a certain limit, according to the direction with respect to a certain fixed line in the crystal, called its optical axis. Suppose, then, to take the simplest case, that the eye-lens of a telescope, instead of glass, were formed of such a crystal (say of quartz, which may be worked as well or better than glass), and of a spherical form, so as to offer no difference when turned about on its centre, other than the inclination of its optical axis to the visual ray. Then when that axis coincides with the line of collimation of the object-glass, one image only will be seen, but when made to revolve on an axis perpendicular to that line, two will arise, opening gradually out from each other, and thus originating the desired duplication. In this contrivance, the angular amount of the rotation of the sphere affords the necessary datum for determining the separation of the images.

(203.) Of all methods which have been proposed, however, the simplest and most unobjectionable would appear to be the following. It is well known to every optical student, that two prisms of glass, a flint and a crown, may be opposed to each other, so as to produce a colourless deflection of parallel rays. An object seen through such a compound or achromatic prism, will be seen simply deviated in direction, but in no way otherwise altered or distorted. Let such a prism be constructed with its surfaces so nearly parallel that the total deviation produced in traversing them shall not exceed a small amount (say 5'). Let this be cut in half, and from each half let a circular disc

be formed, and cemented on a circular plate of parallel glass, or otherwise sustained, close to and concentric with the other by a framework of metal so light as to intercept but a small portion of the light which passes on the outside (as in the annexed figure), where the dotted lines represent the radii sustaining one, and the un

dotted those carrying the other disc. The whole must be so mounted as to allow one disc to revolve in its own plane behind the other, fixed, and to allow the amount of rotation to be read off. It is evident, then, that when the deviations produced by the two discs conspire, a total deviation of 10' will be effected on all the light which has passed through them; that when they oppose each other, the rays will emerge undeviated, and that in intermediate positions a deviation varying from 0 to 10', and calculable from the angular rotation of the one disc on the other, will arise. Now, let this combination be applied at such a point of the cone of rays, between the object-glass and its focus, that the discs shall occupy exactly half the area of its section. Then will half the light of the object lens pass undeviated-the other half deviated, as above described; and thus a duplication of image, variable and measureable (as required for micrometric measurement) will occur. If the object-glass be not very large, the most convenient point of its application will be externally before it, in which case the diameter of the discs will be to that of the object-glass as 707: 1000; or (allowing for the spokes) about as 7 to 10.

(204.) The Position Micrometer is simply a straight thread or wire, which is carried round by a smooth revolving motion, in the common focus of the object and eye-glasses, in a plane perpendicular to the axis of the telescope. It serves to determine the situation with respect to some fixed line in the field of view, of the line joining any two objects or points of an object seen in that field - as two stars, for instance, near enough to be seen at once. For this purpose the moveable thread is placed so as to cover both of them, or stand, as may best be judged, parallel to their line of junction. And its angle, with the fixed one, is then read off upon a small divided circle exterior to the instrument. When such a micrometer is applied (as it most commonly is) to an equatorially mounted telescope, the zero of its position corresponds to a direction of the wire, such as, prolonged, will represent a circle of declination in the heavens and the "angles of position" so read off are reckoned invariably from one

point, and in one direction, viz., north, following, south, preceding; so that 0° position corresponds to the situation of an object exactly north of that assumed as a centre of reference,

90° to a situation exactly eastward, or following; 180° exactly south; and 270° exactly west, or preceding in the order of diurnal movement. When the relative position of two stars, very near to each other, so as to be seen at once in the same field of view, is to be determined in this way, especially if they be of unequal magnitudes, the best form of the instrument consists, not in a single thin wire to be placed centrally across both the stars, but in two thick parallel wires, between which both stars are brought under inspection in a symmetrical situation, by which arrangement the parallelism of the line joining their centers with the direction of the wires can be very much more accurately judged of. It gives great advantage, moreover, to the precision of such a judgment, if the position of the observer be such as to bring the principal section of his eye (that which in his upright position is vertical) into parallelism with the wires.

(204 a.) To see the fiducial threads or wires of an eyepiece or micrometer in a dark night is impossible without introducing some artificial light into the telescope, so as either to illuminate the field of view, leaving the threads dark, or vice versa. To illuminate the field, the light of a lamp is introduced by a lateral opening into the tube of the telescope, and dispersed by reflexion on a white unpolished surface, so arranged as not to intercept any part of the cone of ray going to form the image. For illuminating the wires, direct lamp light is thrown on them from the side towards the eye; the superfluous rays being stifled by falling on a black internal coating, or suffered to pass out to the tube through an opposite aperture opening into a dark chamber.

(204 b.) When the wires are seen dark on an illuminated field, the colour of the illuminating light is of great importance. As a matter of experience, it is certain that a red illumination affords a far sharper and clearer view of the wires than any other.

(204 c.) For observing the sun, darkening glasses are neces

sary. In this case red glasses are inappropriate, because they transmit the solar heat freely, by which the eye would be seriously injured, and even when very deep tinted, render prolonged inspection intolerably painful. Green glasses are free from this objection. Both the light and heat of the sun, however, may be subdued by reflexion at glass surfaces, the light returned by regular reflexion on glass being only about 2 per cent. of that incident on it. A reflecting telescope specifically adapted for viewing the sun may be constructed by making the specula of glass, the object mirror having the form of a double concave lens, whose anterior surface (that producing the image) is worked into a paraboloid of the proper focal length, and the posterior to a sphere of considerably greater curvature to transmit and disperse outwards the refracted rays into the open air behind (for which purpose the telescope should be open at both ends) and to so weaken those reflected by dispersing them as not to interfere with the distinctness of the image. Neither the quality of the glass, nor accuracy of figure in the posterior surface is of any importance to the good performance of such a reflector.* Should the light be not sufficiently enfeebled by the first reflexion, it may be still further reduced (to about 1-900th part of its original intensity) by making the small speculum of glass also in the form of a prism; the reflection being performed on one of its exterior surfaces, and the refracted portion being turned away and thrown out at the other.

(204 d.) Advantage may be also taken (as in Sir D. Brewster's polarizing eye-piece) of the properties of polarized light, which may be diminished in any required degree by partial reflexion in a plane at right angles to that of its first oblique reflexion. Or without polarization, the light may be enfeebled by successive reflexions between parallel surfaces to any extent.

(204 e.) When the object in view is to scrutinize, under

*I would take this opportunity earnestly to recommend the construction of a helioscope on this principle, first propounded and more fully described in my Cape observations (p. 436) to the attention of the practical optician. (Author.)

high magnifying powers, minute portions of the solar disc; the light and heat of the general surface may be intercepted by a metallic screen placed in the focus where the image is formed, and pierced with a very small hole, allowing that minute portion only to pass through and be examined with the eye-piece; the observer being thus defended from the glare. By this arrangement, Mr. Dawes, to whom the idea is due, has been enabled to observe some very extraordinary peculiarities in the constitution of the sun's surface, discernible in no other way, an account of which will be found in their proper place.

(204 f.) Since the use of large reflectors has become common among astronomers, the necessity of supporting the ponderous masses of their specula without constraint or undue pressure in any direction (which would distort the figure of their polished surfaces), renders the use of some ready method of verifying, from instant to instant, the adjustment of their lines of collimation (or the optical axis of the reflectors), and of readjusting it, when shifted, indispensable. For this purpose, a small collimating telescope (Art. 178.), illuminated by reflexion from a lamp outside, is fixed within the tube of the reflector, its object-end being turned towards the speculum. Upon the image of the crosswires of this telescope formed in the focus of the reflector, and seen through its eye-piece as a real object, the transits and altitudes of celestial objects may be observed as if it consisted of actual wires; for these, it is manifest, if once placed so as to bisect a star, will continue to do so, whatever amount of tilting the reflector might be subjected to, either in a lateral or vertical plane. The rays from the star and the axis of the collimator remaining parallel, the latter axis, and not that of the reflector, becomes in fact the real line of collimation or optic axis of the instrument, when objects are thus directly referred to it. Should convenience of micrometric measurement, or the observation of faint objects in a very feebly illuminated field, preclude such direct reference, the position of the speculum must from time to time be examined, and if faulty, readjusted by bringing the micrometer