1882.] On the Spectra of Carbon and its Compounds. 125 



The apparatus "used in the experimeuts has been already described 

 in our paper on the spectrum of the compounds of carbon with 

 hydrogen and nitrogen (" Proc. Roy. Soc," vol. 30). No attempt 

 was made to use perfectly pure gases or to remove all traces of 

 nitrogen from the vessels employed; the object being to study the 

 variation of the groups of lines, perfect purity in the gases was not 

 required. 



The arc discharge between graphite poles in carbonic acid shows the 

 triple set, beginning about 4380, with traces of the other sets of 

 cyanogen bands at 4218 and 3883. If the carbonic acid gas is dis- 

 placed by air, the triple set are very much weakened and are some- 

 times invisible, while the other fluted series at 4218 and 3883 are 

 greatly strengthened. 



The spark discharge does not show the cyanogen sets in carbonic 

 acid, but a series of five groups appear between the limits of the lines 

 S and N of the solar spectrum, which may possibly be due to carbonic 

 acid or carbonic oxide. The carbonic oxide flame, however, does not 

 show this set. If the spark discharge is taken between graphite 

 poles in nitrogen all the cyanogen series appear; but in hydrogen 

 they sometimes disappear. As a rule, however, they remain faint even 

 when a current of the gas is kept continuously passing through the 

 bulb in which the discharge is taken. With the arc discharge in 

 hydrogen the triple set are well marked, while the series at 4218 dis- 

 appear, and the ultra-violet group are just visible ; the hydrocarbon 

 set, however, at 4310 come out strong. 



In order to ascertain how the pressure of the surrounding atmos- 

 phere affected the emissive power of the cyanogen produced synthe- 

 tically in the arc discharge, a series of observations on the spectra 

 obtained under diminished pressure of the gaseous atmosphere was 

 undertaken. The pressure in different gases was reduced to a mean 

 value of about 1 inch of mercury, and under such conditions the 

 intermittent discharge of the De Meritens machine was examined. 

 The arc in air, under these circumstances, showed the blue hydro- 

 carbon set ; all the cyanogen series of bands together with a nitrogen 

 series near H. Carbonic acid at the same pressure had the triple set 

 of lines strongly marked, while the others were decidedly weaker, as 

 in the experiments with the gas at atmospheric pressure. 



In hydrogen, at equal pressure, the triple set disappear, and the 

 hydrocarbon set at 4310 occur, which series is not generally seen in 

 the photographs of the arc spectrum. But what is very remarkable 

 is the appearance of two lines of carbon, viz., 2836'3 and 2837 - 2 (also 

 2506 and 2508) in the spectrum of the discharge, whereas these 

 spark lines are not generally found in the arc spectrum. 



The following lines of carbon have been observed in the arc dis- 

 charge of the Siemens continuous current machine, as well as in that 



