the Solvent in Electrolytic Conduction. 389 



mines at different times and temperatures its electrical con- 

 ductivity. He makes a special study of ferric chloride, and 

 shows that there is a considerable fall in its resistance-value ; 

 he does not chemically estimate the amount of the change, 

 nor give the products of the change. Such a chemical decom- 

 position I had suggested previously as the cause of the con- 

 ductivity-values for ferric-chloride solution not being inversely 

 proportional to the dilution. Foussereau considers this 

 change under the influence of temperature, and my results 

 would show that the change increases with the dilution of the 

 solution. Such, too, may be the reason for the peculiar 

 behaviour of the alcohol solutions to which I have already 

 drawn attention. 



This decomposition of the chlorides, and probably of other 

 salts in solution, greatly complicates the measurement of 

 conductivities of salt-solutions. It affects also the question of 

 the change of conductivity with temperature. 



There are many salts for which such decomposition is 

 entirely negligible. I have determined the temperature- 

 coefficients of a number of different solutions ; the result of 

 these experiments I hope to present on a future occasion. The 

 resistances were measured up to temperatures of from 80° to 

 100°; and after cooling down, the resistance-value for the 

 temperature of the air was found not to have altered ; in 

 these cases there could have been no definite chemical change 

 in the solution ; the conductivity-values were found to be 

 proportional to the temperature. 



This decomposition of the salt by the solvent perhaps throws 

 light on the alteration in the value of the temperature-coeffi- 

 cient with dilution. 



Dr. Armstrong considers this change to be due to the fact 

 that in concentrated solutions complex molecules exist, which 

 are decomposed in dilute solutions ; for certain salt-solutions 

 the temperature-coefficients decrease with dilution. Bouty*, 

 in the Journal de Physique, gives the case of lead nitrate ; 

 and I have found the same for mercuric-chloride solutions. 



If, under the action of the solvent, there are at high tempe- 

 ratures definite chemical changes, it seems impossible to hold 

 the view of Arrhenius, that the action of the solvent is to 

 cause the simple dissociation of the salt ; it is difficult to believe 

 that in a moderately dilute solution 90 per cent, of hydro- 

 chloric acid should be dissociated, the atoms of H and CI exist- 

 ing free, and that only 1 per cent, of acetic acid should be so 

 dissociated. 



* Jovrnal de Physique, Jan. 1887. 



