268 REPORT— 1880. 



traces of moisture, and to the fact that this spectrum is well developed 

 under circumstances "where we know hydrocarbons to be present. Finally 

 we give the wave-lengths of .the least refracted lines of the most con- 

 spicuous bands. According to Angstrom and Thalen they are : 5633"0 ; 

 5164'0; 4736'0. Watts gives slightly different values, via.: 5634-7; 

 5165'5 ; 4739-8. 



Compounds of Carhmi and Nitrogen. — The flame of cyanogen, which 

 had already been examined by Faraday and Draper, before the days of 

 Spectrum Analysis, shows a series of bands in the red, reaching into the 

 green, which are not seen in any other flame. Pliicker and Dibbits have 

 given drawings of these bands, which have their sharp edge on the most 

 refrangible side. There is no doubt that they are due to a compound of 

 carbon and nitrogen. The same bands are also seen, when cyanogen is 

 burnt in oxygen, although, according to Morren and Watts, they are less 

 developed, a fact which they ascribe to the smaller quantity of undecom- 

 posed cyanogen at the higher temperature of the flame in oxygen. Ac- 

 cording to Pliicker and Hittorf, and also Dibbits, the bands in the red 

 become more brilliant when cyanogen is burnt in oxygen. There seems 

 to be a conflict between the increased brilliancy due to a higher tempera- 

 ture and the decrease of luminosity due to the more rapid decomposition 

 in the oxygen flame. 



Besides the red and yellow bands, a cyanogen flame shows a series 

 of bands in the blue, violet, and ultra-violet. These bands have been, 

 until quite recently, ascribed to carbon, as they have also been observed in 

 carbon compounds not containing nitrogen, but according to the experi- 

 ments of Professors Liveing and Dewar, they can in those cases always 

 be traced to impurities containing nitrogen. 



Thus, according to Watts, the bands are seen when a spark is taken 

 in carbonic oride at the atmospheric pressure. According to Professors 

 Liveing and Dewar this is true, if the carbonic oxide has been prepared 

 from ferrocyanide of potassium. When the gas, however, was made by 

 the action of sulphuric acid on dried formiate of sodium, a faint trace of 

 one of the bands only could be detected. When the gas was pi'epared 

 by heating a mixture of quicklime with pure and dry potassium oxalate, 

 no trace whatever of the bands in question appeared. 



Similarly Watts and Lockyer had observed the bands in a tube con- 

 taining- carbon tetrachloride, but, according to Professors Liveing and 

 Dewar, these bands do not appear when the tetrachloride has been well 

 purified, and when all traces of air have been expelled from the tube. 



Experiments with naphthalene gave the same results ; the bands did 

 not appear when the air had been properly expelled from the tube. 



Tlaese experiments seem conclusive as to the chemical origin of the 

 spectrum in question. It seems remarkable, however, that this spectrum 

 should be reversed in the solar spectrum ; for a photograph taken by 

 Lockyer shows a decided coincidence of one of the flutings with a series 

 of dark lines in the solar spectrum ; and Professors Liveing and Dewar 

 consider the reversal of another set of flutings still further in the ultra- 

 violet as probable. The spectrum we have been discussing consists chiefly 

 of three sets of bands ; the first set consists of seven fluted bands (wave- 

 lengths 4600 to 4502, Watts), the second set of six bands (\=4220 to 

 4158, Watts), and the third set in the ultra-violet of five bands (X = 3883-5 

 to 3850, Liveing and Dewar). According to Professors Liveing and 



