UNDERGROUND WATER SUPPLY 457 
In our text-books it is frequently stated that rain water follows one of 
three courses—it is evaporated and absorbed by vegetation, it runs over the 
surface to the sea, or it percolates below ground. And this is followed by 
the statement that the surface run-off is determined by measuring the dis- 
charge of rivers. But a little consideration shows that, neglecting the small 
fraction which may flow directly into the coastal strip, the total river dis- 
charge gauged near the river mouths measures the surface run-off, plus 
the percolation which emerges at the surface in springs and seepages, to 
find its way into rivers and the sea. So that river discharge, subtracted 
from rainfall, gives the loss which is evaporated and absorbed by vegetation. 
With regard to the surface, the water stored in lakes and reservoirs may 
be regarded as more or less permanent, but not absolutely stationary, the 
inflow from surface streams and springs, together with that supplied directly 
by rain, being balanced by their outflowing streams, and by evaporation 
from their surfaces. 
In like manner the underground storage in the rocks may be thought of 
as more or less permanent, though not stagnant, the replenishment from 
percolation balancing the natural outflow in springs and seepages, together 
with that which may be artificially pumped to the surface. But in the 
ultimate analysis, it is the replenishment by rainfall of both surface and 
underground storage which balances the total run-off, plus absorption and 
evaporation loss. 
Since our concern is with underground water, it is evident that it would 
be of value to determine, if possible, the fraction of the rainfall in any area 
which replenishes the underground reservoir, and emerges at the surface 
in natural springs and seepages, or is available for pumped water supplies. 
As regards rainfall, we are fortunate in having available records all over 
the country, mainly through the work of the British Rainfall Organisation, 
which collects, analyses, and publishes every year the gaugings of some 5,000 
observers. Assomeof these records go back for seventy-five years and even 
longer, we are able to abstract from them, though as yet imperfectly, the 
long and short cycles of wet and dry years. With more complete records 
and a better understanding of the incidence, and possibly the causes, of these 
cycles, we may some day be in a position to anticipate and make provision 
for these periods of excessive rainfall and drought. 
When we turn to the records of river gaugings, we have to confess that we 
in this country have been sadly negligent in the past, and far behind the 
United States, for example. The Inland Water Survey Committee, already 
referred to, is seeking to make good this deficiency, but obviously it will 
take many years of systematic work before the data we so badly need are 
available. 
How then is it possible to estimate the quantity of underground water 
which is available for use in any area? What fraction of the rainfall nor- 
mally percolates downwards from the surface to replenish the under- 
ground reservoir ? 
If, in any watershed or basin, we know the total run-off, which includes 
percolation, then of course this figure, subtracted from the rainfall, will give 
us the loss by evaporation. And it so happens that in this country this 
loss, arrived at by difference, is practically the same as the figure we get 
by measuring the loss from artificially constructed evaporation tanks. But 
then we do not know the fraction of the total run-off which is due to 
percolation. 
It is the practice of water engineers to assume an evaporation loss, which 
in this country varies widely. It may be as low as 10 or 11 in. of rainfall 
in northern hilly districts, and perhaps as high as 18 or 19 in. in the south 
R 
