228 SCIENCE PROGRESS 



We are now in a position to consider what deductions 

 can be made from a spectrum with regard to the nature of 

 the substance producing it. Let us turn first to the continuous 

 spectrum proceeding from a soHd body. Inasmuch as precisely | 

 similar continuous spectra are yielded by all hot solid or liquid * 

 bodies, it is clear that, from the point of view of chemical 

 analysis, they are of no use whatever. Nevertheless, they have 

 one property which is of considerable significance. The 

 spectrum, it must be remembered, is an ordered array of waves 

 of different lengths, carrying energy from the source of radia- 

 tion to the body on which the spectrum is received. We 

 have said nothing as yet as to the manner in which the total 

 quantity of energy is distributed among the different waves. 

 Investigation shows that this question of distribution is of 

 very great importance. It is not a difficult matter to deter- 

 mine, in any given spectrum, in what waves the greatest 

 amount of energy is stored. When this is done for spectra 

 emanating from bodies at known temperatures, it is found 

 that the length of those particular waves, multiplied by the 

 absolute temperature of the source, is a constant quantity. 

 We have, therefore, only to find the wave-length of maximum 

 energy in any continuous spectrum in order to discover the 

 temperature of the radiating body. To continue our former 

 simile, we might say that, though our furniture vans, being 

 of all possible kinds, give us no clue to the identity of the 

 sender, yet, by the manner in which the furniture is distributed 

 among them, we can estimate his degree of agitation. This 

 method is of great use for the determination of very high tem- 

 peratures, or of temperatures which cannot be measured 

 directly, such as those of the stars. 



The continuous spectrum, therefore, is not without meaning. 

 But the greatest revelations proceed from the line and band 

 spectra. It is there that spectrum analysis has its roots. This 

 exceedingly useful process depends on two fundamental prin- 

 ciples. First, the spectrum given by the vapour of any sub- 

 stance is absolutely characteristic of the substance, and can 

 arise from nothing else ; second, the spectrum varies with the 

 conditions of excitation, but still remains characteristic of its 

 chemical origin. Furthermore, if any number of substances 

 are mixed together and vaporised, the lines of all of them 

 appear in the spectrum when the appropriate exciting agency is 

 employed. The advantages of this method of analysis are 

 obvious. The whole of the investigation is carried out by a single 

 operation, and, moreover, it is not necessary for the unknown 

 substances to be seen or handled. The presence of hydrogen 

 in a star millions of miles away is demonstrated as clearly as 

 though the gas were captured and imprisoned in a test-tube 



