in the Pacific, ^c. 99 



had once been girt by such a wall, the conclusion was not 

 unreasonable that the constant pressure of the water, poured 

 over from the windward side of the reef, would eventually 

 form a breach in the weakest portion of its leeward side. But 

 in truth such a condition of things never existed, there being 

 in most Paumotus to the present day, on one side or other, a 

 considerable portion of the reef where there is no ridge to bar 

 the escape of the water from the lagoon, and yet a passage is 

 found. Nor, as I have said, is this always on the leeward 

 side. At Raraka, which is situated in the full strength of 

 the south-east trades, it is on the southern, or windward side. 

 At Clermont Tonnerre, and Tooa, the westernmost of the 

 Disappointment Islands, the entrance, if any exists, must be 

 to windward, as the coast in that direction is very low and 

 broken, and none was discovered in running down the lee 

 shore. At Waterland Island the opening is in the eastern or 

 weather reef. 



Moreover, there are reefs, like that of Ocean Island, as yet 

 wholly submerged, save in two or three small spots, which 

 are intersected by no less than four canals. At this island, 

 there is one entrance on the south-east side, and three about 

 half a mile apart in the south-west portion of the reef. In 

 these last, it is worthy of notice that the depth is eight or ten 

 feet greater than that of the lagoon, and therefore could not 

 have been excavated by the outward rush of its waters, as 

 this at most could have worn a passage in the reef to the level 

 of their bed. 



On the other hand, at Aitoho, the eastern Disappointment 

 Island, no passage was met with, in pulling all round the isl- 

 and, although portions of it are so low, that the surf at high 

 water must break heavily into the lagoon.* 



At Rose Island, where, judging from the drift line on the 

 beach, there is a rise and fall of about five feet, the entrance 

 is to leeward, and the same at Christmas Island. Through 

 all these channels, the ebb pours out with great velocity, 



• This peculiarity, with the apparent great depth of its waters, and its circular 

 outline, render it possible that this lagoon is one of those occupying the site of 

 a submerged crater. 



