Bible Encyclopedias
Telescope

Eye-pieces

The eye-pieces or oculars through which, in case of visual observations, the primary images formed by the objective are viewed, are of quite secondary importance as regards definition in the central portion of the field of view. If an eye-piece blurs the definition in any degree in the centre of the field it must be very badly figured indeed, but the definition towards the edge of the field, say at 20° away from the centre of the apparent FIG. 4. field of view, depends very in timately upon the construction of the eye-piece. It must be so designed as to give as flat an image as is possible consistently with freedom from astigmatism of oblique pencils. The mere size of the apparent field of view depends upon obtaining the oblique pencils of light emerging from it to cross the axis at the great possible angle, and to this end the presence of a field-lens is indispensable, which is separated from the eye-lens by a considerable interval.

The earlier arrangement of two lenses of the Huygenian eye-piece (see Microscope) having foci with ratio of 3 to I, gives a fairly large flat field of view approximately free from distortion of tangential lines and from coma, while the Mittenzw ? y variety of it (fig. 4) in which the field-lens is changed into a meniscus having radii in about the ratio of +I to - 9 gives still better results, but still not quite so good as the results obtained by using the combination of two convexo-plane lenses of the focal ratio 2 to I.

In the Ramsden eyepiece (see Microscope) the focal lengths of the two piano-convex lenses are equal, and their convexities are turned towards one another. The field-lens is thus in the principal focal plane of the eye-lens, if the separation be equal to -f f2). This is such a practical drawback that the separation is generally 4ths or $ths of the theoretical, and then the primary image viewed by the eye piece may be rather outside the field-lens, which is a great practical advantage, especially when a reticule has to be mounted in the primary focal plane, although the edge of the field is not quite achromatic under these conditions.

Kellner Eye-piece

In order to secure the advantage of the principal focal plane of the eye-piece being well outside of the field-lens and at the same time to obtain a large flat field of view with oblique achromatism and freedom from coma and distortion, there is no better construction than the modified Kellner eye-piece (fig. 5) such as is generally used for prismatic binoculars. It consists of a piano-convex field-lens of crown glass. and an approximately achromatic eye-lens, some distance behind it, consisting of an equi-convex crown lens cemented to a concavoplane flint lens, the latter being next to the eye.

There are also other eye-pieces having the field-lens double or achromatic as well as the eye-lens.

In cases where it is important to get the maximum quantity of light into the eye, the field-lens is discarded and an achromatic eye-lens alone employed. This yields a very much smaller field of view, but it is very valuable for viewing feeble telescopic objects. and very delicate planetary or lunar details. Zeiss and Steinheil's monocentric eye-pieces and the Cooke single achromatic eye-piece (fig. 6) are examples of this class of oculars. (H. D. T.) Reflecting Telescope. The following are the various forms of reflecting telescopes: The Gregorian telescope is represented in fig. 7. A A and B B are concave mirrors having a common axis and their concavities facing each other. The focus of A for parallel rays is at F, that of B for parallel rays at f - between B and F. Parallel rays falling on A A converge at F, where an image is formed; the rays are then reflected from B and converge at P, where a second and more enlarged image is formed. Gregory himself „A FIG. 7. - Gregorian Telescope.

showed that, if the large mirror were a segment of a paraboloid of revolution whose focus is F, and the small mirror an ellipsoid of revolution whose foci are F and P respectively, the resulting image will be plane and undistorted. The image formed at P is viewed through the eye-piece at E, which may be of the Huygenian or Ramsden type. The focal adjustment is accomplished by the screw S, which acts on a slide carrying an arm to which the mirror B is attached. The practical difficulty of constructing Gregorian telescopes of good defining quality is very considerable, because if spherical mirrors are employed their aberrations tend to increase each other, and it is extremely difficult to give a true elliptic figure to the necessarily deep concavity of the small speculum. Short appears to have systematically conquered this difficulty, and his. Gregorian telescopes attained great celebrity. The use of the Gregorian form is, however, practically abandoned in the present day. The magnifying power of the telescope is = Ff /ex, where F and f are respectively the focal lengths of the large and the small mirror, e the focal length of the eye-piece, and x the distance between the principal foci of the two mirrors (=Ff in the diagram) when the instrument is in adjustment for viewing distant objects. The images are erect.

The Cassegrain telescope differs from the Gregorian only in the substitution of a convex hyperboloidal mirror for a concave ellipsoidal mirror as the small speculum. This form has two distinct advantages: (I) if spherical mirrors are employed their aberrations have a tendency to correct each other; grain' (2) the instrument is shorter than the Gregorian, caeteris paribus, by twice the focal length of the small mirror. Fewer telescopes have been made of this than perhaps of any other form of reflector; but in comparatively recent years the Cassegrain has acquired importance from the fact of its adoption for the great Melbourne telescope, and from its employment in the 60-in. reflector of the Mount Wilson Solar Observatory (see below). For spectroscopic purposes the Cassegrain form has peculiar advantages, because in consequence of the less rapid convergence of the rays after reflection from the convex hyperboloidal mirror, the equivalent focus can be made very great in comparison with the length of the tube. This permits the employment of a spectroscope furnished with a. collimator of long focus. The magnifying power is computed by the same formula as in the case of the Gregorian telescope.

The Newtonian telescope is represented in Fig. 8. A A is a concave mirror whose axis is a a. Parallel rays falling on A A converge on the plane mirror B B, and are thence reflected at Ne w right angles to the axis, forming an image in the focus of i the eye-piece E. The surface of the large mirror should be a paraboloid of revolution, that of the small mirror a true optical plane. The magnifying power is = Fle. This form is employed in the FIG. 5.

1 ?; ? 6a_....?

FIG. 6.

construction of most modern reflecting telescopes. A glass prism of total reflection is sometimes substituted for the plane mirror. „9 A FIG. 8. - Newtonian Telescope.

The Herschelian or front view reflector is represented in fig. 9. A A is a concave parabolic mirror, whose axis a c is inclined to the axis of the tube a b so that the image of an object in the focus of the mirror may be viewed by an eye-piece at E, the angle b a c being tier equal to the angle c a E. This form was adopted by the - elder Herschel to avoid the loss of light from reflection in the small mirror of the Newtonian telescope. The front view telescope, however, has hardly been at all employed except by the Herschels. But at the same time none but the Herschels have swept the whole sky for the discovery of faint nebulae; and FIG. 9. - Herschelian Reflector.

probably no other astronomers have worked for so many hours on end for so many nights as they did, and they emphasize the easy position of the observer in using this form of instrument.

Construction of Specula. The composition of metallic specula in the present day differs very little from that used by Sir Isaac Newton. Many different alloys have been suggested, some including silver, nickel, zinc or arsenic; but that which has practically been found best is an alloy of four equivalents of copper to one of tin, or the following proportions by weight: copper 252, tin 117.8. Such speculum metal is exceedingly hard and brittle, takes a fine white polish, and when protected from damp has little liability to tarnish. The process of casting and annealing, in the case of the specula of the great Melbourne telescope, was admirably described by Dr Robinson in Phil. Trans., 1869, 1 59, p. 1 35. Shaping, polishing and figuring of specula are accomplished by methods and tools very similar to those employed in the construction of lenses. The reflecting surface is first ground to a spherical form, the parabolic figure being given in the final process by regulating the size of the pitch squares and the stroke of the polishing machine.

Soon after Liebig's discovery of a process for depositing a film of pure metallic silver upon glass from a salt of silver in solution, Steinheil (Gaz. Univ. d'Augsburg, 24th March 1856), and later, independently, Foucault (Comptes Rendus, vol. xliv., February 1857), proposed to employ glass for the specula of telescopes, the reflecting surface of the glass speculum to be covered with silver by Liebig's process. Those silver-on-glass specula are now the rivals of the achromatic telescope, and it is not probable that many telescopes with metal specula will be made in the future. The best speculum metal and the greatest care are no guarantee of freedom from tarnish, and, if such a mirror is much exposed, as it must be in the hands of an active observer, frequent repolishing will probably be necessary. This involves refiguring, which is the most delicate and costly process of all. Every time, therefore, that a speculum is repolished, the future quality of the instrument is at stake; its focal length will probably be altered, and thus the value of the constants of the micrometer also have to be redetermined. Partly for these reasons the reflecting telescope with metallic mirror has never been a favourite with the professional astronomer, and has found little employment out of England.' In England, in the hands of the Herschels, Rosse, Lassell and De la Rue it has done 1 There is a noteworthy exception in the case of the i 8-in. speculummetal mirror employed by Sir William Huggins at Tulse Hill, with which a large part of his remarkable and important series of astrospectroscopic results have been obtained. So far as we know, this mirror has never been repolished since its first installation in 1870, and still retains its admirable surface. One of Short's mirrors, made about 1760 or 1770, of 6-in. aperture, now in the possession of Sir William Huggins, has surfaces which still retain their original perfection although they have never been repolished.

splendid service, but in all these cases the astronomer and the instrument-maker were one. The silver-on-glass mirror has the enormous advantage that it can be resilvered with little trouble, at small expense, and without danger of changing the figure. Glass is lighter, stiffer, less costly and easier to work than speculum metal. Silvered mirrors have also some advantage in light grasp over those of speculum metal, though, aperture for aperture, the former are inferior to the modern object-glass. Comparisons of light grasp derived from small, fresh, carefully silvered surfaces are sometimes given which lead to illusory results, and from such experiments Foucault claimed superiority for the silvered speculum over the object-glass. But Sir David Gill found from experience and careful comparison that a silvered mirror of 12-in. aperture, mounted as a Newtonian telescope (with a silvered plane for the small mirror), when the surfaces are in fair average condition, is equal in light grasp to a first-rate refractor of 10-in. aperture, or area for area as 2: 3. This ratio will become more equal for larger sizes on account of the additional thickness of larger object-glasses and the consequent additional absorption of light in transmission.

Mounting of Telescopes. The proper mounting of a telescope is hardly of less importance than its optical perfection. Freedom from tremor, ease and delicacy of movement and facility of directing the instrument to any desired object in the heavens are the primary qualifications. Where accurate differential observations or photographs involving other than instantaneous exposures have to be made, the additional condition is required that the optical axis of the telescope shall accurately and automatically follow the object under observation in spite of the apparent diurnal motion of the heavens, or in some cases even of the apparent motion of the object relative to neighbouring fixed stars.

Our limits forbid a historical account of the earlier endeavours to fulfil these ends by means of motions in altitude and azimuth, nor can we do more than refer to mountings such as those employed by the Herschels or those designed by Lord Rosse to overcome the engineering difficulties of mounting his huge telescope of 6 ft. aperture. Both are abundantly illustrated in most popular works on astronomy, and it seems sufficient to refer the reader to the original descriptions.2 We pass, therefore, directly to the equatorial telescope, the instrument par excellence of the modern extra-meridian astronomer. The article Transit Circle describes one form of mounting in which the telescope is simply a refined substitute for the sights or pinules of the old astronomers. The present article contains a description of the mounting of the various forms of the so-called zenith telescope. In its simplest form the mounting of an equatorial telescope consists of an axis parallel to the earth's axis, called" the polar axis "; a second axis at right angles to the polar axis called" the declination axis "; and the telescope tube fixed at right angles to the declination axis.

In Fig. to A A is the polar axis; the telescope is attached to the end of the declination axis; the latter rotates in bearings which are attached to the polar axis and concealed by the telescope itself. The telescope is counterpoised by a weight attached to the opposite end of the declination axis. The lower pivot of the polar axis rests in a cup-bearing at C, the upper bearing upon a strong metal casting M M attached to a stone pier S. A vertical plane passing through A A is therefore in the meridian, and the polar axis is inclined to the horizon at an angle equal to that of the latitude of the place of observation. Thus, when the declination axis is horizontal the telescope moves in the plane of the meridian by rotation on the declination axis only. Now, if a graduated circle B B is attached to the declination axis, together with the necessary verniers or microscopes V V for reading it (see Transit Circle), so arranged that when the telescope is turned on the declination axis till its optical axis is parallel to A A the vernier reads 0° and when at right angles to A A 90°, then we can employ the readings of 2 Herschel, Phil. Trans., 1795, 85, p 347; Rosse, Phil. Trans., 1840, p. 503; 1861, p. 681.

FIG. to. - Equatorial Telescope. English form.

this circle to measure the polar distance of any star seen in the telescope, and these readings will also be true (apart from the effects of atmospheric refraction) if we rotate the instrument through any angle on the axis A A. Thus one important attribute of an equatorially mounted telescope that, if it is directed to any fixed star, it will follow the diurnal motion of that star from rising to setting by rotation of the polar axis only. If we now attach to the polar axis a graduated circle D D, called the" hour circle,"of which the microscope or vernier R reads o h when the declination axis is horizontal, we can obviously read off the hour angle from the meridian of any star to which the telescope may be directed at the instant of observation. If the local sidereal time of the observation is known, the right ascension of the star becomes known by adding the observed hour angle to the sidereal time if the star is west of the meridian, or subtracting it if east of the meridian. Since the transit circle is preferable to the equatorial for such observations wherein great accuracy is required, the declination and hour circles of an equatorial are employed, not for the determination of the right ascensions and declinations of celestial objects, but for directing the telescope with ease and certainty to any object situated in an approximately known position, and which may or may not be visible to the naked eye, or to define approximately the position of an unknown object. Further, by causing the hour circle, and with it the polar axis, to rotate by clockwork or some equivalent mechanical contrivance, at the same angular velocity as the earth on its axis, but in the opposite direction, the telescope will, apart from the effects of refraction, automatically follow a star from rising to setting.

Types of Equatorials. - Equatorial mountings may be divided into six types. (A) The pivots or bearings of the polar axis are placed at its extremities. The declination axis rests on bearings attached to opposite sides of the polar axis. The telescope is attached to one end of the declination axis, and counterpoised by a weight at the other end, as in fig. io. (B) The polar axis is supported as in type A; the telescope is placed between the bearings of the declination axis and is mounted symmetrically with respect to the polar axis; no counterpoise is therefore requisite. (C) The declination axis is mounted on the prolongation of the upper pivot of the polar axis; the telescope is placed at one end of the declination axis and counterpoised by a weight at the other end. (D) The declination axis is mounted on a forked piece or other similar contrivance attached to a prolongation of the upper pivot of the polar axis; the telescope is mounted between the pivots of the declination axis. (E) The eye-piece of the telescope is placed in the pivot of the polar axis; a portion or the whole of the axis of the telescope tube coincides with the polar axis. (F) The telescope is fixed and the rays are reflected along its axis from an external mirror or mirrors. Mountings of types A and B - that is, with a long polar axis supported at both ends - are often called the" English mounting,"and type C, in which the declination axis is placed on the extension of the upper pivot of the polar axis, is called the" German mounting,"from the first employment of type C by Fraunhofer. A description of some of the best examples of each type will illustrate their relative advantages or peculiarities.

Type A. - Fig. io may be taken as a practical example of the earlier equatorials as made by Troughton in England and afterwards by Gambey for various Continental observatories. In the Phil. Trans. for 1824 (part 3, pp. 1-412) will be found a description by Sir John Herschel and Sir James South of the equatorial telescope which they employed in their measurements of double stars. The polar axis was similar in shape to that of fig. 10 and was composed of sheets of tinned iron. In Smyth's celebrated Bedford telescope the polar axis was of mahogany. Probably the best example of this type of mounting applied to a refractor is that made by the elder Cooke of York for Fletcher of Tarnbank; the polar axis is of cast iron and the mounting very satisfactory and convenient, but unfortunately no detailed description has been published. In recent years no noteworthy refractors have been mounted on Great this plan; but type A has been chosen by Grubb for the Me!- great Melbourne reflector, of 48-in. aperture, with marked ingenuity of adaptation to the peculiar requirements telescope. of the case. Fig. I i shows the whole instrument on a small scale with the telescope directed to the pole, and the hour circle set 6" from the meridian.

Type B. - The most important examples of type B are Airy's equatorial at Greenwich (originally made to carry a telescope of 13-in. aperture, but now fitted with a telescope by Grubb of 28-in. aperture), and the photographic equatorials of is-in. aperture employed at Paris and other French observatories, of which the object-glasses were made by the brothers Henry and the mountings by Gautier of Paris.

These instruments have done admirable work in connexion with the great international undertaking, the Carte du Ciel. The general construction will be understood from fig. 12. The double polar axis is composed of hollow metal beams of triangular section. The hour circle has two toothed circles cut upon it, one acted upon by a worm screw mounted on the pier and driven by clockwork, the other by a second worm screw attached to the polar axis, which can be turned by a handle in the observer's hand and thus a slow movement can be given to the telescope in right ascension inde FIG. 12. - Paris Observatory Instrument.

After an illustration in La Nature, by permission of Masson et Cie.

pendently of the clock. Slow motion in declination can be cornmunicated by a screw acting on a long arm, which latter can be clamped at pleasure to the polar axis. An oblong metallic box fitted with pivots, whose bearings are attached to the triangular beams, forms the tube for two parallel telescopes; these are separated throughout their length by a metallic diaphragm. The chromatic aberration of the object-glass of one of these telescopes is corrected for photographic rays, and the image formed by it is received on a highly sensitive photographic plate. The other telescope is corrected for visual rays and its image is formed on the plane of the spider-lines of a filar micrometer. The peculiar form of the tube is eminently suited for rigid preservation of the relative parallelism of the axes of the two telescopes, so that,;i the image of a certain selected star is retained on the intersection of two wires of the micrometer, by means of the driving clock, aided by small corrections given by the observer in right ascension and declination (required on account of irregularity in the clock movement, error in astronomical adjustment of the polar axis, or changes in the star's apparent place produced by refraction), the image of a star will continue on the same spot of the photographic film during the whole time of exposure. In these telescopes the photographic objectglass has an aperture of 13 in. and the visual object-glass of to in. Both telescopes have the same focal length, viz. 11.25 ft., so that, in the image produced, 1 mm. is = 1 1 of arc. An excellent mounting of type B, made by T. Cooke & Sons of York, has been employed by Franklin Adams for making his maps of the sky.

Type C. - Many more telescopes have been made of type C than of any other, and this form of mounting is still most generally employed for the mounting of modern refractors. Fraunhofer's chef-d'oeuvre, the great Dorpat refractor, made for Otto Struve about 1820, had a mounting of this type, and was the first equatorial of any importance to be provided with clockwork. The instrument, shown in fig. 13, is described in detail by Struve (Beschreibung des auf der Sternwarte zu Dorpat befindlichen grossen Refractors von RIH? ?

_ - FIG. I I. - Melbourne Reflector.

Fraunhofer, Dorpat, 1825), and was an enormous advance upon all previous telescopes for micrometric research. In the hands of Struve results were obtained by it which in combined quality and quantity had never before been reached. Its success was such that the type of Fraunhofer's telescope became stereotyped for many years not only by Fraunhofer's successors but throughout Germany. When, twelve years afterwards, Struve ordered the is-in. refractor for the new observatory at Pulkovo, the only important change made by Fraunhofer's successors was, at Struve's suggestion, the substitution of a stone pier for the wooden stand in the original instrument.

Both the Dorpat and the Pulkovo refractors are defective in rigidity, especially in right ascension. The declination circle is most inconvenient of access, and slow motion in declination can only be effected when the instrument is clamped by a long and inconvenient handle; so that, practically, clamping in declination was not employed. The slow motion in right ascension is defective, being accomplished in the Dorpat refractor by changing the. rate of the clock, and in the Pulkovo FIG. 13. - Dorpat Refractor.

refractor by a handle which, when used, affects very injuriously the rate of the clock for the time being. Struve's skill as an observer was such that he used to complete the bisection on the fixed wire of the micrometer by a pressure of the finger on the side of the tube - a method of proved efficiency in such hands, but plainly indicative of the want of rigidity in the instrument and of the imperfection of the slow motions (see Micrometer).

The driving circle is also much too small, so that a very slight mechanical freedom of the screw in the teeth involves a large angular freedom of the telescope in right ascension, while its position at the lower end of a too weak polar axis tends to create instability from torsion of that axis. Strange to say, the wooden tube long retained its place in German telescope-mountings.

About 1840 a great advance was made by the Repsolds of Hamburg in the equatorial mounting of the Oxford heliometer. The driving circle was greatly increased in diameter and placed at the upper end of the polar axis, and both the polar and declination axes were made much stronger in proportion to the mass of the instrument they were designed to carry. (A figure of the instrument is given in the Oxford Observations for 1850.) About 1850 Thomas Cooke of York began his career as a maker of equatorial telescopes. The largest example of his work is the refractor of 24-in. aperture, originally made for the private observatory of Robert Stirling Newall at Gateshead, Northumberland, and afterwards presented by him to the University Observatory, Cambridge. Cooke's mounting is admirable for its symmetry and simplicity of design, its just apportioning of strength, and a general suitability of means to ends. It is not a little curious that the obvious improvement of trans ferring the declination axis as well as the declination-clamp to the telescope end of the declination axis was so long delayed; we can explain the delay only by the desire to retain the declination circle as a part of the counterpoise. We believe the first important equatorials in which the declination was read from the eye-end were the 15-in. by Grubb and the 6-in. by Cooke, made for the observatory of Lord Crawford (Lord Lindsay) at Dun Echt, Aberdeenshire, about 1873. The plan is now universally adopted. Telescopes of such dimensions can be conveniently directed to any object by the circles without the observer being under the necessity to climb a special ladder. But when much larger instruments are required the hour circle becomes inaccessible from the floor, and means have to be devised for reading both circles from the eyeend. This was first accomplished by Grubb in the great refractor of 27-in. aperture which he constructed for the Vienna observatory, represented in section in fig. 14. The observer's eye is applied to the small telescope E, which (by means of prisms numbered I, 2, 3, 4) views the vernier attached to the cross-head simultaneously with the hour circle attached to the upper end of the polar axis. Light to illuminate the vernier and circle is thrown from the lamp ?i? L upon prism 4 by the prisms 6 and 5. Prism i is in the axis of the declination circle and always reflects rays along? that axis, whatever the position of the telescope may be, whilst the prisms 2, 3, 4, 5 and ?. ?% 6 are attached to the crosshead ?;y; and therefore FIG. 14. - Grubb's 27-in. Refractor (Vienna).

preserve their relative positions to each other. Through the eyepiece of the bent 1 telescope E' another hour circle attached to the lower end of the polar axis can be seen; thus an assistant is able to direct the telescope by a handle at H to any desired hour angle. A slight rotatory motion of the telescope E on its axis enables the vernier of the declination circle to be read through prism 1. The leading features of this fine instrument represent those of all Grubb's large telescopes. The mode of relieving the friction of the declination axis is similar to that employed in the Melbourne telescope and in the account of the Vienna telescope published by Grubb. The end friction of the polar axis is relieved by a ring of conical rollers shown in section beside the principal figure.

From this point we must condense farther description into critical remarks on a few typical modern instruments.

(I) Telescopes of Moderate Size for Micrometric Research Only. - Fig. 15 shows the mounting of the 8-in. refractor, of 9-ft. focal length, at the private observatory of Dr Engelmann, Leipzig. The object-glass is by Messrs Clark of Cambridge, Mass., the mounting by the Repsolds of Hamburg. The declination circle reads from the eye-end, and four handles for clamping and slow motion in right ascension and declination are situated near the observer's hands. The tube is of sheet steel, light, stiff, and free from tremor. The eye-end carries the micrometer with an illuminating apparatus similar to that described under Micrometer. The lamp near the eye-end illuminates the field or the wires at pleasure, as well as the position circle of the micrometer and the declination circle; a separate lamp illuminates the hour circle. An excellent feature is the short distance between the eye-piece and the declination axis, so that 1 In the bent telescope refracting prisms are employed at the corners to change the direction of the rays.

of so large an instrument. We have the authority of Otto Struve for stating that in practice they are all that can be desired. There is in this instrument a remarkably elegant method of relieving the friction of the polar axis. Let A A (fig. 17) be a section of the polar axis; it is then easy to adjust the weight P attached to its lower end so that the centre of gravity X of the whole moving parts of the instrument shall be in the vertical (V V) of a line passing through the apex of the hollowed flange p q at q, which flange forms part of the polar axis. If now a wheel W is forced up against q with a pressure equal to the weight of the moving part of the instrument, the whole weight of the moving part would rest upon W in unstable equilibrium; or if a pressure R, less than W, is employed, we have the end friction on the lower bearing removed to an extent = R sin 4),4), and the friction on the bearings of the upper pivot removed to the extent of F cos 4), - where 4 is the latitude of the place. The wheel W is therefore mounted on a guided rod, which is forced upwards by suitable levers and weights, and this relief of pressure is precisely proportional to the pressure on the respective bearings. The Repsolds find it unnecessary to relieve the friction of the declination axis.

In such large telescopes it becomes a matter of the first importance to provide means of convenient access to the eye-end of the instrument. This the Repsolds have done in the Pulkovo telescope by means of two platforms, as shown in fig. 16. These platforms are capable of easy motion so that the astronomer may be conveniently situated for observing an object at any azimuth or altitude to which the telescope may be directed. For the great refractor more recently erected at Potsdam, Messrs Repsold arranged a large platform mounted on a framework which is moved in azimuth by the dome, so that the observer on the platform is always opposite the dome-opening. This framework is provided with guides on which the platform, whilst preserving its horizontality, is V the observer has to follow the eye-end in a comparatively small circle; another good point is the flattening of the cast-iron centrepiece of the tube so that the flange of the declination axis is attached as near to the axis of the telescope tube as is consistent with free passage of the cone of rays from the object-glass. The substitution of small incandescent electric lamps is an improvement now universally adopted.

(2) Telescopes for General P urposes. - T he modern equatorial should, for general purposes, be capable of carrying spectroscopes of considerable weight, so that the proportional strength of the axes and the rigidity of the instrument have to be considerably increased. The original mounting of the Washington refractor of 26-in. aperture and 32z-ft. focal length (described in Washington Observations, 1874, App. 1) was in these respects very defective, the polar and declina tion axes being only 7 in. in diameter.

The great Pulkovo refractor (fig. 16) erected in 1885 is of so-in. aperture and 45 ft. focal length. The object-glass is by Clark, the mounting by the Repsolds. The tube is cylindrical, of riveted steel plate, graduated in thickness from the centre to its extremities, and bolted by very powerful flanges to a strong short cast-iron central tube, in which, as in Dr Engelmann's telescope (fig. 15), the attachment to the flange of the declination axis is placed as close as it can be to the axis of the tube without interfering with rays converging from the object-glass to any point in the field of view. A new feature in this instrument is the platform at the lower end of the polar axis, where an assistant can view the hour circle by one eyeFIG. 15. - Dr Engelmann's piece and the declination circle by another (looking up the per forated polar axis), and where he can also set the telescope to any hour angle by one wheel, or to any declination by a second, with the greatest ease. The observer at the eye-end can also read off the hour and declination circles and communicate quick or slow motions ,to the telescope both in right ascension and declination by conveniently Pulkovo , placed handles. The eye end presents an refractor appearance too complicated to be figured here; it has a micrometer and its illumination for the position circle, a micrometer head, and a bright or dark field, clamps in right ascension and declination and quick and slow motion in the same, a finder, microscopes for reading the hour and declination circles, an illuminated dial showing sidereal time and driven by an electric current from the sidereal clock, and counter weights which can be removed when a spectroscope or other heavy appliance is added. All these, although making up an apparently complicated apparatus, are conveniently FIG. 16. - Pulkovo Refractor.

arranged, and are all necessary for the quick and easy working raised and lowered nearly in an arc of a circle of which the point of intersection of the polar and declination axes is the centre. The rotation of the dome, and p _z with it the platform-framework, is accomplished s by means of electric motors, as also is the raising and lowering of the platform on its = =_ framework. The current is supplied by accumulators, and the switch-board is ^ q attached to the platform in a position convenient for use by the astronome his assistant. = -= - / '. In the original design sup plied for the 36-in. telescope of the Lick Observatory at Mount Hamilton, California, Grubb suggested that the whole floor, 70 ft. in diameter, should be raised and lowered by water power, under control of the observer by means of electric keys which act on secondary mechanism that in turn works the valves and reversing gear of the water engines. Other water engines, similarly connected, with keys at the observer's hands, rotate the dome and perform the quick motions in right ascension and declination. (An illustration showing these arrangements appeared in The Engineer of July 9, 1886.) Grubb's suggestion of the "rising floor" was adopted, although his original plans for the mounting were not carried out; the construction of the mounting, dome, floor, &c., having been entrusted to Messrs Warner & Swasey of Cleveland, Ohio, U.S.A. It has been contended that it is undesirable to move so great a mass as a floor when a platform alone is required to carry the observer. But a floor, however heavy, suspended by three wire ropes and properly balanced over large, well-mounted pulleys, requires an amount of energy to work it which does not exceed that required to operate a platform of moderate dimensions, and there is a freedom, a safety and a facility of working with a complete floor which no partial platform can give. A floor can be most satisfactorily operated by hydraulic means, a platform cannot be so well worked in this way. The best floor mounting we know of is that designed by O. Chadwick for the Victoria Telescope of the Cape Observatory. An account of it will be found in the History and Description of the Cape Observatory. This floor can be raised at the rate of 1 ft. per second or as slowly as the observer desires - whilst in all the large platforms we have seen (Potsdam and Paris), the rate of shift is tedious and time-consuming.

The largest refracting telescope in active use is the Yerkes telescope, with an object-glass of 40-in. diameter by Alvan Clark & Son of Cambridge, U.S.A., and with a mounting, dome and rising floor by Warner & Swasey of Cleveland, Ohio, U.S.A. The reader will gather a good general idea of the design from fig. 16. The eye-end is shown on the plate, fig. 25.

The chief defect in equatorial mountings of type C is that in general they are not capable of continued observing much past the meridian without reversal. This is an unquestionable drawback when long exposures near the meridian are required. By the 8-in. Refractor.

use of an overhanging polar axis the difficulty can be overcome; it has been successfully adopted by Repsolds for their astrographic equatorials of is-in. aperture and 11.25-ft. focus, and on a much smaller scale by Warner & Swasey for the Bruce telescope of io-in. aperture and 50-in. focus, made for the Yerkes Observatory. The ably the most perfect yet proposed for modern astrophysical research.

The declination axis is here represented by what are practically the trunnions or pivots of the tube, resting in bearings which are supported by the arms of a very massive cast-iron fork bolted to the upper end of the polar axis. This axis is a hollow forging of nickel steel, of which the accurately turned pivots rest on bearings attached to cast-iron uprights bolted upon a massive cast-iron base plate. The base plate rests upon levelling screws which permit the adjustment of the polar axis to be made with great precision. The combined overhanging weight of the cast-iron fork, the mirror and tube is so great, that without a very perfect relief-friction system the instrument could not be moved in right ascension with any approach to practical ease. But a hollow steel float, 10 ft. in diameter, is bolted to the upper end of the polar axis just below the fork. This float dips into a tank filled with mercury so that practically the entire instrument is floated by the mercury, leaving only sufficient pressure on the bearings to ensure that the pivots will remain in contact with them. The 60-in. silver-on-glass mirror (weighing about one ton) rests at the lower end of the tube on a support-system consisting of a large number of weighted levers which press against the back of the glass and distribute the load. Similar weighted levers around the circumference of the mirror provide the edge support. The telescope is moved in right ascension and declination by electric motors controlled from positions convenient for the observer. The driving clock moves the telescope in right ascension by means of a worm-gear wheel, 10 ft. in diameter, mounted on the polar axis.

The 60-in. mirror is of 25-ft. focus, but for certain classes of work it is desirable to have the advantage of greater focal length. For this purpose the telescope can be used in the four different ways shown in fig. 21.

(I) As a Newtonian reflector, fig. 21 (a), the converging rays from the 60-in. mirror being reflected to the side of the tube where FIG. 19. - Bruce Telescope, made for the Yerkes Observatory.

From Professor Hale's The Study of Stellar Evolution, by permission of the University of Chicago Press.

latter is shown in fig. 19. Stability in this method of mounting can only be secured by excessive weight and rigidity in the support of the overhanging axis. In the case of the Victoria telescope (24-in. aperture and 222-ft. focus) mounted at the Cape of Good Hope on this plan, it has been found necessary to add supporting stays where great rigidity is required, and thus to sacrifice continuous circum-meridian motion for stars between the zenith and the elevated pole.

Type D. - The first important equatorial of type D was the 4-ft. reflecting telescope of Lassell (Mem. R.A.S., xxxvi. 1-4), and later Lord Rosse's 36-in. reflecting telescope at Birr Castle (Phil. Trans., clxxi. 153), and A. Common's 36-in. reflecting telescope mounted by him at Ealing xlvi. 173-182). In Lassell's instrument (a reflector of the Newtonian type) the observer is mounted in the open air on a supplementary tower capable of motion in any azimuth about the centre of motion of the telescope, whilst an observing platform can be raised and lowered on the side of the tower. In Lord Rosse's instrument (also of the Newtonian type) the observer is suspended in a cage near the eye-piece, and the instrument is used in the open air. Common's telescope presents many ingenious features, especially the relief-friction by flotation of the polar axis in mercury, and in the arrangements of the observatory for giving ready access to the eye-piece of the telescope.

Type C seems indeed to be the type of mounting most suitable for reflecting telescopes, and this form has been adopted for the 60-in. reflector completed by G. W. Ritchey, under the direction of Professor G. E. Hale, for the Mount Wilson Solar Observatory. The instrument is shown in fig. 20, and its design is unquestion FIG. 21.

From Professor Hale's The Study of Stellar Evolution, by permission of the University of Chicago Press.

the image is formed, and where it may be photographed or viewed with an eye-piece. In this case the image is formed without secondary magnification and the focal length is 25 ft.

(2) As a Cassegrain reflector, fig. 21 (b), in which case the upper section of the tube bearing the plane mirror is removed and a shorter section substituted for it. This latter carries a hyperboloidal mirror, which returns the rays towards the centre of the large mirror and causes them to converge less rapidly. They then meet a small plane mirror supported at the point of intersection of the polar and declination axes, whence they are reflected down through the hollow polar axis as shown in fig. 2, and come to focus on the slit of the powerful spectroscope that is mounted on a pier in the chamber of constant temperature as shown in fig. zo. In this case the equivalent focal length is 150 ft.

(3) As a Cassegrain reflector, for photographing the moon, planets or very bright nebulae on a large scale, as shown in fig. 21 (c), with an equivalent focal length of loo ft.

(4) As a Cassegrain reflector, for use with a spectroscope mounted in place of the photographic plate, fig. 21 (d); in this case a convex mirror of different curvature is employed, the equivalent focus of the combination being 80 ft.

Type E. - In the Comptes Rendus for the year 1883, vol. 96, pp. 735-74 1, Loewy gives an account of an instrument which he calls an "equatorial coude," designed (I) to attain greater stability and so to measure larger angles than is generally possible with the ordinary equatorial; (2) to enable a single astronomer to point the telescope and make observations in any part of the sky without changing his position; (3) to abolish the usual expensive dome, and to substitute a covered shed on wheels (which can be run back at pleasure), leaving the telescope in the open air, the observer alone being sheltered. These conditions are fulfilled in the manner shown in fig. 22. E P is the polar axis, rotating on bearings at E and P. The object-glass is at 0, the eye-piece at E. There is a plane mirror at M, which reflects rays converging from the object-glass to the eye-piece at E. A second mirror N, placed at 45° to the optical axis of the object-glass, reflects rays from a star at the pole; but by rotating the box which contains this mirror on the axis of its supporting tube T a star of any declination can be observed, and by combining this motion with rotation of the polar axis the astronomer seated at E is able to view any object whatever in the visible heavens, except circumpolar stars near lower transit. An hour circle attached to E P and a declination circle attached to the box containing the mirror N, both of which can be read or set from E, complete the essentials of the instrument.

There must be a certain loss of light from two, additional reflections; but that could be tolerated for the sake of other advantages, provided that the mirrors could be made sufficiently perfect optical planes. By making the mirrors of silvered glass, one-fourth of their diameter in thickness, the Henrys have not only `. succeeded in mounting them with all necessary rigidity free from flexure but have given them optically true plane surfaces, notwithH standing their large diameters, viz., II and 15.7 in. Sir David Gill tested the equatorial coude on double stars at the Paris Observatory in 1884, and his last doubts as to the practical value of the instrument were dispelled. He has never seen more perfect optical definition in any of the many telescopes he has employed, and certainly never measured a celestial object in such favourable conditions of physical comfort. The easy position of the observer, the convenient position of the handles for quick and slow motion, and the absolute rigidity of the mounting leave little to be desired. In a much larger /!w/ - Shutter '60' inch Reflecting Telescope Dome and Building, Solar Observatory Mount Wilson,Calif. Designed G.W.Ritchey Air tight Building Sheet hen sun Building Scale of Feet FIG. 20.

From Professor Hale's by permission of the University of Chicago Press.

Loewy's ial FIG. 22. - Loewy's Coude Equatorial.

instrument of the same type subsequently mounted at Paris, and in like instruments of intermediate size mounted at other French observatories, the object-glass is placed outside the mirror N, so that both the silvered mirrors are protected from exposure to the outer air.

A modification of Loewy's equatorial coude has been suggested by Lindemann (Asir. Nachr., No. 3935); it consists in placing both the mirrors of Loewy's "equatorial coude" at the top of the polar axis instead of the lower end of it. By this arrangement the long cross tube becomes unnecessary, and neither the pier nor the observatory obstruct the view of objects above the horizon near lower transit as is the case in Loewy's form. The reflected rays pass down the tube from the direction of the elevated pole instead of upward towards that pole. The observer is, therefore, at the bottom of the tube instead of the top and looks upward instead of downward. The drawbacks to this plan are (1) the necessarily large size of the upper pivot (viz. the diameter of the tube) and of the lower pivot (which must be perforated by a hole at least equal in diameter to the photographic field of the telescope), conditions which involve very refined arrangements for relief of friction, and (2)ythe less comfortable attitude of looking upward instead of downward. The plan, however, would be a very favourable one for spectroscopic work and for the convenient installation of an underground room of constant temperature. The difficulties of relief friction could probably be best overcome by a large hollow cylinder concentric with the polar axis fixed near the centre of gravity of the whole instrument and floated in mercury, on the plan adopted in the Mount Wilson 60-in. reflector already described, but in this case the floating cylinder would be below and not above the upper bearing.

In 1884 Sir Howard Grubb (Phil. Trans. R. Dub. Soc., vol. iii. series 2, p. 61) proposed a form of equatorial telescope of which an The excellent example was erected at Cambridge (Eng). in Grubb 1898. The instrument in some respects resembles the equatorial equatorial coude of Loewy, but instead of two mirrors Camp there is only one. A flanged cast-iron box, strongly r i bbed and open on one side, forms the centre of the polar axis. One pivot of the polar axis is attached to the lower end of this box, and a strong hollow metal cone, terminating in the other pivot, forms the upper part of the polar axis. The declination axis passes through the two opposite sides of the central box. Upon an axis concentric with the declination axis is carried a plane mirror, which is geared so as always to bisect the angle between the polar axis and the optical axis of the telescope. If then the objective tube is directed to any star, the convergent beam from the object-glass is received by the plane mirror from which it is reflected upwards along the polar axis and viewed through the hollow upper pivot. Thus, as in the equatorial coude, the observer remains in a fixed position looking down the polar tube from above. He is provided with quick and slow motions in right ascension and declination, which can be operated from the eye-end, and he can work in a closed and comfortably heated room. A large slot has to be cut in the cone which forms the upper part of the polar axis, in order to allow the telescope to be pointed nearer to the pole than would otherwise be possible; even so stars within 15° of the pole cannot be observed. An illustrated preliminary description of the instrument is given by Sir Robert Ball (Mon. Not. R.A.S., lix. 152). The instrument has a triple photo-visual Taylor object-glass of 122 in. aperture and 19.3-ft. focal length.

Type F. - In all the previously described types of telescope mounting the axis of the instrument is either pointed directly at the object or to the pole; in the latter case the rays from the star under observation are reflected along the polar axis by a mirror or mirrors attached to or revolving with it. Equatorials of types A, B, C and D have the advantage of avoiding interposed reflecting surfaces, but they involve inconveniences from the continual motion of the eye-piece and the consequent necessity for providing elaborate observing stages or rising floors. In those of type E the eye-piece has a fixed position and the observer may even occupy a room maintained at uniform temperature, but he must submit to a certain loss of light from one or more reflecting surfaces, and from possible loss of definition from optical imperfection or flexure of the mirror or mirrors. In all these types the longer the telescope and the greater its diameter (or weight) the more massive must be the mounting and the greater the mechanical difficulties both in construction and management.

But if it be possible to mount a fixed telescope by which a solar or stellar image can be formed within a laboratory we give the following advantages: - (1) There is no mechanical limit to the length of the telescope; (2) the clockwork and other appliances to move the mirror, which reflects the starlight along the axis, are much lighter and smaller than those required to move a large telescope; (3) the observer remains in a fixed position, and spectroscopes of any weight can be used on piers within the laboratory; and (4) the angular value of any linear distance on a photographic plate can be determined by direct measurement of the distance of the photographic plate from the optical centre of the object-glass. The difficulty is that the automatic motion of a single mirror capable of reflecting the rays of any star continuously along the axis of a fixed horizontal telescope, requires a rather complex mechanism owing to the variation of the angle of reflexion with the diurnal motion.

Foucault appears to have been the first to appreciate these advantages and to face the difficulty of designing a siderostat which, theoretically at least, fulfils the above-mentioned conditions. A large siderostat, constructed by Eichens after Foucault's design, was completed in 1868 - the year of Foucault's death. It remained at the Paris Observatory, where it was subsequently employed by Deslandres for solar photography. The largest refracting telescope yet made, viz., that constructed by Gautier for the Paris exhibition of 1900, was arranged on this plan (type F), the stars' rays being reflected along the horizontal axis re rac or of a telescope provided with visual and with photo graphic object-glasses of 49-in. diameter and nearly 200-ft. focal length. Up to 1908 neither the optical qualities of the images given by the object-glasses and reflecting plane nor the practical working of the instrument, have, so far as we know, been submitted to any severe test. It is, however, certain that the Foucault siderostat is not capable, in practice, of maintaining the reflected image in a constant direction with perfect uniformity on account of the sliding action on the arm that regulates the motion of the mirror; such an action must, more or less, take place by jerks. There are farther inconveniences in the use of such a telescope, viz., that the image undergoes a diurnal rotation about the axis of the horizontal telescope, so that, unless the sensitive plate is also rotated by clockwork, it is impossible to obtain sharp photographs with any but instantaneous exposures. In the spectroscopic observation of a single star with a slit-spectroscope, this rotation of the image presents no inconvenience, and the irregular action of a siderostat on Foucault's plan might be overcome by the following arrangement: A B (fig. 23) is a polar axis, like that of an equatorial telescope, rotating in twenty-four hours by clockwork. Its lower extremity terminates in a fork on which is mounted a mirror C D, capable of turning about A on an axis at right angles to A B, the plane of the mirror being parallel to this latter axis. The mirror C D is set at such an angle as to reflect rays from the star S in the direction of the polar axis to the mirror R and thence to the horizontal telescope T.

The mirrors of Lindemann's equatorial coude reflecting light downwards upon the mirror R would furnish an ideal siderostat for stellar spectroscopy in conjunction with a fixed horizontal telescope.

Coelostat

If a mirror is mounted on a truly adjusted polar axis, the plane of the mirror being parallel to that axis, the normal to that mirror will always be directed to some point on the celestial equator through whatever angle the axis is turned. Also, if the axis is made to revolve at half the apparent diurnal motion of the stars, the image of the celestial sphere, viewed by reflection from such a moving mirror, will appear at rest at every point - hence the name coelostat applied to the apparatus. Thus, any fixed telescope directed towards the mirror of a properly adjusted coelostat in motion will show all the stars in the field of view at rest; or, by rotating the polar axis independently of the clockwork, the observer can pass in review all the stars visible above the horizon whose declinations come within the limits of his original field of view. Therefore, to observe stars of a different declination it will be necessary either to shift the direction of the fixed telescope, keeping its axis still pointed to the coelostat mirror, or to employ a second mirror to reflect the rays from the coelostat mirror along the axis of a fixed telescope. In the latter case it will be necessary to provide means to mount the coelostat on a carriage by which it can be moved east and west without changing the altitude or azimuth of its polar axis, and also to shift the second mirror so that it may receive all the light from the reflected beam. Besides these complications there is another drawback to the use of the coelostat for general astronomical work, viz., the obliquity of the angle of reflection, which can never be less than that of the declination of the star, and may be greater to any extent. For these reasons the coelostat is never likely to be largely employed in general astronomical work, but it is admirably adapted for spectroscopic and bolometric observations of the sun, and for use in eclipse expeditions. For details of the coelostat applied to the Snow telescope - the most perfect installation for spectroheliograph and bolometer work yet erected - see The Study of Stellar Evolution by Prof. G. E. Hale, p. 131.

The Zenith Telescope The zenith telescope is an instrument generally employed to measure the difference between two nearly equal and opposite zenith distances. Its original use was the determination of geographical latitudes in the field work of geodetic operations; more recently it has been extensively employed for the determination S of variation of latitude, at fixed stations, under the auspices of the International Geodetic Bureau, and for the astronomical determination of the constant of aberration. The instrument is shown in its most recent form in fig. 24. A is a sleeve that revolves very freely and without shake on a vertical steel cone. This cone is mounted on a circular base b which rests on three levelling screws, two of which are visible in the figure. The sleeve carries a crosspiece on its upper extremity to which the bearings of the horizontal axis c are attached. A reversible level d rests on the accurately turned pivots of this axis. The telescope is attached to one end of this axis and a counterpoise e to the other. The long arm f serves to clamp the telescope in zenith distance and to communicate slow motion in zenith distance when so clamped. On the side of the telescope opposite to the horizontal axis is attached a graduated circle g, and, turning concentrically with this circle, is a framework h, to which the readers and verniers of the circle are fixed. This frame carries two very sensitive levels, k and 1, and the whole frame can be clamped to the circle g by means of the clamping screw m. The ob j ect-glass of the telescope is, of course, attached by its cell to the upper end of the telescope tube. Within the focus of the object-glass is a right-angled prism of total reflection, which diverts the converging rays from the object-glass at right angles to the axis of the telescope, and permits the observing micrometer n to be mounted in the very convenient position shown in the figure. A small graduated circle p concentric with A is attached to the circular base b and read by the microscopes q r, attached to a. The instrument is thus a theodolite, although, compared with its other dimensions, feeble as an apparatus for the measurement of absolute altitudes and azimuths, although capable of determining these co-ordinates with considerable precision.

In practice the vertical circle is adjusted once for all, so that when the levels k and l are in the centre of their run, the verniers read true zenith distances. When the instrument has been set up and levelled (either with aid of the cross level d, or the levels k and 1), the reading of the circle p for the meridional position of the telescope is determined either by the method of transits in the meridian (see Transit Circle), or by the observation of the azimuth of a known star at a known hour angle. This done, the stops s and t are clamped and adjusted so that when arm r comes in contact with the screw of stop t the telescope will point due north, and when in contact with s, it will point due south, or vice versa. A pair of stars of known declination are selected such that their zenith distances, when on the meridian, are nearly equal and opposite, and whose right ascensions differ by five or ten minutes of time. Assuming, for example, that the northern star has the smaller right ascension, the instrument is first, with the aid of the stop, placed in the meridian towards the north; the verniers of the graduated circle g are set to read to the reading 40-2(Sn+Ss) where 0 is the approximate latitude of the place and Sn, Ss the declinations of the northern and southern star respectively; then the level frame h is turned till the levels k and I are in the middle of their run, and there clamped by the screw m, aided in the final adjustment by the adjoining slow motion screw shown in the figure. The telescope is now turned on the horizontal axis till the levels read near the centres of these scales and the telescope is clamped to the arm f. When the star enters the field of view its image is approximately bisected by the spider web of the micrometer n, the exact bisection being completed in the immediate neighbourhood of the meridian. The readings of the levels k and 1 and the reading of the micrometer-drum are then entered, and the observation of the northern star is complete. Now the instrument is slowly turned towards the south, till the azimuth arm is gently brought into contact with the corresponding stop s, care being taken not to touch any part of the instrument except the azimuth arm itself. When the southern star enters the field the same process is repeated.

Suppose now, for the moment, that the readings of the levels k and l are identical in both observations, we have then, in the difference between the micrometer readings north and south, a measure of the difference of the two zenith distances expressed in terms of the micrometer screw; and, if the "` value of one revolution of the micrometer screw" is known in seconds of arc we have for the resulting latitude FIG. 24. - Zenith Telescope (by Warner & SwaseY)4 - 2{(n -)+(Sn+ss)I, where - t 5 is the difference of the micrometer readings converted into arc - it being assumed that increased micrometer readings correspond with increased zenith distance of the star.

If between the north and south observation there is a change in the level readings of the levels k and 1, this indicates a change in the zenith distance of the axis of the telescope. By directing the telescope to a distant object, or to the intersection of the webs of a fixed collimating telescope (see Transit Circle), it is easy to measure the effect of a small change of zenith distance of the axis of the telescope in terms both of the level and of the micrometer screw, and thus, if the levels are perfectly sensitive and uniform in curvature and graduation, to determine the value of one division of each level in terms of the micrometer screw. The value of "one revolution of the screw in seconds of arc" can be determined either by observing at transit the difference of zenith distance of two stars of known declination in terms of the micrometer screw, the instrument remaining at rest between their transits; or by measuring at known instants in terms of the screw, the change of zenith distance of a standard star of small polar distance near the time of its greatest elongation.

The reason why two levels are employed is that sometimes crystals are formed by the decomposition of the glass which cause the bubble to stick at different points and so give false readings. Two levels are hardly likely to have such causes of error arise at exactly corresponding points in their run, and thus two levels furnish an independent control the one on the other. Also it is impossible to make levels that are in every respect perfect, nor even to determine these errors for different lengths of bubble and at different readings with the highest precision. The mean of two levels therefore adds to the accuracy of the result.

Attempts have been made to overcome the difficulties connected with levels by adopting the principle of Kater's floating collimator (Phil. Trans., 1825 and 1828). On this principle the use of the level is abolished, the telescope is mounted on a metallic float, and it is assumed that, in course of the rotation of this float, the zenith distance of the axis of the telescope will remain undisturbed, that is, of course, after the undulations, induced by the disturbance of the mercury, have ceased.

S. C. Chandler in 1884 constructed an equal altitude instrument on this principle, which he called the almucantar, and he found that after disturbance the telescope recovered its original zenith distance within j i of a second of arc. R. A. Sampson at Durham (Monthly Notices R.A.S. lx. 572) and H. A. Howe (Ast. Jahrb. xxi. 57) have had instruments constructed on the same general principle. It is, however, obviously impossible to apply a micrometer with advantage to such instruments, because to touch such an instrument, in order to turn a micrometer screw, would obviously set it into motion. The almucantar was therefore used only to observe the vertical transits of stars in different azimuths over fixed horizontal webs, without touching the telescope.

By the use of photography, however, it is possible to photograph the trail of a star as it transits the meridian when the telescope is directed towards the north, and another trail be similarly photographed when the telescope is directed towards the south. The interval between the true trails, measured at right angles to the direction of the trails, obviously corresponds to the difference of zenith distance of the two stars. This principle has been applied with great completeness and ingenuity of detail by Bryan Cookson to the construction of a "photographic floating zenith telescope," ??

NIU?II1010 which he has erected at Cambridge (Eng.) and applied to an investigation of the change of latitude and a determination of the constant of refraction. A description of the instrument, and some preliminary results obtained by it, is given by him (Monthly Notices R.A.S., lxi. 314). [D. GI.]