surface.

In No. 2 the water could rise rapidly until it reached the straw, then
it was stopped almost entirely. But the straw being coarse, the air
circulated in it more or less freely and there was a slow loss by
evaporation. In jar No. 3 the water could rise only to the sand, which
was so coarse that the water could not climb on it to the surface, and
the air circulated in the sand so slowly that there was not sufficient
evaporation to affect scales weighing to one-quarter ounce. No. 4 lost
less than No. 1 because, as in the case of the sand, the water could
not climb rapidly to the surface on the coarse crumbs of soil. The
loss that did take place from No. 4 was what the air took from the
loosely stirred soil on the surface with a very little from the lower
soil. Simply stirring the surface of the sod in No. 4 reduced the loss
of water to less than half the loss from the hard soil in No. 1.

This experiment gives us the clew to the method of checking loss of
water from the soil by evaporation. It is to keep the water from
climbing up to the surface, or check the power of the soil to pump the
water to the surface by making it loose on top. This loose soil is
called a soil mulch. Everything that we do to the soil that loosens
and crumbles the surface tends to check the loss of water by
evaporation from the soil below.

[Illustration: FIG. 30.--TO SHOW THE EFFECT OF A SOIL MULCH
1. Packed soil, lost in 7 days 5.5 ozs. water, equal to 75 tons per
acre.
2. Packed soil, covered with straw, lost in 7 days 2 ozs. water, equal
to 27 tons per acre.
3. Packed soil, covered with sand, lost in 7 days 0 ozs. water, equal to