tional yield can be obtained by extending the spring from the
point where it breaks out along the edge of the water-bearing stratum on
each side. This extension or gathering conduit can be made by building
rough stone walls on each side of the ditch, covering with flat stones
so as to form a pervious channel to intercept the water and lead it to
the chamber from which the supply pipe to the house leads out. The
ground-water level will then be altered as shown by the broken line in
the draining.

More simply it may be made by digging a trench along the hillside at the
same level as the spring, or into the spring if necessary to find the
water, and then laying draintile surrounded by coarse gravel or broken
stone in the trench.

In the western part of the country much knowledge has been gained by
investigating and experimenting on this kind of spring water
development, only there the springs have been made artificially by
digging down to meet the underground flow of water. For example, in the
Arkansas River Valley, California, where it was suspected that water was
flowing underground, a trench was dug transversely across the valley,
and at a depth of six feet sufficient water was found to amount to
200,000 gallons per day for each one hundred feet of trench. On the
South Platte River, near Denver, much the same thing has been done, and
in a trench eighteen feet deep, water is collected at the rate of a
million and a quarter gallons per day for each one hundred feet of
trench. Other examples of the same sort might be given.

For a single house, the spring need usually only be extended by means of
a short trench, and three-inch terra-cotta tile should be laid in the
trench and surrounded by gravel and then covered over. The spring
receiving water from these tiles should be inclosed, as will be
described in a later chapter.

_Supply from brooks._

Whenever a spring is not available and at the same time a supply of
running water by gravity is determined on for a house, recourse is
generally had to brooks which may find their way down the hillsides in
the vicinity. In many instances the water in such brooks is practically
spring water and is the overflow of actual springs. Where the brook is
not subject to contamination between the spring and the point at which
the supply is taken, the latter is as truly spring water as the former,
and if a long length of pipe is saved, there can be no objection to the
brook supply. On t