566 



first cut out, and then immediately inserted again, the luminescence 

 quickly continues without it being necessary to take the electrode 

 out of the liquid and immeising it again. If, however, we wait a 

 sliort time after the cutting out, a hissing sound is lieai'd after about 

 2 or 3 sec, and now the chromoscope is not at once luminescent 

 again when it is inserted. Moreover the experiments of Riesen- 

 FELD and Pfützer (11) have described, plead still more in favour 

 of the existence of a gas layer. There a small of light arc is formed 

 between cathode and liquid, and I could verify that the same spectral 

 phenomena are obtained as in the chromoscope. On use of Pt- or 

 Ir-cathodes, the metal to be detected being present in the liquid, 

 a spark-spectrum is obtained; when, however, the metal is used as 

 cathode with pure acid, and arc-spectrum. 



^ 4. Probably the following idea must be formed about the origin 

 of these cathodic luminescence phenomena. Between electrode and 

 electrolyte there is formed a gas envelope containing hydrogen, 

 water-vapour and some oxygen-, within this layer lies almost the 

 whole fall of potential ot the cell. The cations not being able 

 to traverse this layer, there a current of rapid cathode rays is formed, 

 which discharge these cations. The discharged metal atoms now 

 get into the gas layer, and are excited to the emission of a 

 spark-spectrum by collision with similar tlying electrons. 



The spraying of the cathode is greatly promoted by the impact 

 with positive particles. If, as in the metal chromoscope, the current 

 density and the strength of the current intensity are relatively high, 

 also uncharged atoms of the electrode metal get into the gas layer 

 — either because the spraying consists primarily in a scattering of 

 molecular particles, or because locally a sufficiently high tempera- 

 ture arises — , and then an light-arc is formed and hence an arc- 

 spectrum is observed. 



If the electrolyte at the same time contains a sufficient number 

 of ions of the electrode- or another metal, a spark spectrum of the 

 second metal can of course appear by the side of the arc-spectrum 

 of the first metal. This is, however, not necessary. Depending 

 upon the nature of the electrode metal, the arc-spectrnm is more or 

 less apparent. Thus Moksiï showed already that the spectrum of a 

 platinum cathode is intense in solutions of acids and alkalies, but 

 very faint in solutions of earth alkalies, while a strong aluminium 

 (arc-)spectrum appears with an aluminium electrode in almost any 

 electrolyte. The relations that are valid here must, however, still 

 be examined ; possibly the greater or less tendency to spraying of 



I 



