IRLF CHART OF SPECTRA C D Eb F SCa, G.Sr 7Ba 8T1 9 In 12. H. MANUALS OP ELEMENTAEY SCIENCE. THK SPECTROSCOPE AND ITS WOEK. BY KICHAKD A. PBOCTOK, \v Author of "Saturn and its System," "The Sun," "The Moon" " The Universe," $c., $c. PUBLISHED UNDER THE DIRECTION OF THE COMMITTEE OF GENERAL LITERATURE AND EDUCATI APPOINTED BY THE SOCIETY FOR PROMOTINO CHRISTIAN KNOWLEDGE. LONDON: SOCIETY FOB, PROMOTING CHRISTIAN KNOWLEDGE, NORTHUMBERLAND AVENUE, CHARING CROSS, W.C.J 43, QUEBN VICTORIA STREET, E-C.; 26, ST. GEORGE'S PLACE, HYDB PARK COUNEE, S.TV. BRIGHTON: 135, NORTH STREET. YOBK: E. & J. B. YOUNG Si, CO. 1885. [All lights reserved.] PEEPACB. I HAVE endeavoured to make this little treatise on Spectroscopic Analysis clear and simple for beginners, but at the same time accurate, and as complete as pos- sible within so limited a space following, in this respect, the plan already adopted in my elementary treatises on Astronomy and Physical Geography. In order to make room for subjects properly belonging to Spectroscopy, several subjects which are very fully treated of in large works on the Spectroscope (and indeed in one not containing more letterpress than the present) have been dealt with very briefly. It seems to me that to mention but one such subject full accounts of the various contrivances for obtaining in- tense heat and light, such as that with which Schellen occupies the first fifty-two pages of his treatise on Spectrum Analysis, are unsuitable even for large works intended for the general student, and are altogether out of place in an elementary treatise, necessarily limited in size. I doubt, indeed, whether anything can be usefully said in textbooks respecting details of the con- struction of instruments which the observer (who alone could profit by such explanations) must possess and employ. A few minutes devoted to the examination of the instrument itself will, in such a case, be of more use than many hours' study of textbook explanations. On the other hand, I have endeavoured to give a full account of all the principles on which the application of Spectroscopy depends, as also of all the chief methods of observation and their results. I am greatly indebted to Mr. J. BROWNING, the emi- nent optician, for the use of many cuts illustrating various forms of Spectroscopes, and their adjuncts. CONTENTS, CHAPTER I. ANALYSIS OF LIGHT THE SOLAR SYSTEM. Spectroscopic Analysis Defined, p. 5 Ref rangibility of Light, ib. Reflection of Light, 6 Refraction of Light, 7 Rays which cannot be Reflected, 8 Experiment, ib. Refraction through a Medium, 9 The Prism, 10 Refraction through a Prism, 11 First Case, ib. Second Case, 12 Third Case, ib. Prismatic Dispersion of Light, 13 The Solar Spectrum, 14 Dispersion of Coloured Light, 15. CHAPTER II. DARK SPACES IN THE SPECTRUM. Wollaston's Observations, p. 16 Shape of Aperture, 17 Fraun- hofer's Lines, 18 Angle of Minimum Deviation, 21 Collimating Lens, 22 The Spectroscope, 23 Spectroscopic Batteries, 24 Direct Vision Prism, 32 Direct Vision Spectroscopes, 33 Compound Prisms, 34 Half Prisms, 35 Measuring the Spec- trum, 36. CHAPTER III. VARIOUS ORDERS OF SPECTRA. Spectra of Solid and Liquid Bodies, p. 39 Heat Rays and Chemical Rays, 40 Spectrum of Glowing Vapours, 43 Bunsen's Burner, 44 The Electric Lamp, ib. Various Vapour Spectra. 46 Various Orders of Vapour Spectra, 48 Spectrum of Reflected Light, 49 Absorption Spectra, 50 The Micro-spectroscope, 51 Absorption Spectra of Fluids, 52 Spectrum of Blood, 53 Spectrum of Aqueous Vapour, 54. CHAPTER IV. INTERPRETATION OF THE SOLAR SPECTRUM. Researches of Kirchhoff and Bunsen, p. 55 KirchhofE Interprets Solar Dark Lines, 57 The Solar Sodium Lines, 58 Iron Lines, 60 Constituent Vapours of the Sun's Atmosphere, 62 Inter- pretation of Various Spectra, 64 Spectroscopic Analysis of the Sun's Surface, 66 Spectrum of Sunspots, 67. IV CONTENTS. CHAPTER V. THE SOLAR PROMINENCES, CORONA, ETC. Spectroscopic Study of Solar Prominences, p. 70 The Prominence Spectrum, 72 The Sierra Spectrum, 73 Huggins's Method of Observing Prominences, 74 Results, 76 The Real Solar Atmosphere, 78 The Solar Corona, 79 Its Spectrum, 80 The Spectrum of the Zodiacal Light, 82. CHAPTER VI. SPECTRA OF THE STARS, MOON, PLANETS, COMETS, ETC. Star Spectra, p. 83 The Star Spectroscope, 84 Measuring Lines in Star Spectra, 87 Spectra of Aldebaran and Betelgeux, 88 Spectrum of Sirius, 89 Various Orders of Star Spectra, ib. Spectra of New Stars, 92 Of Star Clusters and Nebulae, 94 Gaseous Nebulas, 96 Spectra of Comets, 97 Winnecke's Comet, 98 Coggia's Comet, 99 Spectra of Meteors, ib. Spec- trum of the Moon, 102 Of Venus and Mars, ib. Of Jupiter and Saturn, 103 Of Uranus and Neptune, 104. CHAPTER VII. ATMOSPHERIC LINES IN THE SOLAR SPECTRUM. The Spectrum of the Atmosphere,^?. 104 Of Aqueous Vapour, 105 Of Lightning, 106 Of the Aurora Borealis, 107. CHAPTER VIII. MEASURING MOTIONS OP RECESSION AND APPROACH. Waves of Coloured Light, p. 110 Effect of Motion on Wave-length of Light, 112 Recession and approach of Stars, 114 Of Prominence Matter, etc., 117 Movements in the Solar Atmo- sphere, 118 Uprush and Downrush in the Sun's Atmosphere, 120 Solar Explosions, 121 Study of the Sun's Rotation and of the Motions of Planets by the Spectroscopic Method, 125. ADDENDA. Photographs of Spectra, p. 128 Bright line Spectra of Nebulae, ti. Spectrum of Meteorites, ib. The two forma of Comet Spectra, id. THE SPECTROSCOPE AND ITS WOEK, CHAPTER I. ANALYSIS OP LIGHT THE SOLAR SYSTEM. THE spectroscope is an instrument for analysing light. By its means the rays of various colours forming a beam of white light are sifted so as to be separately discernible by their effects ; coloured light, when com- pound, is analysed into such component colours as by their combination make up its observed tint; light really of a single colour is shown to be such, and its true colour exactly determined. But the spectroscope has other and more important uses. For by its means we can ascertain the elementary structure of bodies shining with particular tints simple or compound, and the nature of bodies which, being only partially trans- parent, absorb particular colours ; while the condi- tion of such bodies, as to heat, pressure, motion, &c., can, in certain cases, be determined. In fine, spectro- scopic analysis that is, research carried on with the spectroscope affords a means of solving many ques- tions respecting the structure and condition of bodies terrestrial and celestial, and respecting most delicate problems of chemical and microscopical investigation, which had appeared altogether inscrutable before this method of research had been invented. The property of light on which spectroscopic analysis principally depends is what is called refrangibility, or the quality by which light, when in its course it passes from one transparent medium into another of different 6 KEFLEXION OF LIGHT. structure or density, has the direction of its path more or less altered. Bays of different colour being differ- ently affected when thus deflected, or refracted, as it is called, can be separated from each other by submitting them to suitable processes of refraction. What, then, we have first to do, rightly to understand the work of the spectroscope, is to examine the nature and laws of the refraction of light, and to distinguish refraction from other processes which light may undergo. When light falls on an opaque polished surface, the rays are reflected according to a simple law. Thus, if the ray B I (fig. 1) fall on the polished surface AB at I, it will be reflected on the course I S, such that the angle B I Q, be- tween I B and the perpendicular I Q, is equal to the angle S I Q between IS and the same per- pendicular. So that if A S B B be a semicircle, having I as centre, the arc A S is equal to the arc B B. A fraction of the light falling on A B, however, illuminates this surface and makes it visible, some of the rays being irregularly reflected from the various particles composing the surface. The less polished the surface the greater will be the proportion of rays thus scattered to those regularly reflected. A portion of the light is also absorbed by the sur- face, the proportion of the rays thus absorbed depend- ing on the nature of the surface. The surface also may absorb more rays of some colours than of others, and thus the rays by which it is seen will not be pro- portioned in the same way, as to colour, as those falling Fig. 1 . Illustrating the law of the reflection of light. KEFEACTION OF LIGHT. on the surface. It is this power of absorbing more or less of different coloured rays which gives to different surfaces their various colours. If the substance on which the light falls is not opaque, but more or less transparent, then while a portion of the rays falling on the surface are regularly reflected, another portion scattered, and some also absorbed, a considerable portion are refracted through the transparent substance, passing onwards through it in straight lines as before, but no longer in the direc- tion in which they were before travelling. Thus let the ray R I (fig. 2), passing through one transparent medium, as air, fall upon another transparent but denser medium, as water, at I, and let Q I P be perpen- dicular to the common surface of both media. Then the portion not reflected at I will not travel onwards along the prolongation of R I, Fig. 2. Illustrating the total reflection, of rays at the surface of water. but be bent into such a path as I S, the angle SIP being less than the angle R I Q. If, on the other hand, S I be a ray passing through a transparent medium, and at I falling on the surface of another transparent medium of less density, then the portion not reflected at I will pass into the rarer medium, not on its former course, but on such a path as I R, the angle R I Q being greater than the angle SIP. And although this deviation due to refraction does not follow a constant law for all sub- stances, like the deviation due to reflection, yet there is a constant law for the refraction of light from one to another of two given transparent media. Thus, let the path R I (fig. 2) be in dry air of given density and heat, and I S in distilled water of given heat ; draw a 8 KAYS WHICH CANNOT BE REFRACTED. circle, Q A S, about I as centre, and R M, S N, square to A B ; then whatever the direction of R I, the ratio of N I to M I remains constant, being nearly as three to four for the substances named. For other two sub- stances the ratio would still be constant for all direc- tions of R I, but the ratio would not be the same as for air and water. If a ray proceeding on the path R I in a rarer medium is refracted into the path I S in a denser, then a ray pro- ceeding on the path S I in the denser will be refracted into the path I R in the rarer medium. It follows from this that whereas rays falling on a denser medium in any direction whatever will be partly refracted through it, rays falling on a rare medium, so as to make a very small angle with its surface, will not be refracted. Thus we have seen that if the denser medium APB is of water, then, IN being taken equal three-fourths of I M and N S drawn square to A B, I S is the direc- tion of the refracted ray. And no matter how close to B, R may lie, the refracted ray can never lie closer to I A than Is, such that (drawing sn square to AB) I n three-fourths of I B. Now the ray S I will be refracted in direction I R, and a ray in any direction between PI and si will be refracted into the air between the directions I Q and I B ; but a ray to I from any direction between s I and A I manifestly cannot be refracted according to the law described. It is found that in such cases the light is entirely reflected by the surface A I B, so that a ray as rl is reflected in direc- tion Ir'. For such rays, the surface AB acts the part of a perfect mirror. All these properties can be tested by experiment. Thus, if r S P (fig. 2) be the bottom of a basin of water, a mark at S will be seen by an eye at R, as though the mark were at X> for the rays from S will reach the eye in the direction I R. Again, an eye under water at S will see an object at R in the direction S I. An eye under water at ,r (fig. 3), and directed towards the part of the surface lying beyond I to the DISPERSION OF LIGHT. 9 right in the figure, will not see any object placed as B above that surface, but will see an object placed as r' below the surface. It will be the same if the eye at r is outside a glass vessel containing the water. Thus, if A P B be a glass bowl containing small fish, an eye at r outside the bowl will see in direction r I the reflected image of a fish at r', which fish will also be visible nearly in its true place by rays proceeding directly from r' to r. Accordingly, the fish and its image in the mirror-like surface of the water (viewed from below) will be seen, just as we may see at the same tune a person himself and his image in a mirror. It is clearly seen that the effect of refraction must be to separate rays of different kind which had been travelling in one direction, if such rays are differently affected by refraction. And in order to effect this, we should only have to allow the rays after refraction, as at I (fig. 2), to travel each on its new course far enough to get them as widely separated as we required, were it not for the fact that light is absorbed in passing through glass, water, or whatever other medium we employ; and before a sufficient separation had been effected, no visible light would remain for us to observe. If we take a transparent medium, as glass, bounded by parallel sides, as M M' (fig. 3), an in- cident ray, F B, will be refracted at B in the direction B I, and on reaching I will be again re- fracted in the direc- tion I S parallel to B F ; for since a ray proceeding in 3 .-Illustrating the refraction of light direction I B WOUld through a medium bounded by parallel foUow the course faces " B F, making with the perpendicular B N' the angle F B N', the ray B I which makes with the surface 10 THE PRISM. at i the same angle which I R makes with the sur- face at R, must be refracted into a course I S, making with the perpendicular I N the same angle SIN. Here again, then, we fail to separate rays of light which falling together at R may be differently refracted ; for though during the portion R I of their course they are diverging, they become parallel at I, and separate there- fore no farther from each other. But if the surface at I be not parallel to the surface at R, there will no longer be this parallelism; the separated rays will continue to diverge, and we shall have made one step at least towards the analysis of light. It was such reasoning as this which led Grimaldi, and later, Newton, to employ media bounded by non-parallel planes in their experimental researches on light. We are thus introduced to the use of triangular prisms for the analysis of light. A prism is a wedge- shaped figure (fig. 4). It is defined by Euclid as a ' solid figure bounded by five surfaces, two of which (ABC and a b c) are triangles, equal, similar, and parallel to each other, Fig. 4.-The prisnT. wnile tne otner tnree (A b, b C, and C a) are rectangles. Of these surfaces the triangular need not concern us as surfaces ; they might be incomplete and unpolished, or they might be wholly or partly changed. So long as the rectangular portions are perfect as sur- faces, at least where the light passes through them, the action of the prism will not be impaired. Moreover, in most uses of prisms for analysing light two of the rectangular faces only are employed. The glass drops of a chandelier are convenient instances of prismatic figures. Fig. 5 illustrates the manner in which the prism acts on light which is refracted in passing through it. ABC represents a section of the prism parallel to its tri- PRISMATIC DEVIATION OF LIGHT. 11 rig. 5. Path of a ray through a prism. angular faces. The ray D e falling on A B at e is bent towards the prolongation of the perpendicular fe in direc- tion e h. At h it is again bent from the perpendicular lig in direction TiE. The deviation from D D', the original course of the light rays, is in this case increased at h, the total deflection being the angle be- tween E k, produced, and D D'. Here A B and A C are the two rectangular faces of the prism called into action. The edge A between these faces is called the refracting edge; the angle between them is called the refracting angle ; and the divergence of the ray after emergence at h from the course it had before incidence at e is called the angle of deviation. The first general case of deviation is illustrated in fig. 6. Here the incident ray SI is refracted in direction I E, or from the refracting edge, and in emerg- ing at E is again refracted in direction E R, or from the refracting edge, the angle of deviation being the sum of the two deviations. When the deviations are equal we have the special case pictured Kg- Bm _ The first general 9ase of rcfrac ti on in fig. 5.* In fig. 7 through a prism. * By a singular mistake the cases illustrated in figs. 6 and 7 with the special case pictured in fig. 5 are dealt with in Schellen as the three general cases, the general case illustrated in fig. 8 being entirely omitted. In the English edition the mistake is left un- corrected, and it has been reproduced in elementary treatises on the spectroscope. 12 PRISMATIC DEVIATION OF LIGHT. the second general case is illustrated. Here the inci- dent ray being perpendicular to the surface at I under- goes no refraction, but passes on in the straight line S I E. On emerging at E it undergoes refraction from the refracting edge, and the angle of deviation is the angle between S E produced and E R. Thus we have in this Fig. 7. The second general case of refrac- -, AI,,* l' j tion through a prism. case tne Sam 6 kind of deviation as the former, that is, deviation from the refracting edges both at incidence and emergence, but the deviation produced at one place only, at emergence or at incidence according to the course pursued. The third general case is illustrated in fig. 8, where we see the incident ray S I deflected towards the re- fracting edge at I, but from it when emerging at E, in direction E R, the angle of de- viation being the excess of the second deviation over the first. As the path S I E R may be followed either in the direction S I E R or in the direction RE IS, all possible cases of refraction through a prism are illustrated in the figs. 6, 7 and 8. We see that in all cases deviation is from the refracting edge of the prism; for in the two first there is only Fig. 8. The third general case of refraction through a prism. PRISMATIC DISPERSION OF LIGHT. 13 deviation from the edge ; and though in the third case there is deviation at I towards the edge, it is manifest that the deviation from the edge at E is much greater.* In every case then the action of the prism causes a deviation from the refracting edge. The prism, therefore, possesses the property required for the separation of rays of light, if they are not all refracted in the same way when passing from one medium into another (or, technically, if they are not all of the same refrangibility). For in that case rays from the same source of light, having different directions when they emerge from the prism, will continue to separate more and more from each other. Thus, let AB (fig. 9) represent a beam of sunlight passing Fig. 9. Illustrating the action of a prism on rays of different coloured light. through a circular aperture in a screen S S', P a prism of glass, and at V R let there be a surface on which the rays of light may fall. If the prism P were removed, the rays would pass to i, and form there a small oval image of the circular opening in the screen S S'. The prism being interposed, the rays will be refracted in the * Because the interior angle A E I is less than the exterior angle E I B. It is easily seen from fig. 2, and the matter explaining i% that the greater the angle between S I and the surface the less will be the divergence of the ray at I from its original course S I. 14 THE SOLAB SPECTRUM. manner already described. If they are all of the same kind, they will all be refracted in the same degree, and the beam of light will only be bent, as shown by the bent bright line A B i! in fig. 9, and will form a small oval white image at i'. But if the rays forming the beam are unequally refrangible, some will be more bent, proceed- ing after emergence to V, or I, or B, others less, proceed- ing to R, or 0, or Y ; and the light which otherwise would have been gathered at i' will be spread along some such strip as V G R. Newton found (as Grimaldi had before noticed, however) that this happens when sunlight is refracted by a prism, and that the rays which thus differ in refrangibility differ in colour also, the rays which are most refracted being violet, those least refracted being red, the colours corresponding to inter- mediate degrees of refrangibility being, in order (after violet), indigo, blue, green, yellow, and orange. In other words, this experiment showed that sunlight con- sists of rays of seven different colours violet, indigo, blue, green, yellow, orange, and red these rays being less and less refrangible in this order. The streak of coloured light thus formed by the action of the prism is called the prismatic spectrum ; the violet end of the spectrum contains the most refracted part of the light, while the red end contains the part least refracted; and the rays towards the violet end are commonly called the more refrangible rays, those towards the red end being called the less refrangible rays. This way of speaking should be carefully noted. If rays of any one colour of the seven were all equally refrangible, we should find in the experiment illustrated by fig. 9 that the violet rays would make one oval violet image of the small circular opening in S S', as at V ; the indigo rays would make an indigo image, as at I ; the blue, a blue image, as at B ; the green, yellow, orange, and red forming respectively a green, yellow, orange, and red image, as at G, Y, 0, and R. But Newton did not find that this happened, On the contrary, the spectrum formed an unbroken rainbow- tinted streak DISPERSION OF COLOURED LIGHT. 15 (like a cross strip from a rainbow), fading away into darkness at the red and violet ends. He thus perceived that all violet rays are not equally refrangible, nor all indigo rays, nor are all the rays of any colour whatever equally refrangible. Hence, if a beam of rays of any colour, obtained as in the above experiment, be allowed to fall on another prism, these rays, being like the rays of sunlight, differently refrangible, may be separated. Thus, if the screen at V R (fig. 9) is perforated where the rays of any colour fall, the rays thus allowed to pass through may be received on another prism, and further dispersed ; so that, instead of the image which would have fallen on the second screen but for the second prism, a lengthened image is thrown higher up. But this image shows no new colours. It is entirely red, for instance, if the hole is made in the red part of the spectrum V E. A good eye can indeed recognise a slight variation of tint in the red, which tends towards orange at the upper or most refracted part of the beam. But there is no new tint, only a slower gradation from the full red of the lower part to the slightly orange red of the upper part. The idea occurred to Newton, however, that though the rays of any given colour are not all equally re- frangible, yet the entire range of refrangibilities between the extreme red and the extreme violet may not be represented in the spectrum. There might, for instance, be a definite difference between the most refrangible red rays and the least refrangible orange rays ; and in this case the extension of the process described in the last paragraph would show a gap in the spectrum between the orange and the red. And even if gaps could not be seen, yet there would not be a regular gradation of light from end to end of the spectrum. Finding after varying the experiment in different ways, and especially modifying the form of the aperture, that the spectrum still showed a uniform gradation of tints from end to end, Newton concluded that all orders of refrangibility from that of the extreme violet end to that of the ex- 16 DARK SPACES IN THE SPECTRUM. treme red end, are represented in the solar spectrum. If that were really so, the method of analysis which forms the subject of the present book could have had no existence CHAPTER H. DARK SPACES IN THE SPECTRUM. WOLLASTON was the first who succeeded in showing that the solar spectrum is not continuous, as Newton had inferred from his experiments. It is probable that Newton had looked only for well-marked gaps. None such exist in the spectrum. Wollaston employed a method by which small spaces wanting in the spectrum had a fair chance of being detected; for instead of admitting light through a circular aperture (or a trian- gular, oblong, or other like aperture, such as Newton used in other experiments), Wollaston admitted the light through a very narrow slit parallel to the refracting edge of the prism. It will be well, while considering the effect of this change, to notice a difference in Wollaston' s method of observing the spectrum. Newton's plan of receiving the spectrum on a screen, however con- venient for exhibiting the spectrum, is not convenient for observing it. Wollaston observed the spectrum through the prism. Let A (fig. 10) be a bright point of light ; P a prism, through which the eye E observes the bright object A. If rays of any colour were all of the same refrangi- bility, the eye at E would see a red image at R by rays which had followed the course Arr'E, a violet image at V by rays which had followed the course A v v' E, and intermediate images, (orange), Y (yellow), G (green), B (blue), and I (indigo), by rays which had followed the intermediate paths indicated in the figure. As matters actually are, instead of a series of seven images at SHAPE OF APEBTUKE. 17 B, 0, Y, G, B, I, V, the eye will see a blended series forming the spectrum R V, the violet appearing lowest, Fig. 10. The direct examination of the prismatic spectrum. the red appearing uppermost. In other respects this image corresponds exactly with the image thrown on a screen. Moreover, this image may be viewed, if neces- sary, with a telescope, instead of the naked eye. Next let us consider the effect of the shape of the source of light, whether an aperture or a luminous object. Suppose there were but one image for each of the seven colours of the spectrum. Then the spectrum would be a set of seven coloured images of the source of light, arranged as shown in fig. 11. Here we see seven circles, seven triangles, seven oblongs, and seven lines of light, as the respective spectra of a circular, triangular, oblong, and linear source of light. It is clear that the space between two successive linear images is much greater than the space between two successive circular, or triangular, or oblong images. In other words, there is room for many more intermediate images. Hence if rays of certain degrees of refrangi- 18 FRAUNHOFER S LINES. istratmg the source of light on the character of the spectrum. bility are really wanting, so that instead of the spec- trum being formed by a perfect series of overlapping images it is really Aperture "^^^""^^^9 incomplete in parts, there will be a much better chance of de- tecting this fact if we use a fine lumi- nous line for the source of light, or admit light through a fine slit, than where the source of light has any considerable width in the direction of the spectrum's length. Wollaston tried the experiment. Admitting light through a narrow slit, parallel to the refracting edge of the prism, he observed the spectrum in the manner illustrated in fig. 11. He found that it was not continuous, but crossed by certain dark lines, lying at right angles to its length. In other words, light of certain definite degrees of refrangibility is absent from the solar beam. He did not carry this inquiry further, supposing, doubtless, that the discovery was singular rather than valuable. He had no reason for suspecting that the quality he had detected indicated any property peculiar to sunlight. Still less would he suppose that when this property was traced to its source it would help to reveal the very constitution of the sun's mass. But Fraunhofer, in 1814, examined these dark lines with such care and attention, that in recognition of his labours they have ever since been called Fraunhofer's lines. Using for source of light a much finer slit than Wollaston had employed, and studying the image, formed as in fig. 11, with a telescope instead of the FRAUNHOFER'S LINES. 19 unaided eye, he found many dark spaces where Wollaston had seen but few. He counted and mapped, in fact, no less than 576 dark lines. The chief lines in the solar spectrum are indicated in fig. 12 ; and as reference is continually made to the lines as here lettered, the student should carefully note their position in the spectrum. A is a strong line close to the red end of the spectrum. B is a strong and rather broad line in the red. Between A and B is a band of several lines called a. C is a dark and well marked line. Between B and C, Fraunhofer counted nine fine lines ; between C and D about thirty. D consists of two strong lines close together. Between D and E, Fraunhofer counted eighty-four lines. E is a band of several lines, the middle one of the set being stronger than the rest. At b are three strong lines, Fig. 12. The dark lines in the solar spectrum. the two farthest from E being close together. Between E and b, Fraunhofer counted twenty-four lines, and be- tween b and F more than fifty. The lines F, G, and H are well defined. Between F and G, Fraunhofer counted 185 lines, between G and H, 190 lines, and he found many lines between H and I, the violet end of the spectrum. I remind the reader of the importance of noting the position of the lettered lines in the spectrum, for these lines are constantly employed for reference. Let him remember, then, that A, B, and C are in the red portion of the spectrum ; D is in the orange-yellow ; E in the yellow-green ; F in the green blue ; G in the indigo ; and H in the violet. Now let us recall what these dark lines really are. They are gaps in the spectrum indicating the absence B 2 20 FBAUNHOFEB'S LINES. of rays of certain refrangibilities from the beam of solar light. The spectrum shown in fig. 12 is formed in reality of a series of images of the fine slit through which sunlight is received. The red part of the light, with its various degrees of refrangibility, makes a series of red images, the orange makes a series of orange images, and so on. But the red light not containing all the degrees of refrangibility within its limits, certain red images which should appear are wanting, leaving dark spaces, as at A, a, B, and C ; so certain images are wanting in the orange part of the spectrum, others in the yellow, and so forth. Fraunhofer next inquired whether the dark lines may not be due to peculiarities in the substance forming the prism. He found, however, that they may be seen with prisms of every kind of glass and crystal, as well as with prisms formed by enclosing various fluids in prism-shaped phials. He next examined sunlight reflected in various ways ; as from the moon, the planets, from clouds, the sky, terrestrial substances, and so forth. In every case he found the same lines which he had seen in the spec- trum of direct sunlight. He studied the spectrum of the sun when close to the horizon, and found that the violet end of the spectrum is then very faint, and several new lines are to be found in various parts of the spectrum. He examined next the light from the fixed stars. He found that though each star gives a spec- trum showing the prismatic colours, none of these spectra are exactly like the solar spectrum. Some lines in the solar spectrum are wanting in star spectra, others are less strongly marked ; and some lines are seen in star spectra which are absent from the solar spectrum. No two stars seem to have exactly the same spectrum. He found that when the flame of a candle or lamp is the source of light, the spectrum shows only two dark lines, or rather one double line, in the same place as the double line D of the solar spectrum. Before proceeding to consider the results of further ANGLE OF MINIMUM DEVIATION. 21 research into these dark lines in the solar and other spectra, it will be well to describe here the methods by which the spectrum is increased in length, while its purity is retained. If light were not variously refrangible, according to its colour, but all refracted in the same degree, then when we looked at an object through a prism we should still see the object blurred, though not with the pris- matic colours. For from every point of the object a diverging pencil of rays would travel, and this pencil, after being bent twice in its passage through the prism, would no longer diverge from a single point ; so that instead of seeing for each point of the object a cor- responding point in the image, we should have a small round spot of light, and the image made up of such spots of light would be correspondingly confused. There is only one exception to this rule, viz., when the pencil passes on such a course as D e h E in fig. 5 (p. 11) the two parts D e and /iE being equally in- clined to the faces of the prism at I and E. Then the emergent pencil proceeds from a single point (or very nearly so). The prism is said to be used at the angle of minimum deviation in such a case, because the deviation is less than for any other position. Now when we are examining a spectrum as in fig. 10, it is manifest that though one part may be examined at this angle of minimum deviation (the part Gr in the case there illustrated), and so be seen clearly, all the rest must be viewed at a different angle, and therefore less distinctly. This was not a matter of great importance in Newton's or Wollaston's experiments, where the pencils did not follow a long course through the prisms, and where the image was not very closely examined. But when greater dispersion of the rays is to be obtained by increasing the prismatic action, and where the pris- matic image is to be examined telescopically, it be- comes highly desirable that some method should be devised for giving a spectrum pure throughout, instead of one merely pure in one particular part. This diffi- COLLIMATING LENS. Fig. 13. Illustrating Simms's device for purifying the solar spectrum. culty was met by Simms, the optician, in 1830. He simply got rid of the divergence of the rays of light by placing a converging lens in such a position as to make the rays paral- lel. Thus let P (fig. 13) be the prism, S a point of the source of light, S L a divergent pen- cil of light pro- ceeding from S. Then if a con- vex lens, L L', be so placed that the pencil after passing through the lens consists of parallel rays, these parallel rays, falling at i i', on the surface i i' of the prism, are equally refracted, and therefore continue parallel as they pass to e e', where they are equally refracted at emergence, and thus the emergent beam e e' R consists of parallel rays. This beam will still continue parallel after being re- fracted through a second prism, or through any number of prisms, and the image of S will be truly seen by beams which have gone through one or more prisms. This being appreciably true for all rays of all colours, the whole series of images of S forming the spectrum will be truly formed ; or, in other words, all parts of the spectrum will be seen with equal distinctness.* The lens L L' used for this purpose is called the collimating lens. Fig. 14 shows how a telescope is used in combination * Persons unacquainted with the laws of optics sometimes sup- pose that Newton's discovery of the advantage of using the angle of minimum deviation is important for modern work with the spectro- scope. Thus I have seen it stated in an elementary treatise on the spectroscope that "our spectroscope depends" inter alia on this discovery. But in reality the invention of the collimating lens removes entirely the difficulty which Newton partially met by using the prism at the angle of minimum deviation, and enables the observer to use the prism at other angles with equal effect. THE SPECTROSCOPE, 23 with a collimating lens. The light emitted from the source of light, after passing through the slit seen at the extremity of the left hand tube, is converged into parallelism by the lens of this tube, and after passing through the prism, falls on the object glass of the tele- scope on the right, and is acted upon precisely as the rays from a distant object are acted on by the object- glass and eye-piece of a telescope. Comparing fig. 13 with figs. 10 and 14, it will be noticed that the use of a collimating lens and observing telescope are simply Fig. 14. A spectroscope of one prism with collimating tube and observing telescope. devices for first purifying the spectrum R V of fig. 10, and then observing it with suitable telescopic power. When the rays have been made parallel by means of a collimating lens, any number of prisms may be used to increase the dispersion of the rays. The effect of several prisms in increasing the dispersion is illustrated in fig. 15, where light received through the slit S S' is supposed to be carried round a battery of four prisms, SPECTROSCOPIO BATTERY. Fig. 15. Showing how several prisms may be used to increase the dispersion. with increase of dispersion at each, until finally it forms a spectrum on the screen a b. In* the figure only seven spectral images of the slit are shown V, the violet image, by rays which have been most refracted; R, the red image, by rays which have been least refract- ed, and the rest in intermediate posi- tions. It will be seen that such a way of using many prisms would in some sense corre- spond with the effect of removing the screen further away from the source of light in Newton's original experiment. Although no screen is used by spectroscopists, but the emergent rays received into a telescope, as shown in fig. 14, it will be con- venient in what follows to refer to a spectrum supposed to be received on a screen, as in fig. 15. It is easy to see why with increase of dispersion the dark lines are more clearly shown. If the increase of dispersion merely made the spectrum longer, without modifying its nature, very little would be gained. But it is easy to see that more than this is done. Let it be remembered that in the spectrum we have a series of images of the slit, and that our chance of detecting the absence of certain rays from the solar beam depends on the recognition of a space in the spectrum where no image is formed. Now let a b c (fig. 16, 1.) represent a small part of the spectrum as it would be if the slit, instead of having definite though small breadth, were an actual line. Then the light corresponding to the two edges of the dark space b forms in the real spec- PUEITY OF THE SPECTRUM. 25 tram two images of the slit, as at B, where these images are of such a width as to touch, leaving no dark space between them. T TI But now suppose we ni increase the dispersive A jj| B c ^ power, and get a spec- trum of greater length, U , . as at II. Then the parts I ^ abc of the true spec- ** b ate irnm fhat i3 nn ah'. The figure illustrating the subject in the above-mention u work is also incorrect. RECESSION OF SIRIUS. 115 next is shown the broad F line in the spectrum of Sirius, displaced towards the red. Next is the F line in the solar spectrum. Below that again is the spectrum of hydrogen at atmospheric pressure. It is important to notice these two lower spectra. They indicate what had been observed by Huggins, in confirmation of observations previously made by Pliicker and Hittorf, respecting the widening of this line of hydrogen under increase of pressure, viz., that the line widens symmetrically on either side of its position as a thin line. Hence the displacement of the F line of Sirius can be explained only by the motion of Sirius ; and as the displacement is towards the red, the star is receding from the earth. The amount of the observed dis- placement indicated a motion of recession at the rate of about forty-one and a half miles per second. But as the earth had at the time a motion of recession from Sirius, due to her own motion in her orbit, at the rate of about twelve miles per second, there remained a balance of almost twenty-nine and a half miles of recessional motion per second. The sun himself is in motion towards the constellation Hercules, though his rate of motion is uncertain. I place no reliance myself on Otto Struve's estimate of the rate of this motion, having been able to show that the assumptions on which his reasoning is based are not sound. According to his estimate the sun's motion from Sirius would be rather more than three miles per second, leaving about twenty- six miles per second for the actual recession of Sirius from the part of space through which we are travelling. I believe that a much larger proportion of the relative recessional motion of Sirius is due to our sun's motion. Be this as it may, however, there can be no doubt that the distance between Sirius and the sun is increasing at the rate of nearly thirty miles per second; for Huggins's result has since been confirmed by himself with better instrumental means, and also by Mr. Christie at the Greenwich Observatory. Subsequent observations by Mr. Huggins have H 2 116 EECESSION AND APPROACH OF STARS, supplied the following indications of stellar motions of recession and approach : STARS RECEDING FROM THE SUN. STAR x . Apparent Motion. Earth's Motion. Motion from Sun. Sirius 26 to 35 10 to 15 18 to 22 Betelgeux 37 15 22 Kigel 30 15 15 Castor 40 to 45 17 23 to 28 Regulus 30 to 35 18 12 to 17 ft Ursae Majoria -\ e * y ( J 60 9 to 13 17 to 21 STARS APPROACHING THE SUN. STAR. Apparent Motion. Earth's Motion. Motion towards Sun. Arcturus 50 4- 5 55 Veea ... 40 to 50 _i_3-9 44 to 54 30 -1- 9 39 Pollux 32 4-17 49 a Ursse Majoris 35 to 50 4-11 46 to 61 Several of these results have been tested and confirmed by observations made at Greenwich. The method thus employed upon the stars was * The community of motion of these five stars was a discovery of special interest to me, as I had predicted it more than a year before it was made. My study of the stellar proper motions having indicated the existence of drifting motions among the stars that is, of cases where groups of stars are travelling onwards together through space I selected a remarkable case, that of the five stars above bracketed, as one which the spectroscopic method might suc- cessfully deal with, all the stars being bright ; and in a lecture delivered before the Koyal Institution in May, 1870, I expressed my confident belief that all these stars would be found to have a common motion of recession or approach. In May, 1870, 1 received a letter from Mr. Huggins, stating that these five stars are all receding at a rate of about seventeen miles per second. PROMINENCE-MATTER, ETC. 117 manifestly applicable to the sun; and so soon as Huggins's method of observing the prominences without an eclipse had shown that the substance of the promi- nences often changes in position at the rate of many miles per second, it was natural that observers should endeavour to ascertain whether the prominence matter gave at times spectroscopic evidence of motions of approach or recession. The application of the spectro- scopic method is simplified in this case by the fact that the lines for comparison are visible at the same time as the lines affected by the motion. Thus let SS' (fig. 65) repre- sent a portion of a " ' < prominence PP' under SpectroSCOpic exami- Kg- 65. Change in the position of promi- ~A 1^4- 1? u nenceF-line due to motion of prominence- nation, and let I be matter. the position of the solar dark F-line, while the bright line shown above is the prominent F-line displaced towards the violet. Thus this displacement is recognised. Moreover a higher dispersive power can be used than when the faint spectrum of a star is examined. In the case here illustrated we should know that the promi- nence matter was moving rapidly towards the observer. This would correspond to the rush of the whole of the prominence matter included within the strip ss' in one direction. Ordinarily we should expect, from what has been seen of the changes affecting the shape of a prominence, to find different parts of the prominence line differently affected. Thus we might meet with such varieties of displacement as are in- dicated in fig. 66. If the prominence F line appears as at I, we infer that the hydrogen in the part of the solar prominence (or sierra) under examination (that within the strip s s') is approaching the eye, but more rapidly at the upper part, and not at all at the sun's surface S S'. If the F line appeared as at II, we should form a similar inference as to 118 MOVEMENTS IN THE SOLAR ATMOSPHERE. motion of the prominence matter from the eye. If the F line appeared as at III, we should infer that the upper half of the prominence matter was receding >v IV Fig 66. Spectroscopic indications of rapid movements taking place in the solar atmosphere. bodily, the lower half approaching, but not so swiftly in its lower portions, and not approaching at all close by the sun's surface. Lastly, if the F line appeared as at IV we should infer that the lower half of the promi- nence matter was neither approaching nor receding, while the upper part within the range of view was stirred by both kinds of motion, perhaps the nearer portion receding and the remoter approaching, or the remoter receding and the nearer approaching. (The differ- ences of thickness of the F line would indicate varieties of pressure and temperature, as already shown.) The results of actual observations upon the promi- nence F line can thus be readily interpreted. The bright F line has been seen as in the first picture of fig. 66, so that the hydrogen matter of the sierra must have been rapidly approaching the observer, the deflec- tion being towards the right, i.e., towards the violet end of the spectrum. This line has been seen as in the second picture when the matter of the sierra was rushing swiftly from the observer. And lastly, when the line was seen bent both ways as in the other pictures of fig. 66, part of the hydrogen of the sierra was rushing swiftly towards, and part was rushing swiftly from, the observer, whence we may infer with considerable pro- bability that a solar cyclone was in progress. In some cases when a prominence-line has been under observa- MOVEMENTS IN THE SOLAR ATMOSPHERE. 119 tion, the spectrum showed a portion of the prominence was neither approaching nor receding from the observer, giving the hydrogen line in its normal position, while a portion was rushing rapidly towards the observer, less swiftly low down, but very rapidly higher up. There is one circumstance which renders these in- dications of the approach or recession of prominence or sierra matter somewhat doubtful. Professor Young, of Dartmouth College, has observed that sometimes when the F line of a prominence has shown marked signs of disturbance, the C line has been upright and in its normal position. This is a very perplexing observa- tion, and almost seems to suggest that hydrogen is not in reality an element but compound, the F line belonging to one component, the C line to another, and that the intense heat of the sun separates these two components. There is, however, another and perhaps more satisfactory way of getting over the difficulty, viz., by supposing that under the conditions existing when solar hurricanes are in progress the C line of the hydrogen disappears, so that only the comparatively quiescent portions give both this line and the F line, the disturbed portions giving the latter only. It is obvious that as the bright hydrogen lines from the prominences or sierra may by their displacement in- dicate rapid motions of approach or recession, so also the dark hydrogen lines belonging to the ordinary solar spectrum may indicate such motions. In the former case, the motions of approach or recession indicate horizontal motions with respect to the sun's surface ; in the latter such motions partake more or less of a vertical character, and where the region observed is near the centre of the solar disc they are almost wholly vertical. But as the ordinary tele-spectroscopic study of the prominences indicate motions both horizontal and vertical with respect to the surface of the sun, we may expect to find spectroscopic indications of both kinds of motion. The displacement of dark lines is indicated in fig. 67. 120 CURRENTS IN THE SUN S ATMOSPHERE. The narrow strip of the sun's surface under examina- tion (see fig. 44, p. 67) lay across a sun spot, and we see in the appearance of the F line of hydrogen the signs of disturbance in the spot. At the upper and lower extremities this line is of its normal width ; then come two parts, one at the top, one at the bottom, where the line is bright, showing the intense heat of the hydrogen at the borders of the spot. Over all the rest of the spectrum's breadth the hydrogen Miles per second. 40 From the Observer. 120 Towards Ked. Miles per second. 40. Towards 80 L the Observer. 120 ) Towards Yiolet Fig. 67. Illustrating the spectroscopic evidence of rapid movements in the solar atmosphere. line is much broader, and is also less defined, than usual. In one place we see it bent towards the red end of the spectrum, indicating a rapid motion of reces- sion, that is to say, the swift down-rush of hydrogen as if into the depths of the spot. The rate of this down- rush is indicated by the numbers placed above (the vertical numerals 0-00048505 mm, indicating the wave- length of light at the place of the F line in millimeters, each equal to a twenty-fifth part of an English inch, very nearly). The motions which are thus indicated in the vaporous SOLAR OUTBURST, 121 atmosphere of the sun are so enormous that some evi- dence other than that derived from the displacement of the spectral lines seems required to confirm results so startling. This therefore seems the place to describe the most striking direct evidence yet obtained of rapid motions in the vaporous matter surrounding the sun. On September 7th, 1871, Professor Young (then of Dartmouth College, Hanover, N.H., now of Princeton), observed by the spectroscopic method, described at p. 75, the long low-lying cloud of glowing hydrogen, depicted in fig, 68. " It had remained with very little Fig. 68. Cloud of glowing hydrogen seen by Young, September 7th, 1871. at 12.30. change," he says, " since the preceding noon a long, low, quiet-looking cloud, not very dense or brilliant, nor in any way remarkable except for its size. It was made up mostly of filaments nearly horizontal, and floated above the sierra with its lower surface at a height of some 15,000 miles, but was connected with it, as is usually the case, by three or four vertical columns brighter and more active than the rest." ... In length it measured some 100,000 miles ; in height, that is from sun's surface to the uppermost edge of cloud, about 54,000 miles. " At 12.30," proceeds Young, "when I was called away for a few minutes, there was no indica- tion of what was about to happen, except that one of the 122 SOLAK OUTBURST. connecting stems of the southern extremity of the cloud had grown considerably brighter and was curiously bent to one side; and near the base of another at the northern end a little brilliant lump had developed itself, shaped much like a summer thunder-head." Fig. 68 represents the prominence at this time, a being the little thunder-head. " What was my surprise, then," he continues, " on Fig. 69. The same prominence region at 1.5 p.m. returning in less than half an hour (at 12.55), to find that in the meantime the whole thing had been literally blown to shreds by some inconceivable uprush from beneath. In place of the quiet cloud I had left, the air, if I may use the expression, was filled with flying debris fcOLAR OUTBURST. 123 a mass of detached vertical fusiform filaments," each from 4,500 to 13,000 miles long by 900 or 1,300 miles wide, brighter and closer together where the pillars had formerly stood, and rapidly ascending. " When I first looked, some of them had already reached a height of nearly 100,000 miles, and while I watched them they rose with a motion almost perceptible to the eye, until in ten minutes (5 min. past 1) the uppermost were more than 200,000 miles above the solar surface. This was ascertained by careful measurement." The mean of three closely agreeing determinations gave about 207,000 miles as the extreme height attained by these wondrous filaments or wisps of glowing hydrogen, the least of which had a surface largely exceeding that of the British Isles. The velocity of ascent, more than 100,000 miles in ten minutes, or 166 miles per second, is greater than any velocity hitherto determined by the displacement of lines in the solar spectrum. Fig. 69 represents the appearance of the hydrogen filaments when some of them had attained the greatest observed height. As the filamants rose, they gradually faded away, and at a quarter-past one only a few filmy wisps, with some brighter streamers low down near the sierra, re- mained to mark the place. But in the meanwhile the little thunder-head before alluded to had grown and developed won- derfully, into a mass of rolling and ever- changing flame, to speak according to appearances. First it Fi ?- 70 ~ The same > IAO P- m - was crowded down, as it were, along the solar surface ; later it rose almost pyramidally 50,000 miles in height ; then its summit 124 SOLAR OUTBURST. was drawn out into long filaments and threads which were most curiously rolled backwards and downwards like the volutes of an Ionic capital ; and finally it faded Fig. 71. The same, 1.55 p.m. away, and by half-past two had vanished like the other. Figs. 70 and 71 show it in its full development, the former having been sketched at Ih. 40m. and the latter at Ih. 55m. p.m. The whole prominence suggested most forcibly the idea of an explosion under the great prominence, acting mainly upwards, but also in all directions outwards, and then, after an interval, followed by a corresponding in-rush ; and it seems far from impossible that the mysterious coronal streamers, which have now been proved to be truly solar, may find their origin and explanation in such events. The evidence which this tremendous outburst gave of rapid motions in the solar atmosphere is found to be stronger even, when carefully examined, than it appears on the face of it. I made a calculation at the time when Professor Young's account was first published, as to the circumstances under which matter expelled from the sun would travel upwards to the height observed by Young in this instance taking ten minutes to pass from a height of 100,000 miles to a height of 200,000 miles and I found that this could only happen if the expelled SOLAR OUTBURST. 125 matter were resisted (as we should expect) by the atmo- sphere through which it passed. Expelled with a velocity sufficient just to carry it to a height of 200,000 miles, it would (I found) take much more than ten minutes to traverse the latter half of its upward flight. It must then have been expelled with a much greater velocity, which, being reduced by atmospheric resist- ance, carried it only to the height to which it would have passed with a smaller initial velocity not reduced by such resistance. The mathematical investigation of the circumstances showed that the initial velocity must have exceeded 300 miles per second, and may have exceeded 500 miles per second. Of course we do not know certainly that matter was actually expelled from the sun on that occasion, though it is difficult to suggest any other explanation. The apparent motion of glowing hydrogen may have been due, and indeed very likely was due, not to the actual motion of the hydrogen itself, but to the rush through the hydrogen of a number of missiles (solid or liquid or very dense vapour) flung forth from the sun's interior. These in their swift rush through the hydrogen of the sun's atmosphere would so heat it, by friction, as to cause it to glow with intense brightness. In this case, the glow of the hydrogen filaments would give evidence of the resistance demon- strated by my mathematical investigation of the circum- stances of the motion upwards. The filaments them- selves would resemble the wisps, looking like streaks of floating matter, often seen after the passage of a large meteoric mass through our own atmosphere. Similar remarks apply indeed to other cases of the apparent motion of the solar hydrogen. This may be due to the rapid rush through the hydrogen of matter causing it to glow with intense brightness at the place through which the moving matter is at the moment passing. The whole subject of the spectroscopic evidence of rapid motions of matter towards or from the observer has recently been re-examined, in consequence of doubts 126 MEASURING SUN's ROTATION BY SPECTROSCOPIC METHOD. thrown out by Secchi and others, who have not only failed to obtain such results as have been described above, but appear to question the mathematical prin- ciples on which the method depends. At Greenwich, under the general superintendence of the Astronomer Royal, aided by the great practical experience of Mr. Huggins, this method has been applied afresh to the stars, and also to measure known motions of recession and approach, such as those due to the sun's rotation, to the rotation of Jupiter, to the motion of Venus, and so forth. Secchi had failed to obtain any evidence of the sun's rotation by this method. This is not greatly to be "wondered at, seeing that the sun's equatorial parts travel only at the rate of about one mile per second, which though enormous compared with all the forms of motion with which we are familiar, is insig- nificant compared with those other motions of which this ,method had barely been able to afford evidence. The Greenwich observers, after being long foiled by the difficulty of the task, have at length succeeded in recognising the sun's rotation spin by the spectroscopic method no new discovery, it is true, but a veritable triumph of observational skill. It must not be supposed that they simply succeeded in seeing what they knew they ought to see if the method were sound and their instrumental means sufficient. The observations were so conducted that the observer who had to read the spectroscopic evidence had no knowledge of the direc- tion in which it pointed. For the part of the sun under examination was shifted by an assistant observer from one side, where the rotation was bringing the solar surface towards the observer, to the other, where the rotation was carrying it from him, without any intima- tion of the change. The results were in most satisfactory accordance with theory. Before the Greenwich observers who earliest took up this task had succeeded in it, Professor Young, with the comparatively feeble power of the Dartmouth College telescope to aid him, but employing a spectro- MEASURING SUN'S ROTATION BY SPECTROSCOPIC METHOD. 127 scope of great dispersive power, combining the dispersion of a battery of prisms with that due to the reflexion of light from a closely ruled grating of fine lines,* had succeeded in measuring the rotation of the sun. His results were much more satisfactory, in fact, than those obtained at Greenwich, insomuch that he saw reason for believing that this method may suffice to indicate even the difference of rotation rate between the sun we see and the vapour- strata which produce the absorption lines. Nothing more delicate in spectroscopic research than these observations by Professor Young has probably been ever yet accomplished. The observers at Greenwich have also succeeded in measuring satisfactorily by this method the rotation of Jupiter, and the motion of Venus in the line of sight. The method when applied to the moon gives, as it should, no evidence of recession or of approach ; for the moon has no such motion as this method could measure. The Greenwich observations of the stars indicate motions of recession and approach agreeing fairly with those determined by Huggins ; but as yet much has to be done before any but the most practised spectro- scopists can achieve much in the measurement of stellar motions by this extremely delicate and difficult method of observation. The failure of Secchi and others who have attempted to apply this method must be regarded as due entirely to its difficulty, not, as some among them seem to suppose, to any inherent error in the method itself. Certainly there has been no more striking or more * Dr. Rutherfurd, of New York, has succeeded in ruling glass in this way, forming what are called gratings, so closely, that owing to the property called diffraction (in a manner somewhat too com- plex for explanation in these pages), a spectrum of great purity is formed. A diffraction spectrum thus formed has the advantage over the refraction spectrum of spreading the rays truly according to their wave-length; in fact, it has been by means of such gratings that Angstrom has determined the true wave-lengths of light belonging to different parts of the spectrum. 128 ADDENDA. promising application of spectroscopic analysis than this method of determining motions of approach or recession, though it must be admitted that the method is BO exceedingly delicate that we cannot at present hope to see it applied to measure any motions save those of enormous rapidity. ADDENDA. PHOTOGRAPHS OF STELLAR AND PLANETARY SPECTRA. Dr. Henry Draper, of New York, and Mr. Huggins, in England, have (almost at the same time) succeeded in photographing the spectra of Vega, Altair, and Sirius, among the fixed stars; of Venus, among the planets ; and of the Moon. In the photograph of Vega there are lines (bands or strise) near the violet end of the spectrum. The photographic spectrum of Venus shows a large number of lines. There is in this spectrum a weakening of the light towards H, and above that line, of the same character which Draper has observed to take place in photographs of the solar spec- trum near sunset. BRIGHT LINE SPECTRA OF NEBULA. Mr. Stone, formerly an assistant at Greenwich Observatory, has communicated to the Royal Society a paper questioning the gaseity of those nebulae which give a spectrum of bright lines, but on in- sufficient grounds. He supposes that a remote cluster of suns like our own would show the spectrum of the gaseous envelopes of its component stars. But as I pointed out several years since in a paper on the effect of distance on appearance of star clusters, and as Professor Stokes explained after Mr. Stone's paper had been read, distance would not modify the spectrum of a cluster of luminous bodies. SPECTRUM OF METEORITES. Mr. Wright, of Yale College, U.S., has found that the gases evolved from meteoric masses, give, at moderate temperature, a spectrum similar to that of Win'necke's Comet, 3, fig. 62, p. 98. THE TWO FORMS OF COMET SPECTRA. At Lord Lindsay's Observatory, Dunecht, the spectra of two comets (B & C, 1877) were lately observed, and of these the former gave spectrum 3, fig. 62, p. 98, while the latter gave spectrum 4. The point is interesting, because this is the first observation of the second form of cometic spectra since 1868. THE END. ADDENDA. DISCOVERY OF OXYGEN IN THE SUN. Dr. H. Draper (see p. 128) has detected the bright lines of Oxygen in photographs of the violet and indigo parts of the solar spectrum. He strongly suspects, also, ihe presence of bright lines of Nitrogen. What is said at pp. 63 and 64 will indicate the significance of this important discovery so far as its method is concerned, and will show that the fact discovered accords well with what had been before surmised. The con- cluding sentences of the paragraph on p. 64 illustrate Dr. Draper's inference that the bright background of the solar spectrum is, wholly or partly, made up of the spectra of many gases. A -" NEW STAB " FADING OUT INTO A NEBULA. . The " New Star " described at pp. 92, 93, has gradually faded, the spectrum changing by the gradual disappearance of -the continuous or broad-band portions, until in August, 1877 (as seen by Mr. Backhouse), and early in September (as seen by Dr. Copeland, of Lord Lindsay's observatory), only one bright line remained, which had not been the brightest in the original spectrum. This line seems to be identical with the brightest line of the spectrum of the gaseous nebulae (see pp. 95, 96). The Star has, in fact, faded out into a bluish planetary nebula. Probably the nucleus of suoh a nebula was the part which acquired unusual lustre, and has since returned to its usual condition. The history of the "New Star " of 1876 no longer accords, therefore, with that of the " New Star " of 1866, as stated at p. 93. EEEATUM. ON p. Ill line 11, omit the word " except." PUBLICATIONS OF THE itty for fOTiotiitg Cjjrisfe J SPECIFIC SUBJECTS. Fcap. %vo., 64 pp., Limp Cloth, price ^d. each. ALGEBRA, By W. H. H. HUDSON, M.A., Mathematical Lecturer, and late Fellow of St. John's College, Cambridge. (ANSWERS TO THE EXAMPLES GIVEN IN THE ABOVE, Limp Cloth, &/.) EUCLI D, Books I and 2. Edited by W. H. H. HUDSON, M.A., late Fellow of St. John's College, Cambridge. ELEMENTARY MECHANICS, By W. GARNETT, M.A., Fellow of St. John's College, and Demonstrator of Experimental Physics in the University ot Cambridge. PHYSICAL GEOGRAPHY, By the Rev. T. G. BONNEY, M.A., F.G.S., &c., Fellow, Lecturer in Natural Science, and late Tutor of St. John's College, Cambridge. READING BOOKS. *. d. STANDARD i, containing 192 pp., Pica. i8mo. Cloth o 4 STANDARD 2, 192 pp., Pica. iSmo. Cloth o 4 STANDARD 3, 320 pp., Small Pica. Fcap. 8vo. Cloth o 8 STANDARD 4, 320 pp., Small Pica. Fcap. 8vo. Cloth o 8 STANDARD 5, 256 pp., Long Primer. Post 8vo. Cloth i o STANDARD 6, 256 pp., Long Primer. Post 8vo. Cloth i o PRIMER. An Introduction to Standard i o \ ELEMENTARY GEOGRAPHY. By the Rev. T. D. C. MORSE o 4 SCRIPTURAL READING LESSONS : OLD TESTAMENT. First Book. i8mo....C70//& o 4 N EW TESTAMENT. First Book. 1 8mo. . . . Cloth o 4 ATLASES. A MODERN ATLAS; containing 30 Maps, with Indexes, &c Cloth boards 12 O BIBLE ATLAS (The), containing 12 Maps and Plans, with Explanatory Notes, Complete Index, &c. Royal 4to ! Cloth boards 14 O THE STUDENT'S ATLAS OF ANCIENT AND MODERN GEOGRAPHY, with 48 Maps and a copious consulting Index Cloth boards 7 6 THE WORLD, an ATLAS, containing 34 Coloured Maps and Complete Index. Folded 8vo. Cloth, gilt 5 o THE NATIONAL ATLAS, containing 32 Coloured Maps and Complete Index Cloth 2 6 THE YOUNG SCHOLAR'S ATLAS, containing 24 Coloured Maps, and Index. Imp. 4to Cloth 2 6 THE SHILLING BIBLE ATLAS, containing 16 Coloured Maps Paper wrapper i o THE SHILLING QUARTO ATLAS, containing 24 Coloured Maps Paper wrapper i o THE PREPARATORY ATLAS, comprising 16 Maps, full Coloured. Small 4to Covers o 6 THE THREEPENNY ATLAS, containing 16 C ojoured M aps. Crown 8vo Paper cover o 3 THE PENNY ATLAS, containing 13 Maps. Small 4to Covers o I MAPS, MOUNTED ON CANVAS AND ROLLER, VARNISHED. *. d. EASTERN HEMISPHERE 4ft. loin, by 4ft. 2in. 13 o WESTERN HEMISPHERE ditto. 13 o EUROPE ditto. 13 o ASIA. Scale, 138 miles to an inch ditto. 13 o AFRICA ditto. 13 o NORTH AMERICA. Scale, 97m. to in. ditto. 13 o SOUTH AMERICA. Scale, ditto. ditto. 13 o AUSTRALASIA ditto. 13 o AUSTRALASIA (Diocesan Map) ... ditto. 14 o INDIA. Scale, 40 m. to in. 50 in. by 58 in 13 o NEW ZEALAND 3ft. 6 in. by 2ft. 10 in. 9 o AUSTRALIA ditto. 9 o ENGLAND AND WALES ditto. 9 o IRELAND. Scale, 8 miles to an in., 2 ft. 10 in. by 3 ft. 6 in. 9 o SCOTLAND. Scale, ditto. ditto. 9 o GREAT BRITAIN AND IRELAND, The United Kingdom of ... 6ft. 3 in. by 7ft. 4 in. 42 o ENGLAND AND WALES 4ft. 2 in. by 4ft. 10 in. 13 o ENGLAND & WALES (Diocesan Map) ditto. 16 o BRITISH ISLES 58in. by5oin. 13 o HOLY LAND 4ft. 2 in. by 4ft. 10 in. 13 o HOLY LAND, to Illustrate the Old Testament. Scale, 8 m. to in. , 2 ft. 10 in. by 3 ft. 6 in. 9 o HOLY LAND, to Illustrate the New Testament. Scale, 7 m. to in. ditto. 9 o> SINAI (The Peninsula of), the NEGEB, and LOWER EGYPT. To Illustrate the History of the Patriarchs and the Exodus . . . ditto. 9 o PLACES mentioned in the ACTS and the EPISTLES. Scale, 57 m. to in., 3 ft. 6 in. by 2 ft. 10 in. 9 o SMALL MAPS OF HOLY LAND : To Illustrate Old Testament on sheet I 6 ... millboard, varnished 2 o To Illustrate New Testament on sheet I 6 ...millboard, varnished 2 o Places mentioned in Acts and Epistles on sheet I 6 , millbd, varnshd. 2 o SMALLER SERIES OF SCHOOL MAPS : Eastern Hemisphere 27in. by32in. 6 O Western Hemisphere ditto. 6 o Holy Land, to Illustrate the Old and New Testaments. Scale, 9 miles to an inch , ditto. 6 o MANUALS OF HEALTH, Fcap. 8vo., I28//., Limp Cloth, price is. each. FOOD. By ALBERT J. BERNAYS, Professor of Chemistry at St. Thomas's Hospital. HEALTH AND OCCUPATION. By B. W. RICHARDSON, Esq., F.R.S., M.D. HABITATION IN RELATION TO HEALTH (The). By F. S. B. CHAUMONT, M.D., F.R.S. ON PERSONAL CARE OF HEALTH. By the late E. A. PARKES, M.D., F.R.S. WATER, AIR, AND DISINFECTANTS. By W. NOEL HARTLEY, Esq., King's College. MANUALS OF ELEMENTARY SCIENCE, Fcap. 8vo., 128 pp., with Illustrations, Limp Cloth, is. each. PHYSIOLOGY. By F. LE GROS CLARKE, F.R.S., St Thomas's Hospital. GEOLOGY. By the Rev. T. G. BONNEY, M.A., F.G.S., Fellow and late Tutor of St. John's College, Cambridge. CHEMISTRY. By ALBERT J. BERNAYS. ASTRONOMY. By W. H. CHRISTIE, M.A., Trinity College, Cambridge ; the Royal Observatory, Greenwich. BOTANY. By ROBERT BENTLEY, Professor of Botany in King's College, London. .ZOOLOGY. By ALFRED NEWTON, M.A., F.R.S., Professor of Zoology and Comparative Anatomy in the University of Cambridge. MATTER AND MOTION. By J. CLERK MAXWELL, M.A., Trinity College, Cambridge. SPECTROSCOPE, THE WORK OF THE. By RICHARD A. PROCTOR, Esq. CRYSTALLOGRAPHY. By HENRY PALIN GURNEY, M.A., Clare College, Cambridge. ELECTRICITY. By Professor FLEEMING JENKIN. LONDON : NORTHUMBERLAND AVENUE, CHARING CROSS, W.C. J 43, QUBBJf VICTORIA STREET, K.C. ; 26, ST. GEOEGE*S PLACE, S.W. BRIGHTON : 135, ITOBTH STREET. THIS BOOK IS DUE ON THE LAST DATE STAMPED BELOW AN INITIAL FINE OF 25 CENTS WILL BE ASSESSED FOR FAILURE TO RETURN THIS BOOK ON THE DATE DUE. THE PENALTY WILL INCREASE TO 5O CENTS ON THE FOURTH DAY AND TO $1.OO ON THE SEVENTH DAY OVERDUE. SEP 11 .r'SOB.Z