 rangement no less remarkable for its foresight and 



