52 TPE BERMUDA ISLANDS. 



There is little or no growth of coral immediately along the 

 south shore, doubtless due to the great quantity of sediment 

 that is constantly being washed off from the cliffs. The rocks 

 of Harrington Sound, on the other hand, are largely fringed 

 with patches of Isophyllia, Siderastraea, and Millepora, while 

 in the deep quieter' waters, judging from the number of our 

 hauls; Oculina is by no means scarce. Reference has already 

 been made to the vast development of Diploria and Msean- 

 drina on the projecting platform of Castle Harbor, over which 

 the water is normally in a condition of fair stability. This 

 condition confirms the view expressed by Bourne that the 

 coral growth of the inner waters is much more extensive than 

 is generally stated to be the case. A very large part, however, 

 if, indeed, not the greater part, of the floor of the big lagoon 

 is practicallv barren. 



Although the zone of animal activity in a coral island ceases 

 with the water-line, the actual growth of the island does not stop 

 there, but is continued upward by the mechanical and vegeta- 

 ble forces. The destructive action of the billows carries frag- 

 ments of coral-rock far above the limits of coralline existence, 

 triturating the masses into minute surfaces, and upon this im- 

 provised soil a luxurious vegetation, whose origin lies in the 

 seeds wafted thither by the winds, or deposited by birds, may 

 in course of time spring up. Where the action of the breakers 

 is greatest the coral rock assumes the greatest compactness, 

 since the fragments and particles that are derived from the 

 mechanical wear and tear are here firmly lodged or compacted 

 into tlie spaces of the otherwise comparatively loose coral 

 structure. On the oceanic side of the island we find shallow 

 water — ranging to several hundred feet — for a distance of 

 between 300 to 1500 feet, beyond which the descent becomes 

 rapid, dropping suddenly to several thousand feet. At a dis- 

 tance of less than three-quarters of a mile from the Island of 

 Clermont Tonnerre, ike lead was run out to a depth of 3,600 

 feet, and yet no bottofi was found ; at a distance of seven miles 

 a run of 6,000 feet failed to strike bottom. Off the Cardoo 



