low neatly
carried away; and, in other cases, the surface flow over large masses of
rock should be given easy entrance into the tile; but, in all ordinary
lands, whether swamps, springy hill sides, heavy clays, or light soils
lying on retentive subsoil, all ground, in fact, which needs
under-draining at all, should be laid dry above the level to which it is
deemed best to place the drains;--not only secured against the wetting of
springs and soakage water, but rapidly relieved of the water of heavy
rains. The water table, in short, should be lowered to the proper depth,
and, by permanent outlets at that depth, be prevented from ever rising,
for any considerable time, to a higher level. This being accomplished, it
is of no consequence to know whence the water comes, and Elkington's
system need have no place in our calculations. As round pipes, with
collars, are far superior to the "horse-shoe" tiles, and are equally easy
to obtain, it is not necessary to consider the manner in which these
latter should be used,--only to say that they ought not to be used at all.

The water which falls upon the surface is at once absorbed, settles
through the ground, until it reaches a point where the soil is completely
saturated, and raises the general water level. When this level reaches the
floor of the drains, the water enters at the joints and is carried off.
That which passes down through the land lying between the drains, bears
down upon that which has already accumulated in the soil, and forces it to
seek an outlet by rising into the drains.(7) For example, if a barrel,
standing on end, be filled with earth which is saturated with water, and
its bung be removed, the water of saturation, (that is, all which is not
held by attraction _in_ the particles of earth,) will be removed from so
much of the mass as lies above the bottom of the bung-hole. If a bucket of
water be now poured upon the top, it will not all run diagonally toward
the opening; it will trickle down to the level of the water remaining in
the barrel, and this level will rise and water will run off at the bottom
of the orifice. In this manner, the water, even below the drainage level,
is changed with each addition at the surface. In a barrel filled with
coarse pebbles, the water of saturation would maintain a nearly level
surface; if the material were more compact and retentive, a true level
would be attained only after a considerable time. Toward the end of the
flow, th