Hydrogen and the. Spectrum of Aqueous Vapour. 343 



In working with powerful condenser-discharges, I found 

 that I obtained essentially the same spectrum whether the 

 tube was rilled with hydrogen, rarefied air, or nitrogen, not- 

 withstanding the great care taken in heating the tubes and 

 drying the gases. This characteristic spectrum was that of 

 water-vapour, containing lines of hydrogen, of oxygen, and 

 rarefied air. The employment, therefore, of a condenser-dis- 

 charge in glass tubes apparently drives off aqueous vapour 

 from the glass walls, even if extraordinary care is taken to 

 expel such vapour in the process of exhausting such tubes. 

 My experiments entirely confirm Professor Crookes's state- 

 ments on this point. Glass seems to be of a permeable nature, 

 and in the process of manufacture is doubtless filled with 

 gases which reveal themselves when disruptive electrical dis- 

 charges are employed. It does not seem possible that dry 

 hydrogen can exist as such in a ghiss vessel when it is sub- 

 jected to a condenser-discharge; and I therefore believe that 

 the spectrum of hydrogen called the four-line spectrum is an 

 evidence of the presence of aqueous vapour, for it can be 

 produced with the greatest ease when we are certain that this 

 vapour is present. This spectrum, to my mind, is an evidence 

 of the presence of oxygen in the sun's atmosphere. 



The true spectrum of hydrogen seems to be the white 

 spectrum. This is developed by a steady current which 

 should not exceed a certain strength, for a powerful one 

 drives off also rarefied air from the walls of the tube ; and 

 there is a process of occlusion set up if large metallic terminals 

 are employed. Such large terminals are necessary if very 

 strong currents are used. With large aluminium electrodes, 

 under the action of strong currents, spectra of the com- 

 binations of nitrogen with oxygen and hydrogen are pro- 

 duced in tubes which have been filled with dry and pure 

 hj drogen. When such tubes have been subjected to powerful 

 currents, it seems impossible to refil them with pure hydrogen. 

 In one case I rilled a tube with pure diy nitrogen, and passed 

 a current through it sufficiently powerful to melt down one of 

 the aluminium hollow cylinders which formed an electrode. 

 The aluminium was filled with cavities as if gases had bubbled 

 from the interior. When this tube was filled with dry 

 hydrogen, it gave again spectra of compounds of nitrogen 

 with no trace of the hydrogen spectrum. A powerful con- 

 denser discharge, however, produced the aqueous-vapour 

 spectrum. Hydrogen is thus extremely elusive when sub- 

 jected in glass vessels with metallic electrodes, and in the 

 presence of other gases, to electrical dissociation. Indeed, 

 a steady strong battery-current can be employed at a certain 



