cuttings or buds. Self-fertilisation always implies that the flowers in
question were impregnated with their own pollen.

My experiments were tried in the following manner. A single plant, if it
produced a sufficiency of flowers, or two or three plants were placed
under a net stretched on a frame, and large enough to cover the plant
(together with the pot, when one was used) without touching it. This
latter point is important, for if the flowers touch the net they may be
cross-fertilised by bees, as I have known to happen; and when the net is
wet the pollen may be injured. I used at first "white cotton net," with
very fine meshes, but afterwards a kind of net with meshes one-tenth of
an inch in diameter; and this I found by experience effectually excluded
all insects excepting Thrips, which no net will exclude. On the plants
thus protected several flowers were marked, and were fertilised with
their own pollen; and an equal number on the same plants, marked in a
different manner, were at the same time crossed with pollen from a
distinct plant. The crossed flowers were never castrated, in order to
make the experiments as like as possible to what occurs under nature
with plants fertilised by the aid of insects. Therefore, some of the
flowers which were crossed may have failed to be thus fertilised, and
afterwards have been self-fertilised. But this and some other sources of
error will presently be discussed. In some few cases of spontaneously
self-fertile species, the flowers were allowed to fertilise themselves
under the net; and in still fewer cases uncovered plants were allowed to
be freely crossed by the insects which incessantly visited them. There
are some great advantages and some disadvantages in my having
occasionally varied my method of proceeding; but when there was any
difference in the treatment, it is always so stated under the head of
each species.