810 Mr. H. Smith on the Spectroscopy of the Electric 



being in general strongest near the point, these bands are 

 weak close to the point while they are strongest some little 

 distance from the point ; they are also very long. These 

 bands are the spectrum o£ the bubble (figs. 9, 6 a, 7 b, 

 PI. XL). 



This spectrum, I find, is identical with the spectrum of 

 the "afterglow " of hydrogen discovered by Prof. Strutt*. 

 Strutt observed that these bands only appeared when sulphur 

 was present as an impurity, and that their appearance was 

 the most sensitive test for the presence of sulphur in hydrogen. 

 He also observed the spectrum in hydrogen sulphide (after- 

 glow) in the carbon bisulphide flame, and also when sulphur 

 or its compounds was acted on by chemically active nitrogen. 

 When experimenting on the afterglow in hydrogen, he tried 

 unsuccessfully to produce the spectrum in pure sulphur 

 vapour. The same spectrum, I find, had already been 

 observed byHimstedt and von Dechendf. They found that 

 in the glow-discharge in hydrogen, in the region remote from 

 the point there appeared a series of bands degraded towards 

 the red. I have given their values for the wave-lengths of 

 the brightest edges, together with Strutt's for the carbon 

 bisulphide flame and my own for the bubble-discharge 

 spectrum. There can be little doubt that the three spectra 

 are the same. 



Table III. 



Himstedt & von Dechend] An± q Qz1 o Q o on a oc A qkf; 



(glow-discharge). ) 404 394 383 3 ' 4 "" 364 "■ 355 



^flamir^ 011 biSUlphide } 405 ° 3940 3835 3740 3680 3645 3590 3557 

 Bubble spectrum 4046 3939 3828 3741 3680 3645 3589 3556 



The bands at 3680 and 3590 are much fainter than the 

 others. Strutt also gives wave-lengths of the fainter bands 

 extending towards the red and violet. I have not been able 

 to obtain very great accuracy in my own measurements, as 

 the dispersion of the apparatus was so small. The presence 

 of a fainter band a few Angstrom units shorter in wave- 

 length than each bright band also made measurements 

 rather confusing. Strutt suggests that the spectrum may 

 be due to sulphur, as it appears when sulphur compounds 

 are acted on by chemically active nitrogen. The spectrum, 

 however, may be as sensitive to a trace of hydrogen in an 

 excess of sulphur, as Strutt as shown the converse to be true. 

 It seems more probable that the spectrum is due to hydrogen 



* Proc. Roy. Soc. vol. lxxxvi. pp. 529-33 (1912). 

 f Phys, Zeit. vol. ix, pp. 852-3 (1908), 



