MoREow — On the Influence of Self-induction, Sfc. 



615 



hydrogen, when none of the gold lines appears at all, demonstrating the fact 

 that the hydrogen and not the vapour of the electrodes conveys the current 

 across the spark-gap. Hydrogen being electropositive evidently acts in this 

 case like the metals, the opposite effect being observed with electronegative 

 elements. There are five hydrogen lines seen in the photographs, all of 

 which are broad and nebulous, and three of these are exceedingly strong. 

 The so-called cyanogen band A 3883-8, which remains in the spectra when 

 carbon is sparked without self-induction in nitrogen and in the other non- 

 metallic elements, has disappeared in the case of hydrogen. But considering 

 the facility with which hydrogen carries the current to the complete elimi- 

 nation of the lines due to the metallic electrodes, it does not seem remarkable 

 that this band, which in my belief is due to carbon, has disappeared. 



Principal Lines of Hydrogen. 



Oxygen. 



Oxygen was obtained by heating potassium permanganate. The gas 

 was purified and dried by being passed through water, tubes containing 

 solid caustic potash, pumice moistened with sulphuric acid, and then finally 

 bubbled through sulphuric acid. 



Many lines due to oxygen show in the spectrum, but a great number 

 of these disappear or are considerably weakened by the introduction of the 

 self-induction coil. With both carbon and gold electrodes some lines show 

 in the spectra which must be due to oxygen, although no record of their 

 measurements has been found. There are gold lines whose wave-lengths 

 correspond with these, but there being no gold present in the carbon 

 used it seems as if these so-called gold lines had been wrongly identified, 

 and that they were really due to the oxygen of the air in which the gold 

 had been sparked. These lines are A 2982-9 ; A 2382-5 ; A 2365-0 ; A 2352-7 ; 

 A 2300-6. These all occur with no self-induction, and in some cases faintly 

 with self-induction.' 



'Compare Hartley and Adeney's "Measurements of the Wave-lengths of Lines of High 

 Eefrangibility in the Spectra of Elementary Suhstances." Phil. Trans. Eoy. Soc, 1884, p. 63. 

 j)3^.n' SCIENT. PROG. E.D.S., VOL. XIII., NO. XXXIX. 4 Y 



