JACQUES LOEB 467 



chloride, when the acid is HCl. The water inside the pores of a film 

 of gelatin chloride is negatively charged while the gelatin is positively- 

 charged. This is a consequence of the theory of Donnan's membrane 

 equilibrium.^ 



In these experiments the gelatin with which the membrane was 

 coated was originally isoelectric, but the film of gelatin was brought 

 into equilibrium with water of pH 3.0 by putting the collodion bags 

 for several hours into HCl of pH 3.0 before the beginning of the ex- 

 periment. Since the addition of salts influences both the p.d. across 

 the membrane {i.e. the value of E) as well as the p.d. between the 

 liquid inside the pores and the wall of the pore {i.e. the value of e) 

 it is necessary to measure the influence of salts on these two p.d. 

 separately and then try to use the results for the analysis of the curves 

 in Fig. 1. The measurements of the p.d. across the membrane were 

 made with the aid of a Compton electrometer and the two electrodes 

 used were calomel electrodes with saturated KCl solution. Measure- 

 ments of the P.D. across the membrane were made at the beginning 

 of the experiment (Table I) and at the end; i.e., after 20 minutes 

 (Table II). The figures for the p.d. in Table II are lower than in 

 Table I for the reason that during the experiment part of the salt 

 diffused into the outside solution so that the concentration of the 

 salt solution inside the bag diminished while that in the outside in- 

 creased; hence the value of E diminished. In addition the hydrogen 

 ion concentration which was the same inside and outside the salt 

 solution at the beginning changed and this added a complication which 

 has been discussed in a previous paper.® Table I shows that the 

 p.d. across the membrane (£) increases with the valency of the 

 cation and inversely with the valency of the anion, E being a maximum 

 for CeCls, being lower for CaCl2, and still lower for NaCl; for Na2S04 

 it becomes about zero or even slightly negative (Table II). It is, 

 therefore, obvious that in a semiquantitative way the results of Tables 

 I and II {i.e. the values of E) explain the difference in the ascending 

 branches of the curves in Fig. 1 up to a concentration of about m/32. 



It can be shown that this p.d. is at least partly due to diffusion 

 potentials. In measuring the diffusion potentials the principle of a 



«Loeb, J., J. Gen. Physiol., 1921-22, iv, 213. 



