122 TRANSACTIONS OF THE [.JAN. 18, 



ing during the accumulation of the sandstone. The problem is 

 perplexing, so perplexing that no matter which explanation the 

 student accepts he soon will wish that he had accepted another. 



In any event, however, the period of subsidence was succeeded 

 sooner or later b}^ one of elevation, during which the land was 

 raised to approximately its former height. Evidently the old 

 lines of subterranean drainage were reopened and the poorly 

 consolidated sandstone was acted upon very rapidly. Without 

 doubt, many of the old depressions were emptied of sandstone, 

 while new lines of subterranean drainage led to the formation of 

 new groups of " sinks." It is difficult now to distinguish be- 

 tween old and new for sandstone forms the walls of all sinks 

 above water level. 



This period of elevation was long enough to admit of reerod- 

 ing or of forming mostly by subterranean action the great de- 

 pressions of Harrington sound, Pembroke marsh and others of 

 like character in various portions of the archipelago, extending 

 fifty to one hundred feet below the present water level. 



A period of subsidence followed and continued to not a great 

 while prior to the discovery of the islands. There is little rea- 

 son to suppose that any subsidence is now in progress. Professor 

 Rice* has discussed this question carefully from the standpoint 

 of the cartographer, while Professor Agassiz could find no evi- 

 dence of any change in the shallow channels, which, for a cen- 

 tury, have afforded bare passage for fishing boats. Tbe subsi- 

 dence reached almost the maximum of the previous subsidence, 

 the marine deposits being within less than ten feet at most of 

 their original position — a rather interesting fact in view of the 

 other fact that within at least three hundred years no subsidence 

 has occurred. The sinking was very slow, so slow that the for- 

 mation of peat,"}" in Pembroke, and possibly in Devonshire marsh, 

 kept pace with it, those being inclosed spaces not communica- 

 ting with the sea by any surface channel; the great marsh of 

 Mangi'ove creek now communicates with the sea by a shallow 

 channel, being a breached sink; similar swamps exist in Paget 

 and Southampton, as well as in southern Smith and Devonshire, 

 but, so far as could be ascertained, no data exist respecting the 

 thickness of their peat. In other, deeper and perhaps older de- 

 pressions the water gained control through underground chan- 

 nels, as in Harrington sound, the basins within the "hook," 

 those of the lagoon and perhaps some of those in Castle harbor. 



Subsidence must have been very slow in the later portion of 



*Loc. cit.,p. 19. 



t Governor liCfroy informed Professor Rice that the peat in Pembroke marsh is 40- 

 50 feet deep. Loc. cit., p. 7. 



