288 
GEOGRAPHY: W. M. DAVIS 
into the lagoon, where they form a narrow alluvial plain on which most 
of the population dwells. A barrier reef rises at just about such a dis- 
tance off shore as to suggest that its foundation is on the edge of the 
abraded platform that fronts the submerged base of the cHffs; volcanic 
detritus is abundant in the enclosed lagoon. Tahiti thus seems to have 
been clift, about as in sectors L, M or N, and then to have subsided as 
in sector O ; since then it has stood still long enough for abundant delta 
advance and for partial enclosure by a barrier reef, the further growth 
of which is now threatened at certain points by the advancing deltas. 
New Caledonia is peculiar in being a non-volcanic island, and in being 
cHft only on the northeastern side: hence certain special assumptions 
must be made in order to bring it under the explanatory scheme here 
proposed; but the necessary assumptions do not involve any events that 
are not made probable by what is known of the geological and physio- 
graphical history of the island. 
The great majority of reef-encircled volcanic islands in the Pacific are 
volcanic, Uke Reunion and Tahiti; but they differ from these two sub- 
maturely dissected cones in being maturely or more than maturely dis- 
sected, so that hardly a trace of their initial conical slopes remains ; fur- 
thermore, their spurs usually taper to non-clift points and their valleys 
open into broadening bays, as shown in the dotted high-level shoreline 
of sector 0. If the reef-encircled islands of the Pacific have formerly 
passed through an introductory, almost reefless stage, like Reunion, and a 
soon-following reef -encircled stage with partly drowned cliffs and out- 
growing deltas, like Tahiti, they have since then continued their sub- 
sidence as well as their dissection and their reef growth, so that at pres- 
ent they have no trace of the smooth slopes of their initial conical form, 
and no trace of the presumable spur-end cHffs of their early reefless stage: 
they have gained forms that are best explained on the supposition that, 
whether they past through an introductory reefless stage or not, they 
have long suffered dissection and intermittent subsidence, while protected 
from wave attack by an upgrowing, encircling barrier reef. 
A theory of coral reefs is certainly commended when a reasonable ex- 
tension of its postulated conditions and processes enables it to account 
for the cliffs of exceptional reef-encircled islands, for which no adequate 
explanation has been previously given. 
