the Luminosity of Gases. 257 



2. Glowing gases can be obtained of a temperature below 

 150° 0. (cold flames). 



3. Sodium salts emit light in flames only as a consequence 

 of chemical processes. 



4. Metallic sodium heated in neutral gases emits light only 

 as a consequence of chemical processes. 



5. The assumption that gases can emit light by mere ele- 

 vation of their temperature is a hypothesis demanded neither 

 on experimental nor theoretical grounds. 



These conclusions, which go to the root of spectrum analysis, 

 are, it must be admitted, of the most serious importance. I 

 am not satisfied that they are warrantable, and wish to direct 

 attention to the following points : — 



(I.) The fifth, and in all probability the first, conclusion 

 {supra) of Pringsheim is contradicted by the experiment with 

 iodine vapour which I have already described (p. 252). 



(II.) The second conclusion of Pringsheim is unwarrant- 

 able, for the reasons stated in the first part of this paper. It 

 may be that a certain disulphide air-flame can be obtained 

 having an average temperature below 150° 0. This tells us 

 nothing of the temperature of the molecules of CCKj and S0 2 

 which are formed in the flame. Theory, on the other hand, 

 would assign to them a very high temperature if the heat 

 produced in their birth is assumed to be stored in them for 

 an instant. The ready inflammability of the mixture only 

 points to the relatively great distance at which the com- 

 bining molecules of CS 2 may be apart, and to the inter- 

 polation of a large number of non-burning molecules which 

 tend enormously to lower the average temperature of the flame. 



(III.) Pringsheim's third and fourth conclusions are un- 

 doubtedly the most serious. Metallic sodium, he says, heated in 

 neutral gases emits light only as a consequence of chemical pro- 

 cesses. This conclusion is reached only by indirect evidence. 

 In the first place, I would remark on the fact that according to 

 Pringsheim (loc. cit. p. 444) the admission of carbon dioxide 

 to the hot porcelain tube, in which the sodium spectrum has 

 been developed by the action of hydrogen on sodium com- 

 pounds, does not destroy the spectrum, while air does so 

 immediately. Why is this ? Hot sodium vapour and carbon 

 dioxide are quite incompatible ; surely C0 2 and air ought 

 both to extinguish the spectrum ; both are powerful oxi- 

 dizing agents to sodium. The important experiments are 

 those with the movable iron or nickel boat, and in these 

 another difficulty with C0 2 appears. After withdrawing the 



