“It would be well bolted out,” says he, “whether great refractions may not be made upon reflexions, as well as upon direct beams. For example, take an empty basin, put an angel or what you will into it; then go so far from the basin till you cannot see the angel, because it is not in a right line; then fill the basin with water, and you shall see it out of its place, because of the refraction. To proceed, therefore, put a looking-glass into a basin of water. I suppose you shall not see the image in a right line or at equal angles, but wide. I know not whether this experiment may not be extended, so as you might see the image and not the glass, which, for beauty and strangeness, were a fine proof, for then you should see the image like a spirit in the air. As, for example, if there be a cistern or pool of water, you shall place over against it the picture of the devil, or what you will, so as that you do not see the water. Then put a looking-glass in the water; now if you can see the devil’s picture aside, not seeing the water, it would look like the devil indeed. They have an old tale in Oxford, that Friar Bacon walked between two steeples, which was thought to be done by glasses, when he walked upon the ground.” [Illustration: _Fig. 10._] Kircher also devoted himself to the production of such images, and he has given in the annexed figure his method of producing them. At the bottom of a polished cylindrical vessel AB, he placed a figure CD, which we presume must have been highly illuminated from below, and to the spectators who looked into the vessel in an oblique direction there was exhibited an image placed vertically in the air as if it were ascending at the mouth of the vessel. Kircher assures us that he once exhibited in this manner a representation of the Ascension of our Saviour, and that the images were so perfect that the spectators could not be persuaded, till they had attempted to handle them, that they were not real substances. Although Kircher does not mention it, yet it is manifest that the original figure AB must have been a deformed or anamorphous drawing, in order to give a reflected image of just proportions. We doubt, indeed, if the representation or the figure was ever exhibited. It is entirely incompatible with the laws of reflexion. [Illustration: _Fig. 11._]