GEOGRAPHY: W. M. DAVIS 
287 
of supply from the cliffs or the streams, will be slowly shifted seaward 
and the rock platform will be at least in part exposed, especially along 
the 'edge' at its outer margin, where the heavy outside storm waves 
make their first attack on the bottom: here in particular, the abraded 
rock of the platform may be laid bare and a favorable site for coral col- 
onies provided; here, if the subsidence has not been too great, a barrier 
reef may grow up, as in the side section of Sector O. 
As soon as the barrier reef is fairly well established, wave work in the 
lagoon is greatly weakened. If the island now stand still again, the 
lagoon will be gradually aggraded by overwashed detritus from the reef, 
by organic deposits formed in the lagoon, and by outwashed detritus 
from the island: thus a delta-and-reef plain will in time replace the la- 
goon, and when this stage is approached and reached, the stream-borne 
volcanic detritus will be again delivered on the reef face, where it will 
retard the growth of corals and their associates. But if subsidence be 
renewed and the shore line rise to the broken line of sector O favor- 
able conditions will be re-introduced, and the barrier reef will grow 
upward again. The cliffs that were cut before subsidence began will in 
time be completely submerged; the shoreline will then have the pattern 
shown by a dotted Hne in sector O, and this pattern will persist as long 
as subsidence continues, until in the almost-atoll stage the diminishing 
island is reduced to so small a size that it has no room for valleys; only 
a central summit will then remain above sea level; in the next stage 
this disappears, and a perfect atoll results. 
Actual examples of islands and reefs should be found in various stages 
of the sequence of events here deduced, if the scheme by which the de- 
ductions are guided is a true counterpart of natural occurrences. Re- 
union in the western Indian Ocean offers confirmation for sectors K and 
L. As described, figured and mapped by Maillard, and especially by 
von Drasche, it has no embayments; part of the shoreline is clift with 
hanging valley mouths; much of the shoreline is delta-fronted and 
beached; and coral reefs form only fragmentary fringes along about one- 
tenth of its circuit. 
Tahiti, a volcanic doublet in the Society group of the South Pacific, is 
clift nearly all around its figure-8 circuit. The cliffs, well described by 
Agassiz, are highest on the eastern or windward side where some of them 
rise 1000 feet above the sea; a number of the smaller valleys make mere 
notches in the cliff top, and the streams cascade down the cliff face; a few 
of the larger valley mouths are still embayed as all of them once were; 
hence the island has subsided and the cliff base is submerged Most of 
the embayments are now filled with deltas, which are indeed advancing 
