76 Mr. J. Brown. [June 16, 



The probability of sach actions does not, I think, usually receive 

 sufficient consideration in electrochemical investigations. 



Again there is, in connection with the solvent, the uncertainty 

 whether, in calculating the mechanical equivalent of the action in 

 the cell, we should include the heats of solution of substances formed 

 by the action. Where the product is easily soluble, the heat of 

 solution is probably adjuvant ; where insoluble, there is of course no 

 heat of solution adjuvant; where it is soluble with difficulty, the 

 heat of solution may be partly or to some unknown extent ad- 

 juvant. 



The consideration of these difficulties, and especially the complica- 

 tion probably caused by the action of the water on magnesium and 

 aluminium, as mentioned above, led me to the consideration of a 

 much simpler form of cell, in which the solvent with all its com- 

 plications is abolished, and the simple electrolyte in the fused state 

 used alone. 



The number of possible combinations of this kind is limited by 

 the considerations that the electrolytes must be fusible at reasonable 

 temperatures ; and that the boiling point of any component in a cell 

 must not be below the fusing point of the electrolyte in contact with 

 it. A few of the haloid compounds of the metals are suitable ; and 

 the chlorides were chosen to begin with, as in some respects best 

 known. 



The results obtained show at all events the direction in which to 

 look for an exact accordance between theory and experiment. 



I. Cells with Simple Chlorides. 



The cells were of course of the two-fluid type, each metal being in 

 contact with its own chloride only, while the two chlorides were in 

 good contact with each other. The type may be represented for in- 

 stance as 



Zinc | ZnCl 2 | MgCl 2 | Magnesium. 



We cannot say the fused electrolyte is always free from the com- 

 plication of dissolved oxygen, for in some cases this makes itself 

 evident by the formation of oxychlorides. This very formation of 

 oxychlorides may, however, remove the oxygen from the sphere of 

 action. 



We must admit also that in some of the combinations tried, action 

 goes on with open circuit, since several of the metals, notably iron 

 and copper, were somewhat corroded after immersion in the fused 

 salt. 



Further, while the difficulty about the heat of solution of the 

 nascent salts is avoided, there is a question as to whether the heat of 

 solution of one fused chloride in the other, i.e., the heat of combina- 



