l generally is of an impervious character,
and where the road is more or less wet and weeping after long rains, a
continuous system of under-drains is required. If the trouble is local,
here and there in spots, and is obviously caused by the breaking up of
springs from the road-bed, such partial work may be adopted as will tap
the sources of these springs, and lead their water harmlessly away.
Gisborne, one of the best agricultural writers of England, put the case
tersely and well when--objecting to the system of circumventing
springs--he said, "_Hit him straight in the eye_, is as good a maxim in
draining as in pugilism." It is best not to pass up at the side of a
spring, and so creep around behind it to head off its water; but to
drive the drain straight through it, and far enough beyond it to tap and
lead away at a lower level the water which causes it. These drains, as
well as all others intended simply to remove subsoil water, and not to
cut off a weeping stream, are best made with common drain-tiles laid as
before directed, and covered immediately with well-packed earth. Water
enters an under-drain, not from above, but from below; that is to say,
as water, from whatever source, fills the subsoil, it rises therein
until it reaches the floor of the drain, when it enters and is led away,
just as water falling into a cask which stands on end flows off at the
under side of the bung-hole when it reaches its level. Even if the cask
be filled to the top with earth, the rain falling upon it will descend
perpendicularly to the bottom, and will flow off at the bung only when
the soil to that level has become saturated. It will descend through the
soil by the straightest course, and will raise the general level. It
will not violate the laws of gravitation, and run diagonally toward the
point of outlet, as seems to be the general supposition when the
perplexing question, "How does water get into the drain?" is first
considered. When we drive a drain through a spring and into the
water-bearing stratum which feeds it, we simply make it easier for the
water to escape by the drain than to keep on at the higher level, and
break out at the surface of the ground.

As in the case of the sidewalk illustrated in Figure 1, in cutting off a
continuous weeping or ooze from higher land, it is best to introduce a
vertical filling of porous material through which the water will descend
and enter the drain; but, excepting this single instance,