set eisai eget ta Pate 
47 
But at some level or other there must be a downward limit 
beyond which rock-water can no longer percolate. A thin bed of 
clay, or watertight rock of any kind, will always effectually prevent 
the downward flow of water through pervious rock. Therefore 
the water tends to be held up by such a stratum, and to accumulate 
above it until a sufficient head of water has gathered in the stone 
to force some of the water out to the surface, where it usually 
issues along the junction of the pervious rock with the impervious. 
This is how most springs arise. Another cause determining the 
downward limit of rcck-water, is the proximity of surface water. 
Where a river flows over a porous stratum it tends to sink into 
the rock, and, indeed, does do so, until all the adjoining rock below 
its own level, and to the right and the left of it, is saturated also. 
Therefore, near a river, the lowest level it is possible for the rock- 
water to reach is the level of the river adjoining. Now, water 
travels very slowly through most rocks—perhaps it would be more 
correct to say that it oozes through them, rather than flows. Asa 
consequence of this cause, acting in conjunction with capillary 
attraction, the surface of the water within the rock, or the plane of 
saturation, as it is called, is always at a somewhat higher level in 
the rock at the sides of a valley than it is nearer the river. As the 
river falls in level during a drought, the level of the water within 
the rock slowly responds to it, and sinks also. When the level of 
the river water rises through long rain, a corresponding rise of 
water level takes place in the adjoining rock, always with extreme 
slowness. The outward percolation of the water in the rock is, in 
fact, so slow, that the surface of saturation roughly conforms to the 
larger features of the surface outside. So that if we are dealing 
with a large hill of porous rock, such as freestone, there will be 
within that hill a hill of water with slopes lower in angle than the 
slopes of the hill enclosing it. 
If we sink a well into the hill near the summit, we should have 
to go to a greater depth to find water than we should if the well 
were sunk nearer to the foot of the hill. But the level of the 
water in the higher well would always be absolutely higher than 
that in the lower; and in proportion as we try nearer to the foot 
