MABcn 1, 1897.] 



KNOWLEDGE. 



77 



sepulchre from which it was taken. The cycle is now 

 completed, and the now gorgeously clad tiger beetle soon 

 reaches the surface ; and after carefully pluming those most 

 important organs the antennje, as well as every other part 

 of its body, it, for the first time, stretches its wings and 



flies away. 



« 



THE CHEMISTRY OF THE STARS. 



By A. Fowler, F.R.A.S. 



ALTHOUGH actual specimens of celestial bodies 

 cannot be subjected to the ordinary processes of 

 chemical analysis, something can be learned by 

 the aid of the spectroscope of the chemical com- 

 position of those bodies which are self-luminous. 

 Quite naturally spectrum analysis was first employed in 

 the investigation of sunlight — or, rather, it was the inves- 

 tigation of sunlight which pointed the way to the beautiful 

 method of analysis which has evolved step by step from 

 the experiments which Kepler and Newton made with a 

 simple prism. 



The discovery by Wollaston of the dark lines which 

 break the continuity of the gloriously coloured band 

 forming the solar spectrum, their subsequent mapping by 

 Fraunhofer. and their interpretation by Kirchhoti', are 

 points which need not now be dwelt upon, except to remark 

 that the matching of the dark Fraunhofer lines with the 

 bright ones seen in laboratory experiments, when the 

 vapours of various substances are rendered incandescent, 

 is a perfectly definite indication of the presence of corre- 

 sponding vapours in the sun's atmosphere. Applying this 

 method of comparing solar and terrestrial lines, the 

 labours of Kirchhofl', Angstrom, Thalen, Lockyer, and 

 others have demonstrated that most of the familiar 

 metals enter into the composition of the sun's absorbing 

 atmosphere. 



The same principles hold good when the spectroscope 

 is applied to the investigation of the chemical constitution 

 of those stars which, like the sun, give spectra characterized 

 by lines of absorption ; while for stars which give out a 

 bright line spectrum, the investigation by comparison 

 spectra is still more direct. 



Fraunhofer led the way in applying the spectroscope to 

 the stars, and although the real significance of what he 

 saw was unknown to him, he clearly recognized that some 

 stars in a general way resembled the sun, while others 

 were vastly different. His earliest observations were 

 made with a theodolite, having a small prism placed in 

 front of the object glass ; but he afterwards obtained a 

 four-inch telescope provided with a prism of the same 

 aperture. 



It was not until the announcement of Kirchhoff's inter- 

 pretation of the Fraunhofer lines that this pioneer work on 

 the stars was duly appreciated. Rulherfurd, Secchi, 

 Huggins, Vogel, and others, successively took up the 

 inquiry and greatly extended our knowledge of stellar 

 spectra, distinct types being clearly recognized. Owing to 

 the dirticulty in procuring large prisms, Fraunhofer's 

 method of observation was not generally followed — Secchi 

 alone utilizing it — and most of the observers named em- 

 ployed some form of spectroscope attached to the eye end 

 of a telescope, in such a way that an image of the star 

 under observation was projected upon the slit. Dr. Huggins 

 was conspicuously successful in his attempts to secure 

 photographs of stellar spectra, and one magnificent out- 

 come of his work was the discovery of a long series of 

 lines, produced by hydrogen, extending far beyond the violet 

 end of the visible spectrum. 



About ten years ago, Prof. Pickering, of Harvard College 



Observatory, recognized the great value of Fraunhofer's 

 star spectroscope for the photographic registration of 

 stellar spectra, and it is to his revival of this method that 

 we owe many of the greatest advances of recent years. 

 When employed for photography, the prism, or train of 

 prisms, is fixed in front of the object glass with its re- 

 fracting edge parallel to the celestial equator, so that the 

 res'Jting spectrum lies along a meridian, and the necessary 

 breadth of spectrum is then obtained by allowing the 

 driving clock to be slightly in error so that the spectrum 

 trails parallel to its own length. Needless to say, the rate 

 of this trail must be carefully regulated in accordance with 

 the brightness and declination of the star. A prism em- 

 ployed in this way is usually designated an " objective 

 prism," and the whole instrument, when used for photo- 

 graphy, has received the name of " prismatic camera." 



One of the instruments employed by Prof. Pickering is 

 capable of recording the spectra of stars as faint as the 

 eighth magnitude, and although the spectra are on a small 

 scale they sufliciently indicate the characteristic features. 

 The immense advantage of such an instrument in a general 

 survey of the stars is its capacity for registering a large 

 number of spectra with a single exposure, as many as two 

 hundred appearing on one plate. With instruments of 

 greater dispersive power only the brighter stars can be 

 photographed, and it usually happens that only one star 

 at a time can be investigated ; such instruments are 

 therefore specially applicable to a minute analysis of typical 

 spectra. 



Prof. Norman Lockyer soon followed in the adoption 

 of this method of work at South Kensington, and he has 

 obtained spectra of great beauty and value with a six-inch 

 telescope provided with a prism of forty-five degrees. Some 

 of these spectra he has very kindly permitted us to use 

 for the illustration of the present article, and this 

 opportunity of inspecting them will doubtless be generally 

 appreciated. 



The plate which we reproduce shows the violet end of 

 the spectra of six of the brighter stars — Bellatrix, a. Cygni, 

 y Cygni, p Arietis, Procyon, and Arcturus — and the spectra 

 are so arranged that lines common to them fall nearly in a 

 vertical line ; lines of hydrogen being marked H. It will 

 at once be noted that some of the stars show only a few 

 prominent lines, while in others strong lines cross the 

 spectrum at short intervals from end to end. But among 

 the spectra with numerous Unes, as well as in the case 

 of those with few lines, great differences are apparent. 



The mere taking of a spectrum photograph, however good, 

 does not advance matters very much. In the first place it 

 becomes necessary to inquire into the chemical substances 

 represented by the various lines, and the most obvious 

 method is to directly compare the spectra of the stars with 

 those of known substances, exactly as in the case of the 

 sun. This method was, in fact, the first employed — by 

 Dr. Huggins — but it is extremely laborious, and to take a 

 photographic comparison with every known substance 

 would be still more so. With the objective prism, indeed, 

 such comparisons are almost impossible. Fortunately, the 

 repetition of certain groups or series of lines renders such 

 an undertaking unnecessary. To take a definite case ; 

 hydrogen lines having once been identified in a star by 

 a direct comparison with the hydrogen spectrum, the 

 recognition of the lines of this gas in a photograph enables 

 the "wave-lengths" of all the other lines to be determined, 

 and it only remains to compare these figures with those 

 tabulated for the spectra of different substances. 



Matters are much simplified by the fact that there are 

 many stars which give spectra practically identical with that 

 of the sun. Such a star is Arcturus ; and a spectrum of this 



