88 NARRATIVE. 
the curious circular depressions, similar to those seen 
near Tarl Dat, but here they were mostly half full 
of brine, and some contained a thick crust of common 
salt. In some places these pits covered many acres 
of ground. 
My theory of the formation of these pits, and 
the only one which seems to explain all the pheno- 
mena, is as follows. In the valley above Tarl Dat, 
I suppose that under the sand and gravel which 
form the surface of the ground, there is a stratum 
of clay, under which the water from the mountains 
finds its way (diagram A), and that the water, flowing 
in very varying quantity at different times, eats away 
this layer of clay, and causes the surface above to 
subside as the sand is allowed to flow out through 
the openings in the clay. Similar depressions, on a 
much larger scale, but fewer in number, are often 
produced in the same way in many parts of the 
Panjab, about Amballa, for instance, and about mid- 
way between Jhelum and Hawil Pindi. When a 
fall ol rain or snow occurs after dry weather, I 
suppose that the air and mud in the spaces formerly 
excavated by the water get churned up into froth, 
and near Tarl Dat this frothy mud gets forced up 
and produces the appearance of mud volcanoes. 
The brine pits, situated with reference to the river 
as in diagram B, are no doubt formed in the same 
manner, and then as the Karakash river attains a 
certain height from the melting of the snow, it 
partially fills the pits with water, and as it ebbs and 
flows every twenty-four hours, the pits are alternately 
filled and emptied of water every day, and this water 
on evaporating leaves salt — for the water of the Kara- 
