36 Professor A. Orum Brown [Jan. 31, 



salt is constant ; therefore any change in the strength of the current 

 means a corresponding change in the molecular conductivity of the 

 dissolved salt. The molecular conductivity, therefore, increases with 

 the dilution, and asymptotically approaches a maximum. 



I cannot here enter into a description of the great experimental 

 difficulties connected with the determination of the conductivity of 

 extremely dilute solutions, but I may refer to one of them, namely 

 the small but variable conductivity of the water used in preparing 

 the solutions. There seems now to be no doubt that water is in 

 itself an electrolyte. But the purest water that has been obtained 

 has a conductivity of only about 10" ^° as compared with that of 

 mercury as unit. The minutest traces of salts greatly increase 

 the conductivity, so that ordinary distilled water has a conductivity 

 of 3 X 10"^'' or more. With solutions of moderate dilution the 

 variation of this very small quantity is of little consequence, but 

 with extremely dilute solutions the conductivity to be measured is of 

 the same order as that of the water. 



For our present purpose the most important conclusion drawn by 

 Kohlrausch from his observations is his law of the independent rate of 

 motion of the ions in dilute solutions. The rate of motion of any ion 

 towards the electrode depends on the gradient of potential. But 

 Kohlrausch shows that the rate of motion of each ion in dilute solution 

 is proportional to a number, the same whatever be the other ion of the 

 electrolyte. Thus the rate at which the cation K moves towards the 

 cathode in dilute solution, is the same in solutions of KCl, KNO3, 

 KC2H3O2, etc. Kohlrausch gives these numbers for six cations and 

 ten anions. The results calculated from these numbers agree well 

 with the observed conductivities. 



Methods have been devised for directly observing and measuring 

 the rate at which ions travel. In this connection I may specially 

 mention the names of Oliver Lodge, Whetham, and Masson. These 

 measurements agree very well with the rates calculated by 

 Kohlrausch. 



I now show an experiment indicating a way in which such 

 measurements can be made. 



The apparatus * consists of a glass U-tube, with a long stopcock- 

 funnel connected to the lower part of it. The tube is nearly half 

 filled with a dilute (about 0*03 per cent.) solution of potassium 

 nitrate, and then about the same quantity of a solution of potassium 

 permanganate, of the same conductivity as the other solution, is 

 slowly introduced by means of the funnel. The permanganate solu- 

 tion is loaded with urea (a non-electrolyte) so as to make it denser 

 than the nitrate solution ; the permanganate solution now lies in the 

 lower part of the U-tube with a sharp interface between it and the 

 nitrate solution above it in each limb of the tube. If now we connect 



* Experiment from a paper by Nernst. in the Zeitschrift fiir Elektrochemie, 

 ill. p. 308 (1897). 



