250 



Dr. W. M. Bayliss. 



[June 30, 



forces to prevent free diffusion of bodies devoid of electric charge. Moreover, 

 the proportion of non-dissociated to dissociated salt is naturally greater 

 within the membrane, owing to the diminished ionisation in the presence of 

 the dye which has an ion (Na") in common with the sodium chloride. This 

 proportion will also be greater the greater the concentration of the dye. 

 The result will be that the total concentration of sodium chloride will be less 

 within the membrane, since less will be necessary in order to provide the 

 same amount of non-dissociated salt. 



On the other hand, Biltz and v. Vegesack* maintain that the concentration 

 of diffusible ions becomes equal on both sides of the membrane, diffusible 

 ions including the cations of the dye. Consideration of the case where the 

 dye solution is of high concentration and the sodium chloride dilute shows 

 that this view does not afford a sufficient explanation. Such a case is the 

 third one in Table V. The chlorine in the inner solution was, in fact, too 

 small to be accurately estimated ; a faint turbidity only was produced by the 

 addition of silver nitrate. The diffusible ions of the inner solution would 

 have a concentration of at least m/30, since at the concentration in question 

 the dye is more than 50 per cent, dissociated. The ions of the sodium 

 chloride outside could not have a concentration greater than m/QO if the salt 

 were completely dissociated. The investigators referred to find that the 

 conductivity is equal on both sides of the membrane, as their theory requires, 

 when a relatively large concentration of sodium sulphate is present along 

 with a dilute solution, - 0007325 molar, of benzo-purpurin within the 

 membrane. I have always found that the conductivity of the inner solution 

 containing the dye is greater than that of the outer one, although, of course, 

 when the proportion of foreign electrolyte is very great, the difference comes 

 very near the limit of detection. 



The actual values of the conductivity of the inner solutions found in my 

 experiments afford, as far as they go, confirmation of the theory put forward 

 by Donnan. For example, in the first experiment of Table V, the con- 

 ductivity of the inner solution was found to be 5920 recip. megohms. That 

 of the dye alone could not be greater than 3935 recip. megohms, so 

 that the ionised sodium chloride must have had at least a conductivity of 

 1985 recip. megohms, which is that of an m/60 solution at the temperature of 

 the experiment. The amount of chlorine found corresponded to m/52 KaCl,. 

 so that (m/52 — m/60 =) m/400 was present non-ionised. It has been 

 assumed, however, that there was no depression of the ionisation of the dye 

 by the sodium chloride. If there were such an effect, the concentration of 

 the ionised sodium chloride would be greater than m/60, and therefore that 

 * 1 Zeits. f. physik. Chem.,' 1910, vol. 73, p. 492. 



