Density of the Electrolyte at the Electrodes. 281 



to stop its evolution, and thus the concentrated acid will be 

 carried up by the current of oxygen and escape observation. 

 But if the electrode has been filled with hydrogen by previously 

 serving as kathode, the evolution of oxygen on reversal of 

 the current is prevented either by its direct union with the 

 occluded hydrogen, or by the union of the ion S0 4 with the hy- 

 drogen. No gas being evolved, the concentrated acid is able to 

 flow downwards in streaks. This explanation is borne out by 

 the experiments with sulphuric acid detailed above, in which it 

 was seen that, on placing freshly-heated palladium electrodes 

 in sulphuric acid, gas is immediately evolved at the anode ; 

 but if that electrode has been previously rilled with hydrogen, 

 the oxygen, on reversal of the current, does not appear for 

 some time at the anode, and the streaks are visible during this 

 period. 



Concentration of the solution at one electrode is accom- 

 panied by a weakening of the solution at the other elec- 

 trode. Therefore, as in acids the weakening takes place 

 at the kathode (at which the gas is absorbed), we should 

 expect to see streaks ascending from that electrode without a 

 previous reversal of the current. On trying the experiment 

 with palladium electrodes placed horizontally in dilute sul- 

 phuric acid, this supposition is found to be warranted. The 

 reason why the streaks ascending from the kathode were not 

 seen in our earlier experiments is, that the electrodes had 

 been placed vertically with their top edges a little below the 

 surface of the electrolyte, and so the weakened solution in 

 ascending had naturally clung to the surface of the electrode 

 and thus escaped detection. 



The late Professor Christiani (in a work* to which we shall 

 refer more fully in a subsequent paper) gives three instances 

 in which he observes streaks from the electrodes, but does not 

 offer any explanation of their mode of formation. He cites 

 the cases of zinc electrodes in concentrated zinc sulphate and 

 copper electrodes in concentrated copper sulphate"}". In these 

 cases, without a previous reversal, he observed streaks de- 

 scending from the anode and ascending from the kathode 

 at the same time. These results are evidently in accordance 

 with the theory proposed above : no gas is given off at either 

 electrode, and so the concentrated solution is allowed to de- 

 scend in streaks from the anode and ascend in streaks from 

 the (horizontal) kathode. 



* " Ueber irreciproke Leitung electrischer Strome," 1876, H. Fried- 

 lander ; Wied. Meet. ii. p. 727. 

 t Christiani, he. cit, p. 100. 



