232 W. M. Davis — Shaler Memorial Study of Coral Reefs. 



suggested, barrier reefs are converted into atolls by the wear 

 ing away of the central islands. 



The two islands which have deltas advancing into their lagoons 

 are both large : one is Viti Levu, the largest of the Fiji group, 

 where certain rivers of a considerable volume and of greatly 

 increased flow after heavy rains have built their deltas forward, 

 outside of the general outline of the island ; the other is Tahiti, 

 the largest of the Society group, where a present still-stand, 

 recognized by Darwin as well as by Murray, is attested by a 

 belt of alluvial flats and deltas that is more or less continuous 

 around the island ; but recent subsidence is well proved for 

 both these islands by the form of their delta plains, which head 

 between advancing spurs in strongly reentrant spaces that were 

 assuredly drowned- valley embayments before the deltas filled 

 them. J^ew Caledonia might he added as a third example of 

 a large barrier-reef island which possesses good-sized deltas, 

 but the deltas there are as a rule contained in the embayments 

 and do not yet extend beyond the partly drowned headlands 

 by which the embayments are bordered. Large deltas are 

 similarly situated between the headlands of the embayed 

 Queensland coast. Submergence unquestionably preceded delta 

 formation in all four of these examples. 



Reefs as Yeneers on Wave-out Platforms. — The explana- 

 tion of harrier reefs as relatively thin veneers on the outer 

 edge of wave-cut platforms around still-standing islands, was 

 long ago suggested by Tyerman and Bennet (quoted by Dar- 

 win) and afterwards advocated by Guppy and Agassiz; it was 

 extended by Wharton to the explanation of atolls, by suppos- 

 ing that in such cases the volcanic islands were completely 

 truncated before corals were established on the resulting plat- 

 form. The explanation is not satisfactory (1), because, as 

 already stated, the occurrence of drowned-valley embayments 

 in barrier-reef islands shows conclusively, as in sector O, fig. 3, 

 that the relation of land and sea level has changed ; (2) because 

 the prevailing absence of sea cliffs, sectors H, J, around cen- 

 tral islands, negatives, as Darwin long ago pointed out, the 

 occurrence of a sea-cut platform of significant width ; (3) 

 because the depth of many lagoons is greater than that of wave- 

 cut platforms of the same width ; (4) because no almost-atolls 

 are known in which the residual central island is a cliff-rimmed 

 stack, as in sector K' ; (5) because no sufficient reason has been 

 found to explain the delay in the establishment of reefs, as 

 here required, while a broad platform was abraded ; (6) because 

 no uplifted reefs have been found in the form of relatively 

 thin veneers on the outer edge of broadly abraded rock 

 benches, as in sections H' and J'. It is, however, true that 

 nearly all the spur ends of the submaturely dissected cone of 



