on the disc, the author was induced to urge Mr. Darker, of Paradise Street, Lambeth, to persevere in the adjustment of the mirrors, lenses, and lighting until perfection was obtained. During the Christmas of 1866 the oxy-hydrogen kaleidoscope was exhibited daily at the Polytechnic with the greatest success, and by its means the principle of the instrument could be better understood.

FIG. 34.

a, Figures obtained by putting a single figure, such as key, into the apparatus; b, c, other figures produced by using the light only with an empty slide.

It is chiefly by the adjustment of the light that the original angular opening is gradually multiplied by reflection eight times, and eight distinct sectors or divisions become visible on the disc. When the tip of the finger is now inserted, eight single reflections or four double ones are the result, and by thrusting in all the fingers the curious figures shown at e, Fig. 35, are obtained.

Not only are transparent bodies, such as glass, exhibited with success, but any opaque object will produce the most distinct and symmetrical figures on the

FIG. 35.

screen; in Fig. 35 the pattern dis chiefly produced with a cell containing only pins and needles. If glass be used, it should always be broken from coloured glass rods with the hammer, in order to secure the conchoidal fracture, as the wedge-shaped figures give gradual tones of colour, which are very pleasing to the eye, and produce fair imitations of the colours and grouping of rubies, emeralds, and sapphires when projected on the screen.

A gentleman, who saw these and other patterns, and especially some obtained by using ferns and other natural objects, was so pleased that he stated it was his intention to have an oxy-hydrogen kaleidoscope fitted up in his calicoprinting establishment, in order to assist the artist who designed the patterns; and he stated that, although they had long used the ordinary kaleidoscope for this purpose, the oxy-hydrogen one gave a much better notion of the effect required to be produced, and would enable the manufacturer to select and decide upon the best patterns for commercial purposes.

The phenomena of light produced by reflection, and the instruments which have been constructed to demonstrate these effects, are too numerous to be detailed here, so that two or three examples must suffice. The property of reflection is affected more by the condition of the surface than by the physical nature of the substance used as a reflector. The kaleidoscope reflectors employed by Mr. Darker are made of the best plate glass, coated with metallic silver, and it is extremely difficult to prevent a slight deposit of moisture upon them. The watery particles greatly impair the kaleidoscopic figures, and demonstrate how thoroughly the power of reflection depends on the state of the surface, as this exquisitely thin film of moisture interferes with the perfect illumination of the kaleidoscopic figure.

Some mirrors made in Japan have a very curious property. The back is usually ornamented with Japanese characters, also with flowers, vases, &c.; the front is polished in the usual manner, like any other metallic speculum, and, if carefully examined, with or without a magnifying power, betrays nothing more than the highly polished surface of the alloy, which appears to be composed chiefly of tin and copper. When, however, the mirror is held in the highly divergent rays emitted from an oxy-hydrogen light, it not only reflects on to a disc the surface of the polished disc, but likewise all the Japanese characters, vases, and flowers, which are in relievo on the back of the mirror.

We have in the above experiment a scientific puzzle that is somewhat difficult to explain. May it be supposed that much of the success of the effect obtained is due to the nature of the alloy used in the casting of the mirror? The figures in relief on the back of the mirror, during the operation of casting, must first enter the mould in the liquid state: are these first and quickly congealed before the whole mass of metal? and does the minute difference in the molecular condition of the metal produced by a greater rapidity of cooling, extend through the thin metal to the front and polished side?

Would careful heating and annealing destroy the effect? Whatever may be the method employed, it is certain that the figures reflected from the surface are wholly invisible, and cannot be observed in the strongest light, and with a good magnifying-glass. In all cases where metals are inlaid with other metals the lines where the metals join are distinctly visible, and therefore it cannot be supposed that the Japanese mirror is made in this manner. Are the mirrors cast in a double mould, one side of which is in intaglio and the other in relievo, and after being cast do they grind down the sides of the mirror in which the figures are sunk, until they get a plain surface, which is then polished, leaving the other side and back of the mirror with the figures in relief? The pattern die, conferred on both sides of the metal whilst soli

difying, might still further determine the molecular difference. It is a curious circumstance that the Chinese mirrors, made in imitation of the Japanese mirrors, do not answer the purpose, the former being much heavier than the latter. Whatever may be the secret of success, it is certain that this is only another instance of the remarkable ingenuity of the Japanese workers in metal.

Sir D. Brewster explains the apparent anomaly by suggesting that the design on the back is dexterously reproduced by careful engraving, which is so lightly done that the figures traced are quite invisible after the mirror is brought to the highest degree of polish, and it is only by submitting the mirror to a powerful light, and casting the reflection of the surface on a wall, that the design becomes apparent. The concealment of the most delicate engraving, unless done in some way by Barton's ruling-machine, would be extremely difficult, if not impossible. The Japanese know nothing of the machine with which Barton ruled his steel patterns, and even if they did the reflected patterns would give evidence of colour, which is not the case.

In the "Journal of the Asiatic Society," vol. i., page 242, there is a very clever paper, by James Prinseps, on "The Magic Mirrors of Japan." He says: "The Japanese mirror is a slightly convex disc of bell metal, about 6 in. in diameter, and a quarter of an inch in thickness on the edge, ground and polished on the convex face, and covered with a thin coating of silver to give it a white colour. (Fig. 38, p. 39.)

"The back of the mirror is deeply curved or indented, with ornamental work in circles and festoons, and it bears an inscription in the Japanese character in high relief upon what may be termed the tympanum of the disc; in the centre there is a projecting knob, perforated laterally to receive a string for suspending the mirror. The metal is highly sonorous when struck as a bell, and is so soft as easily to be indented or scratched on contact with any hard substance. I found its composition to be

Copper 80
Tin

20

100

with no traces or silver or arsenic, and a very slight indication of zinc." Mr. Prinseps then describes the curious property of the mirror, similar in effect to those already mentioned and illustrated at Fig. 37, p. 36. He then proceeds to discuss the cause of this seeming anomaly.

"It then occurred that the various parts of the Japanese mirror might be of different density, supposing the pattern to be made by stamping, and that either the rays of light might be more forcibly repelled by the denser metal than by the lighter, or that parts of the surface would acquire different degrees of polish, sufficient to cause the illusion, although imperceptible to the eye. But in such case the thin parts, from being the hardest, should give the stronger reflection.

"This supposition was also overthrown by experiment. A disc of silver, having been annealed at a red heat so as to be quite soft, was stamped on the back with a circular ring, deeply indented, so as to harden the silver in that part only. The opposite surface was then ground and polished, when it was found to give a clear and uniformly reflected spectrum.

"Another and, I believe, the true explanation is suggested by the well-known

phenomenon of the reflection from a brass button, which every school-boy has remarked when sporting his Sunday 'blue coat with metal buttons' in the sunshine of his tutor's parlour-window. The button throws a radiated irregular image on the wall, exhibiting two bright concentric circles, one on the edge and another about one-third within it, and there is generally a bright spot in the centre: all of this seems but the picture of the stamp on the back of the button: the radii resemble, and indeed coincide with, the letters of 'superfine' or 'trebly gilt' inscribed within a double circle, and the central spot represents the shank. There can be little doubt that the principle is in this case precisely that of the Japanese mirror; and, on a cursory view, the surface looks equally smooth and unsuspicious. On minute examination, however, of several buttons, I found them to be by no means plane; their general surface is slightly convex; there is a hollow in the centre and a projection in the position of the inscription behind, caused no doubt by the blow necessary in stamping it. The polish is probably given by a rotary motion, and consequently does not remove these very small irregularities. To follow up the experimental investigation, I selected one of the buttons which gave a good image, ground it on a flat hone, and polished it: all of the magical figures vanished in a moment, and a plain, bright disc appeared in their stead. Here, then, may be a key to the mystery of the mirror: the deception is entirely produced by irregularities on the surface, which are rendered the less perceptible to the eye because the surface is convex instead of being plane. But it may be objected that the two circles which appear bright in the reflected spectrum of the button represent the indented or thin parts of the metal, whereas the thick parts of the Japanese mirror are those which will appear illuminated. A short analysis of the facts in either case will readily explain to what these discrepancies are attributable; but it will be necessary to have recourse to a diagram.

"Let A B, Fig. 38, be a plain mirror upon which the rays of light R impinge; they will be reflected uniformly to R', forming a clear image. Now let A B C D E F G be another reflecting surface, having two convexities, B C, E F, and one concavity in the centre D (the condition nearly of the brass button). In this case the light reflected from the outer concave flexures of the protruding portion of the surfaces B C, E F, will converge in the foci b c and e f respectively, at distances corresponding to the radius of their curvature; the effect will, of course, be visible within wide limits of the actual focus. In most of the buttons, however, the central depression is so great that it collects the rays in a focus, d, a few inches only in front of the surface; and when the spectrum is thrown farther off, the rays crossing from two less distinct luminous foci, d' d', it follows from analogy that the thin parts or tympanum of the Japanese mirror are slightly convex with reference to the rest of the reflecting surface, which may have been caused either by the ornamental work being stamped or partially carved with the hammer and chisel on its back, or, what is more probable, that part of the metal was by this stamping rendered harder, so that in polishing it was not worn away to the same extent."

Since the above was written, an English brass-finisher appears to have discovered the secret. Taking ordinary brass, he finds that any figure stamped upon it with a proper die, and ground down and polished, will not reflect the figure impressed by the die; but if the process with the same die is repeated three times, so that the figure intended to be projected from the surface is stamped three times in the same place, and subsequently ground down and polished after each stamping, then a molecular difference is established between