1881.] on the Orujin and Heniitij of Spectra. C77 



overpowered by the continuous spectrum, but was regularly seen when, 

 from some variation in the discliarge, the continuous spectrum became 

 less brilliant. This was tlie lirst occasion on whicli we liad seen tlio 

 hydrogen lines in the arc, tliough Secclii * states tluit he liad seen 

 them by the use of moist carbon poles. The hydrocarbon bands in 

 the green and blue were at intervals well seen. Those in the yellow 

 and orange were, owing doubtless to the smaller dispersion of tlio 

 light in that region, overiiowered by the continuous spectrum. 

 "Whereas when air and carbonic acid gas were used, the inside 

 of the globe was quickly covered with dust from the disintegrated 

 poles, scarcely any dust was thrown oH' when the arc was passed in 

 hydrogen. 



In nitrogen a longer arc could be formed, and the indigo, violet, 

 and ultra-violet bands of cyanogen all came out at intervals brilliantly. 

 The green and blue hydrocarbon bands were also well developed. 



On filling tlie globe with chlorine, keeping a current of that gas 

 passing through it, the arc would not pass through a greater dis- 

 tance than about 2 millims. No metallic lines were visible. At first 

 the violet bands, as well as the green and blue hydrocarbon bands, 

 were visible ; but gradually, when the current of chlorine had been 

 passing for some minutes, there was nothing to be seen but a con- 

 tinuous spectrum with the green and blue hydrocarbon bands. 

 Neither of these bands were strong, and at intervals the blue bands 

 disappeared altogether. 



The arc would not pass in a current of carbonic oxide through 

 any greater space than in chlorine. There was much continuous 

 sjDectrum ; the yellow, green, and blue hydrocarbon bauds were well 

 seen, some of the indigo bands were just discernible, the violet had 

 nearly, and the ultra-violet quite, gone from sight. No trace of the 

 carbonic oxide bands, as seen in the spark discharge in that gas, was 

 visible. This is the more remarkable since under similar circum- 

 stances two of the characteristic lines of hydrogen were seen. 



In nitric oxide a very long arc could be obtained. The violet and 

 ultra-violet cyanogen bands were well seen, the indigo bands were 

 seen, but weaker. The blue and green hydrocarbon bands were also 

 seen well when the arc was short, not so well when the arc was long. 

 Many metallic lines of iron, calcium, and magnesium were seen. 



In ammonia only a short arc could be obtained. All the bands 

 were faint, but the indigo and violet and ultra-violet cyanogen 

 bands were always visible. 



These experiments with difi'erent gases eliminate to a large extent 

 the influence of electric conductivity on the character of the spectrum. 



Apart from the relative electric conductivity of gases, it is clear, 

 from the experiments, that the length and character of alternating 

 electric discharges between carbon poles in dificrent gases do 

 not follow the law which we should expect. It will require a prc- 



* ' Compt. Rend.' 1873. 



