the deeper layers,
and not only derive an abundant supply of food from them, but actually
promote the fertility of the surface soil by the refuse portions of them
which are left upon it. Experience has in this respect arrived at
results which tally with theory, and it is for this reason that the
broad-leafed turnip, which obtains a considerable quantity of its
nutriment from the air, alternates with grain crops which are chiefly
dependent on the soil. It is undoubtedly to some such cause that several
remarkable instances of what may be called natural rotations are to be
attributed. It is well known in Sweden that when a pine forest is
felled, a growth, not of pine but of birch, immediately springs up. Now
the difference in composition of the ash of these trees is not
sufficient to explain this fact, and it must clearly be due to some
difference in the distribution of their roots, or the mode in which they
obtain their food.

Whatever weight may be given to these different explanations of
rotation, there is no doubt about the importance of attending to it, and
there are various practical deductions of much importance to be drawn
from the facts with which we are acquainted. Thus it is to be observed
that the quantities of mineral matters withdrawn by plants of the same
class are generally similar, and hence it may be inferred that crops of
the most opposite class ought as much as possible to alternate with one
another, and each plant should be repeated as seldom as possible, so
that, even when it is necessary to return to the same class, a different
member of it should be employed. Thus, for instance, in place of
immediately repeating wheat, when another grain crop is necessary, it
would theoretically be preferable to employ oats or barley, and to
replace the turnip by mangold-wurzel or some other root. It is obvious,
however, that this system cannot be carried out in practice to its full
extent; for the superior value of individual crops causes the more
frequent repetition of those which make the largest return. But
experience has so far concurred with theory that it has taught the
farmer the advantage of long rotations; and we have seen the successive
introduction of the three, four, five, and six-course shift, and even,
in some instances, of longer periods.

Such is the theory of rotation, and while it will always be most
advantageous to adhere to it, it is by no means necessary that this
should be done in an absolut