NA TURE 



,?'7 



T 



THURSDAY, JANUARY 31, 1901. 



THE SCIENCE OF SPECTRUM ANALYSIS. 

 Handbuch der Spectroscopie. By H. Kayser. Professor 



of Physics at the University of Bonn. Vol. i. Pp. 



xxiv + 782. 251 figures. (Leipzig: Hirzel, 1900.) 

 HERE are comparatively few men of science 

 who can accurately handle a spectroscope 

 and interpret its indications with assurance. The 

 number of chemists, for instance, who could look at the 

 spectrum of a Geissler tube, and pick out at once the 

 lines of hydrogen, oxygen, nitrogen or carbon, is 

 probably very small. No one denies the importance of 

 the spectroscopic method, but its practice requires so* 

 long an apprenticeship and so severe a training, while 

 the experimental facts are so numerous and the pit-fdlls 

 so plentiful, that the physicists and chemists are inclined 

 to shirk the whole subject and to leave it to the few who 

 happen to have been brought up in a spectroscopic 

 atmosphere. 



Part of the cause of this apparent neglect is due to the, 

 want of a proper guide to lead the willing but bewildered 

 student through the intricacies of a most diffuse and 

 uninviting literature. We possess only a few short"text- 

 books which are quite insufficient for any serious require- 

 ments, and various catalogues of papers relating to 

 spectrum analysis which have proved absolutely useless. 

 Prof. H. Kayser, well known as an authority on the sub- 

 ject, has undertaken what must prove to be the work of a 

 lifetime. The first volume of his "Treatise of Spectro- 

 scopy" is now completed and will be welcomed by all 

 who desire to know, as well as by those who already 

 know, something of this branch of science. 



This volume covers 750 pages and deals, after an his- 

 torical introduction, with the instrumental methods of 

 producing and examining spectra. There can be only 

 one opinion on the admirable manner in which Prof. 

 Kayser has accomplished his task. He has succeeded 

 in giving a clear and complete account of his subjebt, 

 and at the same time avoided overburdening his book 

 \ with details, which the reader can always find in the 

 original papers, to which complete references are given. 



The first 120 pages are devoted to the history of the 

 subject, which is dealt with in a fair and impartial spirit. 

 The early papers, in which ideas, now so obvious to us, 

 are present in a vague and intangible form, ai-e fully 

 dealt with, but we naturally turn to the exciting time 

 when Kirchhoff and Bunsen finally disposed of all vague- 

 ness and created the science of spectrum analysis. 

 Questions of priority never remain long in an acute stage, 

 and no one would now detract one tittle from Kirchhoff 

 and Bunsen's merit because others may have had some 

 correct ideas before them. Balfour Stewart came very 

 near the truth, but it is very doubtful whether, even if his 

 treatment of the relation between absorption and emission 

 had been as rigid and conclusive as that of Kirchhoff, 

 he would have carried the scientific world with him in 

 the way the Heidelberg philosophers did. In fact, only a 

 small fraction of the chemists and physicists who hailed 

 the new discovery with delight could possibly have 

 appreciated KirchhofTs mathematical deductions. Even 

 KO, 1631, VOL. 63] 



making full allowance for the fact that most men are 

 more easily convinced by an argument which is entirely 

 beyond their comprehension than by one which they 

 partially understand, I cannot believe that the turning 

 point inthehistory of spectrum analysis lay in Kirchhoff s 

 theoretical proof of the cause of the reversal of the bright 

 lines. The most interesting portion of the history of 

 science lies, to my mind, not so much in studymg the 

 evolution of clear ideas from vague forebodings of truth 

 (though that, no doubt, is of great importance), as in 

 tracing the particular theoretical argument or experi- 

 mental fact which carried conviction. In this respect, I 

 should give the foremost place in the history of spectrum 

 analysis to KirchhofiPs experiment, in which he actually 

 obtained the reversal of the sodium and lithium lines, 

 and I should give almost equal value to the clear insight 

 and experimental skill which allowed Kirchhoff" and 

 Bunsen to assign the D lines with certainty to sodium 

 alone For the ubiquitousness of these lines was one of 

 the great stumbling blocks which had prevented every 

 real advance, by suggesting that different elements might 

 emit the same vibrations. Even those who had recog- 

 nised that the yellow lines owed their origin to the 

 presence of a sodium salt had failed to realise that the 

 salt itself was decomposed, and that the lines were due 

 to the metallic element. 



There is an interesting incident connected with this 

 point which may be mentioned here, though private- 

 conversations, unconfirmed by documentary evidence, 

 have no real value in questions of history. The late Prof. 

 Balfour Stewart assigned his own failure to carry his 

 researches to their logical conclusion to his ignorance of 

 the fact that salt was decomposed in the flame. He 

 made an experiment to see whether rock-salt exercised a 

 selective absorption for light emitted by a sodium flame, 

 and failing to discover such an absorption put the matter 

 aside. But I have been carried away by old recollec- 

 tions, and must pass on from Prof Kayser's first chapter, 

 which carries the history of the subject to Zeeman's dis- 

 covery, and the Baltimore experiments on the influence 

 of pressure. 



The second chapter deals with the methods of pro- 

 ducing luminous vapours. Flames, the voltaic arc, elec- 

 tric sparks in various forms and conditions, and vacuum 

 tubes are discussed in succession ; and even those 

 conversant with the subject will find a large amount of 

 valuable information, especially as the author includes in 

 the discussion such questions as the temperatures of 

 different sources, and touches on the theory of the 

 electric discharge. 



The third chapter, dealing with prisms, has been written 

 by Dr. H. Konen of Bonn. The passage of rays through 

 prisms is traced, and full justice is done to Lord Riiy- 

 leigh's investigations, though two propositions in §§ 309 

 and 310, assigned to Wadsworth, are really contained in 

 Rayleigh's first paper. I think that the investigations of 

 this chapter might have been made clearer and shorter 

 by a more frequent application of Fermat's principle. 

 Special attention may be drawn to the reduction of pris- 

 matic measurements to wave-lengths by means of the 

 interpolation formulae, which have been given by Cornu 

 and Hartmann (§§ 327 and 328). Insufficient attention, 

 to which I must plead guilty myself, has been given 



P 



