CIIANGES OF LEVEL IN CORAL-REEF REGIONS. 319 



When the tide is three feet, beach accumulations of lame 

 masses seldom exceed eight feet above high tide, and the finer 

 fragments and sand may raise the deposit to ten feet; but 

 with a tide of six feet twice this height may be attained. 

 With the wind and waves combined, or on prominent points 

 where these agents may act from opposite directions, such ac- 

 cumulations may be fifteen to twenty feet in height, and occa- 

 sionally thirty to forty feet. These latter are drift deposits, 

 finely laminated, generally with a sandy texture, and com- 

 monly without a distinguishable fragment of coral or shell ; 

 and in most of these particulars they are distinct from reef- 

 rocks. 



2. On islands not coral, the existence of sedimentary de- 

 posits, or layers of rolled stones, inter stratified among the 

 layers of igneous or other rocks constituting the hills. 



B. Evidence of subsidence. 



1. The existence of wide and deep channels between an 

 island and any of its coral reefs ; or in other words, the exist- 

 ence of barrier reefs. 



2. The existence of lagoon islands or atolls. 



3. The existence of submerged atolls. 



4. Deep bay-indentations in the coasts of high islands as 

 the terminations of valleys. — In the course of remarks upon the 

 valleys of the Pacific Islands, presented by the author in his 

 Geological Report, it is shown that they were in general 

 formed by the waters of the land, unaided by the sea ; that 

 the sea tends only to fill up bays, level off the coast, and so 

 give it an even outline. When, therefore, the several valleys 

 of an island continue down beneath the sea, and their enclosing 

 ridges standing out in long narrow points, with abrupt sides,' 

 there is reason to suspect that the island has subsided 

 after the formation of its valleys. For such an island as 



