634 
CHEMISTRY: LEWIS AND MINE 
ELECTRICAL CONDUCTION IN DILUTE AMALGAMS 
By Gilbert N. Lewis and Thomas B. Hine 
DEPARTMENT OF CHEMISTRY, UNIVERSITY OF CALIFORNIA 
Received by the Academy. October 17. 1916 
The modern theory of electrical conduction in metals, according to 
which the metal is dissociated to give a positive ion of low mobility, 
which is characteristic of the metal, and a negative carrier of high mo- 
bility, which is common to all metals and presmnably to be identified 
with the electron, has given rise to certain misgivings, despite the com- 
plete analogy between this theory and the accepted theory of electro- 
lytic dissociation. It is true that in the case of solid metals the crystal- 
line forces, which lead to the formation of the various components of 
the metal into symmetrical space-lattices, produce conditions which 
have no counterpart in liquid electrolytes, but in the case of a liquid 
metal like mercury we must assume that it differs from any electrolyte 
— we are tempted to say from any other electrolyte — only in as far as 
one of its ions has properties which differ very greatly in degree, but 
probably not in kind, from the properties of other ions. If therefore 
we should ultimately find that the Arrhenius theory of ionization is 
not applicable to a liquid metal we should be inclined to believe that it 
is not completely valid in the case of electrolytes. 
This theory of electrolytic dissociation, which was originally pro- 
posed to explain the properties of aqueous salt solutions, has since been 
successfully applied without essential modification to many types of 
non-aqueous solution. Moreover the extremely important work of 
Kraus,^ which unfortunately has not yet been published in full, demon- 
strates the applicability of the same theory to solutions of metals in 
non-metallic solvents. Thus, w^hen metallic sodium dissolves in liquid 
ammonia, he shov\^s conclusively that it dissociates into sodium ions 
and into electrons which, like other ions, are to a considerable extent 
combined with the molecules of the solvent. 
Since therefore the theory of ionization has been successfully applied 
to electrolytes and to solutions of metals in electrolytes, it seems desir- 
able to attempt a further extension of these ideas by several methods to 
solutions of metals in metallic solvents, in particular to solutions of the 
alkali metals in liquid mercury, and thus eventually to pure mercury 
itself. It has been shown by Lewis, Adams, and Lanman^ that the 
transference of matter with the electric current, which had previously 
been regarded as a distinguishing characteristic of electrolytic con- 
ductors, can be detected and measured in sodium and potassium amal- 
