ORIGIN OF BARRIER REEFS AND ATOLLS. 26d 
dence the inner patches would disappear. Moreover, after 
the barrier is once begun, it has growing corals on both its 
inner and outer margins, while a fringing reef grows only on 
one margin. Again, the detritus of the outer reefs is, to a 
great extent, thrown back upon itself by the sea without and 
the currents within, while the inner reefs contribute a large 
proportion of their material to the wide channels between 
them. These channels, it is true, are filled in part from the 
outer reefs, but proportionally less from them than from the 
inner. The extent of reef-grounds within a barrier, raised 
by accumulations at the same time with the reefs, is often 
fifty times greater than the area of the barrier itself. Owing 
to these causes, the rate of growth of the barrier may be at 
least twice more rapid than that of the inner reefs. If the 
barrier increases one foot in height in a century, the inner 
reef, according to this supposition, would increase but half a 
foot ; and any rate of subsidence between the two mentioned, 
would sink the inner reefs more rapidly than they could grow, 
and cause them to disappear. There is therefore no objec- 
tion to the theory from the existence of wide channels and 
open seas; on the contrary, they afford an argument in its 
favor. 
From these remarks on the channels and seas within 
barrier reefs, we pass to an illustration of the origin of an 
atoll. The inference has probably been already made by the 
reader, that the same subsidence which has produced the dis- 
tant barrier, if continued a step farther, would produce the 
lagoon island. Nanuku is actually a lagoon island, with a 
single mountain peak still visible; and Nuku Levu, north of 
it, is a lagoon island, with the last peak submerged. The 
Gambier group, near the Paumotus, appears to have afforded 
an early hint with regard to the origin of the atoll, or at 
