506 
Proceedings of the Royal Society 
reef and the shores of the island, and we have a barrier reef. These 
processes have but to be continued some stages further, when the 
island will disappear beneath the ocean, and be replaced by an atoll 
with its lagoon where the island once stood. 
According to this simple and beautiful theory, the fringing reef 
becomes a barrier reef, and the barrier reef an atoll by a continuous 
process of development. 
Object of the Present Paper. — Professor Semper,* during his 
examination of the coral reefs in the Pelew group, experienced 
great difficulties in applying Darwin’s theory. Similar difficulties 
presented themselves to the author in those coral reef regions 
visited during the cruise of the “ Challenger.” 
The object of the present paper is to show, first, that, while it 
must be granted as generally true that reef-forming species of coral 
do not live at a depth greater than 30 or 40 fathoms, yet that there 
are other agencies at work in the tropical oceanic regions by which 
submarine elevations can be built up from very great depths so as 
to form a foundation for coral reefs ; second , that while it must be 
granted that the surface of the earth has undergone many oscillations 
in recent geological times, yet that all the chief features of coral 
reefs and islands can be accounted for without calling in the aid of 
great and general subsidences. 
Nature of Oceanic Islands and Submarine Elevations. — It is 
now known that, with scarcely an exception,! all oceanic islands 
other than coral atolls are of volcanic origin. Darwin, Dana, and 
others have noticed the close resemblance between atolls and ordinary 
islands in their manner of grouping as well as in their shapes. 
In a previous paper the author pointed out the wide distribution of 
volcanic debris over the bed of the ocean in tropical regions, and 
the almost total absence of minerals, such as quartz, which are 
characteristic of continental land.J There is every reason for 
believing that atolls are primarily situated on volcanic mountains 
and not on submerged continental land as is so often supposed. 
* Zeitschr. fiir Wissen. Zoologie, vol. xiii. p. 563. 
t New Zealand, New Caledonia, and the Seychelles have primitive rocks, 
if these can be regarded as oceanic islands. Some of the islands between New 
Caledonia and Australia may have primitive rocks, and the atolls in these 
regions may be situated on foundations of this nature. 
$ Proc. Roy. Soc. Edin., 1876-77, p. 247. 
