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



rocks are hard or soft. But, as has already been said, sea cliffs are 

 of exceptional occurrence around the shores of barrier-reef 

 islands; the spur ends are as a rule very little cut back, and 

 where low cliffs are seen, they are usually fronted by a visible 

 rock platform, showing that the cliffs were cut at present sea 

 level. Hence as far as these two additional lines of evidence 

 go, it must be concluded, as before, that reef-building corals 

 were generally not killed during the glacial period, and that 

 the flanks of preglacial reefs were protected by growing corals 

 at whatever level the waves beat upon them. Sea waves 

 should not, therefore, be appealed to as of dominating impor- 

 tance in the abrasion of the lagoon floors now enclosed by bar- 

 rier reefs or by atolls. 



It should be noted in this connection that Agassiz reported 

 the frequent occurrence of slightly uplifted "Tertiary lime- 

 stones" in the atoll i-ings of the Paumotus: had the corals of 

 those atolls been killed, the uplifted Tertiary limestones 

 should, according to the glacial-control theory, have been cut 

 down to the level of the present lagoon floors, where they 

 would be to-day invisible ; but they have not been cut away, 

 and hence, if the uplifted limestones are really Tertiary, the 

 corals there were not killed. 



Nevertheless, even if the equatorial ocean were not chilled 

 enough during the glacial period to kill its reef-building corals, 

 the ocean surface must have then been more or less lowered ; 

 and, as the outer slope of the reef is steep, the streams of bar- 

 rier-reef islands must have been temporarily impelled to 

 deepen their valleys. The question thus arises whether nor- 

 mal erosional processes, independent of marine abrasion, could 

 during the glacial period have excavated the valleys that are 

 now invaded by the sea in the existing embayments of barrier- 

 reef islands. 



Depth of Valleys in Barrier-reef Islands. — The first step to 

 be taken in attempting to answer the above question is to 

 deduce the essential features of barrier-reef islands, as they 

 existed just before the glacial period, on the assumption that 

 whatever reefs then surrounded them had been formed by out- 

 growth on still-standing foundations, in accordance with the 

 postulates of the glacial-control theory. These features must 

 as a rule include a less or more dissected volcanic cone, 

 descending to the simple, unembayed shore-line that was 

 formed when the volcano was built up; a surrounding reef- 

 plain of lesser or greater breadth, on which a delta advances 

 opposite each valley mouth ; the structure of the reef being 

 that of an advancing talus, lying on a non-eroded, submarine 

 volcanic slope, with coral in place only in the upper 120 feet 

 of its total thickness. 



