as yet fully attained: in 1784 a race of
silkworms was introduced into France, in which one hundred out of the
thousand failed to produce white cocoons; but now, after careful selection
during sixty-five generations, the proportion of yellow cocoons has been
reduced to thirty-five in the thousand.

With plants selection has been followed with the same good results as with
animals. But the process is simpler, for plants in the great majority of
cases bear both sexes. Nevertheless, with most kinds it is necessary to
take as much care to prevent crosses as with animals or unisexual plants;
but with some plants, such as peas, this care does not seem to be
necessary. With all improved plants, excepting of course those which are
propagated by buds, cuttings, &c., it is almost indispensable to examine
the seedlings and destroy those which depart from the proper type. This is
called "roguing," and is, in fact, a form of selection, like the rejection
of inferior animals. Experienced horticulturists and agriculturists
incessantly urge every one to preserve the finest plants for the production
of seed.

Although plants often present much more conspicuous variations than
animals, yet the closest attention is generally requisite to detect each
slight and favourable change. Mr. Masters relates[463] how "many a patient
hour was devoted," whilst he was {200} young, to the detection of
differences in peas intended for seed. Mr. Barnet[464] remarks that the old
scarlet American strawberry was cultivated for more than a century without
producing a single variety; and another writer observes how singular it was
that when gardeners first began to attend to this fruit it began to vary;
the truth no doubt being that it had always varied, but that, until slight
varieties were selected and propagated by seed, no conspicuous result was
obtained. The finest shades of difference in wheat have been discriminated
and selected with almost as much care, as we see in Colonel Le Couteur's
works, as in the case of the higher animals; but with our cereals the
process of selection has seldom or never been long continued.

It may be worth while to give a few examples of methodical selection with
plants; but in fact the great improvement of all our anciently cultivated
plants may be attributed to selection long carried on, in part
methodically, and in part unconsciously. I have shown in a former chapter
how the weight of the gooseberry has been increased b