456 DR. K. B. DENISON AND DR. B. D. STEELE ON THE 



These results show that the method is capable of great accuracy. There is, indeed, 

 a slight difference between the values of the transport numbers for the normal 

 solutions according as acetate or formate is used as anion indicator, the percentage 

 difference being about the same with potassium chloride as with sodium chloride. 

 Although there is only a small difference between HITTORF'S value for these solutions 

 and that found by us with sodium acetate, it is too large to be attributed to errors of 

 experiment. We have measured these solutions repeatedly, and with the utmost 

 precautions, and invariably obtain results which lie between 0'508 and 0'510 for 

 potassium chloride and between 0'613 and 0'615 for sodium chloride. It is, therefore, 

 necessary to recognise the presence of some unknown disturbing factor in these 

 experiments. Whatever this factor is, its effect has disappeared in the more dilute 

 solutions, and for these accordingly we have continued to use sodium acetate as anion 

 indicator. For stronger solutions, however, it is advisable to use sodium formate. 



Sodium benzenesulphonate lias been successfully employed by us as aiiion indicator 

 in some experiments, but when it is used for stronger solutions thann/10, irregularities 

 occur, the cause of which we have not yet succeeded in tracing. 



Experimental Result*. 



We have measured the transport number and ionic velocities of those salts only 

 which give rise to strong ions, that is, salts which undergo little or no hydrolysis in 

 aqueous solution. 



The necessity of paying attention to the possibility of hydrolysis in all electrolytic 

 experiments cannot be too strongly emphasised. Thus, it is of no value to determine 

 the velocity of the ions of a salt which, when dissolved, gives rise to a complicated 

 ionic system. We cannot assume, a priori, that the transport number of any ion as 

 determined by the present method will give us the true fraction of the total current 

 which is carried by that ion in the presence of other ions of the same sign. In 

 HITTORF'S analytical method the actually measured transport number gives us the 

 fraction of the total current carried by the ion in question, whether complex ions are 

 present or not, if we assume that the current is wholly carried by the simple ions. 

 Whether the direct method, and the method of HITTORF, will give the same transport 

 number for a given ion in more complex ionic systems, or whether the presence of 

 complex ions in solutions will affect the results obtained by the two methods to a 

 different extent, are subjects for future experimental and mathematical investigation. 



By means of the present method we have been enabled to obtain for the first time 

 an experimental determination of the transport number for such salts as potassium 

 chlorate, bromate and perchlorate, which present considerable difficulty in the 

 determination by the analytical method. 



In the following Table II., which contains the results of our transport-number 

 determinations, the values given in the last column are taken from KOHLRAUSCH and 

 HOLBORN'S ' Leitvermogen der Electrolyte.' The salts used were obtained from 

 KAHLBAUM and were not submitted to any further purification. 



