FRYER— FORMATION OF ALDABRA, ETC. 409 



the atoll. As an instance of this it may be mentioned that, when the tide has flowed 

 for 3 or 4 hours outside Passe Houareau, it is only just full ebb near Michel island, 

 while at the same time the tide has been ebbing for 3 hours in the west of the 

 lagoon. The currents, especially in the neighbourhood of the passes, are naturally 

 very powerful. 



The bottom of the lagoon is rock, but in the centre it is so covered with sand, 

 and near the land-rim with mud, that it is difficult to determine its nature ; the few 

 observations made pointed to rock of the "champignon" type. The origin of the mud 

 and sand is of great interest, for owing to the strength of the currents the falling 

 tide is often rendered quite milky or white by the amount of suspended matter. Little 

 material is carried in by the tides and there is no doubt that the mud is the product 

 of lagoon erosion. When describing the typical northern section I pointed out the 

 signs of erosion discovered round the lagoon shores there: these signs extend round 

 the entire lagoon, while all the multitude of islands have overhanging cliffs, and the 

 rocks are all mushroom-shaped owing to the removal of material between tide marks. 

 The evidence leaves no doubt that all the lagoon islands were once connected and formed 

 part of the main land-rim, and thus at least half the present lagoon is due to its own 

 erosive powers. The latter are somewhat difficult to explain ; but the low level of 

 the land near the lagoon, when considered in connection with the mangrove swamp, 

 may account for some of them. Mention has already been made of the power of the 

 roots of the mangroves to split off overhanging pieces of rock. A more important factor 

 in lagoon erosion is to be found in the mud of the swamp, which is extremely foul 

 with decaying organic matter ; it is therefore charged with carbon dioxide and various 

 organic acids, all of which dissolve or react with the calcium carbonate. The lagoon 

 shores are thus coated with a substance which has considerable power in breaking down 

 limestone and this may be sufficient to account for the rapidity of erosion. In any 

 case there is no doubt that the land-rim is rapidly decreasing in size and is gradually 

 being split up into islands. At this point therefore it is opportune to consider the 

 formation of passes. 



An examination of the chart will show that sea erosion has acted very uniformly 

 over long stretches of coast, while lagoon erosion on the other hand is very irregular 

 and tends to the production of long creeks and bays. From this fact and from the 

 foregoing remarks on erosion it will be readily understood that fresh passes must form 

 from the lagoon and not from the sea. All stages of pass formation are demonstrated 

 on Aldabra : 



Stage 1. At Camp Frigate a long creek from the lagoon has passed almost through 

 the land-rim, the mangrove swamp extending across and opening into what was formerly 

 a basin in the rock but now is open both to sea and swamp. The sea is thus enabled 

 to sweep right in but yet does not seem to have sufficient erosive power to force a channel, 

 and it remains for the lagoon to dissolve away a little more rock, when a powerful tidal 

 current will sweep through at each tide and speedily cut a pass. 



Stage 2. The next stage is shown by several of the western channels, as for instance 

 Passe Femmes, that nearest to Picard Island. The channel itself is fairly open though 



