848 Variations in tJte Spectra of Carbon Electrode*. [Apr. 



in an atmosphere of oxygen ; this can scarcely be considered 

 cyanogen line. There remain now the two 3584'8 and 3513'5, 

 as these lines are absent from the spectra taken in oxygen and 

 carbon dioxide, it may well be questioned whether their origin 

 elementary carbon. 



From their occurrence in the spectra taken in air and their bei 

 lengthened when moistened electrodes are used, it seems that 

 their production nitrogen is necessary and water vapour ad 

 tageous ; but they are not yielded by cyanides, and, therefore, in 

 absence of any good reason for this, they cannot be attributed 

 cyanogen. A further examination of the list of lines will show 

 there are seven attributed to carbon by Eder and Valenta and t 

 assigned by them to cyanogen which do not always appear w! 

 powerful sparks are passed between graphite electrodes through 

 Then we have four lines of carbon and three attributed to cyan* 

 which do not appear when the spark is passed through 

 dioxide. Liveing and Dewar (' Roy. Soc. Proc.,' vol. 34, p. 

 have shown that mixed vapours do not give precisely the 

 spectra as the substances present in the mixture would give 

 themselves. In certain cases one element renders the lines 

 another more brilliant, while in other instances some of the lii 

 disappear. Chlorides usually have the effect of sweeping out t 

 fainter lines. 



The lines at 3590'5 and 3585'9 and 3584 are closely adjacent 

 certain nitrogen lines which are somewhat strengthened in the carb 

 spectrum. As the carbon spectrum varies remarkably under differe 

 conditions, it may exercise an influence on the nitrogen spectrui 

 and at the same time be modified by the presence of an atmosphe 

 of this gas. In order to test the probability of the carbon 

 nitrogen spectra being subject to variations when the two elem 

 are together in the spark or flame, it is necessary to consider 

 effect of one spectrum on another when the two are produced si 

 taneously from quite different materials. 



In the oxyhydrogen flame the water-vapour lines are promi 

 but only two groups are visible in the spectrum under normal 

 ditions, and with an exposure of half an hour. If, however, 

 sulphur be burnt in the flame, the conditions being othe: 

 unchanged, then the spectrum, in addition to a band of contin 

 rays and flutings characteristic of sulphur vapour, shows the 

 vapour lines wonderfully strong, with groups extending be 

 those portions of the spectrum usually photographed, and not 

 are the lines distinct, but dense, as if their radiating power or th 

 chemical action of their radiations was greatly increased. Thi 

 does not arise from the continuous spectrum merely overlapping an> 

 apparently strengthening the water-vapour lines, since new groups c 



