Still, this one point requires explanation; it is not the sun’s 3 image that is beheld in that glass stick, because it is not capable of expressing it accurately. True enough it tries to reproduce the image, because the material is smooth and suitable for this purpose. But it fails because its shape is unsymmetrical. If it had been suitably constructed, it would reflect as many images of the sun as it had faces. But since the sides are not distinctly separated from each other, and not bright enough to serve as mirrors, the images are only incipient, not fully expressed; they get confused through being crowded together, and are reduced to the appearance of a single band of colour. VIII But to return--why does the bow not complete the full circle in its 1 form, but appear as only a semicircle when stretched to the full extent of its greatest span? Some are of opinion that the reason is that the sun, being much higher than the clouds, strikes them only on the upper side. Hence their lower parts are not touched by his light. Receiving the sun only on one side, the clouds reproduce only one portion of him, and this is never more than a half. There is very little force in this contention. My reason for saying so? The sun, even though he is on 2 the upper side, yet strikes, and therefore colours, the whole cloud. How could it be otherwise? His rays are wont to be transmitted through the clouds and to penetrate any density in them. Further, the proof they advance is flatly in opposition to their main proposition. For if the sun is higher than the clouds, and his beams, therefore, shed only on their upper side, the bow would never come down as far as the earth. Yet it does descend to the very ground. Besides, the bow is never 3 seen except opposite to, not below, the sun. The fact is, the sun’s highness or lowness does not affect the matter: the side of the cloud that faces him is struck by him throughout its whole extent.