GEOLOGY: W. M. DAVIS 
203 
It may be noted that if the surface forms of the mid-Pacific atolls are con- 
sidered alone, they can be accounted for very satisfactorily by the Glacial- 
control theory, which was invented chiefly to explain the special features 
that they present: but as atolls occur in association with barrier reefs in the 
Caroline, Society, Fiji and other groups, and as the central islands within the 
barrier reefs present features which, although perfectly accounted for by 
Darwin's theory of intermittent subsistence, cannot be accounted for by the 
Glacial-control theory, its apparent success in explaining atolls is thereby 
discredited, all the more so in view of the recent discussion by Skeats of the 
boring in the Funafuti atoll (Amer. J. Sci., New Haven, 45, 1918, 81-90). 
The chief value of the ingenious Glacial-control theory may therefore be 
found not so much in its postulate of the prevalent stability of reef-bearing 
islands, or in its assumption that reef corals were killed and that reef-bearing 
islands were abraded while the ocean was chilled and lowered in the Glacial 
period, but in the emphasis that it gives to changes of sea level from climatic 
causes as a factor in the coral-reef problem; for it is manifest that if the post- 
glacial rise of sea level coincide in time with the subsidence of an island, the 
resulting submergence will be at an accelerated rate and of an increased 
amount; while if the fall of sea level occasioned by the oncoming of a glacial 
epoch coincide with a subsidence, the resulting submergence will be at a re- 
tarded rate and of a decreased amount. It cannot however be supposed that two 
processes so unlike in cause as external climatic changes and internal crustal 
deformations should be closely related in time; their coincidences must be 
fortuitous. Throughout the central Pacific the rate and amount of recent 
submergence have not been as a rule too great to be compensated by reef 
upgrowth; witness the abundant atolls and barrier reefs. But in the region 
of the Australasian archipelagoes compensation of submergence by reef up- 
growth has frequently been unsuccessful; witness the rarity of well developed 
barrier reefs and the almost entire absence of atolls. As the climatic changes 
of ocean level must have been everywhere the same, the factors which have 
determined the success or the failure of reef upgrowth would appear to be 
the rate, the amount and the date of subsidence. 
It may be added that submarine banks, of such form that they are best ac- 
counted for as drowned atolls, are rare in the Pacific. A group of ten or more 
of them is known in an island-free space north of Fiji : several extensive banks 
also occur in Tonga. Exception must therefore' be made in favor of a rapid 
submergance only for these relatively few examples of submerged Pacific 
atolls, and the rule that the great majority of Pacific atolls have not been sub- 
merged faster than the rate of reef upgrowth is thereby proved. In the Indian 
ocean, on the other hand, the number of submarine banks bears a larger 
proportion to that of atolls, and the Indian ocean is generally regarded by 
geologists as the seat of greater and more recent movements of depression 
than the Pacific. Recent and rapid subsidences of moderate amounts may 
therefore be plausibly regarded as of more general occurrence in the Indian 
