the Electricity of Flame. 169 



160, and was finally constant at 170. That the former value 

 184 was not again reached after the evaporation of the water 

 may probably be accounted for by a slight impurity of the 

 (commercial) distilled water. 



Still more striking was the phenomenon when a solution of 

 potassium chloride was employed. Here likewise an instan- 

 taneous diminution of the electromotive force from 184 to 74 

 took place; and when the wire -«as once hastily drawn through 

 the flame so that small particles of fused potassium chloride 

 overspread it, the electromotive force fell quite to 16. This 

 value could not be increased by any shifting of the electrode; 

 so that the cause of this great diminution cannot possibly be 

 the unavoidable change of place of the electrode concerned. 



Finally, let us mention one more circumstance belonging 

 to this — namely that freshly annealed wires, used as electrodes 

 in the air stratum, always give higher values than those which 

 have remained a longer time (say ten minutes) exposed to the 

 air — the explanation of which behaviour, even in the case of 

 platinum, can only be found in an alteration of the quality of 

 the surface. 



§ 7. Repetition of the Experiments with Liquid Electrodes. 



In order to completely avoid the contact of the flame-gases 

 with metals, liquid electrodes, of the form represented in fig. 3, 

 were employed. By the pressure of the liquid column in the 

 glass tube B. a drop was pressed out of the fine aperture a, 

 which was then brought into the air surrounding the flame, 

 and as near as possible to its base. A U tube served to put 

 the flame of a Bunsen burner constructed entirely of glass 

 into communication with the earth, one leg of which ascended 

 the inner cavity of the burner. Both the glass electrodes were 

 filled with distilled water, into which clean platinum wires 

 (Pt) dipped. When the two water columns in the electrodes 

 were connected directly with each other, the electromotive 

 force called forth by the heterogeneity of the platinum elec- 

 trodes amounted, at the maximum, to 0*05 claniell. 



Now, in all the experiments, the electrode which was in air 

 was charged in the same sense as a metallic electrode ; it 

 likewise was positive, though the electromotive force was cer- 

 tainly much weaker. The mean values from three series of 

 experiments were the following: — 



E = 51, E = 51, E = 56, 



while two platinum electrodes at the same flame gave values 

 which lay between 150 and 180. 



A similar series of experiments were performed with a 



