I836.| 



BARRlEli-KEEFS. 



471 



rrr. ■>: — 



I. Vanikoro. 2. Gnmbier Islands. 3. Maurua. 



The horizontal shading shows the barrier-reefs and lagoon-channels. The in4:ned 

 shading above the level of the sea (AA), shows the actual form of the land : the inclined 

 ihading below this line, shows its probable prolongation tinder water. 



Are we to suppose that each island is surrounded by a collar- 

 like submarine ledge of rock, or by a great bank of sediment, 

 ending abruptly where the reef ends? If the sea had for- 

 merly eaten deeply into the islands, before they were protected 

 by the reefs, thus having left a shallow ledge round them under 

 water, the present shores would have been invariably bounded by 

 great precipices : but this is most rarely the case. Moreover, on 

 this notion, it is not possible to explain why the corals should 

 have sprung up, like a wall, from the extreme outer margin of 

 the ledge, often leaving a broad space of water within, too deep 

 for the growth of corals. The accumulation of a wide bank of 

 sediment all round these islands, and generally widest where the 

 included islands are smallest, is highly improbable, considering 

 their exposed positions in the central and deepest parts of the 

 ocean. In the case of the barrier-reef of New Caledonia, which 

 extends for 150 miles beyond the northern point of the island, in 

 the same straight line with which it fronts the west coast, it is 

 hardly possible to believe, that a bank of sediment could thus have 

 been straightly deposited in front of a lofty island, and so far be- 

 yond its termination in the open sea. Finally, if we look to other 

 oceanic islands of about the same height and of similar geological 

 constitution, but not encircled by coral-reefs, we may in vain 

 sea'^h for so trifling a circumambient depth as 30 fathoms, except 



