74 riiOFESSOU T. A\'. EDGP^WORTH DAYID AND MR. G. SAVEET. 



though the work of corrosion is in those latter cases less apparent, because as fost as 

 the material is dissolved away from beneath it is partially renewed by additions of 

 new material either fi'om the ocean side or from tlie lagoon side of the reef In fact 

 the reason for the existence of this central hollow or groove nnniing down the 

 centres of the islets is tliat while the whole of the reef rim above low tide is being- 

 dissolved slowly by h\nnic acid, the work of construction on the portions nearest to 

 the lagoon and ocean are nearly sufHcient to keep pace Avith the destructive solvent 

 forces. In the central portions of the islets, however, the destructive forces have it 

 all their own wny down to the level of low tide, as these portions of the islets are 

 too far distant from the Hurricane Bank on the one hand, and the Lagoon Bank on 

 the other, to receive any additions of new material to replace the loss through 

 solution. Obviously the calcium carbonate in solution imist go somewhere. Part 

 of it sinks through crevices and irregular hollows in the reef into the sea beneath 

 (the foundations of the islets as already stated being as porous as a sponge), and 

 part oozes outwards into the beach sands. Heated as these are by the tropical sun 

 they promote rapid evaporation in the calcareous solutions passing through them. 

 Hence calciinn carbonate wo\dd form and be deposited as cement between the sandy 

 particles.* 



Mr. Stanley Gardiner's observations on tlie formation of a similar rock at 

 Rotuma have such an important bearing on this s\d)ject that we c[uote them 

 heret : — -" On the ])each between tide-maiks it (our beach sand) ends with a 

 beach sandstone formation. . . . The toj) of a layer which is the part most 

 exposed to the air and to the waves is the liardest. If it is removed, as is 

 constantlv done bv the natives for iiravestones, it seems to be reljuilt, i>T0Avini>' 

 lip as it were from the broken edge ; fresh layers on the outside seem to be formed, 

 too, in a similar manner." 



It may be added that a somewliat similar calcareous sandstone, of quite recent 

 origin, occurs on the beacli between tide marks at Long Reef, near Sydney. Its 

 occinrence just at this spot is in that case clearly due to a local soakage of humous 

 acid from an adjacent peat swamp. The reason for the calcareous sandstone there 

 and at Funafuti, &c., forming just betAveen tide marks, the upper limit being usually a 

 little below the level of high water, is related probably to the limits of the saturation 

 zone (?.('., the zone saturated by soakage froni the land), the upjier limit of which, of 

 course, varies with tlie tides, but which is normally a little below the level of higli 

 tide 



* An analogous dejjosit is t'orming at Noifolk Island, as (Icscriljed I)y -J. E. Caunk, F.G.S., of the 

 Geological Survey of N. S. Wale.?, 'Annual Keport Department Mines, N. S. Wales,' for 1885, j). 145. 

 In this case what he terms an oolitic limestone has incrusted the chain cable of the " Sirius," a 

 ship which was wrecked at Noi'folk Island in 1795. This incrustation of the chain with hard fragmental 

 limestone has taken place between high and low wiiter, though nearly at the level of low tide. 



Darwin refers to the formation of a similar calcareous sandstone, ' Coral Reefs,' Chaptei' I., Section 1. 



t 'Pruc. Camb. Phil. Sue.,' 1895-1898, vol. 9, p. 443. 



