Measurements were made of electromotive force in the Donnan equilibrium of systems containing dilute solutions of protein and acid. Removal of the membrane produced a decrease of no more than 2 to 4 mv. in electromotive force, while the membrane potentials, as estimated by the usual arbitrary assumption, were of the order of 12 to 34 mv. Ion ratios, as calculated from analyses for total chloride, were definitely greater than those calculated from the electromotive force of cells with salt bridges, as if there had been combination of some of the chloride ion with protein.

By a method involving equilibration of ice and solution, and analysis of the solution, freezing point depressions of solutions of sodium citrate, oxalate, and fluoride have been determined over the range Δ = 0.45 to 0.65°C. Determinations with sodium chloride solutions have confirmed the accuracy of the method. In each case the freezing point depression is given, within 0.002°C., as a linear function of the concentration. By the use of these linear equations it is possible to prepare a solution of any of these four salts isotonic with a given biological fluid of known freezing point, provided the latter falls within the range studied.

Electromotive force measurements of cells without liquid junction, of the type Ag, AgCl, HCl + protein, H 2 , have been made at 30°C. with the proteins gelatin, edestin, and casein in 0.1 M hydrochloric acid. The data are consistent with the assumptions of a constant combining capacity of each protein for hydrogen ion, no combination with chloride ion, and Failey's principle of a linear variation of the logarithm of the mean activity coefficient of the acid with increasing protein concentration. The combining capacities for hydrogen ion so obtained are 13.4 x 10 –4 for edestin, 9.6 x 10 –4 for gelatin, and 8.0 x 10 –4 for casein, in equivalents of combined H + per gm. of protein.

It has been found possible to obtain constant dry weights of sheet gelatin only by drying in vacuo at temperatures below 100°C. for a period of several weeks. Values are given for the ash and nitrogen content, the specific conductivity, and the isoelectric point of a standard gelatin preparation. By the method of E.M.F. measurements of cells without liquid junction, of the type Ag, AgCl, HCl + gelatin, H 2 , it has been found that this gelatin in 0.1 M HCl combines with a maximum of 9.58 x 10 –4 equivalents of H + and 2.0 x 10 –4 equivalents of Cl - . By means of pH measurements with the hydrogen electrode and a KCl junction, the combination curve of this gelatin with H + from HCl and OH - from NaOH has been determined between pH 1.1 and 12.5.

Two samples of a standard gelatin were studied, both prepared according to published specifications and washed free from diffusible electrolytes. The isoelectric point of this material was determined in four ways. 1. The pH values of solutions of gelatin in water approached the limit 4.86 ± 0.01 as the concentration of gelatin was increased. 2. The pH values of acetate buffers were unchanged by the addition of gelatin only at pH 4.85 ± 0.01. This gives the isoionic point of Sørensen, which is the isoelectric point with respect only to hydrogen and hydroxyl ions. 3. Gels of this gelatin made up in dilute HCl or NaOH, or in dilute acetate buffers, exhibited maximum turbidity at pH 4.85 ± 0.03. 4. Very dilute suspensions of collodion particles in 0.1 per cent gelatin solutions made up in acetate buffers showed zero velocity in cataphoresis experiments only at pH 4.80 ± 0.01. No evidence was found for the assumption that gelatin has two isoelectric points at widely separated pH values. It is concluded that the isoelectric point of this standard gelatin is not far from pH 4.85.

Electromotive force measurements of cells without liquid junction, of the type Ag, AgCl, HCl + protein, H 2 , lead to the conclusion that 1 gm. of edestin (or, more probably, edestan) combines with a maximum of 13.4 x 10 –4 equivalents of H + and 3.9 x 10 –4 equivalents of Cl - , when the protein is dissolved in 0.1 M HCl.

1. Cooper's gelatin purified according to Northrop and Kunitz exhibited a minimum of osmotic pressure and a maximum of opacity at pH 5.05 ±0.05. The pH of solutions of this gelatin in water was also close to this value. It is inferred that such gelatin is isoelectric at this pH and not at pH 4.70. 2. Hydrogen electrode measurements with KCl-agar junctions were made with concentrated solutions of this gelatin in HCl up to 0.1 M. The combination curve calculated from these data is quite exactly horizontal between pH 2 and 1, indicating that 1 gm. of this gelatin can combine with a maximum of 9.35 x 10 –4 equivalents of H + . 3. Conductivity titrations of this gelatin with HCl gave an endpoint at 9.41 (±0.05) x 10 –4 equivalents of HCl per gram gelatin. 4. E.M.F. measurements of the cell without liquid junction, Ag, AgCl, HCl + gelatin, H 2 , lead to the conclusion that this gelatin in 0.1 M HCl combines with a maximum of 9.4 x 10 –4 equivalents of H + and 1.7 x 10 –4 equivalents of Cl - per gram gelatin.

The permeability of gelatin-coated collodion membranes, as measured by the flow of water or of dilute solutions through the membranes, has been found to vary with the pH of the solutions. The permeability is greatest near the isoelectric point of the protein; with increasing concentration of either acid or alkali it decreases, passes through a minimum, and then increases. These variations with pH are qualitatively in accord with the assumption that they are due to swelling of the gelatin in the pores of the membrane, the effects of pH being similar to those observed by Loeb on the swelling of gelatin granules. Indications have been found of a similar variable permeability in the case of membranes coated with egg albumin, edestin, serum euglobulin, and serum albumin.

By the application of Poiseuille's law to the rate of flow of water through collodion membranes, it is calculated that the membranes used had pore radii of the order of 0.3 to 2 x 10 –6 cm. On the same basis the number of pores per sq. cm. appears to vary from 270 x 10 10 to 7 x 10 10 , decreasing with increase in pore size. Reasons are given for preferring these figures for the radii to figures, 100 times as large, which were calculated by others. Microscopic examination of the membranes, with dark-field illumination, indicates that they are made up of solid granules or filaments of collodion much less than 1 x 10 –4 cm. in thickness.

In applying Donnan's theory of membrane equilibria to systems where the non-diffusible ion is furnished by a weak acid, base, or ampholyte, certain new relations have been derived. Equations have been deduced which give the ion ratio and the apparent osmotic pressure as functions of the concentration and ionization constant of the weak electrolyte, and of the hydrogen ion concentration in its solution. The conditions for maximum values of these two properties have been formulated. It is pointed out that the progressive addition of acid to a system containing a non-diffusible weak base should not cause the value of the membrane potential to rise, pass through a maximum, and fall, but should only cause it to diminish. It is shown that the theory predicts slight differences in the effect of salts on the ion ratio in such systems, the effect increasing with the valence of the cation.

1. Collodion membranes of high permeability were found to adsorb weighable amounts of gelatin and egg albumin from solution at 37°C. 2. The effect of protein concentration could be expressed fairly well by a hyperbolic equation proposed by Langmuir for the adsorption of gases by a plane surface, while the usual parabolic adsorption equation of Freundlich did not fit the results. 3. In comparing this effect with solutions of varying pH, it was found there was a decided maximum of adsorption in solutions of isoelectric protein. The effects of acids and salts on the amount of gelatin adsorbed were like those observed by Loeb on the viscosity of gelatin solutions, but opposite in direction. The effects of pH on the amount of adsorbed gelatin and on the fluidity of the gelatin solutions were nearly parallel. 4. Membranes made impermeable by long drying took up very little or no gelatin from solution. 5. In the case of membranes of varying permeability the maximum amount of adherent gelatin increased with the permeability and thickness of the membranes, and appeared to be, within limits, a linear function of the relative pore surface of the membranes as calculated from Poiseuille's law. 6. The film of gelatin greatly decreased the permeability of the membranes, as measured by the flow of water through them. The relative cross-section of the pore openings, as calculated from the permeability measurements, was a linear function of the amount of adherent gelatin. These results led to the conclusion that the gelatin formed a film inside the pores.

Measurements have been made of the solubility at 25°C. of tyrosine in hydrochloric acid and in sodium hydroxide solutions varying from 0.001 to 0.05 M , and also in distilled water. The pH of the saturated solutions was measured with the hydrogen electrode. The following values for the ionization constants of tyrosine have been obtained from the measurements: k b = 1.57 x 10 –12 , k a1 = 7.8 x 10 –10 , k a2 = 8.5 x 10 –11 . The changes in solubility with pH can be satisfactorily explained by the use of these ionization constants.

Measurements have been made at 40°C. of the osmotic pressure and viscosity of 1 per cent gelatin solutions containing varying amounts of hydrochloric acid or sodium hydroxide. Each property was found to exhibit a decided minimum near pH 4.7. In the osmotic pressure experiments the pH of the inside solutions was greater than that of the outside solutions at pH values below 4.7, while it was less than that of the outside solutions at values above pH 4.7. These results indicate that gelatin at 40°C. retains its isoelectric point at about pH 4.7.

Titrations have been made, by the conductivity method, of gelatin solutions with hydrochloric and sulphuric acids. The results indicate an end-point at about 8.6 cc. of N /10 acid per gm. of gelatin, or a combining weight of about 1,160. These results are in fair agreement with those previously obtained by the hydrogen electrode method. Better agreement between the two methods was found in the case of deaminized gelatin. The data are in accord with a purely chemical conception of the combination between protein and acid.

1. The analysis of isoelectric gelatin by the Van Slyke method indicates 0.00040 equivalents of amino N per gm. gelatin. 2. If deaminized gelatin is prepared without heating, the product contains less nitrogen than the original gelatin by an amount equal to 0.00040 equivalents N per gm. protein. 3. Deaminized gelatin, prepared either with or without heating, contains no amino nitrogen detectable with certainty by either the Van Slyke or the formol titration method. 4. The isoelectric point of deaminized gelatin prepared without heating is at pH 4.0. 5. The maximum combining capacity of this protein for HCl is 0.00044 equivalents per gm. 6. The maximum combining capacity of gelatin for HCl should be corrected to 0.00089 equivalents per gm. 7. The difference between these maximum combining capacities, 0.00045, is nearly equivalent to the loss in amino or total nitrogen occurring in the deaminizing reaction. 8. This equivalence constitutes a new indication that the combination of protein with acid is chemical combination.

1. By the use of the silver-silver chloride electrode, measurements have been made of the chloride ion concentrations of 1 per cent solutions of five proteins, containing from 0.001 N to 0.1 N hydrochloric acid. The hydrogen ion concentrations of the same solutions have been measured by the use of the hydrogen electrode. 2. The measurements indicate that the chlorides of gelatin, egg albumin, casein, edestin, and serum globulin are highly ionized electrolytes, ionizing to yield chloride ion and a positive protein-hydrogen ion. Their ionization is therefore similar to that of ammonium chloride. 3. The results do not support the idea that a protein chloride does not yield chloride ion on dissociation. They are not in agreement with the idea that the depressing effect of an excess of HCl on the viscosity and other colloidal properties of a protein chloride solution is due to a repression of the ionization of the protein chloride. The results are, however, in complete accord with the theory of colloidal behavior advocated by Loeb.

1. The globulin prepared from ox serum by dilution and precipitation with carbon dioxide has been found, by electrometric titration experiments, to behave like an amphoteric electrolyte, reacting stoichiometrically with acids and bases. 2. The potential difference developed between a solution of globulin chloride, phosphate, or acetate and a solution of the corresponding acid, free from protein, separated from the globulin by a collodion membrane, was found to be influenced by hydrogen ion concentration and salt concentration in the way predicted by Donnan's theory of membrane equilibrium. In experiments with sodium globulinate and sodium hydroxide it was found that the potential difference could be similarly explained. 3. The osmotic pressure of such solutions could be qualitatively accounted for by the Donnan theory, but exhibited a discrepancy which is explicable by analogy with certain experiments of Loeb on gelatin. 4. The application of Loeb's theory of colloidal behavior, which had previously been found to hold in the case of gelatin, casein, egg albumin, and edestin, has thus been extended to another protein, serum globulin.

The amount of HCl combined with a given weight of gelatin has been determined by hydrogen electrode measurements in 1 per cent, 2.5 per cent, and 5 per cent solutions of gelatin in HCl of various concentrations, by correcting for the amount of HCl necessary to give the same pH to an equal volume of water without protein. The curve so obtained indicates that the amount of HCl combined with 1 gm. of gelatin is constant between pH 1 and 2, being about 0.00092 moles.