168 J. T. Jutson: 



their effect on the disintegration of the rocks. The direct action 

 of rain must also be excluded, as it cannot beat on to many sides 

 and roofs. Similarly the wind must take little part in the actual 

 breaking up of the rock; and as regards the lake waters as abrad- 

 ing agents, the lakes are dry, especially at the edges, for very long 

 periods; and the water in the lakes is, as a rule, not more than a 

 few inches deep.' No normal waterworn pebbles are found, and 

 the only conclusion appears to be that the lake waters have no 

 abrasive power. 



The rock flakes further break up on the floor of the caves, and in 

 time the debris is removed, mainly, in the writer's opinion, by 

 deflation, but discussion of this aspect does not come within the 

 scope of this paper. 



Cavities are sometimes scooped out on the face of a cliff at any 

 height from the ground up to perhaps 20 feet, but the hollowing 

 out is chiefly confined to a band rising from the lake floor to a 

 height of about four feet; and where the undermining takes place 

 regularly over a length of some yards, the cliff may, in a very 

 marked way, overhang a regularly hollowed out area, which is 

 about two feet high from the lake floor upwards, one to two feet 

 broad, and several yards in length along the line of junction of 

 the lake floor with the cliff. Lucas2 has noted that in the decay of 

 building stones in Egypt the action is frequently limited to a 

 metre or a metre and a half above the ground level ; or if not actu- 

 ally limited to that extent, it is usually greatest at or near the sur- 

 face of the ground; and liis conclusion is that the disintegration is 

 chiefly due to the crj^stallization of salts by the evaporation of 

 water on the rock surface, that is the process now under descrip- 

 tion. 



White incrustations or etflorescences occur frequently, but not 

 always, on the Egyptian rocks, which are subject to the process de- 

 scribed. In Western Australia no pronounced efflorescences at the 

 lake cliff cavities have been noted by the writer. Tliis is a jjoint 

 requiring further investigation. 



The ground water is usually close to the lake floors, and is very 

 saline. The following are two analyses, made in the Western Aus- 

 tralian Geological Survey Laboratory, of watos, one of which was 

 collected from a trench sunk on Lake Cowan, Norseman, 3 and the 



1 The writer has heen informed that water, 10 feet or more deep, lias been observ ed in a lake ; 

 but if so, it i.s quite exceptional aiid perhaps due to some artificial embankment. Most oliservers 

 agree that the lake waters are generally not more than a few inches deep. 



2 Op. cit., p. 3. 



3 Norseman is 108 miles south-southeast of CoolKardie. 



