468 MAJOR H. DE HAGA HAIG ON THE [Aug. 1895, 



irrigation. The upper cavern is in a solid mass of rock ; its height 

 is about 40 feet, and width 60 or 70 ; it divides into two smaller 

 ones, which can be followed for some distance. The roof appears 

 to have been broken into by a later lava-flow, which flowed in and 

 partially filled up one side ; in its steep, treacly-looking slope are 

 cracks out of which water wells up into the cavern and runs away 

 down the two small ones which it divides into. The other cavern 

 containing a stream of water passes under the road, but cannot be 

 followed far because the fragments broken down from the roof fill 

 up the opening. On the sea-coast close by are a number of inlets, 

 at the end of one of which can be seen the opening of a cave. The 

 water at this end is quite fresh and issues below the surface, from 

 the cave, in a strong stream ; the inlet itself appears to have been 

 formed by the roof falling in, for it is deep and narrow, with vertical 

 sides. 



In the upper part of Flacq there is a very strange dry ravine 

 which extends for about 1| mile; it is caused by the roof of a 

 cavern falling in, leaving broken vertical walls, in some places as 

 high as 80 feet. Every 200 or 300 yards are pieces of the roof 

 remaining in their original position, and forming natural bridges 

 across the gully. The roof in most places has sunk bodily, and the 

 bottom of the gully is often paved with broad lava-sheets showing 

 ripple-marks, while in other parts it is filled with luxuriant vegeta- 

 tion. This is really one of the strangest natural curiosities in 

 Mauritius, but on account of its out-of-the-way position it is little 

 visited. In the ' Caves ' crater (a cluster of seven), from which 

 many of these lava-flows have come, there are five dome-shaped 

 caves, all in solid rock. .From their position and structure there 

 can be little doubt that they have been formed by the columns of 

 molten lava, in the craters, sinking and leaving their cooled crusts 

 as roofs ; lava-drops trickled down the sides, and here and there 

 the molten rock remained at the same level long enough to allow of 

 a crust beginning on its fresh surface, which remained as a horizontal 

 ledge running round the cave. In one of the small craters there is 

 a little conical hill, down the centre of which is a funnel-shaped 

 hole, 54 feet deep. Probably this was one of the last cones thrown 

 up by the dying volcano, and has not yet been long enough exposed 

 to the weather to become obliterated. 



Off the coast of the Riviere du Rempart district is a low island 

 called He d'Ambre ; it is composed of lava-beds, and has a number 

 of strange circular holes, some of them over 100 feet across, 20 or 

 30 feet deep, with rocky, vertical sides, and partly filled with sea- 

 water. They look as if caused by the falling in of roofs of circular 

 caverns, but probably they are only unusually large steam-holes. 



II. Geology. 



Mauritius does not offer a varied field to the geologist, for the- 

 greater part of its surface is composed of a volcanic breccia : that is 

 to say, angular fragments of lava, usually full of bubble-holes 



