DAVID I. HITCHCOCK 605 



have the same salt concentration as the protein solutions. The 

 beakers were set in a water thermostat at 25° ± 1°C. After equilib- 

 rium had been established (12 to 48 hours), the osmotic pressure 

 was measured in terms of millimeters of the solution in the manometer 

 tubes. The p.d. between the inside and outside solutions was deter- 

 mined in each case with the aid of saturated KCl calomel electrodes 

 and a Compton electrometer. The p.d. measurements were made 

 in a room at about 20°C., but the solutions were very nearly at 25°, 

 since the p. d. was determined for each solution within 2 or 3 minutes 

 after removing it from the thermostat at 25°. Finally the pH of 

 the inside and outside solutions, at 33°, was determined with the hydro- 

 gen electrode and potentiometer. The calculated p. d. values were 

 reduced to 25° by multiplying by fff the differences between the 

 E.M.F. readings obtained for the two solutions with the hydrogen 

 electrode. (This amounts to the same thing as using equation (1), 

 but avoids slight arithmetical errors due to rounding off the values 

 for pH.) The p.d. values were read to 0.5 millivolts, but the repro- 

 ducibility was of the order of 1 millivolt. 



The results of experiments with 0.45 per cent edestin chloride, at 

 pH 3, and the four salts mentioned, are given in Tables II to V. 



The excellent agreement of the observed and calculated values for the 

 P.D. in these salt experiments proves that the Donnan equilibrium 

 governs the effect of salt on the p.d. of edestin chloride solutions fully 

 as well as in the case of gelatin or albumin chloride. Tables II 

 and III show that here too it is the anion of the salt which has the 

 depressing effect on the p.d., the sulfate ion being more effective than 

 the chloride ion. An increase in the concentration of an ion of op- 

 posite charge to that of the protein ion tends to prevent the forcing 

 of acid from the inside to the outside solution, and hence decreases 

 the difference in pH and the resulting p.d. This may also be shown 

 clearly from the results with the different chlorides by plotting the 

 p.d. against the equivalent concentration of chloride ion furnished by 

 the salt, assuming complete or equal ionization of the inorganic 

 chlorides. The results in Tables III, IV, and V are so plotted in 

 Fig. 4. It is evident that the effects of NaCl, CaClo, and LaCU are 

 identical, if compared at the same concentration of chloride ion. 

 This proves beyond a doubt that the valence or nature of the cation is 

 of no influence on the p.d. of edestin chloride. 



