1888.] and the Coral Formations of the Indian Ocean. 



455 



places are explained by the behaviour of a current when it meets with 

 an obstacle which presents a sloping wall to it. The lowest parts of 

 a strong and deep current are the first to strike the sloping wall or 

 glacis ; they are in part divided and sweep around the sides of the 

 obstacle, but a large part of the current is deflected upwards over the 

 slope, and its action, combined with that of the more superficial parts 

 of the same current, results in a moderate current flowing upwards 

 over the slope (vide Plate 4, fig. 2). This moderate current favours 

 coral growth on the slope, but at its upper edge the superficial current 

 combines with the upward stream, and the two dash onwards with 

 increased force over the flat upper surface, destroying all c<>ral life 

 there. At the sides of an island the current flows tangentially, with 

 moderated force, especially in its more superficial portions, and affords 

 the necessary conditions on the external slopes there, whilst on the 

 reverse side of the island the backwash affords weak currents, which 

 are highly favourable to coral growth. Thus it is that everywhere 

 around the island the external slopes are covered with a luxuriant bed 

 of corals. 



To fully understand the manner in which currents are moderated 

 in flowing past an obstacle, it is necessary to remember that water 

 possesses a eertain amount of adhesiveness, and tends to cling to 

 the sides of the obstacle, so that the current is always rather stronger 

 at a little distance from the obstacle, whatever it may be, than it is 

 where it runs close against it, and the rougher the surface past which 

 it flows the greater is the adhesion, as is well known to everyone who 

 has had experience in boating or shipbuilding. 



The net result of all this is that the corals are always thickest alonp- 

 the slopes around a coral reef, and the reef tends to increase at its 

 periphery, growing upwards there, whilst it tends at the same time to 

 spread outwards. These principles hold good in the case of a sub- 

 merged bank as well as in the case of a reef that is awash, and a sub- 

 merged bank must tend in the course of time to reach the surface in 

 its circumferential portions, and form an atoll-shaped reef, on the rim 

 of which detritus may be heaped from place to place, forming shingle 

 cays or islets which may temporarily form dry land. In atolls where 

 storms are of frequent occurrence, regular storm-beaches may bo 

 formed, till the fragments piled high upon one another may form low 

 islets standing some 6 or 10 feet above high water mark, upon 

 which vegetation may subsequently find a footing. Atolls are often 

 formed in this way, without any elevation taking place, and such has 

 undoubtedly been the case in the Florida reefs, where atolls (the Tor- 

 tugas) and barrier reefs and islands have been formed in an area of 

 complete rest. No one who has read the admirable work of Alex. 

 Agassiz on the B'lorida reefs can fail to agree with the author's con- 

 clusion that the islets there have been formed by the action of the 



