150 DR. B. D. STEELE, DR. D. McINTOSH AND DR. E. H. ARCHIBALD 



For many solvents the substances which dissolve to form conducting solutions may 

 be broadly designated as those which enter into combination with the solvent. 



Thus the metallic salts as a class are characterised by their tendency to form 

 compounds with water, while non-saline organic bodies as a class are not able to form 

 such compounds.* Many, but not all, salts which form ammonia compounds dissolve 

 in ammonia to form conducting solutions. 



Compounds of the solute with the solvent are also clearly indicated in the case of 

 many conducting solutions in sulphur dioxide, and WALDEN and CENTNERSZWER 

 (' Zeit. Phys. Chem.,' 1903, 42, p. 432) have isolated and investigated two such 

 compounds containing potassium iodide and sulphur dioxide. 



In the halogen hydrides we find that the only substances which conduct are the 

 amines, alcohols, ethers, ketones, &c., all of which are able to enter into combination 

 with the solvents. Many similar cases have been observed amongst organic solvents, 

 and as an example of these reference may be made to solutions in amylamine 

 (KAHLENBERG and RUHOFF, ' Journ. Phys. Ch.,' 1903, 7, p. 254). 



The study of the behaviour of aqueous solutions has led to ARRHENIUS' theory 

 of ionic dissociation and to VAN 'T HOFF'S theory of solutions ; and numerous 

 investigations have been undertaken with the object of testing these theories, when 

 applied to solutions in non-aqueous solvents. 



As a result it has been found that, as required by the theory, most substances, 

 when dissolved in ammonia, sulphur dioxide, hydrocyanic acid, and some other 

 solvents, show an increase of the molecular conductivity, /j., with dilution, but that 

 the opposite change occurs in solutions of a few substances in the same solvents. 

 This difference in behaviour cannot therefore be conditioned by the nature of the 

 solvent only, although if we consider the inorganic hydrides as solvents, we find 

 that n varies normally, that is to say, increases with dilution, in solutions in water 

 and ammonia, hydrides, namely, of elements in the first series of the periodic table, 

 whereas /A decreases with dilution in solutions in the remaining hydrides, the variation 

 therefore being abnormal, t 



The results of the molecular weight determinations in non-aqueous solvents are, as 

 a general rule, not concordant with the conductivity results, many conducting 

 solutions being known in which, contrary to expectations, the dissolved substance 

 appears to be associated. 



* The view that compounds of the solute and the solvent exist also in solutions appears to be steadily 

 gaining ground, see MORGAN and KANOLT ('Jour. Amer. Chem. Soc.,' 1904, 26, p. 635) and JONES and 

 GETMAN (' Zeit. Phys. Chem.,' 1904, 49, p. 390). 



t Amongst others the following cases have been observed of solutions in which the molecular 

 conductivity decreases with dilution : Silver nitrate, cadmium iodide, and ferric chloride in amylamine 

 (KAHLENBERG and RUHOFF, 'Jour. Phys. Chem.,' 1903, 1, p. 284); Antimony bromide and phosphorus 

 pentabromine in bromine (PLOTNIKOFF, 'Jour. Russ. Phys. Chem. Soc.,' 1902, 34, p. 466 ; 1903, 35, p. 794) ; 

 Hydrogen chloride in ether and in amyl alcohol (KABLUKOFF, 'Zeit. Phys. Chem.,' 1889, 4, p. 429); 

 Hydrogen chloride in cineol (SACKUR, 'Ber. D. Chem. Ges.,' 1902, 35, p. 1242), &c. 



