** ATOLLS" OF THE PACIFIC. 387 



the island takes the form of a circular plateau. In 

 either case such an island is called an atoU. 



The recurrence of this phenomenon is too frequent, 

 more especially in the Pactific Ocean — its cause has 

 been too long a hard nut for the savans — for us to 

 pass it by without saying a few words in respect 

 to it. 



The shape of these islanls led to the belief, that 

 the coral insects had built tlieir reefs on the edges of 

 submarine craters. The general resemblance be- 

 tween them and certain volcanic islands had struck 

 the first observers. Such islands (as Ban-en Island 

 and others) presented a circular belt of mountains, 

 interrupted in one spot by a canal forming a com- 

 munication between the sea and the interior lake. 

 The mountains are the sides of a crater, the bottom 

 of which has been filled up with the invading waters 

 of the ocean. The same general character is at first 

 evident in the appearance of an atoll. But how- 

 can the size of craters which correspond to atolls be 

 explained ? — eleven leagues in diameter, such as that 

 of Bow Island, or of double the diameter, as in several 

 of the Maldive Islands ? Again, how is it possible 

 to admit that the Pacific Ocean is studded with so 

 many volcanoes, all of which attain about the same 

 height of 40 yards ? Lastly, how can the theory of 

 submarine volcanoes be applied to the formation of 



