AGASSIZ: BERMUDAS. 275 



archipeJago, and much more, were dry land." But I do not think that 

 "it was at this time, doubtless, that the great sand dunes were ele- 

 vated." On the contrary, I do not imagine the dunes to have been 

 elevated, but to have perhaps been blown to their greatest height in the 

 manner suggested from a broad coral sand beach. 



Heilprin adopts the suggestion thrown out by Rice, that these seolian 

 accumulations could only have been formed at a time when large areas 

 of reef, and not a simple atoll ring, were exposed above the water level,^ 

 while it is perfectly true, as Heilprin says, that all the sand formed at 

 the present day is derived from the destruction of the existing land 

 masses, and not as a product of the disintegration derived from the 

 growing reef, — a statement which by the way hardly agrees with the 

 graphic description of the destruction of the coral reef off the south 

 shore given by him a page or two before. 



It does not seem to me necessary that there should have existed very 

 wide areas of coral reefs for the formation of the Bermudian seolian hills. 

 A reef of a width of 1,200 to 1,800 feet, such as are known to exist, 

 seems to me ample to supply the material necessary for the formation 

 of the seolian hills, especially if, as may hjive been the case judging by 

 the soundings off the reef, there also existed a comparatively wide shal- 

 low bank outside of it. Taking the subsidence of the Bermudas as 

 twelve fathoms, this belt could have been a belt of twelve to twenty 

 fathoms, a shallow bank, the wear and wash from which would alone 

 supply a large part of the material needed for the formation of the 

 Bermudian dunes, and carry the topography now prevailing pretty well 

 over the whole of the bank inside the ten or twelve fathom line. In 

 addition, we should of course also have the supply derived from the reef 

 itself. The amount of material which can be supplied from a compara- 

 tively small reef and its adjacent bank I saw well illustrated at the 

 Sandwich Islands. A number of dunes were constantly travelling inland 

 from the beach at Spreckelsville, and had covered to a considerable ele- 

 vation a great part of the isthmus connecting the two islands which con- 

 stitute Maui. Very high dunes, from 120 to 180 feet, of nearly a mile 

 in width, occur on the shores of the Baltic, the material of which is 

 derived from a comparatively narrow beach range. 



1 fully agree with tliose who before me have examined the Bermudas, 

 and who consider that subsidence has brought about the existing condi- 

 tion of the islands and sounds. But that is a very different thing from 

 assigning to the corals now growing the formation of the islands owino' to 



^ The Bermurla Islands, p. 46. 



