472 
GEOLOGY: W. M. DAVIS 
subsiding islands. But it is not intended to imply that the view here 
announced appHes to all atolls of the Fiji group, much less to all atolls 
elsewhere, most of which have probably been formed by upgrowth dur- 
ing intermittent subsidence without important interruption by uplift. 
The essence of the case is as follows: In the Exploring Isles of the 
Fiji group, a portion of the largest volcanic island, Vanua Mbalavu, of 
well denuded form, 13 miles long and 930 feet high, is shown in the 
apex of sector M, figure 1; it is in part unconformably covered up to 
heights of 500 or 600 feet with heavy reef and lagoon limestones, deeply 
dissected and well embayed. Avea, a smaller member of the same 
cluster, two miles long, capped with limestones resting unconformably 
on denuded volcanic hills, rises 600 feet from the same broad lagoon, 
FIG. 1. 
and is enclosed with the eight other Isles of the cluster by an irregular 
barrier-reef ring, 23 miles in longest diameter, part of which is shown in 
the same sector. All but one of these islands consist partly or wholly 
of elevated limestone: they therefore presumably represent, as Agassiz 
points out, ''fragmentary remains of the land which must have once 
occupied the area of the lagoon.'' Outside of the barrier reef, twelve 
almost-atolls, atolls, and isolated reef patches rise from deep water,^< 
not more than 5 or 10 miles distant: four of these are represented in 
sector M: an almost-atoll in which the small central rocks are partly 
volcanic, partly limestone; a smaller almost-atoll, with limestone rocks 
alone rising from its lagoon; a true atoll; and a small reef on the verge of 
extinction. The islands and reefs are crowded together in the diagram 
to save space : the circular pattern which the barrier reef of the Explor- 
