There is another experiment, made with a concave mirror, which has somewhat puzzled philosophers to account for the phenomena. Take a glass bottle AC, (fig. 27) and fill it with water to the point B; leave the upper part BC empty, and cork it in the common manner. Place this bottle opposite a concave mirror, and beyond its focus, that it may appear reversed, and, before the mirror place yourself still further distant from the bottle, and it will appear in the situation A B C. Now, it is remarkable in this apparent bottle, that the _water_, which, according to the laws of catoptrics, should appear at A B, appears on the contrary at B C, and consequently, the part A B appears empty. If the bottle be inverted and placed before the mirror, its image will appear in its natural erect position, and the water which is in reality at BC (fig. 28) is seen at A B. If while the bottle is inverted, it be uncorked, and the water run gently out, it will appear, that, while the part BC is _emptying_, that of A B in the image is filling, and, what is remarkable, as soon as the bottle is empty, the illusion ceases, the image also appearing entirely empty.--The remarkable circumstances in this experiment are, first, not only to see the object where it is not, but also where its _image_ is not; and secondly, that of two objects which are really in the same place, as the surface of the bottle and the water it contains, the one is seen at one place, and the other at another; and to see the bottle in the place of its image, and the water where neither it nor its image are.