536 Mr. J. Brown on the Difference of Potential 



5. In the course of these experiments it became very 

 evident that concordant and reliable results were difficult to 

 obtain, partly on account of the smallness of the forces to be 

 measured, and partly from a difficulty in ascertaining the 

 exact origin of these forces so as to arrange the experiment 

 in accord with them. Even the sign of the electrification 

 sometimes reversed during the experiment, as in Nos. 12 and 

 21. The results must therefore be taken as only tentative, 

 but in so far as they may be considered approximate they 

 suggest the following conclusions : — 



(i.) There is no clear connexion apparent between the 

 observed electromotive forces and the heat equivalents of the 

 reactions ; but the amount of water present seems to have an 

 effect, either by the dilution of both solutions equally or one 

 more than the other. 



(ii.) The sign of the electrification may depend, however, 

 on the following considerations. When double decomposition 

 occurs at contact of solutions of two substances, one of which 

 consists of an anion a and cation c, and the other of an anion 

 a' and cation c', the products of the reaction being ad and a'c, 

 then the solution ac will be positive if the combining heat 

 a',e is greater than c',a ; and vice versa. 



This hypothesis appears to conform to the following view 

 of the kind of molecular action that may take place. In the 

 case, for instance, of the reaction 



FeCl 2 + K 2 S = K 2 Cl 2 + FeS, 



the heat equivalent K 2 ,C1 2 being greater than Fe,S, the 

 molecules just before actually double-decomposing might be 

 supposed to become polarized by the attraction between K 2 

 and Cl 2 , and so cause an electrification of the opposite ends 

 of the system, 



+ FeCl 3 K 2 S-, 



similar to that commonly associated with such a system as 

 + FeCl 2 potassium — . 



Although this hypothesis receives support from nearly all the 

 experiments in Table II., and also from several of the pre- 

 liminary experiments not given in the table, it can scarcely 

 be regarded as more than tentative, pending results of 

 experiments under conditions of a more varied character. 



6. In most of the numerous investigations already pub- 

 lished dealing with contact potentials of liquids the potentials 

 of the liquids were tested by means of a metallic connecting 

 pole of platinum, mercury, or other such substance. The 



