1882.] On the Origin of the Hydrocarbon Flame Spectrum. 429 



increased resistance in the arc producing temporarily a higher mean 

 temperature. Experiments in course of execution, where the arc will 

 be subject to a sudden increase of pressure, will, we trust, solve this 

 problem. 



Electric Discharge between Graphite Poles in different Gases. 



When pure graphite is employed, instead of the ordinary carbon 

 poles, and the arc discharge is taken in different gases, in the same 

 way as was described in our first paper, " On the Spectra of the Com- 

 pounds of Carbon," we have noticed some slight differences which are 

 worthy of record. In carbonic acid gas, fine channellings are seen 

 covering the whole extent of the spectrum from the low red as far as 

 the blue set of the flame spectrum, the flutings of the one group being 

 observable as far as the next group. The triple set of the cyanogen 

 flame spectrum remained very strong when all the cyanogen groups in 

 the violet had disappeared. When the carbonic acid is displaced by 

 hydrogen, the hydrogen lines appear, the hydrocarbon flame and the 

 triple sets remaining bright; but in this gas the flame group at 431 is 

 particularly well marked, and the carbon line at 4266 keeps flashing in 

 occasionally. Thus we have in the same field of view the hydrocarbon 

 series, the hydrogen lines, and one of the strongest lines of carbon. 

 The De Meritens intermittent arc discharge was employed in these 

 experiments, and it is curious to note that hydrogen, instead of 

 favouring the passage of the arc between carbon poles, really intro- 

 duces some peculiar resistance, perhaps owing to the reduction of 

 temperature by the gas surrounding the arc, or because of the 

 formation of acetylene. The arc is at any rate much shorter and 

 smaller in section than in air, but the temperature seems to be corre- 

 spondingly increased, as we may infer from the fact that the hydrogen 

 and carbon lines are now very marked. The Siemens arc in air does 

 not show the carbon line at 4266, although we have proved that some 

 of the chief ultra-violet lines occur in this discharge. The arc taken 

 in carbonic oxide shows the triple group along with the usual sets of 

 the hydrocarbon flame spectrum, without any trace of the carbonic 

 oxide spectrum being visible. The occurrence of the triple set of lines, 

 in the absence of other groups characteristic of cyanogen, makes us 

 doubt whether this set has anything really to do with nitrogen. We 

 are inclined to think that their previous appearance when the arc was 

 taken in glycerine containing nitro-benzole was really due to some 

 indirect effect, and ought not to be taken as proof of the formation of 

 cyanogen in the absence of other characteristic groups under such 

 circumstances. 



