1882.] On the Origin of the Hydrocarbon Flame Spectrum. 425 



visible in the spectrum, since we know by experience that the electric 

 discharge does not always light up all that is in a tube. Mercury, 

 for instance, must be present in all these tubes, but its lines do not 

 usually show until the exhaustion is carried to a high degree. 



One tube was filled with nitrogen with which was mixed a small 

 percentage of cyanogen obtained by heating a little mercury- cyanide 

 placed at one end of the drying tube. This tube gave well the chan- 

 nellings in the blue and indigo characteristic of cyanogen, but neither 

 the hydrogen lines nor the spectrum of the hydrocarbon flame. The 

 cyanogen seems to be scarcely at all decomposed by the spark when 

 thus diluted, for the carbon lines were not brought out by the use of a 

 jar, and the cyanogen spectrum remained after continued sparking. 

 We remark that Berthelot found it exceedingly difficult to get 

 acetylene from a mixture of cyanogen and hydrogen at atmospheric 

 pressure ; and in order to make this experiment succeed, short and 

 very powerful sparks had to be continued for some hours. 



In a tube filled with carbonic oxide mixed with a little air imperfectly 

 dried, when not too highly exhausted, the carbonic oxide spectrum, 

 that of the flame of hydrocarbons, and that of cyanogen, may all be 

 seen at once superposed when no jar is used. With a jar and a 

 tolerably high exhaustion the carbonic oxide spectrum, the hydro- 

 carbon flame spectrum, and the carbon line spectrum, may all 

 be seen at the same time. All the foregoing observations were made 

 when the tubes were viewed end on and the image of the narrow 

 part of the tube thrown on the slit by a lens. Tubes filled with 

 carbon disulphide and carbon tetrachloride at reduced pressures have 

 been examined by us, but these compounds of carbon are very quickly 

 decomposed by the spark, so that few observations can be made with 

 one tube. If the exhaustion is not carried very far the spectrum of 

 the hydrocarbon flame is seen both in carbon disulphide and in 

 carbon tetrachloride, when all the precautions above mentioned have 

 been taken to remove moisture, and when no trace of the hydrogen 

 spectrum is visible. At higher exhaustions the spectrum is a faint 

 continuous one together with that of sulphur or chlorine as the 

 one or other is present. The spark taken without condenser between 

 electrodes near together in wide tubes filled with saturated vapour of 

 carbon disulphide or carbon tetrachloride dried with phosphoric 

 anhydride and deprived as completely as possible of air by pumping or 

 boiling out, shows the spectrum of the flame of hydrocarbons brightly ; 

 but in tubes filled with carbonic acid gas from ignited sodium car- 

 bonate and boric anhydride and sealed off at a high temperature, the 

 spectrum is that of carbonic oxide together with that of oxygen. 



Spectrum of the Spark in Compounds of Carbon at Higher Pressures. 

 In the spark taken between poles of purified graphite in hydrogen, 



