These appearances have been a source of great perplexity to former observers: for, as they were remarked chiefly in the open air, where they commonly appeared blue, they were attributed to a certain inherent blue or blue colouring quality in the air. The inquirer can, however, convince himself, by the experiment with the candle in a room, that no kind of blue light or reflection is necessary to produce the effect in question. The experiment may be made on a cloudy day with white curtains drawn before the light, and in a room where no trace of blue exists, and the blue shadow will be only so much the more beautiful. 74. De Saussure, in the description of his ascent of Mont Blanc, says, "A second remark, which may not be uninteresting, relates to the colour of the shadows. These, notwithstanding the most attentive observation, we never found dark blue, although this had been frequently the case in the plain. On the contrary, in fifty-nine instances we saw them once yellowish, six times pale bluish, eighteen times colourless or black, and thirty-four times pale violet. Some natural philosophers suppose that these colours arise from accidental vapours diffused in the air, which communicate their own hues to the shadows; not that the colours of the shadows are occasioned by the reflection of any given sky colour or interposition of any given air colour: the above observations seem to favour this opinion." The instances given by De Saussure may be now explained and classed with analogous examples without difficulty. At a great elevation the sky was generally free from vapours, the sun shone in full force on the snow, so that it appeared perfectly white to the eye: in this case they saw the shadows quite colourless. If the air was charged with a certain degree of vapour, in consequence of which the light snow would assume a yellowish tone, the shadows were violet-coloured, and this effect, it appears, occurred oftenest. They saw also bluish shadows, but this happened less frequently; and that the blue and violet were pale was owing to the surrounding brightness, by which the strength of the shadows was mitigated. Once only they saw the shadow yellowish: in this case, as we have already seen (70), the shadow is cast by a colourless light, and slightly illumined by a coloured one. 75. In travelling over the Harz in winter, I happened to descend from the Brocken towards evening; the wide slopes extending above and below me, the heath, every insulated tree and projecting rock, and all masses of both, were covered with snow or hoar-frost. The sun was sinking towards the Oder ponds[1]. During the day, owing to the yellowish hue of the snow, shadows tending to violet had already been observable; these might now be pronounced to be decidedly blue, as the illumined parts exhibited a yellow deepening to orange.