PROCEEDINGS OF SECTION C. 



407 



Spreading from a central nucleus in round or linear form, the 

 budding reef rises towards the surface. If th^ axis of a linear reef is 

 parallel to the south-east trade its shape persists ; but if it lies across 

 the wind, then, as it comes within the influence of the waves, the pre- 

 vailing south-easter moulds it to a crescent, with the convexity to 

 windward. By a continuation of the crescent horns to leeward, the 



Diagram to sJioio how a Linear Beef lying across the Wind is formed into a Horseshoe. 



characteristic horseshoe shape is attained. As elevation would affect 

 a river so submergence, at the present stage, would rejuvenate the reef. 

 Conversely, maturity or senility in a reef imply quiescence. 



This explanation differs from that of Sir J. Murray, who considers 

 the atoll form to be assumed by abundant growth of well-fed corals 

 on the margin and solution of dead coral rock in the interior. But if 

 solution be so destructive, how can a reef form at all ? 



In its next stage the horseshoe reef reaches the surface as a ridge 

 of growing coral. On this summit the corals die, and the crest is rubbed 

 down level by the surf. The top of the reef then resembles a roadway, 

 with borders of living coral for the side-walks. With age each arm of 

 the horseshoe broadens indefinitely ; the fringes of living coral on 

 either side remain, but an expanse of a mile or more of dead coral 

 intervenes. The broadening of reefs in the central Pacific has been 

 described by one of us.(/) Erosion and solution destroy much material, 

 but the loss is at once made good by fresh matter thrown upon the reef. 



(/) Hedley.— Natural Science, xn., 1898, pp. 174-178. 



