In Part IV., Chapter X., under the title of “The Shapes of Flowers,” I have, after describing their several kinds of symmetry, as habitually related to their positions, made some remarks by way of interpretation. The truth that flowers exhibit a radial symmetry when they are so placed as to be equally affected all round by incident forces, having been exemplified, and also the truth that they assume a bilateral symmetry when they are so placed that their two sides are conditioned in ways different from the ways in which their upper and lower parts are conditioned; I have gone on to inquire (in § 234) by what causes such modifications of form are produced. I have stated that, originally, I inclined to ascribe them entirely to differences in the relations of the parts to physical forces—light, heat, gravitation, etc.; but that I found sundry facts stood in the way of this interpretation. And I have said that “Mr. Darwin’s investigations into the fertilization of Orchids led me to take into account an unnoticed agency.” Continuing to recognize the physical forces as factors having some influence, I have concluded that the most important factor is the action of insects; which, aiding most the fertilization of those flowers which most facilitate their entrance, produce, in course of generations, a form of flower specially adapted to the special position.

Though still adhering to this interpretation, I have since found reason to think that the original interpretation contains a larger portion of truth than I supposed at the time when I was led thus to revise it. While staying at Mürren, in Switzerland, in 1872, I observed some modifications in a species of Gentian, which proved to me that the action of incident physical forces on flowers is, in some cases, very rapid and decided. The species furnishing this evidence was the Gentiana Asclepiadea; which I found in a copse formed of bushes that were here wide apart and there close together. In some places not near to the bushes, the individuals of the species grew vertically; in other places, partially shaded, their inclined shoots curved in such directions as to get the most light; and in other cases their shoots were led to take directions almost or quite horizontal. That, along with these modifications in the directions of their shoots, there went adjustments in the attitudes of their leaves, was a fact not specially worthy of remark; for plants placed inside the windows of houses habitually show us that leaves quickly bend themselves into attitudes giving them the greatest amounts of light. But the fact which attracted my attention was, that the flowers changed their attitudes in an equally-marked manner. The radial distribution passed into a bilateral distribution with the greatest readiness. Comparison of the annexed figures will show the character of this change.

Figure I. represents part of a vertically-growing shoot. This belonged to an individual growing unimpeded by bushes, and getting light on all sides. Here it is observable that the pairs of leaves, placed alternately in directions transverse to one another—one pair pointing, say, north and south, and the next pair pointing east and west—maintain, taking them in the aggregate, a radial distribution; and it is also observable that the alternate pairs of flowers are similarly arranged.

Figure II. is a sketch from a shoot which leaned towards one side, and of which the higher part, as it bent more and more, got its upper side more and more differently conditioned from its lower side. Here we find that not only the leaves, but also the flowers, have adjusted themselves to the changed conditions. The leaves of the lowest pair hang out in the normal way, on the opposite sides of the axis, so that a plane passing through their surfaces will cut the axis transversely; and their two axillary flower-buds, c and d, are similarly placed on opposite sides of the axis. But at the other part of the shoot, we see both that the leaves have adjusted themselves so that their planes, no longer cutting the axis transversely, keep a fit adjustment with respect to the light; and also that the flowers, no longer on opposite sides of the axis, have bent round to the upper side, as at a and b.

Figure III. shows us this re-arrangement carried still further. The shoot it represents was growing in a direction nearly horizontal, and therefore receiving the light only on one side. And here, besides seeing that the leaves have so adjusted themselves that they all lie in approximately the same plane, which is parallel to the axis instead of transverse to it, we see that the two pairs of flower-buds have both come round to the upper side of the axis. So that in this shoot, the original radial symmetry in the arrangement of leaves and flowers, is completely changed into a bilateral symmetry.

These facts do not, it is true, prove any modification in the forms of the flowers themselves: they only prove modification in the grouping of the flowers. But beyond showing, as they do conclusively, how readily a bilateral arrangement of flowers is producible out of an arrangement that was not bilateral, by the action of light, etc.; they give increased probability to the belief that changes in the shapes of flowers are producible by the same agencies. Doubtless this change in the attitudes of the flower-buds is due to the action of light on their calyces and peduncles more than to its action on their unfolding corollas. But along with an action so decided on the growth of these sheathing and supporting organs containing chlorophyll, it is scarcely probable that there is no action on the growth of the petals, containing other colouring matter; considering that in both cases the development of the colouring matter depends on the action of light, and considering also the effect of light on petals, familiarly shown by their opening and closing. And if even but a small effect is producible on the growth of the corolla, then it is to be expected that light will be an agent in changing the form of the corolla, when the attitude of the flower causes its parts to be differently exposed. For a small effect on the individual flower will become a great effect in the flowers of remote descendants; provided the changed attitudes of the flowers preserve considerable constancy throughout the succession of individuals.

Be this as it may, however, the facts I have here described, which I doubt not other observers have seen paralleled in other plants, are instructive, as showing how quickly certain metamorphoses are produced, and as implying the easy establishment of such metamorphoses as permanent characters in a species, if the modifying conditions become permanent. The changes of arrangement I have pointed out, do not become permanent in this species because its individuals are variously affected by the modifying forces: on some they do not act at all, on some a little, on some much; and even on the same individual the different shoots are quite differently affected. But if the habit of this plant were greatly changed—if, for instance, by spreading into habitats yielding abundant nutriment, the plant became very luxuriant, and, multiplying its branches, grew shrub-like; it is clear that, being shaded by one another, these branches would be habitually circumstanced in a way like that which we here see produces bilateralness in the distribution of the flowers, if not in the flowers themselves; and being thus permanently affected, would become permanently bilateral. Accumulating by inheritance, what is here only an individual peculiarity, would become a peculiarity of the species—a specific character.