284 
GEOGRAPHY: W. M. DAVIS 
CLIFT ISLANDS IN THE CORAL SEAS 
By W. M. Davis 
DEPARTMENT OF GEOLOGY AND GEOGRAPHY. HARVARD UNIVERSITY 
Received by the Academy. April 8, 1916 
The great majority of reef -encircled islands in the Coral Seas are of 
maturely dissected form, with slopes of moderate declivity that descend 
to an indented shoreline, where the tapering spurs end in salient points 
and the widening valleys open into branching bays, thus proclaiming 
that the islands have suffered partial submergence since they were dis- 
sected. It is generally agreed by those who accept this interpretation of 
insular rehef in relation to shoreline pattern that the encircling barrier 
reefs have grown upwards from their beginning as fringing reefs during 
the subsidence which gave the shorehne its indented outline. Reasons 
for this conclusion were given in a previous article 'The Origin of Coral 
Reefs' in these Proceedings 1, 146-152, 1915; they have been more fully 
stated in 'A Shaler Memorial Study of Coral Reefs' recently published 
in the American Journal of Science, 40, 223-271, 1915. 
The object of the present article is to call attention to certain excep- 
tional reef-encircled islands, Reunion, Tahiti, and New Caledonia, in 
the Coral Seas, which present clift shore lines, and to suggest an explana- 
tion by which their exceptional features may be simply accounted for. 
If all the islands of the Coral Seas were clift like these three, they would 
support either the theory of veneering barrier reefs on wave-cut plat- 
forms, or the Glacial-control theory of coral reefs. But clift islands are 
exceptional. Some means of explaining them must be found which will 
not be in discord with the theory of coral reefs that is supported by the 
features of non-clift islands, namely, Darwin's theory of intermittent 
subsidence. It will appear from what follows that the suggested means 
of explaining the exceptional clift islands of the Coral Seas gives unex- 
pected support to this theory. 
Imagine a young volcanic island, from which detritus is washed so 
abundantly into the sea that corals and the associated reef-making or- 
ganisms can establish themselves if at all, only in short and discontinuous 
fringing-reef patches around its shore. So long as the island continues 
to grow by eruption, its height and perimeter will increase; its earlier 
fringing reef patches will be buried under newer lava flow?, agglomerate 
beds and ash falls. The completed volcanic island will have a shoreline 
of simple pattern without pronounced salients and reentrants, because 
the materials ejected from the summit crater are spread rather equably 
on all sides of the cone, and thus give it a nearly circular base. The sim- 
