1882.] On the Spectra of Carbon and its Compounds. 129 



or picoline, cyanogen can always be recognised in the flame gases. 

 Chloroform nnder such circumstances yields the largest amount. 

 When a mixture of carbonic oxide and ammonia is passed through a 

 porcelain tube heated in a furnace, large quantities of hydrocyanic acid 

 are produced, especially when the moist gases are employed. Ammonia 

 passed over perfectly pure graphite at a white heat produces hydro- 

 cyanic acid, and the vapour of chloride of ammonium is equally 

 efficient in bringing about this reaction. It appears to result from 

 the experiments that hydrocyanic acid can always be separated 

 from the interior of flames such as we have employed, provided 

 that portion of the flame which, in carbon compounds, is characterised 

 as reducing, be selected. That stage of combustion during which 

 free carbon or dense hydrocarbon vapours containing very little 

 hydrogen are formed, is favourable to the formation of hydrocyanic 

 acid, as ammonia can at this stage react on the carbon. It is quite 

 possible, however, that hydrocyanic acid may exist in small quantity 

 in some of the flames which tested according to this method appear 

 to contain none, and that notion is favoured by the consideration of 

 the dissociation phenomena which are known to occur in flames. 

 This led us again to spectroscopic examination as the most delicate 

 test for the presence of cyanogen, but instead of trusting to the 

 eye as in former experiments, photographs were taken of the spectra 

 of flames by means of a quartz and calcspar train, and the exposure of 

 the plate purposely prolonged. Thus examined, it was found that 

 coal gas well supplied with oxygen gave only the hydrocarbon groups, 

 together with the two interesting additional lines first discovered 

 by Dr. Huggins, having the wave-lengths 3872 and 3890 ; but when 

 -the coal gas passed through ammonia the photographs revealed the 

 characteristic cyanogen groups at 3883 and 4218, the most refrangible 

 group being the strongest. The cyanogen spectrum can then be pro- 

 duced synthetically from nitrogen compounds in flames along with the 

 hydrocarbon spectrum, so that the appearance of the groups of 

 cyanogen is not always associated with a very high temperature such 

 as we have in the electric arc. Cyanogen once formed gives its pecu- 

 liar spectrum at the relatively low flame temperature produced by 

 burning cyanogen mixed with carbonic acid. Of course the mean tem- 

 perature of a flame is very different from the temperature of individual 

 molecules, and this complicates the problem we are discussing. The 

 thermal equivalents of cyanogen and acetylene being highly negative, 

 it is certain that these substances yield on combustion the highest tem- 

 perature of any two compounds burning in oxygen; and we have 

 shown in a former paper that burning cyanogen in nitric oxide gas, 

 which probably induces a still higher temperature, does not bring about 

 any marked change in the character of the spectrum. Spectroscopic 

 analysis can thus detect very small quantities of cyanogen under 



VOL. XXXIV. K 



