This was observed with Desmodium gyrans and Mimosa pudica.  With this latter plant, moreover, the pinnae converge in the evening by a steady movement, whereas during the day they are continually converging and diverging to a slight extent.  In all such cases it was scarcely possible to observe the difference in the movement during the day and evening, without being convinced that in the evening the plant saves the expenditure of force by not moving laterally, and that its whole energy is now expended [page 412] in gaining quickly its proper nocturnal position by a direct course.  In several other cases, for instance, when a leaf after describing during the day one or more fairly regular ellipses, zigzags much in the evening, it appears as if energy was being expended, so that the great evening rise or fall might coincide with the period of the day proper for this movement.

The most complex of all the movements performed by sleeping plants, is that when leaves or leaflets, after describing in the daytime several vertically directed ellipses, rotate greatly on their axes in the evening, by which twisting movement they occupy a wholly different position at night to what they do during the day.  For instance, the terminal leaflets of Cassia not only move vertically downwards in the evening, but twist round, so that their lower surfaces face outwards.  Such movements are wholly, or almost wholly, confined to leaflets provided with a pulvinus.  But this torsion is not a new kind of movement introduced solely for the purpose of sleep; for it has been shown that some leaflets whilst describing their ordinary ellipses during the daytime rotate slightly, causing their blades to face first to one side and then to another.  Although we can see how the slight periodical movements of leaves in a vertical plane could be easily converted into the greater yet simple nyctitropic movements, we do not at present know by what graduated steps the more complex movements, effected by the torsion of the pulvini, have been acquired.  A probable explanation could be given in each case only after a close investigation of the movements in all the allied forms.

From the facts and considerations now advanced we may conclude that nyctitropism, or the sleep of leaves [page 413] and cotyledons, is merely a modification of their ordinary circumnutating movement, regulated in its period and amplitude by the alternations of light and darkness.  The object gained is the protection of the upper surfaces of the leaves from radiation at night, often combined with the mutual protection of the several parts by their close approximation.  In such cases as those of the leaflets of Cassia—­of the terminal leaflets of Melilotus—­of all the leaflets of Arachis, Marsilea, etc.—­we have ordinary circumnutation modified to the extreme extent known to us in any of the several great classes of modified circumnutation.  On this view of the origin of nyctitropism we can understand how it is that a few plants, widely distributed throughout the Vascular series, have been able to acquire the habit of placing the blades of their leaves vertically at night, that is, of sleeping,—­a fact otherwise inexplicable.