258 Prof. A. Suiithelis on the Luminosity of Gases. 



iron boat from the hot part of the tube filled with C0 2 , the 

 spectrum does not fade to the degree anticipated but acquires 

 a stationary intensity. This is attributed to the lingering of 

 sodium vapour, which attacks the sodium silicate with which 

 the interior of the tube has become coated in previous experi- 

 ments. A " reciprocal " reaction is said to take place : the 

 sodium vapour attacks the silicate, liberating other sodium 

 vapour, a process at first rapid but gradually attaining equili- 

 brium. After this the process goes on steadily, sodium libera- 

 ting sodium. This seems to me to be a very extraordinary as- 

 sumption, and more a robbing of Peter to pay Paul than the 

 picture of a chemical equilibrium. We are to suppose that 

 a film of sodium silicate — at the most, liquid— in contact with 

 sodium vapour, and above all in an oxidizing atmosphere of 

 carbon dioxide, is constantly giving and taking sodium atoms 

 so as to keep up the chemical action which Pringsheim 

 demands for the development of the sodium spectrum. Such 

 a state of things will be, I imagine, as surprising to chemists 

 as it is novel. It is true that dissociation phenomena are 

 pictured as involving a constant in-and-out movement of the 

 products of dissociation from the compound undergoing dis- 

 sociation. But if sodium acts on sodium silicate at all it will 

 be to liberate silicon, and after that is complete we can only 

 assume the resulting sodium oxide to react with fresh sodium 

 at the temperature at which sodium oxide dissociates. We 

 have no knowledge of the dissociation of sodium oxide. If, 

 as I venture to think, Pringsheim' s explanation of this point 

 is not to be accepted, there is less difficulty in accounting for 

 his other observations, for the sudden reduction of the intensity 

 of the spectral effect in hydrogen when the boat is suddenly 

 withdrawn is only important in comparison with what occurs 

 in C0 2 . I do not see any cause for surprise that the intensity 

 of the spectrum should suddenly diminish in hydrogen when 

 the boat is withdrawn suddenly. One would expect the 

 moving boat to drag its small atmosphere of sodium vapour 

 with it to the cool part of the tube, — and, again, one would 

 not expect the spectrum to persist in air. I can, therefore, 

 see no adequate grounds for the important conclusion which 

 Pringsheim draws from these experiments, namely, that 

 metallic sodium only gives its spectrum when undergoing 

 chemical change. 



I may summarize the views above expressed as follows : — 

 1. There is no evidence for, but much against, the suppo- 

 sition that sodium salts when introduced into a flame are 

 dissociated by heat so as to liberate the metal. 



