684 Frofessor Dewar [June 10, 



In all the foregoing experiments the bands which Angstrom and 

 Thalen ascribe to hydrocarbons were always more or less plainly seen. 

 Much more care than has generally been thought necessary is needed 

 if the last traces of hydrogen and its compounds are to be removed 

 from spectral tubes. Indeed, water cannot be completely removed 

 from apparatus and reagents which do not admit of being heated to 

 redness. 



Thus a mixture of carbonate of sodium and boric anhydride, 

 previously to admixture heated red hot, was introduced into one end 

 of a piece of combustion tube, near the other end of which wires had 

 been sealed, and the open end drawn out; the mixture was then 

 heated, and when it was judged that all the air was expelled, the tube 

 was sealed off at atmospheric pressure. On passing sparks through it 

 carbonic oxide bands and oxygen lines could be seen, but no hydrogen, 

 hydrocarbon, or nitrocarbon bands could be detected. It appears, 

 therefore, that the application of a red heat is likely to prove a 

 more effectual means of getting rid of moisture than the use of any 

 desiccating agent. 



Are the groups of shaded bands seen in the more refrangible 

 part of the spectrum of a cyanogen flame, of which the three which 

 can be detected by the eye are defined by their wave-lengths (4600 

 to 4502, 4220 to 4158, and 3883 to 3850), due to the vapour of 

 uncombined carbon, or, as we conclude, to a compound of carbon with 

 nitrogen ? 



The evidence that carbon can take the state of vapour at the 

 temperature of the electric arc is at present very imperfect. Carbon 

 shows at such temperatures only incipient fusion, and that uncom- 

 bined carbon should be vaporised at the far lower temperature 

 of the flame of cyanogen is so incredible an hyj^othesis, that it 

 ought not to be accepted if the phenomena admit of any other 

 probable exj^lanation. On the other hand, cyanogen or hydro- 

 cyanic acid is generated in large quantity in the electric arc taken 

 in nitrogen, and Berthelot has shown that hydrocyanic acid is 

 produced by the spark discharge in a mixture of acetylene and 

 nitrogen, so that in the cases in which these bands shine out with the 

 greatest brilliance, namely, the arc in nitrogen and the cyanogen 

 flame, we know that nitrocarbon compounds are present. Further, we 

 have shown that these bands fade and disappear in proportion as 

 nitrogen is removed from the arc. Angstrom and Thalen had previ- 

 ously shown the same thing with regard to the discharge between 

 carbon electrodes ; and the conclusion to which they and we have 

 come would probably have commanded universal assent if it had not 

 been for the fact that these bands had been seen in circumstances 

 where nitrogen was supposed to be absent, but where, in reality, the 

 difficulty of completely eliminating nitrogen, and the extreme sensi- 

 bility of the spectroscopic test, had been inadequately apprehended. 



Our argument is an induction from a very long series of observa- 

 tions which lead to one conclusion, and hardly admit of any other 



