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



the less resistant but by no means weak rocks in the last two 

 localities to lowlands of small relief, for submerged lowlands 

 there correspond to submerged valleys elsewhere. The ques- 

 tion before ns is to compare these indications of pre-submerg- 

 ence erosional work in the Pacific with the valleys and lowlands 

 of formerly glaciated regions, in order to estimate whether the 

 observed amount of valley erosion on the Pacific islands could 

 have been accomplished during the glacial period. The com- 

 parison is difficult, because of differences in rocks and in cli- 

 mate, as well as because of a considerable measure of 

 uncertainty as to the preglacial form of the valleys of glaciated 

 regions ; the result of the comparison will therefore be rather 

 an impression than a demonstration. It is to the effect that 

 the hard-rock valleys and weaker-rock lowlands of such gla- 

 ciated regions as the northeastern United States, northern 

 Great Britain, central France, and the marginal areas of the 

 Alps are largely of preglacial origin, that they were not greatly 

 changed by normal erosion during the glacial period, and hence 

 that they have required a much longer time than the glacial 

 period to reach their present erosional development ; therefore 

 the comparable valleys and lowlands of erosion on Pacific 

 islands could not have been excavated during the glacial 

 period, much less during that part of the glacial period repre- 

 sented by the sum of the glacial epochs. 



Hence it is, as above noted, not so much the depth of valley 

 erosion even by slow-working agents such as small streams 

 that is the most important matter to be considered in this com- 

 parison, but the widening of hard-rock valleys and the general 

 degradation of weaker-rock lowlands by the much slower pro- 

 cesses of weathering, washing and creeping. The incision of 

 rather narrow and occasionally cliff-walled valleys with cascad- 

 ing streams of steep descent, while the baselevel of erosion 

 was formerly relatively lower, in the slopes of a relatively 

 young volcanic island, such as Ovalau in the Fiji group, with 

 the resultant development of short and small embay men ts 

 when the island was submerged to its present shoreline, as in 

 the middle sketch of fig. 6, represents only a moderate begin- 

 ning of the erosional work that has been accomplished in more 

 maturely dissected and more intricately embayed islands, like 

 Yanua Levu or Kandavu in the same group, or like fluaheine, 

 Raiatea and Tahaa in the Society group, as in the left-hand 

 sketch ; and only a still smaller beginning in the reduction of 

 a massive volcanic cone to skeleton islands, as in the right-hand 

 sketch, like Borabora, taken as a typical example by Darwin 

 and Agassiz, or Gam bier island, similarly instanced by Dana. 

 Again let it be remembered that the most significant feature 

 here to be considered is not the depth of erosion below present 



