458 DR. R B. DENISON AND DE. B. D. STEELE ON THE 



It will be seen from the table that our results agree excellently with those 

 obtained by HITTORF'S method, the difference in most cases being only a few units in 

 the third decimal place. One of the most marked exceptions is n/l sodium chloride, 

 for which we find 0'621 instead of 0'637. We have repeated this experiment a 

 number of times and always with the same result. Possibly the older value requires 

 confirmation. 



KOHLRAUSCH and HOLBORN give the same value, 0'508, for the anion transport 

 number of potassium chloride, bromide, and iodide. On the other hand, we have 

 found the following values : 



Potassium chloride, p = 0'508. Potassium bromide, p = 0'518. 

 Potassium iodide, p = 0'513. 



As these results were so different from what we expected, we have measured the last 

 two salts several times and invariably with the same result, so that we were at last 

 convinced that our numbers were correct. We have since found a paper by 

 KOHLRAFSCH (' Zeit, fur Electrochemie,' 1902, VIII, p. 630) which had escaped our 

 notice. In this paper KOHLRAUSCH describes experiments showing that the velocities 

 of the ions Cl', Br', and P, as determined from conductivity measurements, are as 



follows : 



Cl = 65-44. Br = 67'63. I = 60'40. 



These figures give for the transport number of the three salts the following 



numbers : 



Potassium chloride = 0'502, Potassium bromide = 0'511, and 

 Potassium iodide = O'SOl, 



which is practically the same ratio as that found by us. The difference between the 

 actual values is due to KOHLRAUSCH'S figures being for infinite dilution, whereas ours 

 are for n/50 solutions. 



Further confirmation of the correctness of our results is afforded by the fact, which 

 will be referred to again later, that the potassium ion has the same velocity in 

 equimolecular solutions of the above three salts if our values for the transport 

 numbers are correct, but its velocity is different if the ordinary values are correct. 



Measurement of the Ionic Velocities. 



The terra " ionic velocity " appears to have been used by different writers in 

 different senses, and as a consequence there is some confusion as to its exact 

 significance. In what follows we shall speak of: 



(1) The actual measured velocity of the ion or ionic margin, which is its velocity 



under the potential gradient of the experiment. 



(2) The actual mobility (U or V) or the velocity of the ion in a given solution 



under a potential gradient of 1 volt per centimetre. 



