The following selection is from: T. Thorne Baker, The Spectroscope and its Uses in General Analytical Chemistry. William Wood and Co. New York (1923).

Copyright © 1998 Richard A. Paselk


Baker - Table of Contents




The prism spectroscope- Different forms of apparatus - Slits, lenses, and prisms - Direct vision spectroscopes; Reading the angular deviation of a ray - Plane and concave gratings and replicas.
A PRISM SPECTROSCOPE suitable for laboratory work consists essentially of a slit, collimator, prism, and telescope, and some means of determining the angular deviation of the telescope from the axis of the collimator when any portion of the spectrum or any line is viewed in the centre of the field of the eye-piece.
It would thus seem that all spectroscopes would be much alike in pattern, but this is by no means the case, and an efficient instrument must possess many working details which deserve description.
The slit of the spectroscope should be adjustable - i.e., should be capable of being made narrower or wider. A bright line seen in a spectrum is, as already explained, an image of the slit; and as we want to resolve any such line as the yellow sodium 'line' into its constituents if it be not actually single, the slit must be very narrow, otherwise closely adjacent lines would overlap and would appear single As fine a line as possible is required to measure accurately the angular deviation, as will be readily appreciated if we consider Fig. 13. C and D are the jaws of the slit, and S a screw for opening it; the jaw C is usually fixed, the screw actuating the sliding jaw 1). Suppose now that the distance between C and D be r, and we see a bright line in the telescope of width R. The cross-wires in the







eye-piece are focussed upon the centre of the line, which therefore projects R/2 each side of the centre of the field. If now the slit be so nearly closed that the width is very small, and the bright line seen in the telescope is, for practical purposes, infinitely narrow, both sides of the slit will nearly coincide, and hence each side of the line will be R/2 nearer the centre of the field than when the slit was wider open. But as C is fixed, and D moves further from C

according as the slit is opened, the centre of the line would have been a distance R/2 too much to one side had the slit been used wide.
The screw S should be fitted with a divided drum-head, as shown in the figure, so that accurate measurement can be made of the width used at any time. The jaws should be made of untarnishable material, so that chemical fumes will not attack them, and care should be taken never to close the slit completely at any time. If dust gets between the jaws - and it will cause dark lines to appear across the




length of the spectrum if it does - the slit should be opened wide enough to take the point of an orange stick, which is gently worked up and down two or three times, and not withdrawn until the slit has been widely opened
A sliding V-shaped wedge is usually fitted to the slit, by means of which the width of the aperture, and therefore of the spectrum, can be regulated. Another adjunct is the comparison prism seen in Fig. 14, which allows of two different spectra being examined simultaneously.
The jaws are sometimes made of platinoid, and occasionally quartz. A very fine and accurate edge to the jaws

can be obtained in the case of quartz, and its transparence does not interfere, since the edge is cut obliquely, and hence each jaw refracts the rays which pass into it out of the path of the rays from the slit aperture.
As it is essential that the jaws be parallel, screws should be provided to correct any want of parallelism that may occur after much usage. When the slit is closed, the line of light should vanish, whilst if the jaws are not parallel it will appear slightly wedge-shaped. Correction of this matter must be done with considerable care.
The comparison prism, which when in use covers half the slit, is a small right-angled prism that can be swung out of position when not in use. A ray of light from P (Fig. 15) suffers total reflection from the prism at R, and enters the







slit S, travelling along ST above the direct rays from the chief light source, which strike the slit normally.
It can be shown from physical considerations* that the greatest width of slit suitable for perfect resolution is given by the expression

where is the angle subtended by the collimating lens at the slit. As is usually of the dimensions of 1/16, it will be

seen that the width of the slit need not be less than about four wave-lengths of the ray under examination - i.e., about 1/500 millimetre for the middle green of the spectrum.
The light from the slit travels to the collimating lens - a lens fixed at a distance from the slit equal to its focal length, so that the rays emerge parallel from it, and in the same phase. It is obvious that this lens should be achromatic, since with an uncorrected lens, such as is sometimes fitted in inferior instruments, the focal length varies for light of different wave-lengths, and the telescope would have to be adjusted for each individual line. The collimator C (Fig. I6)

* Schuster, 'Theory of Optics,' p. 145.






is fixed rigidly to the stand, and the telescope T, the lenses of which must equally be achromatic, is movable round the divided circle D; the telescope is always a radius of the circle whose centre is the centre of the prism table P. The telescope, when in the approximate position required, is clamped to the divided circle with a clamp M, fitted with a fine adjustment of the tangent form. The prism is clamped by some suitable arrangement to the prism table P, which

should itself be capable of rotation, and provided with some means of measuring any angle through which it is turned. A vernier V, travelling with the telescope round the divided circle, enables accurate angular measurements to be taken. The circle is usually divided in half degrees, the vernier reading either to one minute or to thirty seconds of are. In some very massive instruments the telescope is supported b\- an arm rotating round the pillar, as indicated by the dotted lines. Useful elaborations are levelling screws to both the feet





of the spectroscope stand and to the prism table, as seen in the instrument shown in Fig. 17; a reading microscope should be fitted to an arm overriding the vernier. A spectroscope of the construction thus briefly described is suitable for all ordinary qualitative work, but many refinements are possible and desirable for experimental work and research. A particularly sound form of instrument is one in which the telescope is independently supported, as seen in Fig. 18 The collimator is supported by an iron pillar

independent of the divided circle, and the telescope is counterpoised by an iron weight, and revolves with the divided circle about the central pillar. The telescope has a clamp and tangent screw form of fine adjustment, and the divided circle is protected by an upper disc, with two fixed apertures fitted with verniers at opposite extremes of a diameter, through which the angular deviations are read off, and their mean taken. In a well-known, somewhat similar, pattern of instrument, shown in Fig. 19, the prism table revolves, with the protective disc and verniers, so that the one divided circle answers the double purpose. The selection of lenses and prisms depends in some measure upon the nature of the work to be carried out.




The visible spectrum is bounded on the violet side by the ultra-violet region, which is invisible, but which may be examined by photographic means, or with a fluorescent eye-piece (q.v.), and on the red side by the infra-red region, which is also invisible, but may be examined to some extent by photographic means, and still further by the bolometer. If much work is to be done in the ultra-violet, lenses and a prism of quartz must be available, since flint and crown glass absorb the rays of shorter wave-length; quartz lenses

can usually be obtained in mounts of the standard thread used for the ordinary achromatics. Rock-salt lenses and prisms are usually employed for work in the infra-red. It is an advantage to have more than one flint prism, as one made of sufficiently dense glass to give a really great dispersion in the red does not transmit enough light for convenient examination in the extreme violet, whereas a prism giving a brilliant violet-blue spectrum often gives insufficient dispersion in the red; hence a light and an extra dense flint prism will be found very useful. The greater the refractive index, the greater the dispersion - i.e., the angle between two rays of fixed wave-length; hence, where only





one prism is available, as dense a glass should be chosen as possible, so long as it is not of too yellow a tint to interfere seriously with the passage of the violet rays. Two conditions for efficiency are that the collimating lens must be completely filled with light (which is the case when the slit is a distance from its equal to its focal length), and that the depth of the prism face must be sufficiently great to take up the whole beam of parallel rays emerging from the collimator. This means that if the collimating lens have a clear aperture of 1 1/4 inches, the prism face should be at least 1 1/4 inches deep, and about 1 3/4 inches in length, if the usual 60° prism be employed.
The eye-piece of the telescope deserves some consideration at this point. The crossed spider-webs, which are used for obtaining the position of the line measured, will not be visible in the case of a dark spectrum, containing only a few scattered bright lines. Some method of illuminating the threads is therefore necessary. The Gauss eye-piece provides this; it is of the Ramsden construction, containing a plane mirror inclined at about 45° to the optic axis, and placed between the two component lenses of the eye-piece proper. Light enters through an aperture in the side of the tube, and is reflected from the mirror on to the spider-webs and renders them visible.
The shutter eye-piece, made by Adam Hilger, Ltd., is a useful pattern; this has a bright pointer in the field, which is illuminated by means of a small mirror mounted above the eye-piece tube, the light reflected from the mirror passing through a small aperture in the tube and so on to the pointer; there are also two sliding shutters in the focal plane, so that the part of the spectrum viewed may be narrowed down to any degree. Any bright lines which might obscure neighbouring feeble lines can be thereby shut off, The pointer itself is adjustable laterally.




The most common form of prism spectrometer is fitted with one prism, each angle of which is 60°, and the top and bottom of the prism and one face have their surfaces ground. Greater dispersion can be obtained by the use of two prisms, provided the total path traversed be greater than in the case of one; but for general analytical work one prism, of as large a size as is appropriate for the lenses of the instrument, is preferable. For greater accuracy the position of the prism is sometimes adjusted every time the angular deviation of a ray is measured, so that it is in the position of minimum deviation for that ray; it will be seen later that this condition is realized automatically in the case of the Hilger constant deviation spectroscope.
In order to read the angular deviations more accurately, a good reader or microscope is necessary for viewing the vernier. Where a small magnifying-glass is provided, as is the case with some of the cheaper instruments, it will be invaluable to have a small microscope, attached to an arm, fitted. The more elaborate instruments are fitted with microscopes of moderate power to each of the two verniers provided for obtaining the mean reading of angular deviation. The double reading helps also to check any irregularity in the dividing of the circle. A miniature electric lamp run from dry cells or an accumulator will be found very useful for illuminating the verniers.
Two instruments deserving of special notice are those recently produced by J. J. Griffin and Sons, Ltd.* The smaller of these has a 5-inch divided circle protected by a brass disc, to which are attached two verniers at the ends of one diameter; the circle is divided into half-degrees, and the verniers read to one minute. Both collimator and telescope have rack and pinion focussing, the prism table and telescope have fine adjustments, and the telescope

*Kemble Street, Kingsway, London.





is counterpoised. (This spectrometer has been already referred to; see Fig. I9)
A larger instrument, seen in Fig. 20, is of more massive construction, and the prism table may be rotated by means of the large milled head underneath the tripod. The slit has an adjustment for setting the jaws parallel, and is fitted with a comparison prism. The clearance between collimator and telescope is sufficient to admit of polarizer and analyzer

being fitted to the objectives. The latter are of 1-inch aperture and 9 inches focal length.
A more massive instrument, designed for the most refined research work, is shown in Fig. 21. This is constructed by the Societe Genevoise d' Instruments de Physique.* The objectives are of 33.5 millimetres clear aperture and 325 millimetres focal length, and the large graduated circle has a diameter of 263 millimetres. Two large reading microscopes are provided, as shown in the illustration; and by taking the mean of the two readings it is claimed that

* 95, Queen Victoria Street, London.











the systematic errors in the measurement of angles are less than 2 seconds of arc. With one flint prism of 65 millimetres and of refractive index 1.7774, the resolving power

power is 7,620; and the dispersion 13° 45'. A still larger instrument is made by this firm, giving a resolving power of




53,800; in this spectrometer four reading microscopes are provided.
For analytical and laboratory work a fairly massive spectroscope, with 1 1/4-inch lenses and a 2-inch prism of flint, of refractive index about u=1.65 to 1.75, having the telescope either counterbalanced or substantially supported, will be found generally efficient. Though the prism table should be capable of rotation, it will be found to suffice for general qualitative work to have the prism fixed in the position of minimum deviation for some particular line in the green about intermediate between the least and most refrangible rays measured. Many workers have the prism at minimum deviation for the D line.
An instrument specially suited to industrial and experimental work is known as the constant deviation spectrometer. In this instrument (made by Adam Hilger, Ltd.) the collimator and telescope are fixed permanently at right angles (Fig. 22), and it is therefore a fixed-arm type of spectroscope. The prism is rotated in order to pass through the spectrum, and the advantage of this method is that each ray passes through the prism in the angle of minimum deviation.
The prism, though actually made in one piece, may be considered as made up of three prisms, as seen in Fig. 23. A ray of light AB from the collimator meets the first prism, the angles of which are 30°, 60°, 90°, in the minimum deviation position, so that it is refracted along BC, parallel to the imaginary base. It enters the next prism (a 45°, 45°, 90°





prism) normal to the face, and thus passes undeviated along BC, meeting the next face at an angle of 45°, where it is reflected, therefore, at an angle or 45°, passing normally along CD through the third face and entering the next prism, the angles of which are 30°, 60°, 90°, parallel to its base, whence it emerges along DE, again in the position of minimum deviation, and enters the telescope.
The instrument is provided with a divided drum, which,when turned, rotates the prism table. The table is actually

rotated by means of a fine steel screw, the point of which presses against a projecting arm fixed to it; the drum is fixed to the screw. The drum is so made that a steel indicator, sliding in a helical groove cut in it, runs along its length as it is revolved, and a line on the indicator shows the wave-length at a glance, the wave-lengths being engraved on the drum. This will be seen readily in Fig. 24. The focussing of the telescope is done by turning a milled ring which actuates a helical thread, thus doing away with rack and pinion movements. A choice of two prisms is




offered, one of n = 1.65 and another of n = 1.74 ; and while the greater dispersion of the latter admits of increased accuracy of calibration, it only reduces the range from 3,850 A.U. - 8,000 A.U. to 3,900 A.U. - 8,000 A.U. A very neat tripod instrument of the auto-collimating type (q.v.), mounted like a microscope, and fitted with the direct reading drum, has recently been devised by Hilger, and, being a good deal less costly, will doubtless find favour as a laboratory instrument for qualitative work.
The constant deviation spectroscope requires careful adjustment when it is set up, and care should then be taken not to interfere with the prism. The instruments are sent out with lines scratched on the prism table to indicate just where the prism is to be placed. Having put it in this position, a known line should be examined, and the divided drum turned until the pointer indicates the correct wave-length of this line. The line will probably lie to one side of the cross-wires (or pointer); the prism is then gently tapped until the line comes into the exact centre of the field; it is then tested against another known line, and the reading on the drum checked. It is usually only a matter of a few minutes to set the prism so that two different lines give the correct reading, when the scale will be found to be accurate throughout the entire range.
A similar instrument is obtained fitted with a film replica of a Rowland's metal diffraction grating, mounted on a right-angle prism, from the hypotenuse of which the light is totally reflected, so that one can pass through the spectrum by a similar rotation of the prism table, the collimator and telescope being again fixed permanently at right angles. The accuracy of this instrument is about two, to two and a half, times that of the prism spectrometer, and readings may be made to an accuracy of about one Anstrom unit. A direct reading spectrometer comprising many interest-.





ing features has been recently introduced by Bellingham and Stanley, Ltd., in which the use of lenses is obviated, and the prism is used always in the position of minimum deviation. The optical system may be seen from a reference to Fig. 25. Light from the source passes through a slit S, and is reflected by means of a small right-angle prism, the position

of which can be adjusted, upon the surface of a concave glass mirror M. This directs a parallel beam of light on a 30° prism P, with silvered back. The ray enters the prism and is reflected back to the concave mirror, which brings it to a focus in a plane through the slit E. The prism is mounted on a rotating table, which is operated by an arm and micro-

meter screw, whereby the wave-length of light seen in the slit E can be determined. The prism is thus used always in the position of minimum deviation, and there are no errors due to chromatic differences of focus. The instrument is of high precision, and incidentally makes an extremely convenient form of monochromatic illuminator.




Auto-collimating Instruments. - In this type of spectroscope the telescope and collimator are one - i.e., a slit is placed at the eye-piece end of the telescope, and the rays travel down to a prism and are reflected back, passing up the tube again to the eye-piece. The arrangement is seen in Fig. 26. The light enters a slit in the side of the telescope tube, and is reflected by a small right-angle prism down the tube, where it emerges from the lens L upon the prism face, the

prism having an angle of 30°, and being set at the angle of minimum deviation for a ray of given wave-length. It therefore meets normally the opposite face AB, which is silvered, and thus it is reflected back along its original path, and is seen in the upper half of the eye-piece; as only light of one wave-length can meet the prism at minimum deviation a spectrum is produced, and is seen in the eye- piece. In the auto-collimating spectrometer made by Zeiss (Fig. 27) the prism P is of a compound type; F represents





the slit, adjusted by the screw S, and T is the divided circle. The instrument is seen in Fig. 28 fitted with camera attachment, where, as will be seen, an ordinary prism has been substituted for the compound prism, and a light metal camera, counterpoised by a weight, takes the place of the ordinary telescope. A simple and inexpensive

auto-collimating goniometer is seen in Fig. 29. In order to find the angle of a prism, the prism table is turned until the reflection of the slit from one fate is seen coincident with the pointer in the eye-piece. The angleof the prism table is then read. The prism table is now turned until the slit is again seen reflected from the opposite face, and the new angle read on tile divided circle. The angle between the two faces is given by-.
If a combination of two equal prisms be taken, one in the opposite position to the other as regards their bases




(Fig. 30), a given ray of light will eventually emerge from the combination unaltered in direction, its path being displaced but parallel to the original path. But if each prism be made of glass of different refractive index, CD will no longer emerge parallel to AB, except in the case of a ray

of light of one wave-length, for which n is the same in each prism. By combining, therefore, two prisms of suitably chosen refractive index, the light from a slit opposite the first face of the combination will be split up into a spectrum on emerging from the other face; in this way a direct vision

spectroscope may be constructed. Three or five prisms are usually employed, the combination being referred to as the prism train; the prisms are usually of crown and flint glass, and of such refractive indices that one ray is undeviated by its passage through the train, all others being, of course, deviated. In Fig. 31, S is the slit of the spectroscope, the





width usually being variable by means of a collar, C a collimating lens, P the combination of prisms, and E a mngnifying eye-piece. A sliding tube, in which the eye-piece and prism train are mounted, is provided, so that the spectrum

can be distinctly focussed; the distance between the slit and the prisms is generally fixed.
If we assume the refracting angles of the prisms to be very small, then, considering a ray at minimum deviation, we have

where d is the angle of minimum deviation for the ray and n the refractive index of the prism. For two different rays of wave-length m and n, we should have

for the mean ray.
The dispersive power is therefore expressed by -

The angles of the prisms could be chosen such that the dispersion produced by each was the same, but if
the angles be chosen such that the deviation of a given mean ray produced by each is the same, then the dispersion of the prisms is different; the mean ray would be undeviated, but





the rays on either side of it would be deviated, hence a spectrum would be produced at direct vision.
The following are some dispersive powers compared:


The direct vision spectroscope is more or less of an instrument for rough qualitative work, but some large

patterns are made, with which approximate wave-length determinations can be made. One useful little instrument is made by Zeiss, in which a scale of wave-lengths is seen across the spectrum; a small instrument fitted with a photographic scale and rotating mirror, giving the scale in juxtaposition with the spectrum, is made by Hilger, and so on. The scales, of course, only divide the spectrum into com- paratively large areas, but the positions of absorption bands, and bright lines in flame spectra and so on, can be determined with fair accuracy.
Another method of measurement is by means of a travelling slit, actuated by a micrometer screw, so that different parts





of the spectrum can be brought into the field. Such an instrument is made by Hilger, a finely divided drum-head being provided with micrometer screw motion (Fig. 32), and a fixed pointer in the eye-piece, these arrangements 'enabling wave-length determinations to be made with considerable accuracy.' The instrument is seen mounted on a laboratory stand in Fig. 33. Small direct vision spectroscopes, made with a diffraction grating replica mounted on a narrow

angle prism, give still better resolution, and are made by several makers.
We have also to consider the photographic scale method of wavelength determination as used in many patterns of prism spectrometers. The principle is seen in Fig. 34. Here C is the collimator and T the telescope; S is a third tube, illuminated by some light source at L, and possessing a photographed scale, either of wave-lengths or some arbitrary units, an image of which is cast upon the prism face AB,




and thence reflected into the telescope; consequently one sees in the eye-piece the spectrum itself, with the scale overlapping it or in juxtaposition, and by means of the scale a direct reading can be taken of a bright line or absorption hand. If an arbitrary scale is used, the numbers on the scale must be referred to another scale, whereby the wavelength can be estimated and thus no trouble is saved over using a divided circle the readings on which can be referred to a standard scale (q.v.). Where a photographed scale of

wave-lengths is used, some fairly precise means of adjustment should be provided, so that if, for instance, the prism is set at minimum deviation for the D line, the scale may be shifted into such position that 5,893 A.U. is indicated on it, when the whole scale will be in correct position.
The Comparison Spectroscope.-This is an instrument designed for roughly comparing the absorption spectra of two liquids or coloured translucent objects, and for colour matching work. A very convenient form is made by Zeiss, the principle of which is seen in Fig. 35. Here F is the stage of the instrument, with two apertures, G1 and G2 illuminated from beneath by light reflected by mirrors; cells containing the liquids to be compared may be placed over the apertures, and the rays from each coloured source are reflected by the






prisms R1 and R2 into the slit S. The spectroscope is of the direct vision type, C being the eye-piece, and the two spectra are seen in juxtaposition, whilst a photographic scale of wave-lengths is thrown on to the spectra by means of the side tube D, which is illuminated from a separate source.

A similar instrument is also made in which three spectra can be compared. A noteworthy accessory is a cell provided with a micrometer arrangement for varying the depth of the coloured liquid, which is very useful when comparing absorption spectra; the movement reads to .05 millimetre, so that very small differences in the width of liquid traversed by the rays from the light sources can be arranged.




The preliminary adjustment of the instrument is carried out as follows: By turning the milled ring B in Fig. 35 the lens O is focussed until the bright lines in a known spectrum are seen quite distinctly. The width of the slit is regulated by the screw A. The wave-length scale is then brought into perfect adjustment with the lines, and is then locked in position. The two prisms, R1 and R2, are provided with lenses, L1 and L2 the focal length of which (measured in glass) is equal to the distance traversed by the rays reflected to the slit, so that parallel rays illuminate the slit.
The comparison spectroscope is not only a useful instrument for the testing of coloured liquids, light filters, colored gelatin, and so on, but it may also be used for obtaining the approximate limits of absorption bands, etc., by illuminating the one half of the slit with a source of light, such as an electric spark or arc, giving several lines of known wavelength, the other half of the slit being, of course, illuminated by white light passing through the coloured specimen. The scale need not then be used.
The majority of prism spectrometers may be used with a plane diffraction grating as the dispersing medium. A plane grating ruled by a diamond is a costly thing, but the celluloid replicas that can be cast from a ruled grating are so perfect that they may be used for exact wavelength determinations. These replicas are usually cemented to a piece of optically worked flat glass, or to plane or concave speculum metal where reflection spectra are used. They can be purchased for from fifteen shillings upwards, and give very great resolving power.
The resolving power is equal to /whereandare in the wave-length of two radiations which can just be detected as separate. A grating replica, with about 15,000 lines to the inch, will give a resolving power of at least





18,750, while, as we know, a resolving power of only 982 is required to separate the D yellow lines of sodium.
The dispersion produced by the grating is such, as we have seen, that the light is diffracted proportionally to its wavelength, and hence a normal spectrum is produced. The orange and red portions of the spectrum are thus very much better proportioned than in the prismatic spectrum, whilst the blue-violet region is more cramped, but thereby more brilliant. To put it briefly, the diffraction grating admits of greater accuracy in wave-length determination, while the far greater brilliance of the prismatic spectrum

renders it of greater use for the identification of very faint lines.
In order to use a grating replica a holder is necessary, which must be substituted for the prism clamp ordinarily used on the prism table. Such a holder is shown in Fig. 36, where the grating replica is held against a vertical frame by spring clips. The base of the holder is usually fastened to the prism table by means of two upright pillars, which pass through the base, and screw into threaded holes made in the table. Mountings can be obtained which are fitted with three levelling screws, and these can be stood on the prism table.
The plane of the grating should pass through the centre




of the prism table, and it should, of course, be possible to rotate the latter, so that the grating may be used either normal to the axis of the collimator, or inclined at an angle. When normal, and using the first order spectrum, the wavelength of a line is measured by the expression


and b being known, the angular deviationof the telescope readily furnishes the necessary details for the determination. Having read off the angular deviation, the telescope should be turned to the other side of the axis of the collimator, when the same line will be seen in the first order spectrum; if this angle be ', then by taking the average reading

greater accuracy will be obtained, and incidentally any error due to the dividing of the circle will be minimized. When the grating is inclined at an angle, and used in the normal position, the wave-lengthis given by the expression

whereis the angle made by the plane of the grating with the axis of the collimator, as shown on p. 10.
When examining spectra on each side of the normal, it will often be found that one spectrum is brighter than the other; it also happens sometimes that one particular colour or line even will be disproportionately bright. These faults are due to defects or lack of symmetry in the ruling of the grating, which are, of course, reproduced in the replicas.
Very much higher dispersive power may be obtained by means of the Michelson echelon diffraction grating (Fig. 37); with the ordinary grating the first order spectrum is much brighter than that of the second order, and the brightness rapidly diminishes with increasing orders; in the echelon grating, however, the light is concentrated into the higher





orders. An auxiliary spectroscope is necessary to analyze the light, and the echelon grating is placed between a collimator and telescope, the slit of the collimator being placed in the focal plane of the telescope of the auxiliary spectroscope, so that it is illuminated by one spectrum line. The value of the echelon lies chiefly in such investigations as those of the Zeeman effect (q.v.), and of the minute structure or variation of spectrum lines; for details the reader is referred to the literature of Adam Hilger, Ltd., Baly's 'Spectroscopy,' Watts' Study of Spectrum Analysis,' etc.

Prism and grating replica can be used together, the angle of the prism being so arranged that the spectrum is seen at direct vision, and by means of this combination considerable dispersion is obtained. Many useful little direct vision instruments are made on these lines by various makers, and cost very little, although their performance is quite remarkable.
A grating replica mounted on a piece of plane glass may also be cemented with Canada balsam to one face of a 60° prism (the glass of the replica to the prism face), and the dispersion thus obtained will be very great, the brightness of the spectrum being midway between that of a prism and a grating. To take an actual example, a 60° prism was used on a spectrometer, and the angle between the extreme visible violet and extreme visible red was 4° 43'; on cement-




ing a grating to the prism and measuring the same angle it was found to be 19°15', thus giving nearly four times the dispersion.
When a prism grating is used, it is, of course, necessary to plot a wavelength scale for the spectroscope just as if a prism were used.
In order to make wave-length determinations with a plane grating, the number of rulings or lines - actually the combined width of the ruling and the space between it and the next ruling - to the centimetre must be known, or, if not known, it must be ascertained.
Suppose the grating to have 14,450 lines to the inch (the number of lines per inch is usually stated on the replica), then, since

1,000 mm. = 39.37043 inches,


1 inch = 1/(.03937043 mm.);


therefore b = 1/(.03937043 X 14450 mm.),


or b = .0017577 mm.

Hence, if the angular deviation for a certain line were found to be 19°18' with the plane grating normal to the collimator, the wave-length would be found from


The general data for the method will be found in the next chapter.