1332 MCBAIN AND KAM: CXXIV.-The Efect of Salts on the Vapour Presswe and Degree of Dissociation of Acetic Acid in Solution. An Experimental Refutation of the Hypothesis that Neuti-al Salts Increase the Dissocia-tion Constants of Weak Acids and Bases. By JAMES WILLIAM MCBAIN and JAMES EAM. IR 1899 Arrhenius proposed a modification of his classical dis-sociation theory to the effect that salts increase the disaociatio THE EFFECT OF SALTS ON THE VAPOUR PRESSURE ETC. 1833 constants of weak acids present with them in solution as if either the water had acquired greater dissociating power or the salt itself was acting as a dissociating medium. The exppimental evidence he adduced was the rate of inversion of sucrose by weak acids in the presence of salts. Arrhenius’s idea was very generally accepted and developed, particularly in the field of non-aqueous solutions; but in 1914 it was called in question by McBain and Coleman’s re-interpretation of the direct experimental evidence.On recalculation of the data given by Arrhenius in conformity with presenbday conceptions of this reaction they found that the supposed effect was entirely absent thus reversing the significance of the experiments. In other words the dissociation constant of weak acids is not affected by the presence of salts. They followed this up by a review of all the available experi-mental evidence bearing on this subject and they found that it supported only the simple form of the classical dissociation theory. One isolated group of experiments was left outstanding inas-much as in this particular case the measurements were conflicting in their evidence ; these were certain determinations of hydrogen ions by the method of electromotive force.The potential of the hydrogen electrode in solutions of acetic acid was greater than that predicted when sodium chloride was present although this was not the case when sodium acetate was the added salt. (1) Object of t h e Present Tnvestigatkon. The present communication adduces a hitherto unsuspected but general effect of such salts as sodium chloride on undissociated acetic acid which would account for the apparent results derived from the measurements of hydrogen electrode potential. The equilibrium under discussion is HAc H*+Acl, where HAc stands for a weak acid such as acetic acid.The potential of the hydrogen electrode in this solution is admittedly boo great when sodium chloride is present. This has hitherto been interpreted as a real increase in acidity; in other words a displacement of the equilibrium to the right-an enhancement of the dissociation constant itself. The electrical potential of the hydrogen electrode however, measures the product of the chemical potential and the concen-tration of the hydrogen ion. Instead of assuming that the increase in this product is due to increase in concentration we here submi 1334 MCBAXN AND KAM: experimental evidence for the alternative explanation that the other factor the chemical potential has been enhanced. If the concentration of the hydrogen ion has remained unaltered, but its chemical potential or reactivity has been increased it is necessary for the continuance of equilibrium that the chemical potential or reactivity of the substance on the left-hand side of the chemical equation should likewise have increased.Such increase in reactivity or potential of undissociated acetic acid in the solution must be accompanied by a parallel increase in the partial pressure of acetic acid in the vapour phase. This is open t o direct experimental test and we find that a remarkable increase is actually exhibited fully accounting for the electromotive force data observed. This removes the last evidence in favour of Arrhenius’s proposed modification of his classical dis-sociation theory. The effect here discovered has to be taken into account in most determinations of electromotive force.(2) The Experimental Method. The simple experimental method adopted was the distillation of aqueous solutions of acetic acid with and without addition of various salts. Pipettes burettes and measuring flasks were carefully Cali-brated. The distillations were carried out in a flask of fused silica of about 1500 C.C. capacity heated directly by contact with a large Bunsen flame. The neck and upper half of the flask were covered with a lagging of magnesia and asbestos so as to avoid as much as possible fractional distillation in the flask. The d i s tance between burner and flask was kept constant through all distillations as was also the flame itself. The quantity of distilling liquid mas each time 1000 c.c.and the distillate was collected in four to five fractions of about 75 C.C. each. After each fraction the distillation was quickly inter-rupted for the abstraction of a similar quantity of about 75 C.C. from the residue in the flask; the first residue was dbstracted as soon as the liquid began to boil and just previous to the collection of the first fraction of the distillate. The residues were left to cool in glass-stoppered bottles vaseline being used on the stoppers to prevent ingress of carbon dioxide from the air. Samples of 20 C.C. of each of the distillates and residues were titrated against standard solutions of sodium hydroxide of approximately equal strength with phenolphthalein as indicator ; the usual precautions were taken to avoid the vitiating effmij of carbonic acid.Thus for each distillate the ratio R2 could b THE EFFECT OF SALTS ON THE VAFOUR PRESSURE ETC. 1336 determined between the concentrations of acetic acid in that dis-tillate and the mean concentration of acid in the residue in the flask before and after. Thus the ratio R was determined for solutions of acetic acid of concentrations varying from 0.05 to 0.5N. It appeared that, a t least between these limits R increases only very slightly with the concentration as is apparent from the curve No. I of Fig. 1, which shows the value of R plotted against the acid concen-trations, The same operations were repeated with solutions of sodium, potassium and lithium chlorides potassium thiocyanate potassium nitrate sodium sulphate and sodium acetate in 0.05 to 0.4N-solu-tions of acetic acid.The concentration of the salts was deter-mined by careful evaporation of 20 C.C. of each residue in a porce-lain evaporating dish in a hot-air oven a t temperatures depending on the nature of the salt in question. The acetic acid was titrated as before in distillate and residue to obtain the ratio R of the concentration. The values of the ratios R are dependent on the concentration of the salts but they are independent of the concentration of acetic acid. Sodium acetate differs from the other salts investi-gated in that it has scarcely any effect on the distillation of acetic acid. In this manner values of R derived from a very large number of distillations of solutions of pure acetic acid became the standard of comparison for a number of distillations of acetic acid contain-ing added salts.(R1 - R') gives the percentage increase of The expression 100 the ratio R caused by the addition of salt. The experimental data here presented comprise well above one hundred such determinations. R2 (3) Method of Calculation of the Distillation Data. Distillation was selected for the measurement of the partial oapour pressure of acetic acid merely for the sake of convenience and accuracy. It must be borne in mind that the composition of the distillate shows only the relative proportion of acetic acid and water vapour in the vapour phase above the solution. What is required is the absolute magnitude of the partial vapour pressure of acetic acid a t a definite temperature say looo.The polymerisa-tion of the acetic acid vapour may be neglected for the present purpose since it amounts to only a few per cent. a t these lo 1336 UCBAIN AND KAM: partial presaureg. Even this slight effect ia largely eliminated in comparing R with R,. Ordinary variations in barometric pressure and the concomitant alterations in boiling point have no appreciable influence on the composition of the distillate. Hence all the distillations may without error be regarded as having been in effect carried out a t loo* even where much salt has been added. In order to obtain the absolute instead of the relative magnitude of the partial pressure of acetic acid in the vapour distilling over a t looo the actual partial pressure of the water has to be evaluated.Now the partial pressure of the water which in pure water wits 760 mm. has been diminished by two effects for which allowance has to be made. The first correction may be termed the (( osmotic correction.” The vapour pressure of the solvent has been reduced in familiar fashion through the osmotic activity of the substance in solution. Hence in all cases the observed concentration of the acetic acid in the distillate must be diminished accordingly before use. This is readily done with sufficient accuracy for the present purpoee by taking the lowering of vapour pressure of the solvent to be 1.80 per cent. per mol. of total solute (ions and undissociated acid and salt). It results from the effect of the appreciable partial pressure of the acetic acid in lowering the pressure a t which the water actually distils over instead of this occurring a t a partial pressure of water vapour equal to 760 mm.Here again then in order to base the relative magnitude of the partial pressure of acetic acid on the constant value of 760 mm. for that of water vapour throughout the vola-tility correction has to be applied so as to diminish the observed concentration of the acetic acid in the distillate accordingly. This consists in the reduction of the latter by 0.12 per cent. for a 0 . W -solution of acetic acid and taking this correction as proportional to the concentration of the acetic acid in the distillate. Since the acetic acid was at most N / 2 this correction in no case exceeded 0.6 per cent. A third and final correction had tc be made this time in the apparent composition of the solution undergoing distillation.This is the ((correction for dissociation” of the acetic acid in the solu-tion. The actual ratios measured were those between the con-centrations of distillates and the corresponding solutione in the distilling flask (the (I residues ”). What is required is the ratio based on the actual concentration of undissociated acetic acid in the flask. In the case of solutions containing only acetic acid. The second correction is the ‘( volatility correction. THE EFFECT OF SALTS ON TEB) VAPOUB PRESSURE ETC. 1337 this consisted simply in subtracting the known amount of dis-sociated acetic acid from the total concentration of acetic acid in the flask. The dissociation constank of acetic acid was taken to be 1.11 x 10-5 a t looo.The correction for dissociation involves much calculation where salt is present since in order to determine the actual concentration of undissociated acetic acid it is necessary to calculate the amounb of the various ions and undissociated Salk present including those formed by metathesis. For example with common salt the follow-ing molecular species were present H’ Act Na’ Cl’ HAc HCI, NaC1 and NaAc. This was done by Arrhenius’s method which is based on the principle of isohydrism (loc. cit.) and agrees with the method of Sherrill ( J . Amer. Chem. SOC. 1910 32 741). The calculation is laborious and involves successive approximations Fortunately, the exact degree of dissociation of the various salts has but little influence on the results since the really important values appear in both numerator and denominator of Arrhenius’s equations ; con- I ductivity data a t the ordinary temperature could theref ore be employed failing the existence of such a t higher temperatures and concentrations.Indeed the calculation for solutions of one salt in acetic acid might have been applied to the case of any other salt of the same concentrations except of course in the case of sodium acetate. The effect is chiefly dependent on the relative concentration of acid and added salts. ( 4 ) The Distillation of Solutions of Pure Acetic Acid. Following the method already described fifteen distillations were carried out with N / 20- to A7 / 2-solutions of acetic acid involving more than seventy determinations.The object in view was to determine the ratio R, that is the ratio concentration of acid in distillate t o concentration of acid in residue. C c r Thus R,= 2, Cd being the concentration of the distillate C being the mean value of the concentrations of the residue before and after the separation of the distillate. For any one concentration the results agreed to within 1 per cent. The values of R were corrected as described in Section (3) above and they increase by only about 1 per cent. over the whole range of concentration. The relative concentration of double molecules, which has not been allowed for changes from about 5 to 8 per cent. over this same range. The experimental evidence is summarised in table I 1338 MoBAIN AND ?CAM: TABLE I.(Curve I.) Ratio of Concentration i n Distillate and Residues in Aqueous Acetic Acid. Concentration of acid in Rz R2 flask. (uncorrected). (corrected). N/20 0.662 0.67 1 N/10 0.666 0.67 1 0.673 0.674 0.680 0.677 0-686 0.678 N15 N /3 N12 The corrected values of R increase very slowly indeed. Plotting them against the concentrations we obtain a straight line (Curve I in Fig. 1). FIG. 1. I. CH,*CO,H COW. (B=0*6S!; C=0.678). 11. NaCl uncorr. Percentage ancreme due to salt. 111. NaCl corn. 9 ) ?1 9 ) 9 ) IV. KCI ? 9 9 * 1 1 V. KCNS , 9 9 I 9 1 9 ?1 VIA.Czcrue logcJ = [{1-BR,).log3.2-3026 + R . K . Co 2.3026 01 A0 = v,,- 1. Ratios of acetic acid in distillate and reeidue with and without ao?ded salts. The corrections for volatility and osmotic effect can be taken from graphs; for example in the case of N/20-acetic acid the THE EFFECT OF SALTS ON THE VAPOUR PRESSURE ETC.1339 amount to -0.06 and -0.09 per cent. respectively. The third correction for degree of dissociation of the acetic acid is here 1.50 per cent. Thus the total correction in this case is an increase of 1'35 per cent. on the observed ratio R,=0.662. Hence we obtain This value is smaller than the one arrived a t by Lord Rayleigh (Phil. Mag. 1902 [vi] 4 535) R2=0.73. Corrected it reduces sligh€ly to 0.725 which is still considerably greater (the reduction in this case being caused by the method of calculation of R from the residues). The discrepancy is however probably due to the different manner of heating.Lord Rayleigh kept the neck and upper part of the distilling flask hotter than the boiling liquid in order to prevent condensation. We found however that this involved the quantitative evaporation of drops splashing up f roan the boiling liquid which produces the same error as if they had splashed directly into the distillate. It was for this reason that we relied on good heat insulation and fairly rapid distillation. R (corrected) = 0.662 x 1.0135 = 0.671. T o t e on the Calculation of the Composition given Stage of the Distillation of Following Lord Rayleigh (Zoc. cit.) but residues expressed in mols. per litre, where C0 and C are the concentrations and in the distilling flask before and after the For example for R,=0.671 and --O- c 1 c,- 9 of the Residues at any Acetic Acid.using concentration of: Vo and Vl the volumes distillation, 5 = 8 . 2 2 1 . Hence in v, order to double the concentration of the resiiual acid solution in 7.921 the flask -5.- =On878 or almost 88 per cent. of the volume of the 6.221 solution must be distilled over. For R = 0.73 (Rayleigh's value in Nll0-solution) this quantity would amount to 92 per cent. Again if 40 per cent. of the original volume of solution is dis- c tilled off that is if Vl=O*6 Vo,+ =1*183 which means that the concentration of the residue will have increased by 18.3 per cent. 6 0 (5) The Experimental Data for Added Salts. Having determined the ratios R for solutions of acetic acid the corresponding ratios R, were obtained in exactly the same manner, but with a salt added to the solution.VOL. cxv. 3 I! 340 MCBAIN AND KAM: The salts used were the purest obtainable in 1913 and are named in Section (2) (above). The distillates were in each case tested for traces of the stronger acid formed by metathesis but these were found to be negligible. The corrections were calculated and applied in the same manner as before except that the acetic acid destroyed by metathesis was calculated according to Arrhenius's principle of isohydrism (Zeitsch. physikal. Chem. 1899 30 208). The results are summarised in tables eII-VIII which require no further explanation except to note that in tables I1 and VTI space is saved by averaging the figures for all the values obtained over certain ranges of concentration.The number of experimental values so averaged is given in the last column. With the exception of sodium acetate all these salts cause a remarkable increase of the ratio cd and sodium chloride shows this increase more than any other of the salts investigated. Sodium acetate on the other hand appears to have no appreciable effect even in l.0N-solution. TABLES I1 .-VIII. Ratios R of Concentration of Distillates and Residues of Apue0u.e Acetic Acid with added Salts and Increase thereof over R, of Table I . c,' TABLE 11. (Curves 11. and 111.). Sodium Chloride. HAc. 0-1-0*2 0.1-0.3 0.1-0.3 0.1-0.3 0.13 0.076 0.078 0.077 Sdt. (0.12) (0.21) (0-29) (0-41) 0.661 1-04 1.38 2-30 RP 0.688 0.708 0.721 0.742 0.790 0.830 0,904 1-076 R2-0.669 0.67 1 0-672 0.67 1 0-668 0.663 0.664 0.664 Rl (corr.). 0.690 0.7 10 0.716 0.721 0.715 0-820 0.885 1.033 R2 %-3 (corr.). [ R ] Expts. 0.673 0.673 0.673 0.673 0.67 1 0.671 0.67 1 0.671 2.6 ~ 7 5.4 16 6.6 15 9.6 4 16.9 1 22.2 1 32.1 1 53.9 1 TABLE 111. (Curve IV. Fig. 1). Potassium Chloride. 0.22 0.2-0.3 0.7146 0.673 0.7111 0.674 5.55 4 0.2540 0.4673 0.7389 0.675 0.7309 0.676 8.20 1 0.2157 (0.70) 0.7741 0.673 0.759 0-674 13-55 2 0.2053 0.9228 0-8121 0-673 0.7930 0.674 17.70 1 0.22 1.17 0.8383 0-673 0.8133 0.674 21.55 2 0.2194 1.953 0.9363 0.673 0.8872 0.674 31.60 1 TABLE IV. (Curve V. Fig. 1). Potassium Thiocyanate. 0.2-0.23 0.2-0.34 0.6900 0.673 0.6868 0.674 1.90 6 0.2673 0.6590 0.7297 0.675 0.7181 0.675 6.40 1 0.2366 1-663 0.7693 0.674 0.7343 0-674 8.90 1 0.2596 2.451 0.8710 0.675 0.8112 0.676 20-10 THE EFFEUT OF SALTS ON THB BAkpOUR PRESSURE ETC.1341 HAc. 0.22 0.22 0.25 0.25 0.27 0.29 0-21 0.22 TABLE V. (Curve VI. Fig. 2). Sodium Sulphtc. S d t . (0.07) (0.13) (0.22) (0-35) (0-61) 1-10 1.40 1.64 RIB 0.669 0.674 0.674 0.683 0.7 10 0.762 0.807 0.839 R,* 0.673 0.673 0.675 0.675 0.677 0.678 0.673 0.673 Rl. (COAT.). 0.660 0.673 0.67 1 0.678 0.700 0.743 0.786 0.813 R,. (con.). 0-674 0.674 0.674 0,674 0.675 0.675 0.674 0.674 - 0.2 8 - 0.4 7 + 0.5 6 + 3.8 2 + 10.0 1 + 16.6 1 + 20.6 1 TABLE VI. (Curve VII. Fig. 2). Lithizcm Chloride.0-2-0.22 0.2-0.25 0.7143 0.673 0.7116 0,674 5.57 3 0.2269 0.2799 0.7188 0.673 0.7147 0-674 6.00 1 0,2469 0.3930 0.7450 0.674 0.7375 0.674 9.40 1 FIG. 2. .i % VI. Na,,SO, cow. VIII. D O s cow. VII. LiCl , IX. NaCH,'COB GOW. Percentage invreaae in ratios of acetic acid in diabillate and residue due to added s&. TABLE VTI. (Curve VIII. Fig. 2). Potassium Nitrate. 0.21 0.22 0.6944 0.673 0-6940 0.674 1.13 2 0.24 0.3-0.5 0.7198 0.674 0.7110 0.674 2-32 2 TABLE VIII. (Curve IX. Fig. 2). Sodium Acetate. 0.11 (0.14) 0.6822 0.667 0.6769 0.671 0.8 3 0*1-0*23 0.4-0*67 0.6948 0.671 0.6799 0.673 0.9 7 0.14.28 0*9-1*1 0.7001 0.673 0,6799 0.673 1.3 3 0.2900 1.810 0.7244 0.678 0.6834 0.675 1-3 1 The increase of R, the relative concentration of the distillate, for the same sollation of aoetic acid that is the expression 3 ~ 1342 M c B m AND KAM: 100 1- amounted to no less than 62 per cent.(observed) (” isR2) in the c a k of 2*3N-sodium chloride. The increase for most salt8 seems proportional to the concentra-tion of the salt and independent of the concentration of acetic acid. Curves 11-V of Fig. 1 and VI of Fig. 2 show this percentage increase plotted against the concentrations and they point con-vincingly towards a straight line function between these two values. Sodium aulphate differs from all the other salts in that the experi-mental evidence shows a slightly negative effect a t lower concen-trations up to about 0.3N. Note on the Calculation. of Residues in the Presence of Salts which exhibit a Straight Line Function of the Ratio Increase.The calculation is similar to the one for the pure acid solution, but we must introduce the functional relation of R with regard to the concentration of the salt. We have found experimentally for all salts except the sulphate and acetate 100(Bd2) = lOOKp or R =R2(Kp + I) 4 in which p is the concentration of the salt and K is a constant. Hence R,=R2 - + 1 where V is the volume in litres containing 1 mol. of salt. I f there are y mols. of acetic acid in V litres of the solution in the flask and a quantity dv of the solution containing dy of the acetic acid distils over we may set the concentration of the dis-(; ) tillate equal to dY - d V = RIP= Y R2(; + 1);. Since c = 3 and by v‘ Integrating, En 5 = R*X (F) + (R - qzrt 5, C1 v Vl Vl or where Co and C are the concentrations and Vo and V the volumes: in the distilling flask before and after distillation THE EFFEUT OF SALTS ON THE VAPOUR PRESSURE ETC.1343 An alternative formula deducible in the same way is Either of these expressions can be used to calculate the con-centration of the acetic acid in the distilling flask if the initial value of the salt concentration and either its increase or the relative volume of the residual solution are given. They contain two constants. lOOR is the one which is peculiar to the added salt and it is simply the percentage increase by 1-ON-salt. The other R, is the ratio of acetic acid in distillate and residue for the same solution in the absence of salt. It is convenient t o express VT1 as a fraction of V, the volume a t the beginning of the distillation containing 1 mol.of salt ; but the initial concentra€ion 1 of salt in mols. per litre= -. vo Inspection of the equations show that they are identical with that deduced above for solutions of pure acetic acid except for the correcting factor R,lp('"- "I) This factor of course dis-appears for large values of Vo that is for very low concentration of added salt. Conversely for very high concentrations of salt it is predominant as is evident from the consideration that for say, 2-6N-potassium chloride no separation takes place owing to the concentration of acetic acid in the distillate having been so raised as to equaI that of the residue. Above such concentrations the residue becomes weaker instead of stronger.The peneral behaviour of a distillation is shown graphically in curve VIA of Fig. 1 which assumes that 1.0N-sodium chloride was initiallv present ( Vo= 1). AS the distillation proceeds the salt accumulates and the concentration of acetic acid in the residue slowly rises to a maximum where 50 per cent. of the liquid has C distilled over (V,=0*5). At this point l o g 2 = 0.03 whence the increase of concentration is 7-2 per cent. At V,=0.27 when 73 per cent. has been distilled over the concentration of the acid is again the same as it was before the ,distillation. Beyond this point the value for log CJ assumes rapidly negative values, owing to the high concentration of salt in the residue. The general equation may be tested by one of our experimental resulte.Taking the last pair of values in table I11 for potassium chloride 1 *ON-pota8sium chloride increases the ratio of acetic acid by 18.5 per cent. hence K=0.185. Further R,=0.674, v0.v1 Q, C 1344 McBAIN AND KAM: 1.128 V,, Vo=-. 1 These values inserted in the equation 1.1 28 v1 = - 1.953 lead to the prediction that C’,=1.O78C0. Experimentally Co was 0.2118N (less 1 per cent. for metathesis) and C after distillation 0.2194N (plus 5 per cent. for the three corrections). Hence C = -Go = 1.0950,. In general the concentrations observed appear to agree with the predicted values within about 2 per cent. To sum up the process of concentration or separation of the constituents of a binary mixture by means of distillation may thus be considerably accelerated or retarded by the addition of a salt, and will largely depend on the values of the constants R and R, that is on the nature of the mixture and of the added salt.0.230 0*210 (6) Discussion of the Results. The remarkable effed of a salt on the partial vapour pressure of acetic acid must be evident from the preceding. Comparing the slopes of the various graphs showing the relationship between salt concentration and percentage ratio increase it; appears that the effect is greatest for the chlorides of lithium sodium and potassium and least for sodium acetate. The effect is in all cases independent of the concentration of the acetic acid. The increase of the concentration of the residue in the flask for acetic acid in aqueous solution during the distillation not very rapid in itself is still less if salt is also present.At a concentra-tion of 2-3N-sodium chloride the residue becomes weaker ; in other words the vapour phase in the flask contains more acetic acid than the liquid from which it originates. For sodium sulphate up to about 0*35N the experimental evidence for the ratio increase seems somewhat complicated. I f anything there is a negative effect as the course of curve VI in Fig. 2 indicates. Beyond 0*35N the effect is decidedly positive. Between O.35N and l-lN the graph is practically a srtraicht line, but beyond l * l N its slope appears to increase until at 1*55N the limit of solubility is approached. Sodium acetate up to 1.8N shows only a very slight effect;. Although the effect of the cation is undeniable (note for example, the greater slope of the ratio increase for sodium chloride as com-pared with the one for potassium chloride) the influence of the anion seems t o be the predominating factor.The series sodium acetate potassium nitrate potassium thiocyanate sodium eulphate, pofsssium chloride lithium chloride sodium chloride shows the inaresse in a progressive degree. It is evident that this order is uot that of the Hofrneister or lyotropic series THE EFFECT OF SALTS ON THE VAPOUR PRESSURE ETC. 1345 It is well known that there is a general qualitative similarity between the effect of neutral salts on such various phenomena as solubilities of gases and non-electrolytes surf ace tension compressi-bility maximum density of water viscosity dielectric constant, imbibition and gelatinisation of gels and increase or decrease of rate of catalysis.The explanation of this undoubted parallelism is wholly unknown and in each individual case there are pro-nounced exceptions. I n the present instance the exceptions are the acetate and the sulphate although they may be paralleled by certain cases of catalysis. Possibly in the case of the sulphate our correction for metathesis (formation of HSO,’) has not been suffici-ently great. The effects are reconcilable with a solvate form of the dissociation theory. With regard t o the main thesis of this paper it has now been proved that the reactivity of the undissociated acetic acid is increased by addition of such salts as sodium chloride. A 0-2N-solution is affected to the extent of 5.5 per cent.by 0.2N-sodium chloride. It was pointed out in the earlier paper (Zoc. cit.) that Walpole’s measurements of electromotive force for this particular case gave a result for the hydrogen ion which was 7 or 8 per cent. too high. These two effects 5.5 and 7 or 8 per cenk. are equal within the experimental error and thus the effect on the dissocia-tion constant deduced for acetic acid cancels out and leaves that constant unchanged by the presence of the salt. Thus the electromotive force data may be regarded as agreeing with all the other data bearinq on this subject i n d the experi-mental evidence all points to the conclusion that the dissociation constants of weak substances are not appreciably affected by the addition of salts.One point we have not investigated namely the effect of salts on the chemical potential of the acetate ion. Summary. (1) It is shown experimentally that many salts enhance the partial vapour pressure of acetic acid in aqueous solution by very appreciable amounts. I n the case of 2-3N-sodium chloride the increase amounts to no less than 62 per cent. (2) Since this partial pressure is a measure of the reactivity of the undissociated acid in the solution. the undissociated acid must be regarded as exhibiting enhanced chemical potential in the presence of such salts. ‘This is parallel with the available data for the effect of such salh on the measurement of hydrogen ion by electromotive force. The enhancement is thus discovered to be operative on both sides of the chemical equation and hence to leave the dissociation constant of acetic acid sensibly unaltered 1346 RIVETT AND O’CONNOR SOME TERNARY These experiments remove the only remaining evidence (apart from the ambiguous behaviour of certain insufficiently investigated colloids) for the view that salts might have been regarded as increasing the strength of weak acids. (3) Whereas a number of salts increase the partial pressure of acetic acid to an extent proportional to the concentration sodium sulphate exhibits a more complicated behaviour whilst sodium acetate has only a very slight effect. All electromotive force data on weak acids in the presence of salts other than sodium acetate require to be corrected for the effeck here described. THE CHEMIOAL DEPARTMENT, UNIVERSITY OF BRISTOL. [Received September 17th 1919.