ACCURATE MEASUREMENT OF IONIC VELOCITIES, ETC. 451 



ments the solution to be measured was separated from the solutions containing the 

 slower indicator ions by two partitions of a gelatine solution of the same indicators, 

 and in this way it was possible to place the solutions in position without any 

 appreciable mixing taking place at the surfaces of contact. 



It has been shown by DENISON (loc. cit.) that the deviations referred to are due, 

 largely, if not entirely, to the occurrence of electric endosmose at the gelatine 

 partitions. DENISON measured the amount of endosmose, and applied the correction 

 to the ratio U/V with satisfactory results in the case of many salts. 



The present research has been undertaken with the object of devising a method by 

 means of which the solutions could be superposed without mixing, and which would 

 avoid the use of membranes of any sort during the progress of the experiment. 



The use of gelatine especially was to be avoided, not only on account of electric 

 endosmose, which would be caused by any membrane, but also on account of the ease 

 with whicli it melts on the passage of even small currents, and on account of the 

 impossibility of obtaining it free from saline impurities. Even the purest obtainable 

 gelatine contains a quantity of saline matter which is an appreciable fraction of the 

 concentration of a moderately dilute salt solution, and this is probably the reason 

 why our previous attempts to measure such solutions have been unsuccessful, and it 

 appears to be impossible to measure transport numbers in dilute solutions in a system 

 containing gelatine. Now we know that it is only in the case of the simple salts of 

 the alkali metals that the transport number is practically independent of concen- 

 tration ; with other salts the general tendency is for the anion transport number to 

 increase with increasing concentration. This gives, of course, different values for the 

 velocity of the same ion at different dilutions, and, moreover, the velocity of one and 

 the same ion is found to be different for the same concentration when measured in 

 different salts. These differences, however, vanish in dilute solution* in which, in 

 accordance with the theory of KOHLRAUSCH and ARRHENIUS, a given ion has the 

 same velocity in different salts. Hence an extension of the direct method of 

 measuring ionic velocities to dilute solutions is much to be desired on account of its 

 great simplicity and ease of manipulation. 



Before this could be accomplished, however, it was necessary to devise some means 

 of observing the margins. If gelatine be employed as membrane, the margins become 

 invisible in solutions more dilute than about 0'2 N. 



It was thought that an electrical method could be used to indicate the position of 

 the boundary by taking advantage of the difference in conductivity of the indicator 

 and measured solutions. This method, however, proved impracticable. Coloured 

 salt solutions could not be used as indicators, as in solutions of about 0'02 normal the 

 colour of solutions of salts such as copper sulphate, cobalt sulphate, nickel sulphate, &c., 

 is much too faint to be serviceable. The intensely coloured ions of some of the 

 organic dyes were tried, but, although certain acid dyes are excellent anion indicators, 

 * STEELK and DENISON, ' Journ. Chem. Soc.,' 1902, LXXXL, p. 456. 



3 M 2 



