212 W. M. DAVIS 
below sea-level necessary to explain such contacts is greater than 
could have been accomplished during the glacial epochs of lowered 
sea-level, and wherever the depth of such erosion as indicated by 
the inclosing slopes of the drowned valley embayments is greater 
than 240 feet, or 40 fathoms, the submergence there involved 
cannot be explained by the postglacial rise of ocean-level, and 
local subsidence must be called on to account for the additional 
submergence, unless, indeed, the additional submergence is ascribed 
to a general rise of ocean-level due to recent uplifting of some 
other part of the sea bottom; but the combination of these two 
uniform changes of level would produce the same submergence 
on still-standing islands everywhere; whereas the sea-level and 
elevated reefs of the Pacific archipelagoes call for submergences 
varying in amount, date, and place, and alternating with emergences 
in varying order, thus producing a complication of changing levels | 
that cannot be accounted for without local uplifts and subsidences 
such as Darwin’s theory of coral reefs postulates. Thus at the 
outset of our inquiry the observed structures of sea-level and of 
elevated reefs appear to correspond very well with the hypo- 
thetical structures deduced from the theory of intermittent sub- 
sidence. 
Prolonged crustal stability as postulated in the glacial-control 
theory.—In view of the citations from Darwin’s book given above 
it seems fair to regard his theory of intermittent subsidence as 
adaptable to many different conditions, some of which are veri- 
fied by the examples already adduced; not that every quoted 
opinion is correct, but that the careful consideration given to the 
different aspects of the fundamental postulate of subsidence shows 
that it was carefully examined, and that the theory which is based 
upon it is so elastic that it can accommodate a large variety of 
conditions and processes. Darwin’s theory is in this respect 
strongly unlike the glacial-control theory, which is narrowly limited 
in its fundamental assumption of a ‘“‘long period of nearly perfect 
stability for the general ocean floor,” or, in other words, a “general 
crustal stability in the coral sea’’; not that its processes require 
this narrow limitation for their operation, but that certain observed 
features, namely, the nearly level surface of submarine banks and 
