GEOLOGY: W. M. DAVIS 
653 
The yielding and slow sinking of the volcanic islands under the influence 
of gravity must be regarded as the cause of the downward movement of 
large amount and long duration which must be assumed in order to 
explain the formation of barrier-reefs and atolls in true oceanic regions."^ 
This appears to me an important suggestion, and one that is likely 
to remove the objections to Darwin's theory of coral reefs in so far as 
they are directed against a great subsidence of broad ocean-floor areas : 
for although Darwin himself was led by his coral-reef theory to infer 
the subsidence of such areas, it is clear from the original exposition of 
his theory that local subsidence of reef foundations will serve all its 
needs. It may be added that the accumulation of the great limestone 
masses of atolls upon slowly sinking volcanic foundations must aid and 
prolong their sinking; also that no comparable sinking of volcanic cones 
upon continents need be expected, not only because of the differences 
supposed to exist between the earth's crust in continental and oceanic 
areas, to which Molengraaf calls attention, but also because continental 
volcanoes suffer erosion, whereby their waste is carried away and widely 
distributed, while oceanic volcanoes, even if they rise for a time above 
sea level and suffer erosion, retain the waste from their summits on 
their flanks. 
But it is particularly the relation of Molengraaf 's hypothesis to changes 
of ocean level that I desire to emphasize. Let it be imagined that the 
ocean floor suffers no deformation apart from that associated with vol- 
canic action, and that a thousand great volcanic cones are built up from 
it in the coral zone, one after the other and at such intervals of time 
that their formation stretches through the Tertiary and Posttertiary 
periods. The building of the first cone would cause a slight rise of ocean 
level. As the cone slowly subsided the ocean surface would return to 
its normal stand, were it not that the subsiding cone is reconstituted 
in an atoll as fast as it subsides,'^ and that other cones are built up as the 
first one sinks. Later formed cones would prolong the changes thus 
initiated, and the slow rise of the ocean would continue, particularly if the 
isostatic sinking of some of the cones were incompletely accomplished. 
Even if some of the cones sank so fast that reef -making corals could not 
reconstitute them, the net result of this process, after many cones had 
been built up and sunk again, would be, not a lowering of the ocean sur- 
face such as according to the usual interpretation of the theory of sub- 
sidence has been supposed to accompany the upgrowth of coral reefs, 
but a slow, small, and long continued rise of the ocean surface. 
The rise of the ocean surface thus caused would be much less in total 
amount than the sinking suffered by any one of the volcanic reef founda- 
