NA TURE 



265 



THURSDAY, JANUARY 



1903. 



RADIATION AND SPECTROSCOPY. 



Handbuch der Spectroscopic. Vol. ii. Yon H. Kayser. 

 Pp. xi + 696. (Leipzig : Hirzel, 1902.) Price 2/. net. 



THE second volume of this important work follows 

 the first after a remarkably short interval of time. 

 Being essentially a book of reference, the reviewer's task 

 is an easy one, as the value of the work is best indicated 

 by means of a short summary of its contents. Generally 

 speaking, we may say that this volume deals with the theory 

 of molecular radiation and those facts of spectroscopy which 

 throw some light on the theory. The discussion therefore 

 embraces the question of multiple spectra, the observed 

 regularities in the spectra of bodies and the Zeeman effect. 



The first chapter deals with the connection between 

 emission and radiation. The early work of Balfour Stewart 

 and of Kirchhoff, which had already been touched upon in 

 the first volume from the historical side, is now discussed 

 as regards its logical stringency. It is, perhaps, to be 

 regretted that this chapter was written before Lord 

 Rayleigh's defence of Stewart's proof had been published ; 

 but Kayser has added a footnote in which he quotes 

 Rayleigh's opinion, having already given in the text the 

 verbal transcription of the passage in Stewart's writing 

 on which his claim to independent and previous discovery 

 of the fundamental law of spectroscopy rests, so that 

 every reader may form his own judgment. 



The paragraphs dealing with the experimental verifica- 

 tion of the relation between absorption and radiation are 

 of considerable interest, and reveal the great need of 

 further work in this direction. No one doubts the 

 accuracy of the theoretical law in he case for which it 

 applies, which is that of thermic equilibrium. But ex- 

 perimental investigation of the absorption in flames and 

 Geissler tubes is very likely to throw some fresh light on 

 the mechanism of luminescence due to chemical or 

 electric action. The results of Gouy's investigation are 

 sufficiently curious to render their repetition and extension 

 desirable. 



The second chapter, which treats of the radiation of 

 solid bodies, is a most valuable and complete summary of 

 our knowledge of the radiation of hot solids, first as 

 regards their total radiation, leading up to Stefan's law, 

 and secondly as to the partition of the total energy into 

 its elementary portions depending on wave-length. Great 

 progress has recently been made in our knowledge of the 

 emission of black bodies, but the radiative properties of 

 gases, dealt with in the third chapter, present greater 

 difficulties. In the first place, we are not able experi- 

 mentally to render a gas luminous by a purely thermal 

 process, and it is well known that some writers have gone 

 so far as to assert that a purely thermal radiation cannot 

 give a discontinuous spectrum. There is no doubt that 

 such a view would get over some theoretical difficulties! 

 but at present the facts seem against it. At any rate, so 

 far as experiments go, the relation of radiation to absorp- 

 tion has not been found to be materially different in the 

 flame and in the arc than it is in a body which is in 

 thermal equilibrium. We may mention in this connection 

 an observation of Gunther quoted by Kayser (p. 182), in 

 NO. 1734, VOL. 67] 



which sodium rendered luminous in a Bunsen flame 

 shows absorption lines when the light from a platinum 

 wire made incandescent in the same flame is sent 

 through it. 



The question of the mechanics of radiation is at 

 present in a state of transition, the electron theory rapidly 

 gaining ground. A certain inequality of treatment in 

 this respect in different parts of the book could not be 

 avoided, but the modern theory is not neglected, though 

 not pressed forward with that sympathy which some 

 might perhaps have wished. 



One small matter I should like to set right. Prof. 

 Kayser quotes me as supporting E. Wiedemann's 

 calorimetric measurement of the heat necessary to dis- 

 sociate the molecule of hydrogen into its constituent 

 atoms. I was indeed surprised to find, on looking up 

 the reference given by Kayser, how strongly I had ex- 

 pressed myself in this respect, a few months after 

 Wiedemann's paper had appeared. I have long since 

 become convinced that the experiments were inconclusive, 

 and I therefore quite agree with Kayser's own views on 

 the subject. 



The questions relating to the variability of spectra, in- 

 cluding the effects of pressure, temperature and mode of 

 incandescence, are fully discussed in the fourth and fifth 

 chapters. 



Our knowledge of the influence of the various forms 

 of electric discharge and the different modes of pro- 

 ducing incandescence is gradually becoming more com- 

 plete, but there is still considerable difference of opinion 

 as to the interpretation of the facts. The additional lines 

 which, e.g., are introduced into the spectra of metals 

 when a Leyden is introduced into the discharge of an 

 induction coil may be interpreted as due to an increase 

 of temperature merely, or as due to some peculiarity of 

 the form of discharge. 



The general opinion of spectroscopists, towards which 

 Prof. Kayser seems also to lean, is that there are effects 

 which may be peculiar to the method of discharge and 

 cannot be explained by mere changes of temperature ; 

 but, on the other hand, all classifications of star spectra 

 are more or less based on the supposition that the tempera- 

 ture is the chief, if not the only, cause of the differences 

 observed. Little has been done of late years to obtain 

 a direct answer to the question whether a spectrum is 

 completely defined by temperature and pressure without 

 reference to electrical effects. If we limit ourselves to 

 temperatures not higher than that of the electric arc or 

 the oscillatory discharge obtained by means of a con- 

 denser and self-induction, experimental evidence seems 

 to show that the method of producing incandescence is 

 immaterial. At .any rate, identical spectra may be 

 obtained (a) in an electric arc produced by a constant 

 current, (6) by an oscillatory discharge from metallic 

 poles, (c) in the cone of a Bunsen burner. At present 

 there seems, therefore, no reason to suppose that high 

 temperature is not in itself sufficient to produce all the 

 effects observed when incandescence is produced by a 

 high-tension spark. 



Some of the older researches, which were carried out 

 at temperatures sufficiently low to admit of measurement 

 or approximate estimation, might with advantage be 

 repeated or extended. It is now generally admitted that 



N 



