256 VOYAGE TO THE 



v?n P ' P or tio n only of the width of the island, which slopes 

 — *v~J on both sides, by an almost imperceptible inclination 



1826. to the first ledge, where, as I said before, its descent 

 is very steep ; but this is greatly altered by circum- 

 stances, and the growth or age of the island. Those 

 parts of the strip which are beyond the reach of the 

 waves are no longer inhabited by the animals that 

 reared them, but have their cells filled with a hard 

 calcareous substance, and present a brown rugged 

 appearance. The parts still immersed, or which 

 are dry at low water only, are intersected by small 

 channels, and are so full of hollows, that the tide as 

 it recedes leaves small lakes of water upon them. 

 The width of the plain or strip of dead coral, in the 

 islands which fell under our observation, in no in- 

 stance exceeded half a mile from the usual wash of 

 the sea to the edge of the lagoon, and in general was 

 only about three or four hundred yards. Beyond 

 these limits, on the lagoon side in particular, where 

 the coral was less mutilated by the waves, there was 

 frequently a ledge, two or three feet under water at 

 high tide,* thirty to fifty yards in width ; after 

 which the sides of the island descended rapidly, 

 apparently by a succession of inclined ledges formed 

 by numerous columns united at their capitals, with 

 spaces between them in which the sounding-lead 

 descended several fathoms. This formation, though 

 not clearly established as applying to all the islands, 

 was so conspicuous in some as to justify the conclu- 

 sion with regard to others. At Bow and Matilda 

 Islands, I have been tolerably minute in my descrip- 

 tions of them, and it will be unnecessary here to 

 repeat what has been said there ; but these two, as 



* At Bow Island, on the sea side, it was more. 



