573 



anionic electrons, wliicii fly through the anode layer with strong 

 acceleration. By this the matter in this layer is brought to Inmines- 

 conce in the same way, hence with emission of the same spectrum, 

 as tliis would happen by means of cathode rays. If the layer becomes 

 too thick, higher potentials will be required to bring about a passage 

 of the current, and finally current could only pass in certain 

 cases when the layer is traversed by sparks (limiting tension with 

 anodes). When on the other hand the layer is attacked by the 

 electrolyte in some way or other, it is very well possible that also 

 the light emission at the anode can vary locally, and in particular 

 the periodic oscillations of the intensity along the anode become 

 possible. Increase of temperature will always hamper the lumines- 

 cence, either because the solubility of the anode product is in 

 general increased by it, or because the layer is rendered less stable 

 by it in mechanic respect. If the anode layer has little mechanical 

 stability in itself (e.g. mercury iodide), a certain minimum current 

 density will be required to form a coherent layer with sufficient 

 velocity, and to allow this to continue to exist, in spite of continued 

 decomposition. 



By this conception also the analogy between the anodic and 

 catliodic luminescences is clearly brought ont. 



Summarizing we may say that also in these anodic luminescence 

 phenomena, as this was earlier shown for ordinary chemi-lumines- 

 cence (5), not the anode-reaction in itself takes place with light 

 emission. 



It must rather be admitted that first reaction products are formed 

 which are brought to emission, in this case by means of the electric 

 energy of a source of current outside the system examined '). Hence 

 there is no question of an ion reaction, which takes place with light 

 emission, and of a departure from the general rule that it is just 

 these reactions, which proceed practically with iiilinite velocity, that 

 are never accompanied by a light-emission. 



The above considerations show further that VoLTA-luminescence 

 occurs very frequently, but also that it can be of a very different 

 character. On further investigation of these phenomena it will be 

 necessary to distinguish these kinds of VoLTA-luminescence scrupu- 

 lously. The present investigation may be considered as a first attempt 

 at reconnoitring the ground in this respect. 



') In cases of common chemi-luminescence the reaction itself furnishes the 

 energy necessary to excite to light emission some of the kinds of molecules present 

 in the system, (see 5). 



