Aside from the ordinary equipment of oculars various accessories form an important part of the observer’s equipment, their number and character depending on the instrument in use and the purposes to which it is devoted.

First in general usefulness are several special forms of eyepiece equipment supplementary to the usual oculars. At the head of the list is the ordinary star diagonal for the easier viewing of objects near the zenith here shown in Fig. 126. It is merely a tube, A, fitting the draw tube of the telescope, with a slotted side tube B, at a right angle, into which the ordinary ocular fits, and a right angled prism C with its two faces perpendicular respectively to the axes of the main and side tubes, and the hypothenuse face at 45° to each. The beam coming down the tube is totally reflected at this face and brought to focus at the ocular. The lower end of the tube is closed by a cap to exclude dust.

One looks, by help of this, horizontally at zenith stars, or, if observing objects at rather high altitude, views them at a comfortable angle downward. The prism must be very accurately made to avoid injury to the definition, but loses only about 10% of the light, and adds greatly to the comfort of observing.

Of almost equal importance is the solar diagonal devised by Sir John Herschel, Fig. 127. Here the tube structure A, B, is quite the same as in Fig. 126 but the right angled prism is replaced by a simple elliptical prism C of small angle, 10° or less, with its upper face accurately plane and at 45° to the axes of the tubes, resting on a lining tube D cut off as shown. In viewing the sun only about 5% of the light (and heat) is reflected at this upper surface to form the image at the eye piece.

Any reflection from the lower polished surface is turned aside out of the field, while the remainder of the radiation passes through the prism C and is concentrated below it. To prevent scorching the observer the lower end of the tube is capped at E, but the cap has side perforations to provide circulation for the heated air. Using such a prism, the remnant of light reflected can be readily toned down by a neutral tinted glass over the ocular.

In the telescopes of 3 inches and less aperture, and ordinary focal ratio, a plane parallel disc of very dark glass over the ocular gives sufficient protection to the eye. This glass is preferably of neutral tint, and commonly is a scant 1/16 inch thick. Some observers prefer other tints than neutral. A green and a red glass superimposed give good results and so does a disc of the deepest shade of the so-called Noviweld glass, which is similar in effect.

With an aperture as large as 3 inches a pair of superimposed dark glasses is worth while, for the two will not break simultaneously from the heat and there will be time to get the eye away in safety. A broken sunshade is likely to cost the observer a permanent scotoma, blindness in a small area of the retina which will neither get better nor worse as time goes on.

Above 3 inches aperture the solar prism should be used or, if one cares to go to fully double the cost, there is nothing more comfortable to employ in solar observation than the polarizing eye piece, Fig. 128. This shows schematically the arrangement of the device. It depends on the fact that a ray of light falling on a surface of common glass at an angle of incidence of approximately 57° is polarized by the reflection so that while it is freely reflected if it falls again on a surface parallel to the first, it is absorbed if it falls at the same incidence on a surface at right angles to the first.

Thus in Fig. 128 the incident beam from the telescope falls on the black glass surface a at 57° incidence, is again reflected from the parallel mirror b, and then passed on, parallel to its original path, to the lower pair of mirrors c, d. The purpose of the second reflection is to polarize the residual light which through the convergence of the rays was incompletely polarized at the first.

The lower pair of mirrors c, d, again twice reflect the light at the polarizing angle, and, in the position shown, pass it on to the ocular diminished only by the four reflections. But if the second pair of mirrors be rotated together about a line parallel to b c as an axis the transmitted light begins to fade out, and when they have been turned 90°, so that their planes are inclined 90° to a and b (= 33° to the plane of the paper), the light is substantially extinguished.

Thus by merely turning the second pair of mirrors the solar image can be reduced in brilliancy to any extent whatever, without modifying its color in any way. The typical form given to the polarizing eyepiece is similar to Fig. 129. Here t_2 is the box containing the polarizing mirrors, a b, fitted to the draw tube, but for obvious reasons eccentric with it, t_1 is the rotating box containing the “analysing” mirrors c, d, and a is the ocular turning with it.

Sometimes the polarizing mirrors are actually a pair of Herschel prisms as in Fig. 126, facing each other, thus getting rid of much of the heat. Otherwise the whole set of mirrors is of black glass to avoid back reflections. In simpler constructions single mirrors are used as polarizer and analyser, and in fact there are many variations on the polarizing solar eyepiece involving about the same principles.

In any solar eyepiece a set of small diaphragms with holes from perhaps 1/64 inch up are useful in cutting down the general glare from the surface outside of that under scrutiny. These may be dropped upon the regular diaphragm of the ocular or conveniently arranged in a revolving diaphragm like that used with the older photographic lenses.

The measurement of celestial objects has developed a large group of important auxiliaries in the micrometers of very varied forms. The simplest needs little description, since it consists merely of a plane parallel disc of glass fitting in the focus of a positive ocular, and etched with a network of uniform squares, forming a reticulated micrometer by which the distance of one object from another can be estimated.

It can be readily calibrated by measuring a known distance or noting the time required for an equatorial star to drift across the squares parallel to one set of lines. It gives merely a useful approximation, and accurate measures must be turned over to more precise instruments.

The ring micrometer due, like so much other valuable apparatus, to Fraunhofer, is convenient and widely used for determining positions. It consists, as shown in Fig. 130, of an accurately turned opaque ring, generally of thin steel, cemented to a plane parallel glass or otherwise suspended in the center of the eyepiece field. The whole ring is generally half to two thirds the width of the field and has a moderate radial width so that both the ingress and the egress of a star can be conveniently timed.

It depends wholly on the measurement of time as the stars to be compared drift across the ring while the telescope is fixed, and while a clock or chronometer operating a sounder is a desirable adjunct one can do pretty well with a couple of stop watches since only differential times are required.

For full directions as to its use consult Loomis’ Practical Astronomy, a book which should be in the library of every one who has the least interest in celestial observations. Suffice it to say here that the ring micrometer is very simple in use, and the computation of the results is quite easy. In Fig. 130 F is the edge of the field, R the ring, and a b, a'b', the paths of the stars s and s', the former well into the field, the latter just within the ring. The necessary data comprise the time taken by each star to transverse the ring, and the radius of the ring in angular measure, whence the difference in R. A. or Dec, can be obtained.[20]

Difference of R. A. = ½ (t'-t)½ (T'-T) where (T'-T) is the time taken for transit of second star. To obtain differences of declination one declination should be known at least approximately, and the second estimated from its relative position in the ring or otherwise. Then with these tentative values proceed as follows.

Put x = angle aob and x' = angle a'o'b'
Also let d = approximate declination of s and
d' = approximate declination of s'

Then sin x = (15/2r) cos d (T'-T)

sin x' = (15/2r) cos d' (t'-t) and finally
Difference of Dec. = r (cos x'-cos x), when both arcs are on the same side of center of ring. If on opposite sides, Diff. = r (cos x' + cos x).

There is also now and then used a square bar micrometer, consisting of an opaque square set with a diagonal in the line of diurnal motion. It is used in much the same way as the ring, and the reductions are substantially the same. It has some points of convenience but is little used, probably on account of the great difficulty of accurate construction and the requirement, for advantageous use, that the telescope should be on a well adjusted equatorial stand.[21] The ring micrometer works reasonably well on any kind of steady mount, requires no illumination of the field and is in permanent working adjustment.

Still another type of micrometer capable of use without a clock-drive is the double image instrument. In its usual form it is based on the principle that if a lens is cut in two along a diameter and the halves are slightly displaced along the cut all objects will be seen double, each half of the lens forming its own set of images.

Conversely, if one choses two objects in the united field these can be brought together by sliding the halves of the lens as before, and the extent of the movement needed measures the distance between them. Any lens in the optical system can be thus used, from the objective to the eyepiece. Fig. 131 shows a very simple double image micrometer devised by Browning many years ago. Here the lens divided is a so-called Barlow lens, a weak achromatic negative lens sometimes used like a telephoto lens to lengthen the focus and hence vary the power of a telescope.

This lens is shown at A with the halves widely separated by the double threaded micrometer screw B, which carries them apart symmetrically. The ocular proper is shown at C.

Double image micrometers are now mainly of historical interest, and the principle survives chiefly in the heliometer, a telescope with the objective divided, and provided with sliding mechanism of the highest refinement. The special function of the heliometer is the direct micrometric measurement of stellar distances too great to be within the practicable range of a filar micrometer—distances for example up to 1½° or even more.

The observations with the heliometer are somewhat laborious and demand rather intricate corrections, but are capable of great precision. (See Sir David Gill’s article “Heliometer” in the Enc. Brit. 11th Ed.). At the present day celestial photography, with micrometric measurement of the resulting plates, has gone far in rendering needless visual measurements of distances above a very few minutes of arc, so that it is somewhat doubtful whether a large heliometer would again be constructed.

The astronomer’s real arm of precision is the filar micrometer. This is shown in outline in Fig. 132, the ocular and the plate that carries it being removed so as to display the working parts. It consists of a main frame aa, carrying a slide bb, which is moved by the screws and milled head B. The slide bb carries the vertical spider line mm, and usually one or more horizontal spider lines, line mm is the so-called fixed thread of the micrometer, movable only as a convenience to avoid shifting the telescope.

On bb moves the micrometer slide cc, carrying the movable spider line nn and the comb which records, with mm as reference line, the whole revolutions of the micrometer screw C. The ocular sometimes has a sliding motion of its own on cc, to get it positioned to the best advantage. In use one star is set upon mm by the screw B and then C is turned until nn bisects the other star.

Then the exact turns and fraction of a turn can be read off on the comb and divided head of C, and reduced to angular measure by the known constant of the micrometer, usually determined by the time of passage of a nearly equatorial star along the horizontal thread when mm, nn, are at a definite setting

apart. (Then r = (15(t'-t) cos d)/N where r is the value of a revolution in seconds of arc, N the revolutions apart of mm, nn, and t and d as heretofore.)

Very generally the whole system of slides is fitted to a graduated circle, to which the fixed horizontal thread is diametral. Then by turning the micrometer until the horizontal threads cut the two objects under comparison, their position angle with reference to a graduated circle can be read off. This angle is conventionally counted from 0° to 360° from north around through east.

Figure 133 shows the micrometer constructed by the Clarks for their 24 inch equatorial of the Lowell Observatory. Here A is the head of the main micrometer screw of which the whole turns are reckoned on the counter H in lieu of the comb of Fig. 132. B is the traversing screw for the fixed wire system, C the clamping screw of the position circle, D its setting pinion, E the rack motion for shifting the ocular, F the reading glass for the position circle, and G the little electric lamp for bright wire illumination. The parts correspond quite exactly with the diagram of Fig. 132 but the instrument is far more elegant in design than the earlier forms of micrometer and fortunately rid of the oil lamps that were once in general use. A small electric lamp with reflector throws a little light on the spider lines—just enough to show them distinctly. Or sometimes a faint light is thus diffused in the field against which the spider lines show dark.

Commonly either type of illumination can be used and modified as occasion requires. The filar micrometer is seldom used on small telescopes, since to work easily with it the instrument should be permanently mounted and clock-driven. Good work was done by some of the early observers without these aids, but at the cost of infinite pains and much loss of time.

The clock drive is in fact a most important adjunct of the telescope when used for other purposes than ordinary visual observations, though for simple seeing a smooth working slow motion in R. A. answers well. The driving clock from the horological view-point is rudimentary. It consists essentially of a weight-driven, or sometimes spring-driven, drum, turning by a simple gear connection a worm which engages a carefully cut gear wheel on the polar axis, while prevented from running away by gearing up to a fast running fly-ball governor, which applies friction to hold the clockwork down to its rate if the speed rises by a minute amount. There is no pendulum in the ordinary sense, the regularity depending on the uniformity of the total friction—that due to the drive plus that applied by the governor.

Figure 134 shows a simple and entirely typical driving clock by Warner & Swasey. Here A is the main drum with its winding gear at B, C is the bevel gear, which is driven from another carried by A, and serves to turn the worm shaft D; E marks the fly balls driven by the multiplying gearing plainly visible. The governor acts at a predetermined rotation speed to lift the spinning friction disc F against its fixed mate, which can be adjusted by the screw G.

The fly-balls can be slightly shifted in effective position to complete the regulation. These simple clocks, of which there are many species differing mainly in the details of the friction device, are capable of excellent precision if the work of driving the telescope is kept light.

For large and heavy instruments, particularly if used for photographic work where great precision is required, it is difficult to keep the variations of the driving friction within the range of compensation furnished by the governor friction alone, and in such case recourse is often taken to constructions in which the fly balls act as relay to an electrically controlled brake, or in which the driving power is supplied by an electric motor suitably governed either continuously or periodically. For such work independent hand guiding mechanism is provided to supplement the clockwork. For equatorials of the smallest sizes, say 3 to 4 inches aperture, spring operated driving clocks are occasionally used. The general plan of operation is quite similar to the common weight driven forms, and where the weights to be carried are not excessive such clocks do good work and serve a very useful purpose.

An excellent type of the simple spring driving clock is shown in Fig. 136 as constructed by Zeiss. Here 1 is the winding gear, 2 the friction governor, and 3 the regulating gear. It will be seen that the friction studs are carried by the fly balls themselves, somewhat as in Fraunhofers’ construction a century since, and the regulation is very easily and quickly made by adjusting the height of the conical friction surface above the balls.

For heavier work the same makers generally use a powerful weight driven train with four fly-balls and electric seconds control, sometimes with the addition of electric motor slow motions to adjust for R. A. in both directions.

Figure 135 is a rather powerful clock of analogous form by the Clarks. It differs a little in its mechanism and especially in the friction gear in which the bearing disc is picked up by a delicately set latch and carried just long enough to effect the regulation. It is really remarkable that clockworks of so simple character as these should perform as well as experience shows that they do. In a few instances clocks have depended on air-fans for their regulating force, something after the manner of the driving gear of a phonograph, but though rather successful for light work they have found little favor in the task of driving equatorials. An excellent type of a second genus is the pendulum controlled driving clock due to Sir David Gill. This has a powerful weight-driven train with the usual fly-ball governor. But the friction gear is controlled by a contact-making seconds pendulum in the manner shown diagrammatically in Fig. 137. Two light leather tipped rods each controlled by an electro magnet act upon an auxiliary brake disc carried by the governor spindle which is set for normal speed with one brake rod bearing lightly on it. Exciting the corresponding magnet relieves the pressure and accelerates the clock, while exciting the other adds braking effect and slows it.

In Fig. 137 is shown from the original paper, (M. N. Nov., 1873), the very ingenious selective control mechanism. At P is suspended the contact-making seconds-pendulum making momentary contact by the pin Q with a mercury globule at R. Upon a spindle of the clock which turns once a second is fixed a vulcanite disc ?, d, e, s. This has a rim of silver broken at the points ?, d, e, s, by ivory spacers covering 3° of circumference. On each side of this disc is another, smaller, and with a complete silver rim. One, ??, is shown, connected with the contact spring V; its mate ?'?', on the other side contacts with U, while a third contact K bears on the larger disc.

The pair of segments s, ?, and d, e, are connected to ? ?, the other pair of segments to ?' ?'. Now suppose the discs turning with the arrows: If K rests on one of the insulated points when the pendulum throws the battery C Z into circuit nothing happens. If the disc is gaining on the pendulum, K, instead of resting on ? as shown will contact with segment ?, s, and actuate a relay via V, exciting the appropriate brake magnet.

If the disc is losing, K contacts with segment ?, d, and current will pass via ?'?' and U to a relay that operates the other brake magnet and lets the clock accelerate. A fourth disc (not shown) on the same spindle is entirely insulated on its edge except at points corresponding to ? and e, and with a contact spring like K.

If the disc is neither gaining nor losing when the pendulum makes contact, current flows via this fourth disc and sets the relay on the mid-point ready to act when needed. This clock is the prototype of divers electrically-braked driving clocks with pendulum control, and proved beautifully precise in action, like various kindred devices constructed since, though the whole genus is somewhat expensive and intricate.

The modern tendency in driving apparatus for telescopes, particularly large instruments, is to utilize an electric motor for the source of power, using a clock mechanism merely for the purpose of accurately regulating the rate of the motor. We thus have the driving clock in its simplest form as a purely mechanical device worked by a sensitive fly-ball governor. The next important type is that in which the clock drive is precisely regulated by a pendulum clock, the necessary governing power being applied electrically as in Fig. 137 or sometimes mechanically.

Finally we come to the type now under consideration where the instrument itself is motor driven and the function of the clock is that of regulating the motor. A very good example of such a drive is the Gerrish apparatus used for practically all the instruments at the various Harvard observatory stations, and which has proved extremely successful even for the most trying work of celestial photography. The schematic arrangement of the apparatus is shown in Fig. 138. Here an electric motor shown in diagram in 1, Fig. 138, is geared down to approximately the proper speed for turning the right ascension axis of the telescope. It is supplied with current either from a battery or in practice from the electric supply which may be at hand. This motor is operated on a 110 volt circuit which supplies current through the switch 2 which is controlled by the low voltage clock circuit running through the magnet 3. The clock circuit can be closed and opened at two points, one controlled by the seconds pendulum 5, the other at 7 by the stud on the timing wheel geared to the motor for one revolution per second. There is also a shunt around the pendulum break, closed by the magnet switch at 6. This switch is mechanically connected to the switch 2 by the rod 4, so that the pair open and close together.

The control operates as follows: Starting with the motor at rest, the clock circuit is switched on, switches 2, 6 being open and 7 closed. At the first beat of the pendulum 2, 6 closes and the current, shunted across the loop containing 5, holds 2 closed until the motor has started and broken the clock circuit at the timer. The fly-wheel carries on until the pendulum again closes the power circuit via 2, 6, and current stays on the motor until the timer has completed its revolution.

This goes on as the motor speeds up, the periodic power supply being shortened as the timer breaks it earlier owing to the acceleration, until the motor comes to its steady speed at which the power is applied just long enough to maintain uniformity. If the motor for any cause tends to overspeed the cut-off is earlier, while slowing down produces a longer power-period bringing the speed back to normal. The power period is generally ¼ to ½ second. The power supplied to the motor is very small even in the example here shown, only 1 ampere at 110 volts.

The actual proportion of a revolution during which current is supplied the motor is therefore rigorously determined by the clock pendulum, and the motor is selected so that its revolutions are exactly timed to this clock pendulum which has no work to do other than the circuit closing, and can hence be regulated to keep accurate time. The small fly-wheel (9), the weight of which is carefully adjusted with respect to the general amount of work to be done, attached to the motor shaft, effectively steadies its action during the process of government. This Gerrish type has been variously modified in detail to suit the instruments to which it has been applied, always following however the same fundamental principles.

An admirable example of the application of this drive is shown in Fig. 139, the 24 inch reflector at the Harvard Observatory. The mount is a massive open fork, and the motor drive is seen on the right of the mount. There are here two motors, ordinary fan motors in size. The right hand motor carries the fly-wheel and runs steadily on under the pendulum control. The other, connected to the same differential gear as the driving motor, serves merely for independent regulation and can be run in either direction by the observer to speed or slow the motion in R. A. These examples of clock drive are merely typical of those which have proved to be successful in use for various service, light and heavy. There are almost innumerable variations on clocks constructed on one or another of the general lines here indicated, so many variations in fact that one almost might say there are few driving clocks which are not in some degree special.

The tendency at present is for large instruments very distinctly toward a motor-driven mechanism operating on the right ascension axis, and governed in one of a considerable variety of ways by an actual clock pendulum. For smaller instruments the old mechanical clock, often fitted with electric brake gear and now and then pendulum regulated, is capable of very excellent work.

The principle of the spectroscope is rudimentarily simple, in the familiar decomposition of white light into rainbow colors by a prism. One gets the phenomena neatly by holding a narrow slit in a large piece of cardboard at arms length and looking at it through a prism held with its edge parallel to the slit. If the light were not white but of a mixture of definite colors each color present would be represented by a separate image of the slit instead of the images being merged into a continuous colored band.

With the sun as source the continuous spectrum is crossed by the dark lines first mapped by Fraunhofer, each representing the absorption by a relatively cool exterior layer of some substance that at a higher temperature below gives a bright line in exactly the same position.

The actual construction of the astronomical spectroscope varies greatly according to its use. In observations on the sun the distant slit is brought nearer for convenience by placing it in the focus of a small objective pointed toward the prisms (the collimator) and the spectrum is viewed by a telescope of moderate magnifying power to disclose more of detail. Also, since there is extremely bright light available, very great dispersion can be used, obtained by several or many prisms, so that the spectrum is both fairly wide, (the length of the slit) and extremely long.

In trying to get the spectrum of a star the source is a point, equivalent to an extremely minute length of a very narrow slit. Therefore no actual slit is necessary and the chief trouble is to get the spectrum wide enough and bright enough to examine.

The simplest form of stellar spectroscope and the one in most common use with small telescopes is the ocular spectroscope arranged much like Fig. 140. This fits into the eye tube of a telescope and the McClean form made by Browning of London consists of an ordinary casing with screw collar B, a cylindrical lens C, a direct vision prism c, f, c, and an eye-cap A.

The draw tube is focussed on the star image as with any other ocular, and the light is delivered through C to the prism face nearly parallel, and thence goes to the eye, after dispersion by the prism. This consists of a central prism, f, of large angle, made of extremely dense flint, to which are cemented a pair of prisms of light crown c, c, with their bases turned away from that of f.

We have already seen that the dispersions of glasses vary very much more than their refractions so that with proper choice of materials and angles the refraction of f is entirely compensated for some chosen part of the spectrum, while its dispersion quite overpowers that of the crown prisms and gives a fairly long available spectrum.

The cylindrical lens C merely serves to stretch out the tiny round star image into a short line thereby giving the resulting spectrum width enough to examine comfortably. The weak cylindrical lens is sometimes slipped over the eye end of the prisms to give the needed width of spectrum instead of putting it ahead of the prisms.

A small instrument of this kind used with a telescope of 3 inches to 5 inches aperture gives a fairly good view of the spectra of starts above second or third magnitude, the qualities of tolerably bright comets and nebulÆ and so forth. The visibility of stellar spectra varies greatly according to their type, those with heavy broad bands being easy to observe, while for the same stellar magnitude spectra with many fine lines may be quite beyond examination. Nevertheless a little ocular spectroscope enables one to see many things well worth the trouble of observing.

With the larger instruments, say 6 or 8 inches, one can well take advantage of the greater light to use a spectroscope with a slit, which gives somewhat sharper definition and also an opportunity to measure the spectrum produced.

An excellent type of such an instrument is that shown in Fig. 141, due to Professor AbbÉ. The construction is analogous to Fig. 140. The ocular is a Huyghenian one with the slit mechanism (controlled by a milled head) at A in the usual place of the diaphragm. The slit is therefore in the focus of the eye lens, which serves as collimating lens. Above is the direct vision system J with the usual prisms which are slightly adjustable laterally by the screw P and spring Q.

At N is a tiny transparent scale of wave lengths illuminated by a faint light reflected from the mirror O, and in the focus of the little lens R, which transfers it by reflection from the front face of the prism to the eye, alongside the edge of the spectrum. One therefore sees the spectrum marked off by a bright line wave-length scale.

The pivot K and clamp L enable the whole to be swung side-wise so that one can look through the widened slit, locate the star, close the slit accurately upon it and swing on the prisms. M is the clamp in position angle. Sometimes a comparison prism is added, together with suitable means for throwing in spectra of gases or metals alongside that of the star, though these refinements are more generally reserved for instruments of higher dispersion.

To win the advantage of accurate centering of the star in the field gained by the swing-out of the spectroscope in Fig. 141 simple instruments like Fig. 140 are sometimes mounted with an ordinary ocular in a double nose-piece like that used for microscope objectives, so that either can be used at will.

Any ordinary pocket spectroscope, with or without scale or a comparison prism over part of the slit, can in fact be fitted to an adapter and used with the star focussed on the slit and a cylindrical lens, if necessary, as an eye-cap.

Such slit spectroscopes readily give the characteristics of stellar spectra and those of the brighter nebulÆ or of comets. They enable one to identify the more typical lines and compare them with terrestrial sources, and save for solar work are about all the amateur observer finds use for.

For serious research a good deal more of an instrument is required, with a large telescope to collect the light, and means for photographing the spectra for permanent record. The cumulative effect of prolonged exposures makes it possible easily to record spectra much too faint to see with the same aperture, and exposures are often extended to many hours.

Spectroscopes for such use commonly employ dense flint prisms of about 60° refracting angle and refractive index of about 1.65, one, two, or three of these being fitted to the instrument as occasion requires. A fine example by Brashear is shown in Fig. 142, arranged for visual work on the 24 inch Lowell refractor. Here A is the slit, B the prism box, C the observing telescope, D the micrometer ocular with electric lamp for illuminating the wires, and E the link motion that keeps the prism faces at equal angles with collimator and observing telescope when the angle between these is changed to observe different parts of the spectrum. This precaution is necessary to maintain the best of definition.

When photographs are to be taken the observing telescope is unscrewed and a photographic lens and camera put in its place. If the brightness of the object permits, three prisms are installed, turning the beam 180° into a camera braced to the same frame alongside the slit.

For purely photographic work, too, the objective prism used by Fraunhofer for the earliest observation of stellar spectra is in wide use. It is a prism fitted in front of the objective with its refracting faces making equal angles with the telescope and the region to be observed, respectively. Its great advantages are small loss of light and the ability to photograph many spectra at once, for all the stars in the clear field of the instrument leave their images spread out into spectra upon the photographic plate.

Figure 143 shows such an objective prism mounted in front of an astrographic objective. The prism is rotatable into any azimuth about the axis of the objective and by the scale i and clamping screw r can have its refracting face adjusted with respect to that axis to the best position for photographing any part of the spectrum. Such an arrangement is typical of those used for small instruments say from 3 inches to 6 inches aperture.

For larger objectives the prism is usually of decidedly smaller angle, and, if the light warrants high dispersion, several prisms in tandem are used. The objective prism does its best work when applied to true photographic objectives of the portrait lens type which yield a fairly large field. It is by means of big instruments of such sort that the spectra for the magnificent Draper Catalogue have been secured by the Harvard Observatory, mostly at the Arequipa station. In photographing with the objective prism the spectra are commonly given the necessary width for convenient examination by changing just a trifle the rate of the driving clock so that there is a slight and gradual drift in R. A. The refracting edge of the prism being turned parallel to the diurnal motion this drift very gradually and uniformly widens the spectrum to the extent of a few minutes of arc during the whole exposure.

When one comes to solar spectroscopy one meets an entirely different situation. In stellar work the difficulty is to get enough light, and the tendency is toward large objectives of relatively short focal length and spectroscopes of moderate dispersion. In solar studies there is ample light, and the main thing is to get an image big enough to be scrutinized in detail with very great dispersion.

Especially is this true in the study of the chromospheric flames that rim the solar disc and blaze over its surface. To examine these effectively the spectroscope should have immense dispersion with a minimum amount of stray light in the field to interfere with vision of delicate details.

In using a spectroscope like Fig. 142, if one turned the slit toward the landscape, the instrument being removed from the telescope and the slit opened wide, he could plainly see its various features, refracted through the prism, and appearing in such color as corresponded to the part of the spectrum in the line of the observing telescope. In other words one sees refracted images quite distinctly in spite of the bending of the rays. With high dispersion the image seen is practically monochromatic.

Now if one puts the spectroscope in place, brings the solar image tangent to the slit and then cautiously opens the slit, he sees the bright continuous spectrum of the sky close to the sun, plus any light of the particular color for which the observing telescope is set, which may proceed from the edge of the solar disc. Thus, if the setting is for the red line of hydrogen (C), one sees the hydrogen glow that plays in fiery pillars of cloud about the sun’s limb quite plainly through the opened slit, on a background of light streaming from the adjacent sky. To see most clearly one must use great dispersion to spread this background out into insignificance, must keep other stray light out of the field, and limit his view to the opened slit.

To these ends early solar spectroscopes had many prisms in tandem, up to a dozen or so, kept in proper relation by complicated linkages analogous to the simple one shown in Fig. 142. Details can be found in almost any astronomical work of 40 years ago. They were highly effective in giving dispersion but neither improved the definition nor cut out light reflected back and forth from their many surfaces.

Of late simpler constructions have come into use of which an excellent type is the spectroscope designed by Mr. Evershed and shown in diagram in Fig. 144. Here the path of the rays is from the slit through the collimator objective, then through a very powerful direct vision system, giving a dispersion of 30° through the visible spectrum, then by reflection from the mirror through a second such system, and thence to the observing telescope. The mirror is rotated to get various parts of the spectrum into view, and the micrometer screw that turns it gives means for making accurate measurement of wave lengths.

There are but five reflecting surfaces in the prism system (for the cemented prism surfaces do not count for much) as against more than 20 in one of the older instruments of similar power, and as in other direct vision systems the spectrum lines are substantially straight instead of being strongly curved as with multiple single prisms. The result is the light, compact, and powerful spectroscope shown complete in Fig. 145, equally well fitted for observing the sun’s prominences and the detailed spectrum from his surface.

In most of the solar spectroscopes made at the present time the prisms are replaced by a diffraction grating. The original gratings made by Fraunhofer were made of wire. Two parallel screws of extremely fine thread formed two opposite sides of a brass frame. A very fine wire was then wound over these screws, made fast by solder on one side of each, and then cut away on the other, so as to leave a grating of parallel wires with clear spaces between.

Today the grating is generally ruled by an automatic ruling engine upon a polished plate of speculum metal. The diamond point carried by the engine cuts very smooth and fine parallel furrows, commonly from 10,000 to 20,000 to the inch. The spaces between the furrows reflect brilliantly and produce diffraction spectra.[22]

When a grating is used instead of prisms the instrument is commonly set up as shown in Fig. 146. Here A is the collimator with slit upon which the solar image light falls, B is the observing telescope, and C the grating set in a rotatable mount with a fine threaded tangent screw to bring any line accurately upon the cross wires of the ocular.

The grating gives a series of spectra on each side of the slit, violet ends toward the slit, and with deviations proportional to 1, 2, 3, 4, etc., times the wave length of the line considered. The spectra therefore overlap, the ultra violet of the second order being superimposed on the extreme red of the first order and so on. Colored screens over the slit or ocular are used to get the overlying spectra out of the way.

The grating spectroscopes are very advantageous in furnishing a wide range of available dispersions, and in giving less stray light than a prism train of equal power. The spectra moreover are very nearly “normal,” i.e., the position of each line is proportional to its wave length instead of the blue being disproportionately long as in prismatic spectra.

In examining solar prominences the widened slit of a grating spectroscope shows them foreshortened or stretched to an amount depending on the angular position of the grating, but the effect is easily reckoned.[23]

If the slit is nearly closed one sees merely a thin line, irregularly bright according to the shape of the prominence; a shift of the slit with respect to the solar image shows a new irregular section of the prominence in the same monochromatic light.

These simple phenomena form the basis of one of the most important instruments of solar study—the spectro-heliograph. This was devised almost simultaneously by G. E. Hale and M. Deslandres about 30 years ago, and enables photographs of the sun to be taken in monochromatic light, showing not only the prominences of the limb but glowing masses of gas scattered all over the surface.

The principle of the instrument is very simple. The collimator of a powerful grating spectroscope is provided with a slit the full length of the solar diameter, arranged to slide smoothly on a ball-bearing carriage clear across the solar disc. Just in front of the photographic plate set in the focus of the camera lens is another narrow sliding slit, which, like a focal plane shutter, exposes strip after strip of the plate.

The two slits are geared together by a system of levers or otherwise so that they move at exactly the same uniform rate of speed. Thus when the front slit is letting through a monochromatic section of a prominence on the sun’s limb the plate-slit is at an exactly corresponding position. When the front slit is exactly across the sun’s center so is the plate slit, at each element of movement exposing a line of the plate to the monochromatic image from the moving front slit. The grating can of course be turned to put any required line into action but it usually is set for the K line (calcium), which is photographically very brilliant and shows bright masses of floating vapor all over the sun’s surface.

Figure 147 shows an early and simple type of Professor Hale’s instrument. Here A is the collimator with its sliding slit, B the photographic telescope with its corresponding slide and C the lever system which connects the slides in perfectly uniform alignment. The source of power is a very accurately regulated water pressure cylinder mounted parallel with the collimator. The result is a complete photograph of the sun taken in monochromatic light of exactly defined wave length and showing the precise distribution of the glowing vapor of the corresponding substance.

Since the spectro-heliograph of Fig. 147, which shows the principle remarkably well, there have been made many modifications, in particular for adapting the scheme to the great horizontal and vertical fixed telescopes now in use. (For details of these see Cont. from the Solar Obs. Mt. Wilson, Nos. 3, 4, 23, and others). The chief difficulty always is to secure entirely smooth and uniform motion of the two moving elements.

So great and interesting a branch of astronomy is the study of variable stars that some form of photometer should be part of the equipment of every telescope in serious use for celestial observation. An immense amount of useful work has been done by Argelander’s systematic method of eye observation, but it is far from being precise enough to disclose many of the most important features of variability.

The conventional way of reckoning by stellar magnitudes is conducive to loose measurements, since each magnitude of difference implies a light ratio of which the log is 0.4, i.e., each magnitude is 2.512 times brighter than the following one. As a result of this way of reckoning the light of a star of mag. 9.9 differs from one of mag. 10.0 not by one per cent but by about nine. Hence to grasp light variations of small order one must be able to measure far below 0.1^m.

The ordinary laboratory photometer enables one to compare light sources of anywhere near similar color to a probable error of well under 0.1 per cent, but it allows a comparison between sharply defined juxtaposed fields from the two illuminants, a condition much more favorable to precision than the comparison of two points of light, even if fairly near together.

Stellar photometers may in principle be divided into three classes. (1) Those in which two actual stars are brought into the same field and compared by varying the light from one or both in a known degree. (2) Those which bring a real star into the field alongside an artificial star, and again bring the two to equality by a known variation, usually comparing two or more stars via the same artificial star; (3) those which measure the light of a star by a definite method of extinguishing it entirely or just to the verge of disappearance in a known progression. Of each class there are divers varieties. The type of the first class may be taken as the late Professor E. C. Pickering’s polarizing photometer. Its optical principle is shown in Fig. 148. Here the brightness of two neighboring objects is compared by polarizing at 90° apart the light received from each and reducing the resulting images to equality by an analyzing Nicol prism. The photometer is fully described, with, several other polarizing instruments, in H. A. Vol. II from which Fig. 148 is taken.

A is a Nicol prism inserted in the ocular B, which revolves carrying with it a divided circle C read against the index D. In the tube E which fits the eye end of the telescope, is placed the double image quartz prism F capable of sliding either way without rotation by pulling the cord G. With the objects to be compared in the same field, two images of each appear. By turning the analyzing Nicol the fainter image of the brighter can always be reduced to equality with the brighter image of the fainter, and the amount of rotation measures the required ratio of brightness.[24] This instrument works well for objects near enough to be in the same field of view. The distance between the images can be adjusted by sliding the prism F back and forth, but the available range of view is limited to a small fraction of a degree in ordinary telescopes.

The meridian photometer was designed to avoid this small scope. The photometric device is substantially the same as in Fig. 148. The objects compared are brought into the field by two exactly similar objectives placed at a small angle so that the images, after passing the double image prism, are substantially in coincidence. In front of each of the objectives is a mirror. The instrument points in the east and west line and the mirrors are at 45° with its axis. One brings Polaris into the field, the other by a motion of rotation about the telescope axis can bring any object in or close to the meridian into the field alongside Polaris. The images are then compared precisely as in the preceding instance.[25] There are suitable adjustments for bringing the images into the positions required.

The various forms of photometer using an artificial star as intermediary in the comparison of real stars differ chiefly in the method of varying the light in a determinate measure. Rather the best known is the ZÖllner instrument shown in diagram in Fig. 149. Here A is the eye end of the main telescope tube. Across it at an angle of 45° is thrown a piece of plane parallel glass B which serves to reflect to the focus the beam from down the side tube, C, forming the artificial star.

At the end of this tube is a small hole or more often a diaphragm perforated with several very small holes any of which can be brought into the axis of the tube. Beyond at D, is the source of light, originally a lamp flame, now generally a small incandescent lamp, with a ground glass disc or surface uniformly to diffuse the light.

Within the tube C lie three Nicol prisms n, n₁, n₂. Of these n, is fixed with respect to the mirror B and forms the analyser, which n₁ and n₂ turn together forming the polarizing system. Between n_1 and n_2 is a quartz plate e cut perpendicular to the crystal axis. The color of the light transmitted by such a plate in polarized light varies through a wide range. By turning the Nicol n_2 therefore, the color of the beam which forms the artificial star can be made to match the real star under examination, and then by turning the whole system n_2, E, n_1, reading the rotation on the divided circle at F, the real star can be matched in intensity by the artificial one.

This is viewed via the lens G and two tiny points of light appear in the field of the ocular due respectively to reflection from the front and back of the mirror B, the latter slightly fainter than the former. Alongside or between these the real star image can be brought for a comparison, and by turning the polarizer through an angle [alpha] the images can be equalized with the real image. Then a similar comparison is made with a reference star. If A be the brightness of the former and B of the latter then

A/B = sin²a/sin²

The ZÖllner photometer was at first set up as an alt-azimuth instrument with a small objective and rotation in altitude about the axis C. Since the general use of electric lamps instead of the inconvenient flame it is often fitted to the eye end of an equatorial.

Another very useful instrument is the modern wedge photometer, closely resembling the ZÖllner in some respects but with a very different method of varying the light; devised by the late Professor E. C. Pickering. It is shown somewhat in diagram in Fig. 150. Here as before O is the eye end of the tube, B the plane parallel reflector, C the side tube, L the source of light D the diaphragm and A the lens forming the artificial star by projecting the hole in the diaphragm. In actual practice the diameter of such hole is 1/100 inch or less.

The light varying device W is a “photographic wedge” set in a frame which is graduated on the edge and moved in front of the aperture by a rack and pinion at R. There are beside colored and shade glasses for use as occasion requires. The “photographic wedge” is merely a strip of fine grained photographic plate given an evenly graduated exposure from end to end, developed, and sealed under a cover glass. Its absorption is permanent, non-selective as to color, and it can be made to shade off from a barely perceptible density to any required opacity. Sometimes a wedge of neutral tinted glass is used in its stead.

Before using such a “wedge photometer” the wedge must be accurately calibrated by observation of real or artificial stars of known difference in brightness. This is a task demanding much care and is well described, together with the whole instrument by Parkhurst (Ap. J. 13, 249). The great difficulty with all instruments of this general type is the formation of an artificial star the image of which shall very closely resemble the image of the real star in appearance and color.

Obviously either the real or artificial star, or both, may be varied in intensity by wedge or Nicols, and a very serviceable modification of the ZÖllner instrument, with this in mind was recently described by Shook (Pop. Ast. 27, 595) and is shown in diagram in Fig. 151. Here A is the tube which fits the ordinary eyepiece sleeve. E is a side tube into which is fitted the extension D with a fitting H at its outer end into which sets the lamp tube G. This carries on a base plug F a small flash light bulb run by a couple of dry cells. At O is placed a little brass diaphragm perforated with a minute hole. Between this and the lamp is a disc of diffusing glass or paper. A Nicol prism is set a little ahead of O, and a lens L focusses the perforation at the principal focus of the telescope after reflection from the diagonal glass M, as in the preceding examples. I is an ordinary eyepiece over which is a rotatable Nicol N with a position circle K. At P is a third Nicol in the path of the rays from the real star, thereby increasing the convenient range of the instrument. The original paper gives the details of construction as well as the methods of working. Obviously the same general arrangement could be used for a wedge photometer using the wedge on either real or artificial star or both.

The third type of visual photometer depends on reducing the light of the star observed until it just disappears. This plan was extensively employed by Professor Pritchard of Oxford some 40 years ago. He used a sliding wedge of dark glass, carefully calibrated, and compared two stars by noting the point on the wedge at which each was extinguished. A photographic wedge may be used in exactly the same way.

Another device to the same end depends on reducing the aperture of the telescope by a “cat’s eye,” an iris diaphragm, or similar means until the star is no longer visible or just disappearing. The great objection to such methods is the extremely variable sensitivity of the eye under varying stimulus of light.

The most that can be said for the extinction photometer is that in skillful and experienced hands like Pritchard’s it has sometimes given much more consistent readings than would be expected. It is now and then very convenient for quick approximation but by no courtesy can it be considered an instrument of precision either in astronomical or other photometry.[26]

The photometer question should not be closed without referring the reader to the methods of physical photometry as developed by Stebbins, Guthnick and others. The first of these depends on the use of the selenium cell in which the electrical resistance falls on exposure of the selenium to light. The device is not one adapted to casual use, and requires very careful nursing to give the best results, but these are of an order of precision beyond anything yet reached with an astronomical visual photometer. Settings come down to variations of something like 2 per cent, and variations in stellar light entirely escaping previous methods become obvious.

The photoelectric cell depends on the lowering of the apparent electric resistance of a layer of rarified inert gas between a platinum grid and an electrode of metallic potassium or other alkali metal when light falls on that electrode. The rate of transmission of electricity is very exactly proportional to the illumination, and can be best measured by a very sensitive electrometer. The results are extraordinarily consistent, and the theoretical “probable error” is very small, though here, as elsewhere, “probable error” is a rather meaningless term apt to lead to a false presumption of exactness. Again the apparatus is somewhat intricate and delicate, but gives a precision of working if anything a little better than that of the selenium cell, quite certainly below 1 per cent.

Neither instrument constitutes an attachment to the ordinary telescope of modest size which can be successfully used for ordinary photometry, and both require reduction of results to the basis of visual effect.[27] But both offer great promise in detecting minute variations of light and have done admirable work. For a good fundamental description of the selenium cell photometer see Stebbins, Ap. J. 32, 185 and for the photoelectric method see Guthnick A. N. 196, 357 also A. F. and F. A. Lindemann, M. N. 39, 343. The volume by Miss Furness already referred to gives some interesting details of both.