138 DR. B. D. STEELE, DR. D. MoINTOSH AND DR. E. H. ARCHIBALD 



PART III. 



The Transport Numbers of Certain Dissolved Substances. By B. D. STEELE. 



THE strikingly abnormal variation ot molecular conductivity with dilution that we 

 have found to occur in solutions in the liquefied halogen hydrides finds a possible 

 explanation in the assumption that it is the solvent and not the solute which is 

 ionised. As the transport number of the dissolved substance might be expected to 

 yield information not only as to the correctness of this assumption, but also as to the 

 constitution of the electrolyte, the transport numbers of a few substances have been 

 measured, and the results are given in the following pages. 



The only measurements of the migration ratio which have hitherto been made in 

 solvents other than water are those of a few salts in methyl and ethyl alcohol, and ol 

 silver nitrate in pyridine and in acetonitrile. 



Direct measurements of the velocities of certain ions in liquefied ammonia have 

 recently been made by FRANKLIN and CADY ('Journal of Amer. Chem. Soc.,' 1904, 

 vol. 26, p. 499), who used a modification of MASSON'S method (' Phil. Trans.,' A, 1902, 

 vol. 192, p. 331). 



Method of Measurement. 



It has been shown by the author (STEELE, 'Phil. Trans.,' 1902, A, vol. 198, p. 105) 

 that the direct method of measurement gives trustworthy results only when the salt 

 under examination is of the simplest type. Now HITTORF has shown that in 

 alcoholic solution cadmium iodide and certain other salts are dissociated into ions 

 which are much more complicated than those occurring in aqueous solutions of the 

 same concentration. 



The only substances which we have found to be capable of forming conducting 

 solutions in any of the solvents which we have been investigating are certain organic 

 compounds, and although the nature of the ions into which these dissociate is entirely 

 unknown, it is probable that the ionisation is even more complicated than that of 

 cadmium iodide dissolved in alcoholic solution. 



From these considerations it was decided to use HITTORF'S method, notwithstanding 

 the fact that it is much more tedious and presents greater experimental difficulties 

 than the alternative method of direct measurement. 



HITTORF'S method consists in the analysis, after electrolysis, of the solution which 

 surrounds one of the electrodes. The original concentration being known, the actual 



