1898.] Prof. Liveing, On the Flame-spectrum of Mercury. 39 



was introduced into the electric arc. Recently I have found that 

 the same ray was produced by mercury heated in a flame of 

 cyanogen burning in oxygen. As Prof. Dewar and I pointed out 

 long ago, this flame is nearly the hottest known, on account of the 

 endothermic character of cyanogen and the large amount of 

 energy stored in the compound and liberated as heat when it is 

 burnt. Besides the ultra-violet ray, A, 2535*8, I found another 

 mercury ray developed in the cyanogen flame at wave-length 4358 

 in the indigo-blue. 



The cyanogen was made by heating cyanide of mercury, and 

 after passing through a spiral tube kept cool to condense the most 

 of the mercury vapour, it was burnt at a platinum jet in the 

 middle of a wider jet of oxygen. The spectrum was formed by a 

 concave grating and photographed, and the wave-lengths of the 

 rays determined by comparison with the iron spectrum. 



The emission of these two rays in a flame where the stimulus 

 appears to be only the high temperature, since mercury is not 

 known to combine with either oxygen, carbon, or nitrogen at such 

 a temperature, seems to be a fact of some theoretic interest. 

 I cannot help regarding the production of spectra by an electric 

 discharge as essentially a different process from the production by 

 heat. The electric discharge may, not improbably, set up vibra- 

 tions in the molecules directly, and the heat which attends the 

 discharge, often only feeble, may be only a secondary phenomenon. 

 At any rate a great many rays are given out by various elements 

 in an electric discharge which have never been observed to result 

 from mere heating. But the emission of light by hot mercury 

 vapour is of interest in connexion with the fact that the ratio of 

 the specific heats of mercury vapour, at constant pressure and 

 constant volume, is almost exactly what it would be if none of the 

 heat added were employed in producing any vibratory motion, or 

 other form of motion within the molecule. The spectrum emitted 

 by mercury in the cyanogen flame proves that heat, at a suffi- 

 ciently high temperature, is, in part, transformed into vibratory 

 motion which affects the ether ; and the true inference from the 

 ratio of the specific heats appears to be, that, at the temperature 

 at which this ratio was measured, the amount of heat converted 

 into vibratory motion is very small compared with the amount 

 which remains heat. What the temperature may be in the 

 compressed gas during the passage of a sound-wave I do not 

 know, but it cannot be very high. 



In order to find an upper limit I had a tube, similar to those 

 used for finding the velocity of sound in gases, filled, in the dark, 

 with a mixture of hydrogen and chlorine in equal volumes, and 

 then the tube was sounded by drawing it through the fingers 

 covered with a resined glove. The sounding of the tube was 



