I Now pass to the structure and modifications of flowers. A complete flower consists of (I) an outer envelope or calyx, sometimes tubular, sometimes consisting of separate leaves called sepals; (2) an inner envelope-or corolla, which is generally more or less coloured, and which, like the calyx, is sometimes tubular, sometimes composed of separate leaves, called petals; (3) of one or more stamens, consisting of a stalk ox filament, and a head or anther, in which the pollen is produced; and (4) a pistil or an ovary, which is situated in the centre of the flower, and contains one or more seeds or ovules. The pistil consists of a stalk or style; and a stigma, to which the pollen must find its way in order to fertilize the flower, and which in many familiar instances forms a small head at the top of the style. In some cases the style is absent, and the stigma is consequently sessile.

Thus, the pistil is normally surrounded by a row of stamens, and it would seem at first sight a very simple matter that the pollen of the latter should fall on the former. This in fact does happen in many instances, and flowers which thus fertilize themselves have evidently one great advantage - few remain sterile for want of pollen.

Such cases, however, are much less frequent than might at first be supposed, and there are three principal modes by which self-fertilization is prevented. Firstly, in many species, the stamens and pistil are situated in different flowers; such species are called diclinous; when the male and female flowers are on the same plant, they are termed monoecious; when on different plants, dioecious.

Secondly, in other instances, as was first discovered by Sprengel, though the stamens and pistil are both situated in one flower, they are not mature at the same time, and the pollen, therefore, cannot fertilize the stigma. These plants are called dichogamous. Sometimes, as in the Arum, the pistil matures before the anther, and these plants are called proterogynous; but much more frequently the anther matures before the pistil; and such plants are called proterandrous.

Thirdly, there are some plants in which, as was first discovered by Mr. Darwin, the same object is secured by the existence, within the limits of the same species, of two or more kinds of flowers, differing in the relative position of the stamens and pistil, which are so placed as to favour the transference by insects of the pollen from the anther of the one form to the pistil of the other. These plants are termed heteromorphoas; some of them have two kinds of flowers, and are called dimorphous; while others have three forms, and are called trimorphous.

But even in plants which belong to none of these categories we find minor modifications which tend to prevent self-fertilization; and Mr. Darwin is probably right in his opinion that no plant invariably fertilizes itself. Thus in some species where the stamens surround the pistil, and which might, therefore, be supposed to be arranged in such a manner as to ensure self-fertilization, the anthers do not open towards the pistil, but on the sides, and by no means therefore in a favourable position with reference to the transference of the pollen. In most, if not all the Cruciferse, the anthers in young flowers have the side which opens turned towards the pistil; but before the anthers come to maturity they twist themselves round, so as to turn their backs to the stigma. Again, in pendent flowers, where the pistil hangs below the anthers, the stigmatic surface is never the upper one, which would catch any falling pollen; but on the contrary, the lower one, which could hardly be touched by the pollen of the same flower, but which is so placed as to come in contact with any insect or other body approaching the flower from below.

It is also probable that many cases will be found to exist, in which, though the pollen necessarily comes in contact with the pistil of the same plant, fertilization does not take place. However improbable this might a priori appear, it is nevertheless said by Hildebrand to be the case in Corydalis cava and Pulmonaria (Fig. 96), by Gartner in Verbascum nigrum (Fig. 98), and Lobelia fulgens; by Scott in Primula verticillata, Oncidium, etc.

Other cases are recorded in which plants are more or less insusceptible of fertilization by their own pollen. Moreover, even where plants are capable of self-fertilization, the pollen from another flower is often more effective than their own, whence it follows that if a supply of pollen from another plant be secured, it is comparatively unimportant to exclude the pollen of the plant itself; for in such cases the latter is neutralized by the more powerful effect of the former.

Everyone who has watched flowers, and has observed how assiduously they are visited by insects, will admit that these insects must often deposit on the stigma pollen brought from other plants, generally those of the same species; for it is a remarkable fact, as Aristotle long ago mentioned, that in most cases bees confine themselves in each journey to a single species of plant; though in the case of some very nearly allied forms this is not so; for instance, it is stated, on good authority, that Ranunculus acris, R. repens, and R. bulbosus, are not distinguished by the bees, or at least are visited indifferently by them, as is also the case with two of the species of clover, Trifolium fragiferum and T. repens.

Even in the simplest and most regular flowers, where the stamens surround the pistil, and both are mature at the same time, insects may visit the flower, and yet not fertilise it with its own pollen, because they touch the anther with one side of the proboscis and the stigma with the other. There are, however, in flowers a great many admirable and beautiful contrivances, tending to prevent the fertilization of a flower by its own pollen; in consequence of which insects habitually carry the pollen from the anthers of one flower to the stigma of another.