the subsoil. On drying the soil
it becomes lighter in colour and loses its stickiness, but it has not
permanently changed because as soon as water is added it comes back to
what it was before.

[Illustration: Fig. 2. Columns showing what 100 parts of soil and
subsoil were made of]

The dried lumps of soil are now to be broken up finely with a piece of
wood, but nothing must be lost. It is easy to see shrivelled pieces of
plant, but not easy to pick them out; the simplest plan is to burn them
away. The soil must be carefully tipped on to a tin lid, or into a
crucible, heated over a flame and stirred {5} with a long clean nail.
First of all it chars, then there is a little sparkling, but not much,
finally the soil turns red and does not change any further no matter
how much it is heated. The shade of red will at once be recognised as
brick red or terra cotta, indeed "terra cotta" means "baked earth."
When the soil is cold it should be examined again; it has become very
hard and the plant remains have either disappeared or have changed to
ash and crumble away directly they are touched. On weighing a further
loss is discovered, which was in our experiment:--

Weight of top soil after drying but before burning ... 83 decigrams
" " " " " " after " ... 76 "
--
The part that burnt away weighed ... 7 "

Weight of subsoil after drying but before burning ... 87 decigrams
" " " " " after " ... 84 "
--
The part that burnt away weighed ... 3 "

These results are entered on the column in Fig. 2.

The surface soil is seen to contain more material that will burn away
than the subsoil does. When the burnt soil is moistened it does not
become dark and sticky like it did before, it has completely changed
and cannot be made into soil again. It is more like brick dust than
soil.

For further experiments we shall want a fresh portion of the original
soil.

On a wet afternoon something was noticed that enabled us to get a
little further with our studies. The rain water ran down a sloping
piece of ground in a tiny channel it had made; the streamlet was very
muddy, and at first it was thought that all the soil was washed away.
But we soon saw that the channel was lined {6} with grit