PHYSICAL CHEMISTRY ' 197 



deriving yi^, but unfortunately the values obtained are in 

 disagreement with the precision measurements of Washburn 

 and Wieland at high dilutions. Kendall points out that the 

 Ghosh theory fails in all solvents with a low dielectric constant. 

 In such solvents ionisation should proceed exceedingly rapidly 

 with increase in the dilution ; even the specific conductivity 

 should increase with the dilution. In practice, however, 

 exactly the opposite behaviour is exhibited. 



Kraus shows by graphical methods that the equations of 

 Ghosh can fit the experimental results over only a small range 

 of concentrations. Thus Log yj plotted against the cube 

 root of the concentration should he on a straight line if Ghosh's 

 theory were applicable, whereas the experimental points lie 

 on a curve concave towards the axis of concentrations at low 

 concentrations and convex towards this axis at high concen- 

 trations. Another weakness of the theory is that it ignores the 

 possibility of any specific influence of the solute on the solvent, 

 such as any change in the dielectric constant of the solvent 

 which may have occurred after the solution of the electrolyte. 

 Kendall would go further and ascribe conductance to the 

 solvent. He considers that both the Arrhenius theory and 

 the theory of complete ionisation are at fault in referring 

 conductance in solution entirely to the solute. 



These weighty objections do not invalidate the whole of 

 the theory of Ghosh. It would appear, however, that modifi- 

 cations in the original postulates will be necessary if the new 

 line of attack on electrolytic solutions is to prove fruitful. 



The Constitution of Metallic Substances. — Kraus, in a recent 

 communication {J.A.C.S., 1922, 44, 12 16), has thrown new 

 light on the properties of metallic substances. According to 

 modern views, metals owe their characteristic properties 

 chiefly to the presence of freely moving valence electrons. In 

 the formation of a salt from a metal, these electrons are given 

 up to an electronegative element or group, and the metallic 

 properties disappear. According to Kraus, metallic elements 

 are salts in which the electronegative group is replaced by the 

 negative electron. Support may be gathered for this view 

 from the behaviour of metals when dissolved in Hquid ammonia. 

 These solutions have been shown by Kraus (see Science 

 Progress, 1921, 62) to exhibit properties intermediate between 

 those of metals and electrolytes. At low concentrations, even 

 the alkah metals lose their metalHc properties. Thus a dilute 

 solution of a metal is virtually a solution of a salt giving 

 positive and negative ions resembling those of typical electro- 

 lytes. The anion, under these conditions, consists of a nega- 

 tive electron which is rather loosely associated with ammonia 

 molecules. With increase in the concentration of the metal, 



