XX BARRIER-REEFS 499 



Encircling barrier-reefs are of all sizes, from three miles to 

 no less than forty-four miles in diameter ; and that which fronts 

 one side, and encircles both ends, of New Caledonia, is 400 

 miles long. Each reef includes one, two, or several rocky 

 islands of various heights ; and in one instance, even as many 

 as twelve separate islands. The reef runs at a greater or less 

 distance from the included land ; in the Society Archipelago 

 generally from one to three or four miles ; but at Hogoleu the 

 reef is 20 miles on the southern side, and 14 miles on the 

 opposite or northern side, from the included islands. The 

 depth within the lagoon-channel also varies much ; from i o to 

 30 fathoms may be taken as an average ; but at Vanikoro 

 there are spaces no less than 56 fathoms or 336 feet deep. 

 Internally the reef either slopes gently into the lagoon -channel, 

 or ends in a perpendicular wall sometimes between two and 

 three hundred feet under water in height : externally the reef 

 rises, like an atoll, with extreme abruptness out of the profound 

 depths of the ocean. What can be more singular than these 

 structures ? We see an island, which may be compared to a 

 castle situated on the summit of a lofty submarine mountain, 

 protected by a great wall of coral-rock, always steep externally 

 and sometimes internally, with a broad level summit, here 

 and there breached by narrow gateways, through which 

 the largest ships can enter the wide and deep encircling 

 moat. 



As far as the actual reef of coral is concerned, there is not 

 the smallest difference, in general size, outline, grouping, and even 

 in quite trifling details of structure, between a barrier and an 

 atoll. The geographer Balbi has well remarked that an 

 encircled island is an atoll with high land rising out ot 

 its lagoon ; remove the land from within, and a perfect atoll is 

 left. 



But what has caused these reefs to spring up at such great 

 distances from the shores of the included islands ? It cannot 

 be that the corals will not grow close to the land ; for the 

 shores within the lagoon -channel, when not surrounded by 

 alluvial soil, are often fringed by living reefs ; and we shall 

 presently see that there is a whole class, which I have called 

 Fringing- reefs from their close attachment to the shores both 

 of continents and of islands. Again, on what have the reef- 



