two dissimilar pictures into a raised cone, as in the figure, or into a hollow one, if the picture at D is turned round 180°. If we place the images of two diagrams, one like one of those at A, Fig. 31, and the other like the one at B, vertically above one another, we shall then see, at the same time, the raised and the hollow cone, as produced in the lenticular stereoscope by the three diagrams, two like those in Fig. 31, and a third like the one at A. When the prism is good, the dissimilar image, produced by the two refractions at B and C, and the one reflexion at E, is, of course, more accurate than if it had been drawn by the most skilful artist; and therefore this form of the stereoscope has in this respect an advantage over every other in which two dissimilar figures, executed by art, are necessary. In consequence of the length of the reflected pencil DB + BE + EC + CL being a little greater than the direct pencil of rays DR, the two images combined have not exactly the same apparent magnitude; but the difference is not perceptible to the eye, and a remedy could easily be provided were it required.

If the conical tube LD is held in the left hand, the left eye must be used, and if in the right hand the right eye must be used, so that the hand may not obstruct the direct vision of the drawing by the eye which does not look through the prism. The cone LD must be turned round slightly in the hand till the line mn joining the centre and apex of the figure is parallel to the line joining the two eyes. The same line must be parallel to the plane of reflexion from the prism; but this parallelism is secured by fixing the prism and the drawing.

It is scarcely necessary to state that this stereoscope is

glass, the face BC being one of the surfaces of the plate, will answer the purpose. When the prism is placed at a, Fig. 34, at one end of a conical tube LD, and the diagram

D at the other end, in a cap, which can be turned round so as to have the line mn, Fig. 33, which passes through the centre of the base and summit of the cone parallel to the line joining the two eyes, the instrument is ready for use. The observer places his left eye at L, and views with it the picture D, as seen by total reflexion from the base BC of the prism, Figs. 33 and 35, while with his right eye R, Fig. 33, he views the real picture directly. The first of these pictures being the reverse of the second D, like all pictures formed by one reflexion, we thus combine

two dissimilar pictures into a raised cone, as in the figure, or into a hollow one, if the picture at D is turned round 180°. If we place the images of two diagrams, one like one of those at A, Fig. 31, and the other like the one at B, vertically above one another, we shall then see, at the same time, the raised and the hollow cone, as produced in the lenticular stereoscope by the three diagrams, two like those in Fig. 31, and a third like the one at A. When the prism is good, the dissimilar image, produced by the two refractions at B and c, and the one reflexion at E, is, of course, more accurate than if it had been drawn by the most skilful artist; and therefore this form of the stereoscope has in this respect an advantage over every other in which two dissimilar figures, executed by art, are necessary. In consequence of the length of the reflected pencil DB + BE + EC + CL being a little greater than the direct pencil of rays DR, the two images combined have not exactly the same apparent magnitude; but the difference is not perceptible to the eye, and a remedy could easily be provided were it required.

If the conical tube LD is held in the left hand, the left eye must be used, and if in the right hand the right eye must be used, so that the hand may not obstruct the direct vision of the drawing by the eye which does not look through the prism. The cone LD must be turned round slightly in the hand till the line mn joining the centre and apex of the figure is parallel to the line joining the two eyes. The same line must be parallel to the plane of reflexion from the prism; but this parallelism is secured by fixing the prism and the drawing.

It is scarcely necessary to state that this stereoscope is

applicable only to those diagrams and forms where the one image is the reflected picture of the other.

If we wish to make a microscopic stereoscope of this form, or to magnify the drawings, we have only to cement plano-convex lenses, of the requisite focal length, upon the faces AB, AC of the prism, or, what is simpler still, to use a section of a deeply convex lens ABC, Fig. 35, and apply

the other half of the lens to the right eye, the face BC having been previously ground flat and polished for the prismatic lens. By using a lens of larger focus for the right eye, we may correct, if required, the imperfection arising from the difference of paths in the reflected and direct pencils. This difference, though trivial, might be corrected, if thought necessary, by applying to the right eye the central portion of the same lens whose margin is used for the prism.

If we take the drawing of a six-sided pyramid as seen by the right eye, as shewn in Fig. 36, and place it in the total-reflexion stereoscope at D, Fig. 33, so that the line MN coincides with mn, and is parallel to the line joining the eyes of the observer, we shall perceive a perfect raised

pyramid of a given height, the reflected image of CD, Fig. 36, being combined with AF, seen directly. If we now turn the figure round 30°, CD will come into the position AB, and unite with AB, and we shall still perceive a raised pyramid, with less height and less symmetry. If we turn it round 30° more, CD will be combined with BC,

and we shall still perceive a raised pyramid with still less height and still less symmetry. When the figure is turned round other 30°, or 90° degrees from its first position, CD will coincide with CD seen directly, and the combined figures will be perfectly flat. If we continue the rotation through other 30°, CD will coincide with DE, and a slightly hollow, but not very symmetrical figure, will be seen. A rotation of other 30° will bring CD into coalesence with EF, and we shall see a still more hollow and more symmetrical pyramid. A further rotation of other 30°, making 180° from the commencement, will bring CD into union with AF; and we shall have a perfectly symmetrical hollow pyramid of still greater depth, and the exact counterpart of the raised pyramid which was