Theory of Refraction in Gases. All 



along with a nuiiiber of terms each of which is insignificant 

 excei)t at a certain point. These points I identify with the 

 spectral lines ; and this view seems preferable to the view that 

 spectral lines correspond to free periods of vibration of the 

 atom ; for if there are free periods in coordinates not capable 

 of being atfected by temperature, it is ditficiilt to see how 

 radiation could be kept up, or become appreciable as the 

 temperature is raised. On the present view this difficulty is 

 removed. We have motion in coordinates which can be 

 atiected by temperature; nevertheless the frequencies at which 

 ionization occurs are independent of temperature, while the 

 amount of ionization increases with the temperature ; and 

 this is in agreement with the electrical and optical experiments. 

 We have already seen that merely as obstacles the molecules 

 give refraction of the form /x'-^ — 1 cc p. If then we combine 

 this with the refraction produced by orientation in the 

 molecule, we get as the general form 



where co' is proportional to 6, and ki and ko are constants 

 depending on the molecule. 1 would again point out that 

 at certain points additional terms must be introduced. 



(7) General Application of the Formula. 

 We have now to examine whether the formula 



is capable of explaining the general features which were 

 summarized earlier. 



For^^ = the formula becomes 



If the first term is the greater we shall have approxi- 

 matelv K — loc p, but if the second term is greater K — 1 

 will vary more nearly proportional to pjO. We thus have a 

 means of explaining Baedecker's results for denser gases. 



Referring to the curve for < l-EZ-g) V we see that 



re- 



fraction increasing with the frequency may be explained by 

 the first portion of the curve. In order to meet the fact 

 that /A^ — 1 is very nearly proportional to p we require that 



T^ should be small. Further, the values of yu-^ — 1 must be 

 kid 



