SUBSIDENCE IN PACIFIC CORAL REGIONS. 329 



been small compared with that required to submerge all the 

 lands on which the Paumotus and the other Pacific atolls rest. 

 One, two, or five hundred feet, could not have buried the 

 many peaks of these islands. Even the 1,200 feet of de- 

 pression at the Gambier Group is shown to be at a dis- 

 tance from the axis of the subsiding area. The groups of 

 high islands above mentioned, contain summits from 4,000 to 

 14,000 feet above the sea ; and can we believe it possible 

 that throughout this large area, when the two hundred islands 

 now sunken were above the waves, there were none of them 

 equal in altitude to the mean of these heights, or 9,000 feet"? 

 That none should have exceeded 9,000 feet in elevation, is 

 by no means probable. Hence, however moderate our esti- 

 mate, there must still be allowed a sinking of many thousand 

 feet. Moreover, whatever estimate we make that is within 

 probable bounds, we shall not arrive at a more surprising 

 change of level than our continents show that they have un- 

 dergone ; for since the Tertiary began (or the preceding period, 

 the Cretaceous, closed) more than 10,000 feet have been added 

 to the Rocky Mountains, and parts of the Andes, Alps and 

 Himalayas. 



Between the New Hebrides and Australia, the reefs and 

 islands mark out another area of depression, which may have 

 been simultaneously in progress. The long reef of one hun- 

 dred and fifty miles from the north cape of New Caledonia, 

 and the wide barrier on the west, cannot be explained with- 

 out supposing a subsidence of one or two thousand feet at the 

 least. The distant barrier of Australia is proof of great sub- 

 sidence, even along the border of that continent. But the 

 greatest amount of sinking took place, in all probability, over 

 the intermediate sea, called the " Coral Sea," where there are 

 now a considerable number of atolls. 



