148 A NATUBALI8T IN CELEBES ch. vi 



the same positions in some of the Zoantharian corals, al- 

 though I have no evidence to show that they do. There 

 is then some reason for believing that many of the corals 

 may be partially at least vegetable feeders, and herein we 

 see the possibility of a relationship between the mangrove 

 swamp and the coral reef which has not been previously 

 recognised. If the food of corals is to a certain extent 

 supplied by debris from the trees of the mangrove swamps, 

 we can see why coral banks in the neighbourhood of 

 extensive swamps are often so much more vigorous than 

 those which fringe the steeper shores. The old coral reefs 

 of the shore form, as I have previously pointed out, a sub- 

 stratum for the mangrove swamp, and in return the swamp 

 supplies a certain amount of vegetable food for the living 

 corals. 



Perhaps this is not the only service it renders the 

 living corals. Corals are dependent for their normal growth 

 not only upon their food, but also upon the carbonate of 

 lime in solution in the water, which supplies them with the 

 material for building up their skeletons. But carbonate of 

 lime is only slightly soluble in pure water, the amount 

 which any quantity of water will hold in solution depending 

 upon the amount of free carbonic acid it contains. Now 

 the water which flows through the mangrove swamps must 

 become highly impregnated with carbonic acid and thus 

 become capable of dissolving a comparatively speaking large 

 amount of the carbonate of lime which exists in abundance 

 in the millions of foraminiferous shells and lumps of coral 

 skeletons with which the coasts are littered. The sea- water 

 then which flows through the swamps must be highly 

 charged with carbonate of lime, and therefore particularly 

 favourable for the increase and growth of coral. 



Another problem of considerable interest and import- 

 ance in connection with coral reefs is the meaning of the 



