in the Pacific^ 6fc. 89 



Ka lua Pele, and extended if I remember aright, some seven 

 or eight leagues to the coast. 



But to return from this long digression, to the subject more 

 immediately under consideration. 



It is obvious that as the land sinks, and the water reaches 

 the base of any ridge, barrier, or mount, a shore reef may be 

 formed upon or around it, which, if the subsidence continues, 

 will, from the operation of causes already explained, be 

 gradually converted into an encircling or outlying one. There 

 will naturally be intervening channels, in place of some of 

 the ravines, while on those of a trench-like character and upon 

 the ruins of ancient craters, will form lagoon reefs, and on the 

 whole being overflown, there will thus be formed a group of 

 lagoons varying in size and configuration, according as they 

 rest upon a sunken crater, a ridge, or one of the trench-like 

 ravines, and surrounded by a common reef, which is traversed 

 at intervals, as in its primary state, by passages of various 

 breadth and depth. And such, on a large scale, as I have be- 

 fore observed, is the appearance presented by the Paumotu 

 groups and dangerous archipelagos of Polynesia. 



Although the seaward side of the reefs encircling these, 

 has been described as rising in a perpendicular wall, yet it 

 must not be understood, that by this it is meant, that we lit- 

 erally step from an unfathomable ocean, upon the upper sur- 

 face of a reef. They present a succession of terraces or 

 plateaus, the outer having sometimes twelve or fifteen 

 fathoms ; and in one instance, that of Bellinghausen's Island, 

 twentyeight fathoms of water was found upon it. This low- 

 est plateau is of variable breadth, but I think seldom exceeds 

 one hundred and fifty feet ; declines somewhat rapidly sea- 

 ward, and apparently projects beyond the wall like a shelf, as 

 I have known the lead to fall from twelve fathoms on it, to 

 two hundred and no bottom, within a distance of about ten 

 yards. 



These terraces become, as they recede from the sea, nar- 

 rower and shoaler, presenting a like declination with the low- 

 est, and having at their extremity an abrupt descent of several 

 feet. The highest, or last formed, differs in its margin forming 



12 



