388 W. M. DAVIS 
depth of over 50 fathoms, and the bank would then be even more 
difficult of explanation by the glacial-control theory than it is when 
considered without regard to the elevated reef. However, perhaps 
the elevated reef, if it really occurs, is only a remnant of a much 
larger ancient reef which was, after elevation, eroded to its present 
moderate dimensions while the great bank was forming; in this 
case it would cause no embarrassment, provided the formation of 
the bank would not require so long a time as to involve the complete 
removal of any pre-existent reef and the partial erosion of the 
granitic foundation as well. The study of Mahé by the zodlogists 
and oceanographers who have visited it has not included an attack 
on this problem. 
The absence of spur-end cliffs on this granitic island discounte- 
nances the idea that it rises from an abraded platform; and to this 
it may be added that, not only the breadth of its bays, but also the 
strong unconformity between the sea-level fringing reefs and the 
granitic points which they adjoin, indicates a greater measure of 
erosion beneath present sea-level than could have been accom- 
plished while the ocean was lowered during the glacial period; and 
this calls for subsidence from a former higher stand. 
Namuka bank, the largest of several banks in the Tonga or 
Friendly islands of the South Pacific, is peculiar in having a slant- 
ing surface; it is 30 fathoms deep on the eastern side, where it is 
rimmed by a reef that rises toward or to the surface; it inclines 
gradually to the western, reefless margin, where its depth varies 
from 50 to 70 fathoms. Hence this bank has probably been tilted 
since it was formed. Other banks in the Tonga group also suggest 
recent deformation; none of them confirm the assumption that 
their foundations have enjoyed a long-enduring stability. 
As compared with the profound basins of the ocean, submarine 
banks may be all described together as having small and similar 
depths; but when separately examined their individual depths 
vary by large fractions of their mean depth. If large and small 
banks are compared their depths are in a general way found to 
decrease with their diameters, but there are many departures from 
this rough rule. The vast Seychelles bank (40 fathoms) is not so 
deep as the Great Chagos (48 fathoms) and the Macclesfield (55-60 
