256 Professor Arthur Smithells on 



of sodium with traces of oxygen lingering in the nitrogen 

 and the oxygen contained in the siliceous material of the 

 porcelain tube. To test this question he produces a " cool" 

 flame by igniting a current of air charged with just enough 

 vapour of carbon disulphide to give a combustible mixture. 

 The average temperature of the flame is below 150° C. In 

 this flame a fine spray of a sodium salt gives no spectrum ; 

 and he finds that not until enough additional carbon di- 

 sulphide is supplied to give a much higher temperature does 

 the spectrum appear. Comparing this result with the results 

 of the porcelain-tube experiments, Pringsheim concludes that 

 since in the latter case the sodium salts show no spectrum at 

 temperatures which, in the case of a flame, suffice for the 

 display of the spectrum, chemical action and not heat alone 

 must be the cause of this difference ; that, in short, flame- 

 spectra of sodium salts are not due to simple dissociation by 

 heat, but to chemical processes of reduction. 



By way of confirming this conclusion Pringsheim then 

 describes the effect of heating sodium salts in a hot porcelain 

 tube in presence of different gases. In air, nitrogen, or 

 carbon dioxide no spectrum is produced, but in hydrogen the 

 D-line immediately appears. This is not quenched by the 

 addition of carbon dioxide, but immediately disappears on 

 the introduction of air. The reduction of the sodium salts 

 and the production of a spectrum he also effects by iron and 

 carbon. Finally, by an ingenious device, Pringsheim studies 

 the phenomenon in another way. The sodium salt is intro- 

 duced into the tube in a nickel or iron boat which can be 

 suddenly removed from the hot part of the tube by an electro- 

 magnet. In the case of the iron boat the iron acts also as a 

 reducing agent. The point is to see whether, having deve- 

 loped the spectrum with the iron boat in the hot part of the 

 tube (filled with hydrogen), this spectrum fades suddenly or 

 gradually when the boat is withdrawn to the cold part of 

 the tube. If it fades gradually the conclusion is that the 

 liberated sodium vapour is luminous per se, and only ceases 

 to give a spectrum as it migrates to the cool part of the 

 apparatus. If the spectrum fades suddenly the conclusion is 

 that the spectrum is only evinced at the moment of libera- 

 tion of the sodium from its compound, and is therefore the 

 direct result of chemical action. The latter Pringsheim finds 

 to be the case. 



Pringsheim then summarizes his conclusions as follows : — 

 1. There is no gaseous source of light which satisfies the 

 conditions of KirchhofFs law, 



