156 CORAL AND ATOLLS 



and boil across the flats, only to fall back in broken confusion 

 to the hindrance of the next oncoming roller, that there is 

 something in the moulding of the barrier reef itself that makes 

 it so successful a resister of the ocean's terrible and ever-acting 

 force. No wall made by man could ever hope to withstand 

 for long the perpetual bombardment of those great blue rollers 

 of the Trade-driven Indian Ocean. The great secret of the 

 success of the barrier reef is to be found in the peculiar 

 curvature of its wave-swept surface. From the uneven edge 

 on which large rocks are hurled and firmly seated, and on 

 which the first force of the waves is spent, the barrier runs 

 shorewards as a smooth Nullipore-covered platform which has 

 a slight curve upwards. The barrier is a flattened arch, and the 

 peculiar pitch of this arch is suited to the most efficient 

 spending of the force of the waves. The waves that the 

 boulder- guarded edge has broken become further spent in 

 sweeping across the convex seaward margin, and finally ex- 

 hausted by the uneven flats which line the shore ; whilst the 

 backwash from the spent waves seethes among the flats, and 

 pours backwards with increasing force as the outer slope is 

 reached. Every broken wave helps to be the undoing of its 

 successor. 



The only places in the atoll ring in which this rock flat is 

 wanting are on the north-west, or leeward, side of the group ; 

 and here, although no breaking surf is present, and no sea- 

 washed rock flats stretch outwards, the same coral boulders 

 lie at a somewhat greater depth uninfluenced by the waves. 



The flats of the barrier are entirely composed of coral, 

 living or dead ; and the coral that is living is a very small 

 fraction of the whole. The dead shells of molluscs, the cast 

 ectoskeletons of Crustacea, the hard parts of echinoderms and 

 other marine animals, are too unimportant a contribution to 

 invalidate the statement that the barrier is entirely the pro- 

 duct of corals. 



In certain parts of the island ring it is true that the 

 encrusting Nulliporce constitute a very important part of the 

 barrier. This Nullipara growth appears to be a very varying 



