PHYSICAL HISTORY AND GEOLOGY. 37 



There can be no doubt, it appears to me, as Rein* first clearly 

 demonstrated, that Harrington Sound is not the lagoon of a 

 marginal or secondary atoll, but merely a hole that has been 

 cut out of the land by the sea. I think that every one who 

 has seen the working condition of the water in the Sound, the 

 undercut ledges, the scattered islands and rocks, and above 

 all, the precipitous cliffs, which appear on opposite sides of the 

 water, and show an arrangement of lamination or stratifica- 

 tion similar to that which. is observed in the cut cliffs of the 

 south shore, must arrive at the same conclusion. The same 

 is manifestly also true of much, if not the greater part, of Cas- 

 tle Harbor, which still retains a sea-ward border in the belt 

 of disrupted land which forms Castle Point, and Castle, Goat, 

 Nonsuch, and Cooper's Islands. The widening or expansion 

 of this body of water presents itself vividly to the eye of the 

 observer stationed on an eminence, such as that of St. David, 

 whence the field of vision takes in the patches of separating 

 and separated land which are awaiting the hour of their 

 destruction. 



Along the borders of Castle Harbor at least as far as we 

 observed it on the west and south there is a broad flat ledge, 

 over which the depth of water is only from about six to ten 

 feet ; beyond this there is an abrupt drop into the deeper parts 

 of the lagoon. This feature is frequently found in the true 

 atoll-lagoons, where it forms a shore platform similar to that 

 which is formed around the outer surfaces of sea-cliffs. In 

 how far this ledge may represent a simple coral outgrowth 

 from the shore, or determine a measure of subsidence, can- 

 not well be ascertained. Large numbers of giant brain -corals 

 ('Mreandrina and Diploria), measuring three, four, and five 

 feet in diameter, are scattered over it, and form a series of 

 stepping stones in the water. Many of them grow on and 

 over the edge of the platform, so that the latter overhangs in 

 some places. These corals appear to be absent, or at least 



*Bericht Senckenberg. Naturf. Gtsellsch., 1870, p. 153. 



