338 ALEXANDER AGASSIZ 



because of the existence of a depression on the summit 

 which was looked upon as the remains of a lagoon. 

 Agassiz, however, did not believe these cuplike for- 

 mations represented the floors of old lagoons, but con- 

 sidered them similar to the gigautic banana holes, as 

 they are called, found in the Bermudas. He attributed 

 such depressions to causes now going on and looked 

 upon them as the first process in the erosion of the islands. 

 The decaying vegetation, thickest in the interior of a 

 limestone-island plateau, on settling in any inequality 

 or fissure on the flat top of an island, forms acids. These 

 greatly intensify the solvent action of the rains, which, 

 percolating through the mass, carry off the limestone. 

 A drainage from the edffes toward the centre is estab- 

 lished, and we get the beginning of the saucer-shaped 

 basins so characteristic of the elevated limestone islands 

 of Fiji. At first there is but a slight depression ; this 

 gradually deepens, till when the sea finally breaks in 

 we have an island like Fulanga, about whose outer flanks 

 corals have established themselves. A further process of 

 erosion would result in wearing away this land until 

 nothing remained of the original island but a few islets 

 rising from a denuded reef as in Wailangilala. And 

 finally when the process is carried still further, nothing 

 is left of the island but a submarine ridge upon which 

 corals have established themselves, like Reid Haven. 

 (See colored plate, Figs. 1.) 



Where islands, composed either of volcanic material 

 or limestone, have been eroded to form a submarine 

 -platform, upon which corals have obtained a footing, 

 Agassiz would explain the existence of the lagoon as fol- 

 lows : — The great rollers piled up by the trade winds 

 break over the outer rim, protected by a more vigorous 



