500 ^^*- Arthur Schuster [Jan. '28, 



bourhood, and only gradually disappeared. This experiment shows that 

 the spectrum may exist in other parts of the tube, and that it is there- 

 fore due to a peculiar kind of molecule, and not to anything specially 

 related to electric phenomena taking place in the neighbourhood of 

 the negative pole. Other experiments supported this view. 



The classification of spectra, according to the complexity of the 

 vibrating molecule, is of great theoretical importance ; fur by its 

 means we may hope to obtain some information on the nature of the 

 forces which bind together the atoms into one molecule. Our whole 

 life is a chemical process, and a great part of the mysteries of nature 

 would be cleared up if we could gain a deeper insight into the 

 nature of chemical forces. I believe no other line of investigation to 

 be as hopeful in this respect as the one w^hich examines directly the 

 vibrations of the molecules which take place under the influence of 

 these chemical forces. If we could find a connection between the 

 vibrations of a compound molecule and the vibrations of the simpler 

 elements which it contains, we should have made a very decided step 

 in the desired direction. I need not say that various attempts have 

 been made to clear up so important a point; but we have to deal 

 with complicated forces, and the attempts have as a rule not been 

 crowned with much success. 



There are, however, a few exceptions, a few cases of greater 

 simplicity than the rest, where we are able to trace to their mechanical 

 causes, the spectroscopic changes which take place on chemical com- 

 bination. These few and simple cases may serve as the fingerposts 

 which show us the way to further research, and we may hope, to 

 further success. To make the spectroscopic changes of which I am 

 speaking clear to you, I must have recourse to the analogy between 

 sound and light, and remind you of the fact that when the prongs of 

 a tuning-fork are weighted its tone is lowered, which means that the 

 period of vibration is increased, and consequently that the length of 

 the wave of sound sent out is lengthened. Now, suppose a molecule 

 or atom, the spectrum of which I am acquainted with, enters into 

 combination with another. And suppose that the vibrations of the 

 second molecule are weak or lie outside the visible range of the 

 spectrum, then the most simple assumption which I could make would 

 be that the addition of the new molecule is equivalent to an increase of 

 the mass of the other. An increase of mass without alteration of the 

 force of the molecule, will, as in the case of the tuning-fork, lengthen 

 the period of vibration, and increase the wave length. If a case of 

 that kind were actually to happen, I should observe the whole 

 spectrum shifting towards the red ; and this is what is observed in 

 the few simple cases to which I have referred. The first observation 

 to that effect is due to Professor Bunsen, of Heidelberg. Examining 

 the absorption spectra of different didymium salts, he found such a 

 close resemblance between them, that no difference could be detected 

 with instruments of small powers ; but with larger instruments it 

 was found that the bands varied slightly in position, that in the 



