up, and that nearest to the light sinking down, or both twisting laterally.* We may, also, suspect that the extreme sensitiveness to light of the upper part of the sheath-like cotyledons of the Gramineae, and their power of transmitting its effects to the lower part, are specialised arrangements for finding the shortest path to the light.  With plants growing on a bank, or thrown prostrate by the wind, the manner in which the leaves move, even rotating on their own axes, so that their upper surfaces may be again directed to the light, is a striking phenomenon.  Such facts are rendered more striking when we remember that too intense a light injures the chlorophyll, and that the leaflets of several Leguminosae when thus exposed bend upwards and present their edges to the sun, thus escaping injury.  On the other hand, the leaflets of Averrhoa and Oxalis, when similarly exposed, bend downwards.

It was shown in the last chapter that heliotropism is a modified form of circumnutation; and as every growing part of every plant circumnutates more or less, we can understand how it is that the power of bending to the light has been acquired by such a multitude of plants throughout the vegetable kingdom.  The manner in which a circumnutating movement—­that is, one consisting of a succession of irregular ellipses or loops—­is gradually converted into a rectilinear course towards the light, has been already explained.  First, we have a succession of ellipses with their longer axes directed towards the light, each of which

* Wiesner has made remarks to nearly the same effect with respect to leaves:  ‘Die undulirende Nutation der Internodien,’ p. 6, extracted from B. lxxvii. (1878).  Sitb. der k.  Akad. der Wissensch.  Wien. [page 491]

is described nearer and nearer to its source; then the loops are drawn out into a strongly pronounced zigzag line, with here and there a small loop still formed.  At the same time that the movement towards the light is increased in extent and accelerated, that in the opposite direction is lessened and retarded, and at last stopped.  The zigzag movement to either side is likewise gradually lessened, so that finally the course becomes rectilinear.  Thus under the stimulus of a fairly bright light there is no useless expenditure of force.

As with plants every character is more or less variable, there seems to be no great difficulty in believing that their circumnutating movements may have been increased or modified in any beneficial manner by the preservation of varying individuals.  The inheritance of habitual movements is a necessary contingent for this process of selection, or the survival of the fittest; and we have seen good reason to believe that habitual movements are inherited by plants.  In the case of twining species the circumnutating movements have been increased in amplitude and rendered more circular; the stimulus being here an internal or innate one.  With sleeping plants the movements have