in Liquids free from Gas. 149 



true, be directly proved chemically for platinum by Graham ; 

 yet the phenomena of polarization to be described in the sequel 

 seem to indicate that perfectly similar relations subsist for oxy- 

 gen as for hydrogen, and that only the quantity of oxygen that 

 can be occluded by platinum is much less than that of hy- 

 drogen. 



If, then, an electric current traverses a water-decomposition 

 cell whose liquid holds hydrogen in solution, or whose platinum 

 electrodes have occluded it, at that electrode to which the cur- 

 rent urges the oxygen this may again become water, a corre- 

 sponding quantity of dissolved hydrogen from the water, or of 

 occluded hydrogen from the electrode, being consumed for that 

 purpose. On the other hand, instead of this hitherto free 

 hydrogen (at least not chemically combined with oxygen), an 

 equal amount of electrolytically separated hydrogen will reap- 

 pear at the other electrode, and either be dissolved in the liquid 

 or, if there be time and space for it, even be driven into the pla- 

 tinum electrode. Although herein electrolysis therefore takes 

 place in the liquid, yet the two products of the electrolysis do 

 not finally appear, but the last result is that free hydrogen dis- 

 appears at or in one of the electrodes, and at or in the other ap- 

 pears in increased quantity. Tor this process, which plays a 

 prominent part in polarization- currents, I would propose the 

 name electrolytic convection. Hence, in this process, the elec- 

 tromotive force which drives the current has not also to perform 

 the work against the forces of the chemical affinity of hydrogen 

 and oxygen, which must be performed when water is to be finally 

 separated into these its two elements ; and therefore electrolytic 

 convection can be kept up by a feeble electromotive force which 

 is quite inadequate actually to decompose water — for instance, 

 the force of only a single Daniell element. 



The same thing holds when the liquid contains oxygen, or if 

 the platinum plates have occluded it. Then, through electro- 

 lytic convection, free oxygen disappears on one side, while the 

 same amount is forthcoming on the other. 



The hydrogen or oxygen in this way set free in the process of 

 convection at one electrode, is, so far as it is not occluded in the 

 electrode, evidently as free to diffuse itself in the liquid, to be 

 carried away by currents of the same, and to be developed 

 as gas when the liquid is saturated, as the gases which are de- 

 veloped by ordinary electrolysis. While it is being diffused in 

 the liquid it may also arrive again at the other electrode to fall 

 again under electrolytic convection, and in this manner may 

 keep up in a continual circulation a certain intensity of electric 

 current. 



A Daniell element can thus, in a water- decomposition cell 



