MEASUREMENT OF LIQUID RESISTANCES. 393 



consequently 



x + p = x' + p, 

 or 



x-x' = p -p. 



This method is not, however, free from objections; the deposits 

 of gas bring about changes in resistance which it is impossible to 

 allow for, and there is no certainty that the polarization of the 

 electrodes is constant. 



990. The difficulty disappears if we deal with solutions of 

 metals, and use, as electrodes, plates of the metal of the solution. 

 In this case polarization is always very slight, if not null. Pouillet* 

 measured in this way the resistance of solutions of copper sulphate, 

 of zinc sulphate, etc. He used a cylindrical tube divided into 

 parts of equal length. This tube was closed at the bottom by a 

 plate of the same metal, and a wire of the same metal fixed in a 

 glass tube so that only the bottom was in connection with the 

 liquid, could be adjusted at a variable distance from the bottom. 

 Equal displacements of the liquid obviously represented columns of 

 the same resistance. This arrangement forms a convenient rheostate. 



M. Paalzow f has generalised the method. The liquid to be 

 studied is contained in a U-shaped tube, the legs of which dip 

 in porous vessels filled with the same liquid ; the two porous 

 vessels are placed in larger vessels containing a solution of zinc 

 sulphate and two electrodes of amalgamated zinc. In order to 

 compare two liquids, the porous vessel and the siphon are filled 

 successively with the liquid. It is clear that in these conditions 

 the polarization of the metal electrodes is got rid of, but we do 

 not avoid, at all events completely, the variations of electromotive 

 force at the surfaces of contact of the liquid with the zinc sulphate 

 through the porous vessel. 



991. Measurement of the fall of potential between two given 

 points (933) gives a very simple method, and one not open to 

 objection. | The liquid is contained in a cylindrical tube, closed 

 at its ends by metal plates of the same section as the tube, and 

 which serve as principal electrodes. The flow of electricity may 

 be considered uniform, and the equipotential surfaces as perpen- 

 dicular to the axis of the tube. 



* POUILLET. Comptes rendus, Vol. IV., p. 786. 1837. 

 t PAALZOW. Pog, Ann., Vol. cxxxvi., p. 419. 1869. 

 % BRANLY. AnndeTEcole Norm. [2], Vol. n., p. 209. 1873. LIPPMANN. 

 Comptes rendus, Vol. LXXXIII., p. 192. 1876. 



