The Application Of The Spectroscope To The Stars And Nebulae

( Originally Published Early 1900's )

THE subject covered by the above heading has of late years grown to be a very large one, and it will only be possible to exhibit here a bare outline. The spectroscope was first applied to the stars by Fraunhofer about 1814. His apparatus consisted only of a small prism placed in front of the object glass of a telescope belonging to a theodolite. The intervention of the prism changed the image of the star from the bright point which it showed when viewed by the telescope alone, into a narrow, bright line which exhibited all the colours of the rainbow in their customary order, from red atone end to blue at the other. The formation of the spectrum, as this many-coloured line is called, is easily understood. The light from a star consists, not of rays of one colour alone, but of rays of an infinite number of different colours. These, in the ordinary course of things, follow the same path, enter the telescope together, come practically to the same focus, or nearly so, and produce a single and colourless image of the star, because the combination of all the different col-ours yields the sensation which we term white light. But when the light of a star passes through a prism it becomes bent out of its course, and the several different colours are each differently affected, some being more bent from the original straight line than others. Each separate coloured ray then produces a separate coloured image of the star, and these images no longer converge together to the same point, but fall into position side by side, overlapping each other.

Fraunhofer found, however, in the case of the stars, as he had previously found on examining the light of the sun, that the spectra of the stars were not quite complete, and instead of the coloured line being absolutely continuous from the red end to the violet end, it was interrupted here and there by narrow dark spaces. These spaces, in the case of the planets Mars and Venus, corresponded precisely with those spaces which he had already detected in the spectrum of the sun, and this was natural, since the planets only reflect to us the light which they receive from the sun. But the gaps or dark lines in the spectra of different stars were not precisely identical with those to be traced in the solar spectrum, and, moreover, the spectra of different stars were different.

This was an important discovery, for it proved that the source and cause of these dark lines depended on the sun or on the various stars themselves, as the case might be ; and was not due to anything in our atmosphere, or in celestial space, for in such case all the lines would have been alike. Certain particular lines were indeed traced to our atmosphere as they were invariably seen in the spectrum of any celestial body when it was near the horizon, and was therefore being viewed through a great thickness of terrestrial atmosphere.

Fraunhofer did not arrive at any explanation of the cause of these lines, and a generation passed away before Kirchhoff, in 1859, proved that a number of the solar lines were due to the presence in the sun's atmosphere of the glowing vapours of various metals, of which sodium and iron seemed to be the chief.

The presence of a pair of bright lines in the orange-yellow portion of the spectrum of a candle flame had long been noticed. It had been proved that these were due to sodium, and it had been shown that they corresponded precisely in position to a pair of dark lines known as the D lines in the spectrum of the sun. Kirchhoff succeeded in showing that a glowing gas which, at a given temperature, gives off light of a particular tint (or rather of a particular wave-length) possesses also at that temperature the power of absorbing light of that same wave-length. The surface of the sun (the "photosphere," as it is technically called) emits light of every colour, but superposed on it are the luminous vapours of various metals. These vapours, could we but see them alone, would give us only light of certain particular colours their spectra would be spectra of bright lines. But, looking through them at the solar photosphere (which lies below, these gases shut off from us light emanating from the photosphere of precisely the same quality as they themselves emit. We find, therefore, the solar spectrum crossed by dark lines, which correspond td the bright lines of the gases of the solar atmosphere. The conclusion of the whole matter is that whilst the two D lines show the presence of sodium, other lines, known as C, F, G1, and h show the presence of hydrogen ; whilst iron, magnesium, and other elements have also been severally detected in turn.

The same principle has now been applied to the spectra of stars. In their case, as in the case of the spectrum of the sun, the bright background of the continuous spectrum shows the presence of a stellar photosphere, the dark lines crossing it the presence of particular gases in the stellar atmosphere. But the work of identifying these gases in connection with the stars was one of far greater difficulty than it had been in the case of the sun, owing to the light even of the brightest stars being comparatively so feeble. This task was, however, undertaken by Huggins and Miller with the utmost skill and patience, and hydrogen, sodium, magnesium, iron, calcium and other elements which had been previously detected in the sun were shown to exist in the atmospheres of Arcturus, Aldebaran, and several other stars.

For such researches as those of Huggins and Miller the object-glass prism of Fraunhofer was quite unsuited, and a slit spectroscope was adopted. In this a very narrow slit occupies the focus of the telescope, so that the image formed by the telescope falls upon it. The slit is also in the focus of a small object-glass placed behind it, called the collimator, which renders the rays of light coming from the star parallel to each other. The rays then pass through one or more prisms and so become dispersed, the differently coloured rays undergoing a different amount of bending out of their course. Finally the spectrum thus produced is viewed by means of a small telescope. As the normal image of a star is only a point the resulting spectrum is only a line, and a small breadth has to be imparted to it by means of a cylindrical lens before it can be successfully observed.

A labour of a different character was being undertaken by Secchi at about the same time that Huggins and Miller were at work. This distinguished Italian physicist found that though the spectra of different stars differed in character, these differences might easily be reduced to no more than 3 or 4 simple types. Rutherfurd had made a similar suggestion a little earlier, but Secchi was the first to carry out a systematic spectroscopic examination of any considerable number of stars. More recently, other and more detailed classifications have been proposed by Vogel and by Lockyer and--as regards the photographs of stellar spectra by Pickering, but these have in no way superseded Secchi's scheme of classes ; they have supplemented it rather than replaced it.

Secchi divided the stars into 4 principal groups, which he designated " Types " :—(I.) The white or bluish stars, of which Sirius may be taken as the type. These stars yield spectra with the lines of hydrogen very broad and dark, but the lines of the metals faint and difficult to see, or altogether absent. (II.) The yellow stars, of which our Sun, Arcturus, and Capella may be taken as the chief types. The spectra of these show the lines of hydrogen, but not so broadly or prominently as in the case of the Ist type; the metallic lines are, however, on the other hand, numerous and distinct. (III.) The orange stars, of which a Orionis, a Herculis, and the variable star Mira Ceti are types. This class also includes divers variable stars of long or irregular period. The spectra are crossed by a number of dark bands, very dark and sharp on the side nearest the blue, and shading off gradually towards the red end. (IV.) The red stars, none of which are brighter than 5th magnitude. These have spectra crossed principally by 3 dark bands, due to the absorption of carbon, and shaded the reverse way to those of the IIIrd type.

A number of small stars, distributed along the axis of the Milky Way, and commonly called the " Wolf-Rayet" stars, from the two French astronomers who found the first examples, are now considered, in accordance with a suggestion of Pickering's, to form, together with the planetary nebula, a Vth general type. These show very characteristic spectra, the background being of irregular brightness and crossed by two bright lines in the yellow, by another in the light green, and by a distinctive bright band in the blue.

There are also a few stars which can scarcely be brought under any of the foregoing five heads. For in-stance, many of the stars in Orion have the hydrogen as well as the metallic lines narrow and faint ; they can therefore hardly be placed under either the Ist or IInd types. And it may be added that y Cassiopei, B Lyrae, and a few other stars show the hydrogen lines bright.

Secchi's catalogue contained about 500 stellar spectra, but this number has been very largely increased by Vogel, who has informed us concerning the spectra of about 4000 stars ; whilst Konkoly has dealt with about 2000 stars. All the foregoing were the result of direct eye observation, but a fuller survey has since been accomplished by means of photography. Huggins at an early period applied photography to the study of stellar spectra, and discovered thereby a remarkable series of broad, dark lines in the ultra-violet region of spectra of stars of the Sirius type. Dr. Henry Draper worked on similar lines at about the same time, and after his death his widow placed ample funds at the disposal of the Harvard College observatory for further researches to be carried on in memory of her late husband. One of the results of her generosity, and of Pickering's skilful use of it, is the " Draper Catalogue, a classified catalogue of the photographed spectra of more than 10,000 stars. The classification adopted is somewhat more detailed than Secchi's, but proceeds on essentially the same lines.

In a previous chapter (XII.) I have said a good deal about that remarkable class of objects commonly called the temporary stars, or Nava stars which have suddenly come into view and have then rapidly faded away. Only a few instances have occurred since the application of the spectroscope to stellar observation, and the stars have all been much less bright and enduring than Tycho's famous star of 1572, but striking characteristics have been exhibited by each of those which have been spectroscopically treated.

The spectrum of T Coronae in 1866 showed, besides a continuous spectrum crossed by dark lines, a number of bright lines, amongst which those of hydrogen were clearly to be noticed. In Nova Cygni in 1876, again, a number of bright lines were seen superposed on a continuous spectrum. These bright lines appeared on the whole to correspond to those of the solar chromosphere (the narrow red fringe seen surrounding the sun's disc during a total solar eclipse). The hydrogen lines, and a characteristic line in the yellow, near the D lines of sodium, and called D8 (or the " Helium " line), were the most conspicuous. It must be noted in this connection that the hydrogen lines with the D8 line are also the chief lines exhibited by the "red flames," or "prominences," which are often seen to rise from the solar chromosphere to heights of 100,000 miles or more. It follows from this, therefore, that T Corona and Nova Cygni seemed to offer evidence that stars are not only sometimes composed of the same elements as the sun, and, like it, possess photospheres surrounded by absorbing gases, but also that they possess chromospheres and prominences, so that, in point of fact, the sudden development of brilliancy recorded in the case of these 2 stars was really in the nature of a prodigious chromospheric outburst.

Nova Cygni, however, underwent further changes. When its continuous spectrum had nearly faded out the aspect of the spectrum that remained greatly resembled that of the Wolf-Rayet stars. Later still, in the autumn of 1877, the light of the star appeared concentrated in a single bright line, apparently the line characteristic of the nebulae.

Near the centre of the great nebula in Andromeda a new star became visible in August, 1885. Its spectrum was practically continuous.

Two other Novice have yet to be mentioned, Nova Auriga and Nova Normae, the last named apparently a faint copy of the first. Nova Auriga stands out as perhaps the most interesting and most perplexing object yet studied by aid of the spectroscope. Discovered by Dr. Thomas Anderson on January 24, 1892, but recorded by the automatic stellar camera of Harvard College on December io, 1891, it showed, when subjected to spectroscopic analysis, the twofold spectrum seen in T Corona and Nova Cygni, a continuous spectrum crossed by dark lines, and a spectrum of bright lines, amongst which those of hydrogen were conspicuous, together with many of the principal lines of the solar chromosphere.

The star diminished in brightness very quickly after March 16, 1892, and was unfavourably placed for some months. When it was examined afresh on August 17 by the Lick observers, it was found to have undergone a partial revival, and, as in the case of Nova Cygni, they thought its spectrum closely resembled that of a planetary nebula. Huggins, however, did not regard this conclusion as fairly established. The spectrum showed, it is true, two bright bands near the positions of the two chief nebular lines, but the bands were really groups of bright lines, extending over a considerable length of the spectrum. The most striking feature of the spectrum of Nova Aurigae was the displacement of its lines. As first seen, the bright hydrogen lines were accompanied by dark absorption lines, manifestly due to the same element, but displaced towards the violet as compared with the bright lines. Photographs of the spectrum revealed further details. Many of the dark lines carried a fine bright line upon them ; many of the bright lines could be resolved into two or three components. Here, then, there was at least a double hydrogen spectrum : one of dark lines, the other of bright lines, the two displaced with regard to each other. Possibly there were several such distinct spectra. How were their displacement with regard to each other to be explained ?

Doppler, in 1843, had shown that the motion of a source of light towards the observer must cause a shortening of the intervals between the waves of light. In other words, light of a given special wave-length would have that wave-length diminished, and the light would appear to have shifted its place in the spectrum towards the blue end if the source of the light were in motion towards us. If we adopt this explanation of the composite spectrum of Nova Aurigae it follows that that star must have consisted of two or more bodies moving in different directions in the line of sight with the most amazing velocity. The body giving the dark absorption lines would appear to have been approaching our system at a speed of 400 or 500 miles a second, and the body giving the bright lines to have been receding at a speed of about 300 miles a second.

This is scarcely the place to bring forward in detail theories to explain these complicated spectra. The two most favoured are the " tidal theory," which supposes that the near approach of two great stars to each other has given rise to immense tidal waves of highly heated gas, and the "cosmical cloud theory," according to which these Navae are due to the rush of a swiftly moving star through a nebula.

Doppler's principle (as has already been briefly mentioned in a previous chapter) had been applied to a different problem by Huggins in 1867, who carefully compared the position of the green line of hydrogen as given by a vacuum tube, with that of the same line in the spectrum of Sirius. Later on he examined the spectra of a number of stars, and calculated from the amounts and direction of the displacement of the lines in their spectra, the speed at which the separate stars were moving towards us, or away from us in the line of sight. This research was then taken up at Greenwich and at Rugby, but with insufficient means. Lastly, Vogel pressed photography into the service, and made some very successful observations on about 50 stars.

One result of Vogel's work was the discovery of spectroscopic double stars." The variable star, Algol, had long been suspected to have a dark companion. which by transiting before its primary caused a partial eclipse every 69 hours. Vogel now conclusively showed that this was the case, for Algol was moving round the centre of gravity of the pair in precisely the time required, and the diameter, mass, distance from its primary, and speed in its orbit, of the unseen companion, were all computed.

Spica Virginis proved to be another close double, though in this case the companion does not obscure the bright principal star. Indeed, it is possible that it is as bright as the 3rd magnitude.

In some cases a " spectroscopic double " is composed of two stars of nearly equal brightness. This is the case with C Ursa Majoris and ß Aurigae, which were discovered by Pickering a little before Vogel's proof of the existence of the companion of Algol. The two stars which make up ß Aurigae revolve in an orbit which is but little inclined to the line of sight. Consequently at one time one star will be approaching us in its orbit whilst the other is re-ceding. The lines due to the first star are displaced towards the blue, and those of the second towards the red, and the lines in the compound spectrum are therefore double. A little later both bodies are moving across the line of sight, and therefore are neither approaching us nor receding from us, so that the lines of the two stars exactly coincide. The period in the case of this star is nearly 4 days.

Another probable " spectroscopic double " is the variable star ß Lyrte. This star (as we have already seen) goes through its changes in a little less than 13 days, having two maxima and two minima. Its spectrum shows broad, dark bands, due to hydrogen, besides bright lines, which change their appearance and position from time to time. It has been suggested that the system consists of two stars of unlike spectra revolving round each other, and partially eclipsing each other as they cross the line of sight. The changes of the spectrum are, however, very complicated, and have not yet been completely studied, and so simple an explanation appears scarcely adequate.

A very promising and important study is that of the distribution of the different types of stellar spectra. For this the available material is as yet insufficient. Nevertheless, the Draper catalogue, and the catalogues of Vogel and Konkoly, have enabled some first approximations to be made. It appears, from a consideration of such binary stars as have been spectroscopically examined, that the Ist or Sirius type of stars are much less dense relatively to their brightness than the Solar stars, or are intrinsically brighter relatively to their density. The IInd type of stars, i. e., the Solar stars, and to a less degree the IIIrd type of stars, appear to be pretty evenly distributed over the sky. The Ist, or Sirius type, shows a distinct disposition to aggregation towards the Milky Way, whilst, as already pointed out, the Wolf-Rayet stars cluster along its axis. The proper motions of the Sirius stars appear to be smaller than those of the Solar stars, which from this and other reasons may be supposed to be on the average nearer to us than the Sirius stars. If the Solar type stars be divided into two classes, according to their greater resemblance to Capella and Arcturus respectively, the former class appears to have a larger average proper motion than the latter, and may therefore be supposed to be the nearer stars. The entire subject, however, needs much fuller investigation before any great weight can be attached to these provisional conclusions. The completion of the Draper catalogue by the publication of the results of the survey of the southern heavens carried out at Arequipa, in Peru, under the direction of the Harvard astronomers, will constitute the next important forward step.

The first observation of the spectrum of a nebula was made by Huggins in August, 1864. The object examined was the small, bright, planetary nebula in the pole of the ecliptic, 37 IV Draconis, to which some allusion has already been made. The first scrutiny revealed the fact that there existed an immense difference between its spectrum and an ordinary stellar spectrum. In place of the usual continuous spectrum only three isolated bright lines were seen a proof of the presence of luminous gas. In other words, the object was a true nebula, that is, a mass of glowing gas, and not a star cluster, seeming to be nebulous only on account of its distance.

Of the three lines, one, the faintest, was evidently due to hydrogen. The other two have not yet been identified, but the brightest is very near one of a pair of green lines in the spectrum of nitrogen, and has hence been sometimes spoken of as the "nitrogen line." Other lines due to hydrogen have since been observed in various nebular spectra, together with the well-known chromospheric line . Ds, A number of other lines have also been detected in the visual spectrum with extreme difficulty by different observers, and many more by means of photography in the violet and ultra-violet regions. The sources of these lines have not yet been ascertained, and in a great number of the fainter spectra the line in the green near the nitrogen pair, which is especially to be regarded as the typical nebular line, is alone visible.

The problem of the motions of the nebulae in the line of sight has been attacked by Keeler at the Lick Observatory. He has measured the displacement of the chief nebular line in the spectra of the nebulae, and has obtained evidence of movements varying from a speed of about 40 miles per second of approach, to about 30 miles per second of recession.

Several of the nebulae, as, for example, the great nebula in Andromeda, show continuous spectra. But many of those that give a spectrum of bright lines give also a faint, continuous spectrum. The great nebula of Orion is one of the latter class, though Huggins considers that the seemingly continuous spectrum is resolvable into lines other nebulae show bright lines only, without any trace of continuous spectrum.

In the case of the great nebula in Orion Huggins has secured some photographs of exceptional interest, which show that the stars of the "trapezium " are not merely apparently in the nebulae, but really so, for a number of bright lines (one in particular, with a wave-length of 3730 "tenth metres ") were observed both in the continuous spectrum of two of the trapezium stars, and in the spectrum of the nebulae in their immediate neighbourhood. A later photograph, taken in 1889, in which the slit of the spectroscope was pointed near to the trapezium, but not actually across it, failed to show the 3730 line, which would thus appear to be typical only of the regions of the nebulae close to the stars. It would seem probable, there fore, that these stars are involved in the nebulae.

The subject of spectroscopic observations of the stars and nebulae is a growing one, but we have yet much more to do before we can much more learn.