THEORIES OF BARRIER AND CIRCULAR REEFS. 153 



down by the waves so as to form suitable banks. 2. The banks taken 

 possession of by corals are built up to the sea-level. 3. The coral- 

 growth is confined, or at least most rapid, on the outer margin, because 

 exposed to the action of the sea. Thus arises a ring with blank space 

 within. 4. The action of waves beats these rings into a series of 

 islets. 5. Meanwhile the scouring action of currents and the solvent 

 action of sea- water, scoops out the blank area into a more or less deep 

 lagoon. 6. The action of waves breaking the living coral and the reef- 

 rock forms a debris-j>ile or talus, with steep outward slope, on which 

 the corals continue to grow seaward into deep water. Thus the 

 coral riug continues to spread, like a fairy ring, by growing seaward 

 in every direction and dying behind. 7. According to Darwin, atolls 

 grow continually smaller ; according to Murray, they grow continually 

 larger. 



Barriers are similarly explained. They commence as fringes, which 

 grow seaward as far as depth will allow. Then the corals die near 

 the shore, and this part is scoured out into a channel. Meanwhile 

 the reef extends seaward on its own talus, and the channel is pari 

 passu widened. 



In the present condition of the question it is probable that there 

 are more ways than one in which barriers and atolls may be formed, 

 but Darwin's view seems still to hold its own as a general, though not 

 as a universal theory. 



If we adopt Darwin's view as beiug the most probable, then, of 

 course, every atoll marks the site of a sunken volcanic island. It will 

 be interesting, therefore, to make some estimates based on this view. 



Area of Land lost. — Probably several hundred thousand square 

 miles of habitable high land has been lost by this subsidence. The 

 actual extent of atolls known is at least 50,000 square miles. But this 

 is far less than the loss of high land. For — 1. It is certain that the 

 area of an atoll is always less than that of the original fringe or base of 

 the original high island, for the outer wall of an atoll is not perpendicu- 

 lar. The contraction continues as the subsidence progresses, until 

 small atolls or only lagoonless islands remain. 2. An immense lost 

 area is represented by the space between barriers and their high isl- 

 ands. The great Australian barrier extends along that coast 1,100 

 miles, at an average distance of thirty miles, with a ship-channel be- 

 tween of thirty to sixty fathoms deep. This single barrier, therefore, 

 represents a lost land-area of 33,000 square miles. 3. In the blank 

 area already spoken of, probably many islands went down, and left no 

 record behind. 



The large amount of high land thus lost has been replaced only to 

 a small extent by the wave-formed coral islets on the reefs. 



Amount of Vertical Subsidence. — The amount of subsidence may be 



