XXVIL] RESULT OF ADMIXTURES. 389 



Messrs. Liveing and Dewar have employed many arrange- 

 ments of this kind, and the results that they have obtained are 

 entirely in harmony with what I should have expected. And, 

 furthermore, many results which they consider extraordinary, 

 are easily and 1 sufficiently explained by the view I have put 

 forward. The variations brought about by the mixing of the 

 vapours with other vapours or gases have been variations due 

 merely to increased absorption when the conditions have 

 favoured such increased absorption. They have in no case 

 produced inversions. 



It is the more necessary to refer to these observations, because 

 Messrs. Liveing and Dewar have inferred that they confirm " the 

 theoretical view that alterations of temperature cannot put a 

 stop to any of the fundamental vibrations of a molecule ; " 1 and 

 that spectral changes due to varying admixtures may explain 

 the solar phenomena I have already recorded. On this point 

 they write : 



" There is, however, a further point for consideration which is 

 how far the presence of a mixture of molecules of different elements 

 affects the respective vibrations. This is a condition which obtains 

 in most or all of our observations of the arc in crucibles, as well as 

 in the solar atmosphere, so that it is important to see if any effect 

 can be traced to such a condition of matter. Indeed, in order to 

 arrive at any probable explanation of the variations observed in the 

 spectra of sun-spots and of the chromosphere, we require to study 

 the phenomena produced by such mixtures of vapours as exist in 

 our crucibles, and not merely the spectra produced by the isolated 

 elements either in the spark, arc, or flame." 



As far back as 1878 I published the statement that "in 

 encounters of dissimilar molecules the vibrations of each are 

 damped." 2 This damping process will naturally increase the 

 absorption till the absorbing molecule combines. 



1 Proc. Roy. Soc. vol. xxxiii. p. 431 



2 Studies in Spectrum Analysis, p. 140. 



