Spectra and Planck's Laic. 423 



electrons. They would, for example, differ in iheir specific 

 inductive capacities so that the proportion of the different 

 kinds would be different at different places in a rapidly 

 varying electric field. Again, the atoms of one kind might 

 more readily combine with an atom of a different element than 

 those of another. So that if the gas in a spectroscopic tube 

 were gradually absorbed by chemical means, as, for example, 

 when oxygen is absorbed by sodium, the relative intensity 

 of the lines at the end of the process might not be the same 

 as at the beginning ; or, again, when the spectrum of an 

 element is obtained by decomposing a compound, as, for 

 example, when the spectra of the alkali metals are obtained 

 by putting one or other of their salts in a Bunsen flame, 

 the relative intensity of the lines might depend upon the 

 character of the salt. Bevan's experiment, already alluded 

 to, indicated that the number of atoms giving the lines of 

 shorter wave-length in the principal series increased with 

 the temperature of the sodium vapour. I pointed out many 

 years ago, that the magnitude of the refractive index of 

 helium showed that only a small fraction of the helium 

 atoms could vibrate with the frequency of any particular 

 line in the helium spectrum. Many instances of the vari- 

 ability of the relative intensity of different lines in the same 

 spectrum are given in Kayser's ' Spectroscopic/ Since 

 the processes which make the gas luminous also ionize the 

 gas and thus enable an atom of one kind to be converted into 

 one of another, it does not seem probable that by any process 

 of fractionation we should be able to obtain a gas containing 

 nothing but atoms of one kind, and therefore giving out a 

 spectrum consisting of a single line under all conditions of 

 excitation . 



The electric field inside an atom consists, according to the 

 view we have just taken, of alternate shells of attractive and 

 repulsive forces, the places of transition from attraction 

 to repulsion being places where the force vanishes and where 

 an electron could be in equilibrium ; the distances of these 

 places of equilibrium from the centre being in harmonic 

 progression. 



We now pass on to consider the nature of the forces 

 which act upon the electron and cause it to vibrate in the 

 frequencies of the spectral lines. It is possible, as we shall 

 see, to postulate an infinite number of laws of electric force 

 which could give rise to a sequence of vibrations represented 

 by Rvdberg's law. 



We may also imagine that in addition to the electric 

 field inside the atom there is also a magnetic one, and that 



