222 Proceedings of Royal Society of Edinburgh. [may 16 , 
when no direct chemical action takes place between the metal and 
the products of electrolysis, or between it and substances dissolved 
in the liquid. The great difficulty in all experimental inquiry re- 
garding the resistance of electrolytes is the difficulty of distinguish- 
ing between the effects of true resistance and the effects of true 
reverse electromotive force of polarization. For, if E he the direct 
electromotive force producing a current x in a conducting circuit 
which includes an electrolyte and has resistance R, while e is the 
type of the polarization electromotive force, so that 5(e) is the total 
reverse electromotive force, we have 
E - 5 (e) = Raj . 
But, if e contains a term proportional to aj, we can write this equa- 
tion in the form 
E-5(e') = (R + 5(p))^, 
where p is the constant of proportionality. This shows that the 
value of the resistance in the circuit, as determined by any of the 
ordinary methods, may include a term which does not correspond to 
a true resistance actually existing. Hence, usually, in measuring 
the resistance, polarization is prevented by the use of alternating 
currents — as in Kohlrausch’s method ; or an attempt is made to 
eliminate its effects by keeping it constant while the resistance is 
varied — as in Horsford’s method. And, in addition to this, resist- 
ance at the surfaces of the electrodes must be distinguished from 
other resistance in the circuit. Attempts have been made to do this 
by means of measurements of the heat developed at the surfaces; 
but this method is unsatisfactor} r , as there may be development of 
heat at the surfaces from other causes than the presence of resistance. 
Method of Observation. 
In making experiments on the law connecting current-strength 
and time, when one Daniell cell was placed, along with a galvano- 
meter, in circuit with an electrolytic cell having platinum electrodes 
and containing a solution of sulphuric acid, I noticed that the result 
was very different according as the platinum plates had been newly 
heated to redness or had been left for some hours in the liquid. In 
the latter case the strength of current at a given time was, cceteris 
paribus , much weaker than in the former (see first series of experi- 
