278 CORALS AND CORAL LSLANDS. 



north and east sides a compact coral wall of about five feet 

 elevation, which continues for four and a quarter miles, and 

 then becomes a line of detached rocks at tide level. This 

 bare wall, thus described by Capt. Wm. Reynolds, U.S.N. , 

 appears to be an indication that the land was once finished off 

 under a cessation of subsidence, but that a sinking of small 

 amount has since taken place, amounting perhaps to four 

 ox five feet. 



Ocea?i Island^ in 28^ 25' N., 178° 25° W., another of this 

 range, is very similar to Brooks's in its wall of coral rock on 

 the east; and so also is Pearl and He7'mes'' reef, in 27° 50' 

 N., 176'' W., though the wall of the latter is more a series of 

 detached rocks than a continuous parapet. 



Marquesas. — The Marquesas are remarkable for their abrupt 

 shores, often inaccessible cliffs, and deep bays. The absence 

 of gentle slopes along the shores, their angular features, 

 abrupt soundings close alongside the island, and deep indenta- 

 tions, all bear evidence of subsidence to some extent ; for 

 their features are very similar to those which Kauai or Tahiti 

 would present if buried half its height in the sea, leaving 

 only the sharper ridges and peaks out of water. They are 

 situated but five degrees north of tlie Paumotus, where eighty 

 islands or more have disappeared, including one at least fifty 

 miles in length. There is sufficient evidence that they partici- 

 pated in the subsidence of the latter, but not to the same 

 extent. They are nearly destitute of coral, and apparently 

 because of the depth of water about the islands. 



Gambler Group.-^ln the southern limits of the Paumotu 

 Archipelago, where, in accordance with the foregoing views, 

 the least depression in that region should have taken place, 

 there are actually, as we have stated, two high islands, Fit- 

 calnls and Gambler s. There is evidence, however, in the 

 extensive barrier about the Gambler^ s (see cut on page 227), 

 that this subsidence, although less than farther north, was by 

 no means of small amount. On page 126, we have estimated 

 it at 1,150 feet — possibly 1,750. These highlands therefore, 

 although toward the limits of the subsiding area, were still 



