WATER SUPPLY 225 
creeps along them by “ capillary attraction”; it may thus 
rise against gravity, but it cannot be forced through by 
pressure of a “ head ” of water. Larger tubes and spaces are 
super-capillary, and water is driven through them by gravity 
and gas-pressure. In small tubes, which may be compared 
to the small fissures and passages in rocks, the flow of water 
is controlled by friction, which increases directly with a 
decrease in diameter and with an increase in length, increases 
as the square of the velocity, and increases with the roughness 
of the inner surface of the tube. 
In tropical and warm temperate countries with an annual 
rainfall of less than 20 or 25 inches, and a fine-grained 
uniform soil, run-off and evaporation may remove the 
whole of the rainfall, and there be none left for percolation. 
Nevertheless the upper layers of the crust have been often 
represented as so charged with water that a deep well will 
be successful anywhere. A. Delesse (Bull. Soc. Géol. Fr. (2), 
xix, 1861, p. 64) estimated the amount of subterranean 
water as about equal to that in the oceans, and Slichter 
(U.S.G.S., Water Sup. Pap., No. 67, 1902, pp. 14-15) accepted 
a third of this quantity. The existence of this subterranean 
sea was based on the principle that a current of water uses 
the whole channel open to it. Thus when a stream of water 
enters a trough at one point and flows out at the opposite 
point, it does not pass straight across the trough ; the current 
widens and deepens till all the water shares in the movement. 
Hence it was held that water percolating through the crust 
must spread widely downward and sideways until it saturates 
the crust. Deep bore holes and mines, however, after 
passing through a wet zone often reach rocks that are quite 
dry, although they were deposited in the sea and must have 
been saturated with connate water. 
The slope of the water-table depends on the friction. 
Water poured into an empty U-tube rises to the same level 
in both arms, because the friction is negligible. If the tube 
be filled with sand, the friction is appreciable, and water 
poured into one arm rises slowly in the other; if the lower 
part of the tube be filled with clay the water only penetrates 
the clay by surface-tension, and the ‘‘ head ™ or pressure of 
the water has no effect, and none passes through the clay 
into the other arm.