THE FIJIS 339 



growth of coral. The water thus poured in forms a hy- 

 draulic head that can escape only through the openings 

 in the outer reef flats. It becomes charged with particles 

 of lime or other material, derived mainly from the me- 

 chanical disintegration of the corals or substratum form- 

 ing the surface of the reef, and also in part from the 

 chemical disintegration due to the action of sea water 

 which rots and dissolves the limestones of the reef. Soon 

 there exist all the elements of a modified gigantic pot- 

 hole, from which the churned material is carried out by 

 the currents flowing through the entrances into the la- 

 goon. Where corals have established themselves about 

 an island on the submarine platforms formed from it by 

 denudation and submarine erosion, he would explain, in 

 very much the same way, the passages between the 

 islands and the barrier reefs. 



Given a comparatively small volcanic island upon 

 whose eroded platforms corals have established them- 

 selves, the first result of the processes described above 

 would be an island with a barrier reef like Matuku. As 

 the denudation and scouring continued, they would 

 cause the disintegration of most of the land, as in Komo. 

 The final effect would be the total disappearance of the 

 land, leaving a lagoon enclosed by a reef. In this last 

 stage the position of the atoll near volcanic islands 

 would often be the only guide to the character of its 

 formation. (See colored plate, Figs. 2.) 



There is still another method by which some of the 

 atolls have probably been formed. In the group are two 

 islands, Thombia and Totoya, both volcanic peaks into 

 whose craters the sea has broken through some point 

 in their walls, and formed lagoons. Across the opening 



