COEAL EEEFS. 383 



before our readers a summary of the discussion as it at 

 present stands, with such few remarks as we think more 

 especially conclusive on the argument. 



The hypotheses by which alone we can seek to solve the 

 problem just stated are few in number. Either the corals 

 constructing the lower parts of the reefs must be wholly dis- 

 tinct in species and habits from those which work near the 

 surface ; or the reefs, atolls, and islets we see must be mere 

 superficial coverings or cappings of points and ridges of land 

 underneath ; or, thirdly, there must have occurred such 

 subsidence downwards of the land encircled by or supporting 

 coral formations, as to leave the coral summits solely on the 

 surface of the waters ; with means of increment, where the 

 subsidence further continues, by the superinduction of fresh 

 layers, under the conditions of depth favourable to the living 

 actions producing them. We are unable to find any other 

 suppositions than these which will apply to the solution of 

 the problem before us. 



The first of them is negatived in great part by the im- 

 probability that there should be species of corals differing so 

 widely as to one of the most important conditions and ne- 

 cessities of their existence ; and further, by the negative fact 

 that no examination of the dead coral taken up from great 

 depths has disclosed such varieties. 



The second hypothesis is of more plausible kind, and was at 

 one time adopted by Sir C. Lyell, in common with many 

 other naturalists ; but subsequently relinquished by this emi- 

 nent observer in favour of the last of the opinions just stated. 

 It was, in truth, a natural and easy conception that the coral 

 formations incrusting the upper surface might follow and 

 depicture the outline of the submarine bottom, and the peaks 

 and ridges rising from it. And this argument became more 

 specious when considering the coral islets, with their circular 

 and often deep lagoons within, as representing the cones and 



