piece that we use
in propagating is a part of the original plant, and will therefore be
like it under similar conditions. Just as with the Irish potato, it is
important to know how good a yielder you are planting. You should watch
during harvest and select for propagation for the next year only such
plants as yield best.

We should exercise fully as much care in selecting proper individuals
from which to make a cutting or a layer as we do in selecting a proper
animal to breed from. Just as we select the finest Jersey in the herd
for breeding purposes, so we should choose first the variety of plant we
desire and then the finest individual plant of that variety.

If the variety of the potato that we desire to raise be Early Rose, it
is not enough to select _any_ Early Rose plants, but the very best Early
Rose plants, to furnish our seed.

[Illustration: FIG. 47. LAYERING]

It is not enough to select large, fine potatoes for cuttings. A large
potato may not produce a bountifully yielding plant. _It will produce a
plant like the one that produced it._ It may be that this one large
potato was the only one produced by the original plant. If so, the plant
that grows from it will tend to be similarly unproductive. Thus you see
the importance of _selecting in the field a plant that has exactly the
qualities desired in the new plant_.

One of the main reasons why gardeners raise plants from buds instead of
from seeds is that the seed of many plants will not produce plants like
the parent. This failure to "come true," as it is called, is sometimes
of value, for it occasionally leads to improvement. For example, suppose
that a thousand apple or other fruit or flower seeds from plants usually
propagated by cuttings be planted; it may be that one out of a thousand
or a million will be a very valuable plant. If a valuable plant be so
produced, it should be most carefully guarded, multiplied by cuttings or
grafts, and introduced far and wide. It is in this way that new
varieties of fruits and flowers are produced from time to time.

Sometimes, too, a single bud on a tree will differ from the other buds
and will produce a branch different from the other branches. This is
known as _bud variation_. When there is thus developed a branch which
happens to be of a superior kind, it should be propagated by cuttings
just as you would propagate it if it had originated from a seed.

[Illustration: FIG. 48. CURRANT CUTTING]

Mr. Gideon o