567 



the electrode material is plajing here a prominent part. That melting- 

 point and evaporation point of the metal are not decisive, has already 

 been stated bj' Morse. 



(II) Anodic Luminescence. 



§ 5. As might be expected, the phenomena at the anode are much 

 more numerous and much more complicated than those at the 

 cathode. Besides gas layers, also layers of solid substance can esta- 

 blish themselves here between electrolyte and electrode, thus causing 

 luminescence. The sparks which appear in valve cells at the limiting 

 tensions (10), have not been examined in what follows. 



According to the nature of the emitted light and the cause of the 

 luminescence at the anode, the following typical cases of lumines- 

 cence can be distinguished. 



1. Line- and band-spectra; to a certain extent these are very 

 similar to those at the cathode, but they are generally much weaker. 



2. Arc-spectra, equal to those at the cathode, but which can but 

 rarely be obtained, and then only on definite conditions. 



3. Generally a yellowish luminescence — which in so far as this 

 can be ascertained, is spectroscopically continuous, — without forma- 

 tion of a layer of oxide or anything of this kind. The anode metal 

 (or the carbon used as anode) gets shinj' or bright. 



4. For so far as this can be ascertained a continuous emission, 

 with a maximum of intensity in a definite spectrum region; in this 

 case the formation of solid layers at the anode always takes place. 



First of all we will give some instances and some further parti- 

 culars of the phenomena in each of these four classes. 



^ 6. 1. Already Voi.ler and Walter record that at an intermptor 

 anode from platinum in sulphuric acid 1 : 40, they obtained — by 

 the side of the NaD-line — a faint band spectrum. If this cell 

 contained sulphuric acid and also metal salts, the lines of these 

 metals also appeared. The data of these investigators could be fully 

 confirmed; no more than they, did 1, however, succeed in deter- 

 mining more accurately the band-spectrum lying in the green '). The 

 intensity of the phenomenon was, indeed, too small for spectroscopic 

 investigation, though it was always clearly perceptible, also in aqueous 

 potassium hydroxide 1:10, and on use of other anode metals. Special 

 phenomena were obtained on use of platinum anodes in sulphuric 

 acid 1 : 40, containing at once several metallic salts. 



In order to obtain anodic metal lines, greater quantities of metallic 

 salt must in general be dissolved in the acid. Even then mostly a 



1) Very probably these „bands" belong the oxygen spectrum. 



