IN TEE COCOS-EEELING ISLANDS. 37 



sucli plateaux have reached, on their windward side, the level 

 at which corals prosper, that is, some 120 feet below the surface, 

 these coral reefs spring up and flourish,"* and subsisting at a 

 greater depth than all others, a solid foundation is laid by the 

 close compactly growing Astrww ; then on their dense floor, in 

 whose myriad crannies, molluscs and all manner of marine 

 beings have sheltered, died and left their shells compacted by the 

 carbonate of lime let loose from their partial disintegration and 

 solution into a solid limestone conglomerate consisting of coral, 

 of shells and of all that may have fallen on it, which they have 

 raised layer above layer as near the surface as they may, the 

 Brain-corals (JiLeandrina) and the Porites assume and continue 

 the upward task till they " in their turn reach the limit beyond 

 which they are forbidden by the laws of their nature to pass. . . . 

 But the coral wall continues its steady progress ; for here the 

 lighter kinds set in — the Madrepores, the Millipores and a 

 great variety of Sea-Ferns, — and the reef is crowned at last with 

 a many-coloured shrubbery of low feathery growth." f 



This is in its main outlines Murray's, Semper's, and Agassiz's 

 explanation of how a reef originates. Unfortunately for my 

 own satisfaction and guidance when examining the Keeling reef, 

 I had not read Professor Semper's views, and those of the other 

 two naturalists were not then published. I have now pictured 

 the reef as risen to almost the surface of the sea at ebb spring- 

 tides; higher than this the coral polyps, which die when 

 exposed for a very short period only to the air and the sun, 

 cannot raise it ; but as corals flourish best in the battle of the 

 waves, which are better aerated and charged with the pelagic 

 life which sustains them, they can extend only seaward and 

 grow their fastest, checked solely where ocean currents scour too 

 fiercely past them. In this stage such a coral structure (as the 

 Keeling atoll) might be seen to be roughly circular in form,— 

 observable also in all the raised islets of the group as well as 

 in North Keeling, — doubtless by being beaten on all sides. 

 Travelling from the exterior margin of the reef inwards, coral 

 growth from less abundant sustenance is seen to be less 



* ' The Tortnga and Florida Keofs,' by Alexander Agassiz, Mem. Am. 

 Soc. of Arts and Sc., vol. xi. p. 113. .„ n 



+ ' Florida Keefs.'L.Agas^siz, Mem. Mus. Comp. Zoology, p. 49. Iroo. 

 B. 8. Edinb., No. 107, 1880 : " On the Structure and Origin of Coral Kecfs atul 

 Islands." 



