88 CONDITIONS FAVOURABLE TO Ch. IV. 



accumulate in the hollows formed by tearing out the 

 coral's, but not on the broken and projecting- stumps ; 

 and therefore, in the former case, the fresh growth 

 of the coral would be prevented. By this means 

 the inhabitants keep their harbours clear ; and thus 

 the French Governor of St. Mary's, in Madagascar, 

 6 cleared out and made a beautiful little port at that 

 place.' 



In the last chapter I remarked, that fringing- 

 reefs are almost universally breached where streams 

 enter the sea. 1 Most authors have attributed this fact 

 to the injurious effects of the fresh water, even where 

 it enters the sea only in small quantity and during a 

 part of the year. No doubt brackish water would pre- 

 vent or retard the growth of coral ; but I believe that 

 the mud and sand, which is deposited, even by small 

 rivulets when flooded, is a much more efficient check. 

 The reef on each side of the channel leading into Port 

 Louis at Mauritius, ends abruptly in a wall, at the 

 foot of which I sounded, and found a bed of thick mud. 

 This steepness of the sides appears to be a general 

 character in such breaches : Cook, 2 speaking of one at 

 Eaiatea, says, ' like all the rest, it is very steep on 

 both sides.' Now, if it were the fresh water mingling 



1 Lieut. Wellstead and others have remarked that this is the case in 

 the Ked Sea: Dr. Kuppell (Eeise. in Abyss. Band. i. s. 142) says that 

 there are pear-shaped harbours in the upraised coral-coast, into which 

 periodical streams enter. From this circumstance, I presume, we must 

 infer that before the upheaval of the strata now forming the coast-land, 

 fresh water and sediment entered the sea at these points ; and the coral 

 being thus prevented growing, the pear-shaped harbours were produced. 



2 Cook's First Voyage, vol. ii. p. 271. (Ha wkes worth's edit.) 



