MODE OF ORIGIN. 127 
Fig. 1. 
gett we ee 
at the surface, dropping off at its edge to shallow depths, and then 
some distance out, declining more abruptly. Let the same island 
become submerged till II is the water line:—the reef extends itself 
upward, as submergence goes on, and may have the character at the 
surface represented by b'f’b' f’. There is here a fringing reef and alsoa 
barrier reef, with a narrow channel between, such as we have described 
as existing on the shores of Tahiti (see figure, page 41); 0’ is a section 
of the barrier, c’, of the channel, and /” of the fringing reef. Suppose 
a farther submergence, till III is the water line: then the channel 
(c’c'’) within the barrier is quite broad, as in the island of Nairai or 
Angau; on one side (/”’) the fringing reef remains, but on the other 
it has disappeared, owing, perhaps, to some change of circumstance 
as regards currents, which retarded its growth, and prevented its 
keeping pace with the subsidence. With the water at IV, there are 
two islets of rock in a wide lagoon, along with other islets (2’” 2”) of 
reef over two peaks which have disappeared. The coral reef-rock by 
its gradual growth has attained a great thickness, and envelopes nearly 
the whole of the former land. Nanuku, the Argo Reef, and Ex- 
ploring Isles are here exemplified, for the view is a good transverse 
section of either of them. 0’ 6” are sections of the distant enclosing 
barrier, and c’” c’’, and other intermediate spots, the water within. 
The supposed similarity between these ideal sections and existing 
islands is fully sustained by actual comparison. The annexed 
figure (fig. 2) is a sketch of the island of Aiva in the Feejee 
Fig. 2. 
Group. There are two peaks in the lagoon precisely as above; and 
although we have no soundings of the waters in and about it, nor 
