THE PACIFIC OCEAN. 253 



imaginable, where ships may lie, and wood and 

 water, and refresh their crews, in security, though 

 the tempest howl without. It is a scarcely less bene- 

 ficent provision that the position of the openings 

 is in most cases indicated so as to be visible at a 

 great distance. Had there been merely an opening 

 in the coral rock, it could not have been detected 

 from the sea, except by the diminution of the foaming 

 surf just at that spot ; a circumstance that could 

 scarcely be visible, unless the observer were opposite 

 the aperture. But, in general, there is on each side 

 of the passage a little islet, raised on the points of the 

 reef, which, being commonly tufted with cocoa-nut 

 trees, is perceptible as far off as the island itself, and 

 forms a most convenient landmark. 



Notwithstanding that the highest point of these 

 narrow islets is rarely more than a yard above the 

 tide, it is a remarkable fact that fresh water is fre- 

 quently found in them. It is probable that the coral 

 rock acts as a filter, allowing the sea-water to perco- 

 late through its porous substance, but excluding all 

 its saline particles held in solution. 



Though I have described the two parts of a Coral 

 Island, or Atoll, as it is called, as distinct, yet the 

 difference is only in appearance ; the foundation on 

 every side is the same, a coral reef rising to the sur- 

 face ; but the side most exposed to the action of the 

 waves driven in by the trade-winds is invariably the 

 first to be projected, and attaius a higher elevation 

 than the leeward side. Neither must it be supposed 

 that the belt to windward is always continuous, 

 though the interruptions are comparatively few. 



s 



