Brush Discharge in Weak Acids and Solutions. 819 



The negative brush in the salt-solutions of lithium, sodium, 

 and potassium is the same in character as that in dilute acids. 

 The potential required to produce it is smaller than that re- 

 quired to produce the positive brush, the evolution of gas is 

 only slightly more than the calculated amount (see Table II. 

 for *05 per cent, solution K 2 C0 3 ), and its appearance in the 

 rotating mirror is a series of unbroken bands. That there is 

 a change in the electrical conditions of the brush is suggested, 

 however, by the behaviour of the nitrogen spectrum (referred 

 to below). There it will be shown that the negative bands of 

 nitrogen behave in a manner almost identical with that of 

 the elementary line spectrum of oxygen. It was shown by 

 Fulcher * that these bands are produced by more rapidly 

 moving cathode rays than those which produce the positive 

 bands. The connexion between the oxygen spectrum and 

 that of the metal of the dissolved salt (lithium, sodium, or 

 potassium) was pointed out above. This connexion indicates 

 that the spectrum depends in some way on the concentration of 

 the metallic ion at the negative point. In the interval during 

 which no current flows, the metal liberated will form a certain 

 amount of hydrate in solution with the production of hydroxyl 

 ions. As these ions are not swept away from the point imme- 

 diately the brush occurs, and as they are constantly being 

 renewed, the negative brush will take place in a region con- 

 taining a large number of hydroxyl ions. These ions, as the 

 source of oxygen, may possibly play an important part in 

 the production of the spectrum. 



Water Vapour Bands. 

 The ultra-violet bands of water vapour (discovered simul- 

 taneously by Huggins f a nd by Liveing & Dewar J) appear 

 strongly in all the spectra and occur equally well in each 

 brush. There is no correspondence' between the strength of 

 the bands and the spectra of hydrogen and oxygen. They 

 appear equally well in acids and in alkalies, and so are un- 

 influenced by an excess of hydrogen or hydroxyl ions. They 

 do not depend, as the hydrogen spectra do, upon a high 

 potential to produce them strongly. If the current in the 

 transformer primary is kept constant, then as the water 

 increases in conductivity the bands increase in intensity as 

 the current through the brush increases and the potential 

 difference between the electrodes falls. The intensity of the 

 bands is not proportional to the current only ; a larger 



* Astrophysical Jo urn. vol. xxxvii. p. 60 (1913). 

 f Proc. Roy. Soc. vol. xxx. p. 516 (1880). 



% Proc. Roy. Soc. vol. xxx. p. 580 (1880) ; vol. xxxiii. p. 274 (1882) ; 

 vol. clxxix. p. 27 (1888). 



