Experiments on the Flame Spectra of Carbon Monoxide. 219 



experiments, and though dry oxygen surrounded the dry carbon 

 monoxide gas, yet the moisture of the atmosphere may have contri- 

 buted the water- vapour. 



In the original photograph of the carbon monoxide there were no 

 water-vapour lines apparent in the spectrum. Their invisibility 

 may be accounted for by a want of definition, owing to their being 

 photographed with a more widely open slit. That the relative 

 intensities of the two groups of lines are not the same as in groups 

 V and VI of Messrs. Liveiug and Dewar's photographs, furnishes 

 grounds for believing that they belong to the spectra of two different 

 substances. It is not improbable that one of these substances is an 

 oxide of nitrogen, for it has long been known that nitric acid is pro- 

 duced in minute quantity by burning hydrogen, and there is reason 

 to believe that it is also formed on the outside of a carbon monoxide 

 flame, where a greenish-yellow tinge is observed. 



It has been suggested to me by Professor Smithells that the 

 existence of the water-vapour spectrum, in the circumstances just 

 described, may be a confirmation of the observation made by Dixon, 

 that carbon monoxide when dry will .not burn in dry air (' Phil. 

 Trans.,' 1884, Part II, p. 629 ; see also Brereton Baker, ' Trans. 

 Chem. Soc.,' 1894, p. 611). Smithells states, however, that it will 

 burn if the carbon monoxide is heated (* Trans. Chem. Sec.,' 1894, 

 p. 610). This latter fact renders it possible for carbon monoxide to 

 burn and yet not show the water-vapour spectrum, since, though a 

 damp atmosphere may cause the gas to ignite, the flame will heat 

 the jet and thus the gas subsequently will be heated and burned, 

 although both it and the oxygen are dry. It may also serve to 

 account for the original carbon monoxide spectrum not exhibiting 

 the water-vapour lines. 



The water-vapour spectrum was always more feeble in the photo- 

 graphs taken from the carbon monoxide flame under an exposure of 

 from thirty-five to fifty minutes than in the oxyhydrogen flame under 

 an exposure of only two minutes. This indicates the small propor- 

 tion of the water or other substance present. 



The result of these experiments shows that the spectrum of 

 carbon monoxide consists entirely of a continuous spectrum decreas- 

 ing in intensity towards the more refrangible part of the ultra-violet, 

 about wave-length 3000. JSTo carbon bands or any lines or edges of 

 bands were photographed which could not be accounted for as due 

 to other substances than carbon monoxide. 



R 2 



