fluid conductors when transmitting the electric current . 177 
plicated effect, but capable of easy explanation. The oxida- 
tion takes place over the surface of the metal before the last 
portions of sodium are removed. This is easily proved. 
We have only to break the circuit altogether, and the crust 
of oxide will gradually disappear ( unless suffered to go too 
far), being reduced by the sodium beneath it. Were it then 
not for the crust of oxide, the currents, as has been seen, 
would be in a positive direction. But the oxide, acting on 
the stratum of metallic molecules immediately below it, de- 
prives them of their alloy, which it converts into alkali, leav- 
ing a stratum of pure mercury. Now we have seen that in 
this , the rotation, in the circumstances of the experiment, 
would have a negative direction. We have only then to ad- 
mit that the peculiar action by which the rotations are caused, 
is confined to the common surface of the mercury and liquid, 
to have a perfect idea of the mode in which the whole pro- 
cess is carried on. The stratum of pure mercury on the 
surface is removed by a negative current agreeably with its 
natural relations, and immediately succeeded by a stratum of 
the sodiuretted metal from the interior ; this, in its turn,, is 
deprived of its sodium by the oxide in contact with it, and is 
immediately radiated off like its predecessor, and so on till 
the whole crust of oxide is exhausted or swept off, when the 
remaining mercury, still retaining an excess of sodium, and 
instantly rendered homogeneous, is acted on as an alloy in 
the way already described. 
25. That sodium is actually present in the mercury when it 
has acquired the property of producing currents from the posi- 
tive pole, (which for brevity I will hereafter call the positive 
property) by contact with the negative wire, may be shown by 
mdcccxxiv. A a 
