458 UNDERGROUND WATER SUPPLY 
of England. The remainder is taken as run-off, of which a certain fraction 
emerges at the surface after percolating into the ground. This percolation 
fraction will obviously vary greatly in different districts, according to the 
permeability of the rocks, the existence of fissures, the nature of the surface, 
the general topography, and so forth. In practice it is usual to base one’s 
estimate upon experience of the quantities of underground water which 
can be pumped in a given area without lowering the general water-table. 
Percolation Gauges.—A direct and apparently simple method of deter- 
mining the percolation at any place is to use a Percolation Gauge, such as 
those designed by Baldwin Lathom at Croydon, or those at Rothamsted. 
The results obtained from these gauges are useful for agricultural purposes, 
but for measuring the total fraction of rain which percolates to depth in the 
rocks they have their limitations. Without entering into detail, it is obvious 
that because of their shallow depth—s5 ft. at most—there must be some 
evaporation loss after percolation, and their surface cannot be regarded as 
a true replica of the average natural surface of the district. A more serious 
defect, in my opinion, is that the permeability of the rocks, apart from 
joints and fissures, diminishes with depth, for most rocks are more per- 
meable near outcrop than at a depth of,-say, one or two hundred feet. The 
consequence is that some of the percolated water at shallow depths finds 
ready exit to the surface at low-lying places, without a chance to sink to the 
main water-table, which may be much deeper. 
Very few determinations of the permeability of rocks at depth have been 
made. I have recently estimated the permeability of some samples of 
water-bearing sandstones in the Birmingham district, and the results are 
rather surprising. 
In a deep boring through Bunter sandstone, seven sample cores, at 
depths varying from 300 to 750 ft., were selected for porosity and per- 
meability tests, and it was found that the porosity, or percentage of pore- 
space, varied from 13-2 per cent. to 30-3 per cent. The permeability, 
or flow of water in gallons per square foot in 24 hours, varied still more, 
namely, from 0-05 to 17-4. Although the constant head of water under 
the conditions of the experiment was here small, viz. 6 in., the results for 
the different samples are strictly comparable. ‘They serve to show that 
Bunter sandstones in the same boring may vary greatly in their capacity to 
transmit water, and therefore to yield their supplies to wells and boreholes. 
In another set of experiments on a Keele sandstone, varying from 20 to 
40 ft. thick, and underlying impervious marls, samples were taken from 
cores at a depth of about 100 ft. in many different boreholes, and also from 
the same sandstone near outcrop. Here, again, the porosities varied from 
3°58 per cent. to 20-3 per cent., and the permeability in the direction of 
bedding, i.e. perpendicular to the cores, was astonishingly small, but 
distinctly higher near outcrop. In the latter experiments I had the means 
of estimating almost exactly the amount of water which this bed of sand- 
stone was transmitting from a reservoir, and the conclusion was inevitable 
that practically all the water passing through the sandstone was moving in 
joints and fissures. And yet the rock is a medium-grained sandstone with 
fair porosity, and would be described as a water-bearing rock. 
That porosity has no necessary relation to permeability is well known. 
Chalk may have a porosity of nearly 50 per cent., and yet wells and borings 
in it may yield no water unless there are fissures or bands of flint. This is 
also true to some extent of other porous rocks, such as sandstone. 
It would be all to the good to have a large number of such measurements 
of porosity and permeability of samples taken from known water-bearing 
strata at different depths, and not to rely on a figure, taken maybe from a 
