550 W.M.DAVIS SUBSIDENCE OF KEEF-ENCIRCLED ISLANDS 



last inter-Glacial epoch than of the present post-Glacial epoch are thus 

 set aside; "the sea was actively attacking the islands and continental 

 coasts throughout nearly the whole Glacial period. The reef-building 

 corals were largely killed off long before the ice-caps of the first Glacial 

 stage reached their full size" (180) ; broad platforms were therefore 

 abraded. 



In spite of the confidence with which these assertions are made, and 

 notwithstanding the apparent plausibility of the assumption that reef- 

 building corals should have been killed by the temperature of the Glacial 

 ocean, I can not find independent proof that abrasion actually took place 

 as the Glacial-control theory demands. For, as already noted, the spur 

 ends of reef -encircled islands are not clift as they should be if the reefs 

 were killed and if abrasion endured long enough for the truncation of 

 large preglacial islands now supposed to be represented by the Macclesfield 

 and other banks. N"or are they clift as they should be if the shores of such 

 islands were attacked by waves while their now-embayed valleys were 

 deepened, for abrasion by open-ocean waves is a much faster process than 

 valley deepening by small streams and valley widening by subaerial 

 weathering. I have prepared a S23ecial discussion of this aspect of the 

 problem in the article on Tahiti, above mentioned, and need not pursue 

 it further here except to note that the rock-resistance of an island does 

 not affect the conclusion reached. For if the embayments of an island are 

 well opened, it follows that stream erosion and subaerial weathering must 

 have had opportunity of working for a long enough period to deepen and 

 open the embayed valleys to their observed form, whatever the resistance 

 of the rocks may be ; and hence that during such a period the waves must 

 have had abundant time to cut great cliffs on the island margin. The 

 prevailing absence of such cliffs compels me to believe that the organisms 

 on the flanks of the encircling reefs were not so completely killed as to 

 permit abrasion; the exceptional occurrence of cliffs, as on one side of 

 New Caledonia and all around Tahiti, only serves to emphasize the rule 

 of non-clift shores that elsewhere obtains. 



THE EXPECTABLE FORM OF ABRADED PLATFORMS 



However, in order to give the fullest consideration to the Glacial-con- 

 trol theory, let it be accepted that abrasion proceeded as therein assumed. 

 As a result, all coasts of continents and islands then exposed to wave- 

 work would have been cut back in platform and cliffs, of dimensions pro- 

 portionate to the resistance of their rocks, of their exposure to the waves, 

 and of the duration of wave attack. Let the narrowness of abrasion on 

 many reef -encircled islands — such as Kusaie in the Caroline group, where 



