506 Br. Artlmr Schuster [Jan. 28, 



The spectra of zinc and tin fnrnisli ns other examples in the 

 same direction, but the spectra of nearly all bodies show the same law 

 in a more or less striking way. 



If this law which I have given you is a true one,* and I believe 

 it will stand any test to which no doubt it will be subjected, we 

 shall be able to draw some important conclusions from it. In tlie 

 first place, it will be proved that the forces between atoms do depend 

 on their vibrations. If this is true, any change in the vibrations 

 of the spectrum, however small, will entail a corresponding change in 

 all the other properties of the body. On the other hand, any change 

 in the affinities of the element observed by other means will be 

 represented by a change in the spectrum. 



It is also possible that the introduction of forces due to vibratory 

 motion will help us over a considerable difficulty in the molecular 

 theory of gases. Some of the conclusions of that theory are at 

 present absolutely contrary to fact. A spectrosco2:>ist, for instance, 

 who is acquainted wdtli the mercury spectrum and all the changes in 

 that spectrum which can take place, feels more than sceptical when he 

 is told that the molecule of mercury contains only one atom, which 

 neither rotates nor vibrates. 



Nor can it be of advantage to science to pass silently over this 

 difficulty, or to neglect it as unessential, as is often done by modern 

 writers. The late Professor Maxwell, at least, was well aware of its 

 importance, and has often expressed in private conversation how 

 serious a check he considered the molecular theory of gases to have 

 received. This is not the place to enter more fully into this point 

 and to consider how the vibratory forces may affect some of the 

 suppositions on which the theoretical consequences are founded. 



However important the effects of concentration or dilution on the 

 spectra may be, they render the spectroscope less trustworthy as a 

 thermometric instrument ; for if the company in which a molecule is 

 placed changes the spectrum in the same way as temperature would, 

 it will be difficult to interpret our results. But although the dis- 

 cussion of our observations may be rendered more arduous and 

 complicated, we need not on that account despair. It is one of the 

 problems of spectroscopy to find out the composition of bodies, not 

 only qualitatively, but also quantitatively, and when we shall know 

 in what proportion different bodies are distributed in the sun, we may 



* Lockyer, ' Studies in Spectrum Analysis,' p. 140, draws attention to the 

 fact that an admixture of a second element dims tlie spectrum of the first, and he 

 expresses this fact by saying : " In encounters of dissimilar molecules the vibra- 

 tions of each are damped." Later he has shown that the lines of oxygen and 

 nitrogen, which are wide at atmospheric pressure, thin out when the gases are 

 only present in small quantities. Lecoq de Boisbaudran in his 'Atlas' gives 

 several examples of the differences in the relative brilliancy of lines produced by 

 concentrating or diluting the solution from which the spark is taken. The com- 

 plete parallelism of this change to the changes produced by increased temperature 

 has, however, never received sufficient attention. 



