CORAL REEFS AND SUBMARINE BANKS 309 
dissected volcanic islands encircled by close-set barrier reefs 
ought today to be cut off in cliffs, the face of which, where not 
obscured by apposed deposits, should descend some tens of fathoms 
below present sea-level. This argument applies with special 
force if the great Seychelles bank be considered; for there the several 
granitic islands that are still visible strongly suggest that any 
large preglacial island occupying much of the area of the bank 
would have included a good proportion of resistant rocks in its 
constitution, and such an island could have been reduced to a 
platform of abrasion only by strong wave work continued long 
enough to cut spur-end cliffs on smaller volcanic islands elsewhere, 
around which narrow preglacial reefs were the only defense. 
The inconsistency of the glacial-control theory in demanding 
broad abrasion for large atolls and submarine banks and in allowing 
only narrow abrasion around high volcanic islands within fringing 
or close-set barrier reefs is not, to my mind, satisfactorily explained 
by assuming that the atolls and banks were all represented in pre- 
glacial time by the deeply weathered, weak rocks of old, worn- 
down, still-standing islands, while the high islands were composed 
of resistant rocks of younger islands; for there is nothing to support 
such an assumption as to the preglacial representatives of atolls 
and banks except the consequence that is desired to follow from it; 
indeed the granitic islands of the Seychelles group contradict the 
assumption. Still less has a satisfactory explanation been offered 
for the other inconsistency—pointed out on page 294 in the account 
of Murea—of demanding that comparatively broad valleys must 
be eroded with respect to the lowered sea-level of the glacial period 
by the action of slow-working subaérial processes on the resistant 
rocks of high volcanic islands in the mid-Pacific area of assumed 
stability, and of denying that rock platforms and spur-end cliffs 
of significant dimensions can have been abraded in the same rocks 
while they were undefended by living reefs and attacked by the 
vigorous waves of the trade-wind seas during the same period of 
time as that in which the broad valleys were eroded. Nor can a 
satisfactory explanation of this inconsistency be found without 
concluding that the non-clift islands were protected all through the 
glacial period by living reefs. The upshot of this is that submarine 
