From the several cases above given, there can be no doubt that the position of the leaves at night affects their temperature through radiation to such a degree, that when exposed to a clear sky during a frost, it is a question of life and death.  We may therefore admit as highly probable, seeing that their nocturnal position is so well adapted to lessen radiation, that the object gained by their often complicated sleep movements, is to lessen the degree to which they are chilled at night.  It should be kept in mind that it is especially the upper surface which is thus protected, as it is never directed towards the zenith, and is often brought into close contact with the upper surface of an opposite leaf or leaflet.

We failed to obtain sufficient evidence, whether the better protection of the upper surface has been gained from its being more easily injured than the lower surface, or from its injury being a greater evil to the plant.  That there is some difference in constitution between the two surfaces is shown by the following cases.  Cassia floribunda was exposed to a clear sky on a sharp frosty night, and several leaflets which had assumed their nocturnal dependent position with their lower surfaces turned outwards so as to be [page 295] exposed obliquely to the zenith, nevertheless had these lower surfaces less blackened than the upper surfaces which were turned inwards and were in close contact with those of the opposite leaflets.  Again, a pot full of plants of Trifolium resupinatum, which had been kept in a warm room for three days, was turned out of doors (Sept. 21st) on a clear and almost frosty night.  Next morning ten of the terminal leaflets were examined as opaque objects under the microscope.  These leaflets, in going to sleep, either turn vertically upwards, or more commonly bend a little over the lateral leaflets, so that their lower surfaces are more exposed to the zenith than their upper surfaces.  Nevertheless, six of these ten leaflets were distinctly yellower on the upper than on the lower and more exposed surface.  In the remaining four, the result was not so plain, but certainly whatever difference there was leaned to the side of the upper surface having suffered most.

It has been stated that some of the leaflets experimented on were fastened close to the cork, and others at a height of from ½ to 3/4 of an inch above it; and that whenever, after exposure to a frost, any difference could be detected in their states, the closely pinned ones had suffered most.  We attributed this difference to the air, not cooled by radiation, having been prevented from circulating freely beneath the closely pinned leaflets.  That there was really a difference in the temperature of leaves treated in these two different methods, was plainly shown on one occasion; for after the exposure of a pot with plants of Melilotus dentata for 2 h. to a clear sky (the temperature on the surrounding grass being -2o C.), it was manifest