A further consideration must be mentioned, which is just as manifest 1 as the preceding, to prove that the reflection is given back after the fashion of a mirror; it is never given back save from straight opposite to the sun, that is,[27] unless on one side stands the object to be reflected, and on the other the mirror that reveals it. Proofs are adduced by the mathematicians that are not merely convincing but that compel belief of this. Nor can doubt be left in any mind that the rainbow is an image of the sun, imperfectly reflected owing to the defective shape of the mirror. But meantime let us recall other proofs that may, so to speak, be picked up in the street without any 2 reference to mathematics. Among the proofs of this origin of the bow I place the extreme rapidity of its emergence. In a single moment the huge form with its thousand lines is inwoven in the texture of the heavens, and just as rapidly does it fade. Now, nothing is returned so quickly as an image from a mirror. The mirror does not create anything, it merely reveals it. Artemidorus of Parium tells us further even the kind of cloud required to reflect such an image of the sun. If you 3 make a concave mirror, he says, that is, one resembling half of a ball cut through the middle, and take your stand outside the centre, then those who stand beside you will appear in the reflection inverted and nearer to you than to the mirror. The very same thing, according to him, takes place when we look at a round hollow cloud from the side: 4 the image of the sun detaches itself from the cloud, and is nearer us and more turned in our direction. Therefore the red colour is from the sun, the dark blue is from the cloud: the other hues are produced by a blending of these two. [27] In a writer less prone to repetition the words to the end of the sentence would seem the insertion of a copyist. V