114 TRANSACTIONS OF THE [jAN. 18, 



being similar to the shallow outlet of the sound at the " flats ". 

 This shelf, due to undercutting, has the same origin as the 

 " flats " in Castle harbor or the lagoon. 



Castle Harbor. 



Castle harbor, lying beyond Harrington sound at the north- 

 east and separated from it by a narrow strip of land, is open to 

 the sea on the southerly and easterly sides b}' way of man^^ pas- 

 sages among the islands. The map shows that these passages, 

 with one exception, are mere shallows and that the harbor in 

 general character bears very close resemblance to the lagoon 

 itself. The depth varies from shoal to six or even seven fathoms 

 and the rim of coast except at the southwestwardly side consists 

 of islands separated by the shallows. The greater depths are in 

 enclosed basins. By far the most part of the harbor is very 

 shallow, much of it being reef with only a foot or two of water 

 at ordinary low tide — in many places even bare at extremely low 

 tides, as appears from the whitened tops of some large masses of 

 Diploria. In very many localities the coating of coral is very 

 thin. 



The Bounding Reef. 



The reef along the south shore, often spoken of as the Bound- 

 ing reef, is merely an extension southwardh' of the irregular 

 shoals forming the ocean boundary of Castle harbor. It con- 

 sists of planed off islands extending from one-eighth to some- 

 what more than half a mile from the present coast and, for the 

 most part, covered by water so shallow as to be impassable for 

 fishing boats. They are separated from the shore and from each 

 other by from two to four fathoms of water and, for the most 

 part, break even in moderate weather. If the land should subside 

 fifteen feet another line of islands would be added, the shore 

 would retreat inland nearl}- one-fourth of a mile and in time the 

 islands would be planed ofiT to add another strip to the reef. 

 This condition was recognized in 1894 by Prof. A. Agassiz. 



This detailed examination of conditions on the land and in 

 the several bodies of water leads to but one conclusion — that 

 the contour throughout has a common origin. A subsidence of 

 seventy-five feet would unite the two parallel series of "sinks " 

 in Devonshire parish and the j)rojecting hills would give a divid- 

 ing ridge like that seen in Harrington sound, with seven to 

 twelve fathoms on each side and with four to six fathoms sep- 

 arating the ridge islands. Of course, no such subsidence now 

 would duplicate Harrington sound, for not enough land remains 

 in one body to enclose it, but a subsidence of forty feet would 

 give a similar bodj' of water, differing only in depth. 



