shake a sample of it in a bottle of water and let it settle for
several days, we can tell roughly from the layers of soil in the
bottom of the bottle the relative amounts of sand and clay in the
soil. Also if we weigh a sample before and after burning we can tell
roughly the amount of organic matter in the soil. Test a number of
soils and determine roughly the proportions of sand, clay and organic
matter in them.

=Experiment.=--Take the pans of soil used in our first soil experiment
and separate the soils in the pans into two parts by a trench across
the centre on the pan. Now wet the soil in one side of the pan and
stir it with a stick or a spoon, carefully smooth the surface of the
soil in the other side of the pan and pour or sprinkle some water on
it, but do not stir it. Set the pans aside till the soils are dry.
This drying may take several days and in the meantime we will study
the classification of soils.

[Illustration: FIG. 16.
To show that roots need air. Bottle _A_ was supplied with fresh water,
and bottle _B_ with water that had been boiled to drive the air out
and then cooled.]

[Illustration: FIG. 17.
Bottle _A_ contains fresh water, bottle _B_ contains boiled water.
Notice the air bubbles in bottle _A._]

[Illustration: FIG. 18.
Tumblers _A_ and _C_ contained moist sand, _B_ and _D_ contained
puddled clay. Cuttings in _B_ and _D_ died, because there was not
sufficient ventilation in the clay for root-development.]