d. It consists far more in selecting and propagating
the best specimens than in any so-called "breeding." With the majority
of the herbs this is the most likely direction in which to seek success.

Suppose we have sown a packet of parsley seed and we have five thousand
seedlings. Among these a lot will be so weak that we will naturally
pass them by when we are choosing plantlets to put in our garden beds.
Here is the first and simplest kind of selection. By this means, and by
not having space for a great number of plants in the garden, we probably
get rid of 80 per cent of the seedlings--almost surely the least
desirable ones.

[Illustration: Lath Screen for Shading Beds]

Suppose we have transplanted 1,000 seedlings where they are to grow and
produce leaves for sale or home use. Among these, provided the seed has
been good and true, at least 90 per cent will be about alike in
appearance, productivity and otherwise. The remaining plants may show
variations so striking as to attract attention. Some may be tall and
scraggly, some may be small and puny; others may be light green, still
others dark green; and so on. But there may be one or two plants that
stand out conspicuously as the best of the whole lot. These are the ones
to mark with a stake so they will not be molested when the crop is being
gathered and so they will attain their fullest development.

These best plants, and only these, should then be chosen as the seed
bearers. No others should be allowed even to produce flowers. When the
seed has ripened, that from each plant should be kept separate during
the curing process described elsewhere. And when spring comes again,
each lot of seed should be sown by itself. When the seedlings are
transplanted, they should be kept apart and labeled No. 1, No. 2, No. 3,
etc., so the progeny of each parent plant can be known and its history
kept.

The process of selecting the seedlings the second year is the same as in
the first; the best are given preference, when being transplanted. In
the beds all sorts of variations even more pronounced than the first
year may be expected. The effort with the seedlings derived from each
parent plant should be to find the plants that most closely resemble
their own parents, and to manage these just as the parents were managed.
No other should be allowed to flower.

This process is to be continued from year to year. If the selection is
carefully made, the grower will soon rejoice, beca