1888.] and the Coral Formations of the Indian Ocean. 457 



de terrain par an, tani interieurement qu'exterieurenient, excepte 

 aux pointes N.E. et N.O., ou une partie des sables entraines dn fond 

 de la baie par les vents de snd-est, conservent a ces deux pointes leur 

 large nr premiere." 



M. Spurs has overestimated the rate of destruction, but there can 

 be no doubt that it is very considerable. It is most conspicuous along 

 the shores bordering the lagoon. The stumps of coco-nut palms, the 

 newly-made breaches into the land, forming shallow inland lagoons, 

 the vertical faces of old banks of half consolidated sand all attest it. 

 Just above Point Marianne is a road running along the lagoonward 

 shore, which when I left the island had been narrowed by the action 

 of the sea to a mere path, and was in some places almost impass- 

 able, as the sea had made clean breaches across it, and found its way 

 into some shallow fresh water lagoons lying on the other side of the 

 road. 1 was assured that this road had been over 12 feet wide some 

 years previously, and that it was formerly separated from the lagoon 

 by a narrow strip of land of an equal width. Perhaps the best 

 evidence of the destruction of land is afforded by the "barachois " at 

 the southern extremity of the island. These barachois are inland 

 lagoons connected with the main lagoon by a narrow outlet some 

 2 fathoms deep or more. They are filled and emptied every tide, 

 and their floor is intersected by numerous small channels running in 

 every direction. No corals grow within the barachois, and a slight 

 study convinces the observer that the daily scour of the tides is de- 

 nuding their shores and floors very considerably. 



Barachois are formed in the following way : — During unusually 

 high tides, when the waters of the lagoon are dammed back by a 

 north-westerly wind of unusual violence, the water rises to great 

 heights and invades the land in several places. In some instances 

 it actually makes a breach in the lagoonward shore, and fills up the 

 shallow depressions which are often found in the middle of the strip 

 of land. A pool of salt water is thus formed, which kills the coco 

 palms and other vegetation growing in its bed, and as this process is 

 repeated again and again, in the course of a few years a channel is 

 cut out between the pool and the lagoon, which finally becomes so 

 deep that spring tides, and finally even neap tides, run in and out of 

 the pool regularly. As soon as these conditions are established the 

 channel is scoured out and deepened, and the daily tides scour out 

 the bed of the pool, forming a complete barachois. 



It is not ea-sy for one who has not seen it to understand how 

 much of the loose soil of a coral islet can be moved by a siugle 

 tidal encroachment. It happened that I was riding past the very thin 

 strip of land between Minni Minny and Barton Point the day after an 

 abnormally high tide. The strip of land here is not more than 

 30 yards across, and the sea had washed right over it on the pre- 



