C. G. SUMNER 27 COLLOID CHEMISTRY OF GLIADIN SEPARATION PHENOMENA. BY H. L. BUNGENBERG DE JONG and W. J. KLAAR.* Received 3 I st I-ugust I 93 I . When preparing flour suspensions in very dilute acid we observed that after removing the suspended particles of starch and protein by centrifuging solutions were obtained which differed from other protein solutions by their cloudiness. I t appeared that these turbid solutions could be filtered through ordinary filters without any change in their turbidity. This cloudiness indicated an unusual phenomenon and, indeed a microscopical investigation showed that the liquid contained a large number of drops. * Laboratory Maatschappij de Korenschoof Utvecht Holland 28 COLLOID CHEMISTRY OF GLIADIN SEPARATION These drops did not contain any fat as was proved by colouring them with various dyes which would have coloured fat and it was evident that the drops consisted chiefly of protein.Their liquid nature was further confirmed by filtering through porous crucibles. We gave the name separation to this phenomenon of the formation of a new phase dispersed in the initial phase. Thus for example the initial phase may consist of a protein dispersed in water which may separate to a system of water dispersed in gliadin (the drops) which in turn is dispersed in water. As the system flour-water-acid was thought to be too complicated, and because one or two of the proteins of the gluten appeared to be the cause of the effect of separation we continued the investigation of this phenomenon with solutions of acid and purified gluten which had been dried in vacuo.We embodied the results in a preliminary communication,l and pointed out the connection between this phenomenon of separation and the hydration of the colloidal protein. At the same time a certain analogy between the systems phenol-water and colloidal protein-water was shown in the appearance of a super critical mixing temperature. These separated solutions appeared to become water-clear above certain temperatures whereas on cooling down the liquid separated once more. The cloudy separated solutions could be completely cleared up in alcohol of approximately 50 per cent. by weight while in higher alcohol concentrations the solutions became for a second time cloudy. As the solubility in alcohol of 50 per cent.by weight is a characteristic of gliadin (one of the proteins of the gluten) it was suspected that this protein was responsible for the phenomenon of separation. Therefore we have continued our investigations by studying the properties of gliadin.21 Following the directions of other investigators (Osborne Blish and Sandstedt Sharp and Gortner),4 we have made various preparations of gliadin which has been purified with the utmost care. I t was evident that our supposition was true and that the protein or proteins called gliadin showed the phenomenon of separation (Fig. I ) . Purification by different methods did not have any influence on this phenomenon. Just as with solutions obtained from gluten so in this case the separation was caused by change of charge and of hydration of the protein.This can be brought about not only by the addition of electrolytes and (with acid solutions) of alkali-both of which influence charge as well as hydration-but also by substances that only influence the hydration, e.g. alcohol acetone or resorcinol. On closer investigation of the literature i t appeared that Osborne, the investigator of the proteins of wheat also speaks in his book “The Proteins of the Wheat Kernel,” about cloudy solutions of gliadin. In I929 Berliner and Koopman found microscopically protein drops in acid flour suspensions. However neither of these investigators went any further into the matter. By the addition of alcohol to a cloudy aqueous gliadin sol a per-fectly water-clear soh tion was obtained at an alcohol concentration of Cereal Chem 6 ( 5 ) 373 1929.Ibid. 7 (3) 222 1930. 81bid. (6) 587 1930. * M. J . Blish and R. M. Sandstedt Cereal Chem. 3 144 1926 ; T. B. Osborne, “ The Proteins of the Wheat Kernel,’’ 1907 ; F. Sharp and R. A. Gortner J . physical Ciaem. 27. 481 1923. E. Berliner and J. Koopmann 2. gees. Miihlenwesen 6 57 1929 H. L. BUNGENBERG DE JONG AND W. J. KLAAR zg 48 per cent. + by weight. By further increasing the alcohol concentra-tion the cloudiness reappears. The same phenomena could be observed with acetone. To investigate these phenomena quantitatively we decided to use the viscosimetric method. For this purpose two different kinds of sols were made firstly clear colloidal solutions obtained by addition of a quantity of acid sufficient for complete peptization and secondly sols which were all microscopically separated to a more or less extent.These separated sols were made by adding too small a quantity of acid for complete peptization. Before discussing shortly the results of the viscosimeter measure-ments i t is desirable to point out that in colloidal problems the change of hydration of the particles is one of the most important factors. The relative viscosities of these sols were therefore calculated by dividing the viscosities of the alcoholic protein sols (qs+b) by those of the corre-sponding alcohol-water mixtures (7~). The value thus obtained ?$) the relative viscosity indicates the change in the value I + 5 / 2 4 in the Einstein formula i.e. the change in the total volume of the particles including their solvation layers.From our viscosity determinations i t was evident that by increasing the alcohol concentration in the sol the relative viscosity of the sol increases a t first until a maximum is reached in alcohol of about 48 per cent. by weight. By increasing the alcohol concentration beyond this maximum a sharp decrease in the relative viscosity of the sol occurs. Thus for instance in alcohol 80 per cent. by weight a relative viscosity of the sol was reached far beneath the initial value of the aqueous sol. The same effect could be observed with gliadin in acetone-water mix-tures. In this case the maximum was found a t an acetone concentration of 44 per cent. & by volume. The maximal value in the relative vis-cosity for acetone-water gliadin sols was larger than that for alcoholic gliadin sols starting from the same aqueous sol.From the foregoing facts it must follow that by increasing the alcohol or acetone concentration in a gliadin sol an increase in the solva-tion layer ( I + K+) must take place until the maximum is reached in the relative viscosity. (We neglect in this reasoning that there can also occur by changing the composition of the medium from an aqueous to an alcoholic one a change in the electric term in the formula of von Smoluchowski.) The increase of the protecting layer by addition of alcohol could be observed microscopically. By starting from a strongly-separated aqueous gliadin sol the drops could be seen to swell on adding traces of alcohol. By further increasing the alcohol concentration the 011 tlines of the drops become more and more vague and finally an optically void liquid was obtained.Attention must be drawn to the fact that the sols with maximal vis-cosity both in alcohol- (or acetone-) water mixtures show a very high stability. Neither electrolytes nor dehydrating substances such as resorcinol have any visible effect on the appearance of the water-clear sol; this may be contrasted with aqueous solutions where these sub-stances exercise a marked influence on the appearance of the sols. This stability must have its origin in a very special state of the solvation layer in our opinion this stability must be caused by a layer of alcohol or acetone hydrate molecules 30 COLLOID CHEMIS'I'RY OF GLIADIN SEPARATION The decrease in the relative viscosity in higher concentrations of these dispersion media must be caused by a gradual desolvation of the particles with the result that in practically pure alcohol or acetone the particles must have lost their protecting layer.In these concentrations of alcohol or acetone the sol will be only stable on its charge and is transformed into a suspensoid sol as indicated by its blue opalescent colour Tyndall effect etc. It appears that the shape of the corresponding viscosity curves with varying alcohol or acetone concentrations are the same for optically clear as well as for separated sols. However the percentage rise in viscosity (the difference in relative viscosity between the alcoholic sol at the maximum and the aqueous sol divided by the relative viscosity of the aqueous sol minus I i.e.the percentage increase of the total volume of the particles (#)) is much larger for separated sols. We next investigated the manner in which this percentage rise changed with the acidity of the initial sol. We found that starting from a separ-ated sol by increasing the charge of the particles this percentage rise decreases until a minimum was reached a t an acid concentration in the sol whereat the aqueous sol has its maximal viscosity. On increasing the acid concentration beyond this point a slight drop in this value occurs at first. At still higher acid concentrations the percentage rise increases finally reaching very high values. Here in an aqueous medium a second area of separation could be observed while the alcohol sols were all perfectly clear.As the shape of the curve show-ing the percentage increases plotted against the viscosities of aqueous sols at different p H is perfectly regular it follows that separation is a gradual continuous process in the microscopically visible as well as in the amicroscopical area. Therefore separation does not occur abruptly at a definite state of charge of the particles. These pheno-mena are influenced by the following factors (a) the charge of the particles (b) the hydration of the particles (c) the composition of the dispersion medium (d) the temperature ( e ) the presence of surface active material cf) lyotropic effect of electrolytes. We will now briefly discuss the separation of gliadin. The following illustrations will make this clear :-( I ) Variable charge of the particles in aqueous medium constant temperature.-(This effect is coupled with the lyotropic effect (f) in aqueous medium see page 31).(i) By a decrease of charge of the particles in a clear gliadin sol, either by a change in p~ of the sol or by addition of electrolytes the liquid separates into a number of small drops. By further decreasing the charge of the sol an enlargement of these drops will occur. (ii) So long as the particles are charged the visible drops do not run together. (iii) By total discharging (i.e. iso-electrically) the drops stick to-gether and form strings. Conversely starting from a thread of gliadin in water the thread is seen to expel from its surface a number of small drops when dilute acid is added.The drops spread into the surround-ing liquid and gradually melt away. ( 2 ) Variable medium constant temperature and constant charge on the particles. (i) Separated sols gradually become water-clear when substances are added which increase the solvation of the particles (acetone and different alcohols up to a certain concentration) H. L. BUNGENBERG DE JONG AND W. J. KLAAR 31 (ii) Higher concentrations of these solvents exercise a desolvating influence on the particles and a second separation effect can be observed. (3) Variable temperature constant charge on t h particles aqueous medium. (i) Separated sols become water-clear by increase of temperature of the sols. (ii) Clear sols become turbid on cooling down.(4) Variable temperature and charge q u o u s medium. The higher the charge of the sol particles the lower will be the tem-perature at which visible separation occurs. The curve obtained by plotting these turbidity temperatures against the acid concentration, has the same shape as the viscosity acid concentration curve. ( 5 ) Variable temperature and medium constant chargc of the sol. By increasing the solvation of the particles by change of medium as indicated by an increase in viscosity of the sol the temperature a t which visible separation occurs will be decreased. Therefore a sol with maximal viscosity in alcohol-water has a higher turbidity temperature than the same sol with maximal viscosity in acetone-water. (6) Varinble charge constant temperature constant medium (acetone-water).By decreasing the charge of clear acetone-water gliadin sol cooled to low temperatures a turbidity occurs. (7) Surface active substances. Clear aqueous gliadin sols can be separated by addition of somewhat high concentrations of resorcinol. In aqueous solutions containing still higher concentrations of resorcinol a clearing of the turbid liquid takes place. Before passing on to the discussion of the effects of separation of colloids in general there is a matter which deserves attention. On page 28 we compared the separation of colloids with the separation of the system phenol-water. This comparison holds good only superficially, and strictly speaking i t is not true. From solubility determinations it was evident that we cannot apply the phase-rule to a separated gliadin system at least if we consider the gliadin as a one-component system.This fact is in accordance with the effects found by several investigators among others Ostwald and his collaborators,s who have vainly tried to apply the phase-rule to colloid systems. By the discharging of colloidal solutions a t normal temperature two extreme possibilities can occur. ( I ) The remaining hydration of a dis-charged colloid can be so large that the whole system remains in solution ; or (2) the discharged particles are absolutely free from solvate molecules, in which case the system will flocculate into a solid sediment. Between these two extremes a number of intermediate cases can be found. In these cases the sediment will bind in an uncharged condition more or less solvate molecules.The proteins gelatin and albumin represent the first mentioned case. Both these proteins remain in solution iso-electrically under normal conditions ( i e . normal temperature and in the absence of dehydrating substances). On the other hand by the discharging of glutenin and casein a very fine flocculation can be obtained that microscopically shows a more or less solid nature (Fig. 2 ) . 6 Wo. Ostwald und R. Kohler Koll. 2 43 131 1927 32 COLLOID CHEMISTRY OF GLIADIN SEPARATION In gliadin and in general in the prolamins we have a number of pro-teins the properties of which stand between these two extreme cases. Here the semi-liquid nature of the sediment indicates a relatively high degree of hydration of the uncharged particles.This hydration how-ever is not sufficient to keep the protein in solution. In our opinion, this degree of solvation determines whether a protein will flocculate or separate or remain in solution in uncharged condition under normal circumstances. If this supposition is right the possibility must exist of forcing a protein into every possible form of sediment by influencing its solvation. Several phenomena point in that direction. For example, we have succeeded by the addition of substances which enlarge the solvation of the particles in separating glutenin and casein which normally flocculate. On the other hand albumin can be caused either to separate or to flocculate through a decrease of hydration. In gliadin we have a protein in which different forms of sediment can easily be realised by change in solvation.These changes can be brought about a t normal temperature by different alcohol or acetone concentrations. For instance uncharged gliadin which in aqueous medium forms a semi-solid sediment gives on addition of alcohol sediments of an in-creasingly liquid nature. On increasing the alcohol concentration up to 60 per cent. by volume the gliadin will be maximally solvated and will remain completely in solution and will be insensible even to high concentrations of univalent electrolytes. In this medium gliadin has obtained an albumin-like character. Whereas the concentrations of alcohol under 60 per cent. by volume increase the solvation of gliadin, higher concentrations have a desolvating influence.In consequence, starting from an uncharged gliadin sol in alcohol of 60 per cent. it is possible to realise all natures of sediments from liquid to solid. In addition to the influence which change of medium exercises on the nature of the sediment other physical conditions can have the same effect e.g. temperature and charge. Some proteins which normally flocculate can be made to separate a t higher temperatures. In the last few years these phenomena of separation have attracted the attention of several investigators. In particular H. G. Bungenberg de Jong and H. R. Kruyt and their collaborators have investigated a large number of colloidal systems and in most cases they succeeded in producing separation phenomena. We have discussed above the different factors influencing the pheno-menon of separation but even if the conditions are favourable the ques-tion arises How does this separation this forming of drops take place ? In our opinion the essential factors are not only the degree of solva-tion of the particles but also the surface tension of the particles with their layers.As the drops are liquid the solvation of the system must be somewhat high. As we have discussed a t length the discharging of a colloid with a small hydration will result in flocculation. If on the other hand the solvation is too large the discharging will not have any visible effect. In order to separate this last-mentioned sol the addition of a desolvating agent is necessary. As the phenomenon of micro-and macro-separation on the discharging of a charged colloidal gliadin solution is a continuous process (as has been pointed out) we must assume that strictly speaking such a sol of maximal solvation is already a separated system.From this it must 7 H. G. Bungenberg de Jong und H. R. Kruyt KoZZ. Z. 50 (I) 39 1930. We will first consider the solvation FIG. 1.-Gliadin (KI) x 360. [SEE p a p 28. FIG. 2.-Glutenin (iso-electric) x 200. [Seepage 32 FIG. j.--Influence of varying K I KNO, KBr and KCl concentration on an acid 80 per The concentration for every electrolyte decreases from left to right on the photo. cent. acetone gliadin sol. rsLc paRc 4 H. L. BUNGENBERG DE JONG AND W. J. ICLAAR 33 follow that even in a charged and solvated condition a colloid particle surrounded by its solvation-layer has a definite surface tension towards the dispersion medium.The surface tension however must be very small and this for the following reasons. Firstly a charged colloid particle has such a high solvation that a molecule on the outer sphere of the layer is only very weakly attracted towards the centre of the particle. Secondly by charging a particle the surface tension is decreased. From this it must follow that a decrease of the charge of solvated particles will give an in-crease in the surface tension. With regard to the enlarging of the drops by a decrease of charge, we can say that we have never observed any collision and subsequent confluence of drops so long as the protein is charged. The enlarging of the visible drops must therefore take place by collision with sub-microscopical protein particles.This phenomenon can be compared with the instability of highly-dispersed charged emulsions. Influence of Electrolytes on Positive Oliadin Sols. It is desirable briefly to pass in review the already known lyotropic Considering the formula of Einstein-Smolu- phenomena of emulsoids. chowsky in its most simple form w -m 778 -770 -770 we observe that a decrease of the first term can be brought about either (a) as a result of a decrease of fC2 or (b) as a result of a diminution of +. When equivalent concentrations of electrolytes of the same valency are added we observe a difference in the decrease in viscosity of protein solutions. This can be explained by either (a) a decrease in viscosity caused by a decrease of charge (quasi-viscous effect) the same for all electrolytes of the same valency or (6) a decrease in viscosity varying for electrolytes with the same valency (lyotropic influence).So long as a colloid particle is charged the added electrolyte molecules in the solution will direct their oppositely charged ions towards the charged particle. Therefore in the case of the positive protein the principal lyotropic influence will be exerted by the negative ions. For positive sols the lyotropic series for the univalent negative ions in small concentrations runs as follows SO, Fly Cl Br NO3 I CNS, where CNS has the strongest influence on the dehydration of the par-ticles and Fl exerts the least influence. In some special cases however, Rr and NO can change their places in this series.A similar series characteristic for the negative protein is found with the positive ions viz. Li Na K Rb Cs. Here too some of the ions can interchange. Although for the positive protein the principal lyotropic effect is defined by the negative ions the positive ions will nevertheless exert a certain influence on the dehydration capacity of the negative ions. For this reason the lyotropic effect for NaCl and KC1 in equivalent concentra-tion cannot be the same. We may assume that it is generally known that the lyotropic series occur not only in colloid-chemical phenomena but that they can be found in several other physical processes. For instance this series 34 COLLOlD CHEMISTRY OF GLIADIN SEPARATION appears in the change of the surface tension of water in the decrease of solubility of organic substances by addition of electrolytes etc.Many investi-gators consider the degree of hydration of the different ions to be the cause of the differences in lyotropic effect. The water binding capacity of different ions was determined by Remy Washburn,* and other investigators. Although the absolute values determined by these workers did not agree nevertheless generally speaking the sequence of the water-binding capacity of the different ions was the same as in the lyotropic series save for a few exceptions. pointed out emphatically that a hydration layer is not a definite firm shell of molecules but that the molecules in the outer sphere are less strongly orientated and are more or less movable.Consequently, there will be no hard and fast line between bound and unbound mole-cules. This must be the reason why different methods of determining the hydration give different results. Other investigators sought a connection between the lyotropic effects and the difference in intensity of adsorption of the electrolytes. The following series was found by adsorption of electrolytes with the same cation on charcoal CNS I Br C1 SO,. This adsorption is probably very closely connected with the hydration and very likely these effects are both expressions of the same specific property of the ion. There remain to be mentioned the researches of Gortner Sinclair and Hoffman lo carried out more particularly upon wheat proteins. These investigators experimented on the influence of higher electrolyte con-centrations on the peptisation of the total proteins of wheat and found a lyotropic series for the quantity of protein dissolved in different elec-trolyte solutions.This series was very marked for the negative ions, but for the positive ions it was less distinct ; Na and K have changed places. Here (as always in higher electrolyte concentrations) the order of the lyotropic series is the reverse of that found in small electrolyte concentrations. we published some viscosity determinations of the lyotropic influence of electrolytes on aqueous positive protein sols. In this case we found the usual order of the series Cl<Br<NO,<I<CNS. The SO”, however did not take its place. Evidently the discharging effect of the bivalent ion outbalances its lyotropic influence in small concentrations.Therefore i t is impossible to get an impression of the lyotropic influence of this ion by comparing it with the univalent ions. In the present paper we shall deal with the influence of electrolytes on the viscosity of gliadin sols in alcohol and in acetone solutions. These investigations carried out in various media will give us a certain insight into the viscosity effects caused by the lyotropy and the charge of the ions. At the same time they confirm the gradual nature of the separation phenomena and the increase in stability of the particles in definite alcohol- and acetone-water mixtures. The lyotropic series can be explained in several ways. Robinson In an earlier paper AIcohol Medium.In an earlier publication the influence of varying alcohoI concentra-I t was pointed 8 H. Remy Die electrolytische Wasseriiberfuhrung Berlin I 927 ; Washburn, 9 C. Robinson The Lyotropac Series Diss. Utrecht 1929. 10 Gortner Hoffman and Sinclair Colloid Symposium Mon. 5 179 1927. tions on the viscosity of a gliadin solution was studied. J . Am. Chem. Soc. 31 322 1909 ; ibid. 37 674 1915 H. L. BUNGENBERG DE JONG AND W. J. KLAAR 35 separated with different electrolytes and the following experiments will 1.05 make i t clear that in 48 per cent. /* (a) Firstly we sought to ascertain whether the influence of different alcohol concentrations on an acid 0. 10. r5. 20. 25. 30. 35. 40. 45. - 1-855 2.325 2'718 2.949 3'021 2-913 2-644 -1.784 2'233 2.583 2'7So 2'838 2.760 2'555 2'245 -* v.= very ; sl. = slightly ; op. = opalescent ; fzocc. = flocculated. cld. = cloudy ; clr. = clear ; The appearance of the solution separated by electrolytes changes with increasing alcohol concentration. Whilst the aqueous gliadin solution containing 10 mil. mol. KCl is cloudy this turbidity disappears by adding alcohol and on reaching a concentration of 30 to 40 per cent. alcohol by volume the protein solution becomes water-clear. It remains so till about 70 per cent. alcohol by volume. C.C. Alcohol. %+U % 1.030 1.039 1.046 1.052 1.~61 1.065 1-055 1.035 -- 30 53 73 103 117 83 I7 -v. sl.cld. sl. op. clr. clr. clr. sl. o f . sl. cld. jlocc. %+a -la rise per cent. appearanc 36 COLLOID CHEMISTRY OF GLIADIN SEPARATION At alcohol concentrations of more than 70 per cent.by volume the protein solution becomes again opalescent and afterwards cloudy where-as a t a concentration of 80 to go per cent. alcohol flocculation results. These phenomena are perfectly in accordance with the explanation given in a previous paper ie. the alcohol at first exercises a dispersing effect caused by the enlarging of the protecting layer. Coupled with this is a stabilising effect. In higher alcohol concentrations the gliadin loses one of its stability factors and passes in a suspensoid. By the influence of the electrolyte in solution discharging takes place and flocculation results. The curve obtained from the data of Table I. appears to have the same shape as that for a sol without electrolyte.The maxima in both these curves are found a t the same alcohol concentration. The percentage rise in viscosity at the maximum %!&I- " has a very large value because the 1)II - 1 aqueous solution was separatGd just as in the case of sols separated by a quantity of acid too small for complete peptisation. (b) In addition an investigation was carried out upon the change of the percentage rise in viscosity in 60 per cent. alcohol by volume TABLE II.-INFLUENCE OF CHANGING ELECTROLYTE CONCENTRATION ON THE RE-LATIVE VISCOSITY OF AN ACID GLIADIN SOL IN WATER AND IN ALCOHOL 60 PER CENT. BY VOLUME. 0'4. 0-8. 2. 4. 6. 8. 1.087 1.071 1.053 1.036 1.029 1.024 3-184 3'150 3'102 3'057 3.048 3.037 1.124 1.112 1'095 1'079 1'076 1.072 42 57 79 119 162 zoo clr.clr. sl. Op. Op. cld. cld. Mil. Mol. KI Per Litre. I ?la %+a %+a ?a rise per cent. appearance of hydrosol (at the maximum) with different quantities of electrolyte added. For this purpose 2.5 C.C. of a protein sol with varying quantities of KI and 15 C.C. alcohol were diluted in a volumetric flask to 25 C.C. in a thermo-stat at 25" C. After standing one hour the viscosity readings were made. The final concentrations of KI in solution are 0.4 0.8 2 4 6 and 8 mil. mol. per litre. Table 11. gives the results of the measurements. I t is evident that with increasing separation by addition of more electrolyte the percentage rise in viscosity increases. Since the shape of the curve (obtained by plotting the values of the percentage risesin viscosity in 60 per cent.alcohol by volume against the viscosities of the corresponding hydrosols) is perfectly regular it follows that the separa-tion is a continuous process and does not occur suddenly at a critical charge of the particles. This is in accordance with the previous measure-ments carried out with separated solutions obtained by peptisizing with too small a quantity of acid. (c) In order to investigate the influence of alcohol on the lyotropic effect of different monovalent electrolytes the following experiment was done :-In a volumetric flask of 12.5 C.C. capacity I C.C. of 0.15 N KC1 or KI was added to 10 C.C. of a-gliadin sol containing respectively 0 20 3 H. L. BUNGENBERG DE JONG AND W. J. KLAAR 37 KCl KI KCI K I K C1 KI and 45 C.C.alcohol per 50 C.C. liquid. After making up to the mark at 25' C. with distilled water the viscosities of these solutions were deter-mined as well as those of the corresponding alcohol-gliadin solutions without electrolyte and of the corresponding alcohol-water mixtures. The relative viscosities 'aA of these three sols (KCl KI and without electrolyte) in these different alcohol media were then calculated by dividing the viscosities of the alcohol sols by the corresponding visco-sities of the alcohol-water mixtures. By interpolation from an alcohol-water viscosity graph these alcohol-water mixtures were found to con-tain approximately 0 26 50 and 65 per cent. alcoho! by weight-Table 111. gives the results of these measurements. VA 0. 26. 50. 65.1 Per Cent. Alcohol by Weight. 1.088 2'640 3'164 2.897 I'O_?O 2'495 3'000 2'761 1 %+a 1.009 2.470 2.996 2'757 1'000 2.396 2'833 2.633 1 ?a 1.088 1.102 1.117 IVOI 1'030 r*o41 1.059 1,049 1'0og 1'031 1.058 1.048 7s+a - 1 va } rise per cent. - 36-6 96.6 63-3 244' 544' 433' -TABLE III.-INFLUENCE OF DIFFERENT UNIVALENT NEGATIVE IONS ON THE RE-LATIVE VISCOSITY OF AN ACID GLIADIN SOL WITH VARYING ALCOHOL CON-CENTRATION. These data indicate ( I ) Two gliadin solutions with the same charge on the particles have a different percentage rise in viscosity when the same quantity of alcohol is added to both. (2) The percentage rise in viscosity by adding the same quantity of alcohol is larger for an electro-lyte with a strongly dehydrating ion (lyotropic ion) than for one con-taining a less strong ion.(3) By increasing the alcohol concentrations till about 50 per cent. the relative viscosities of the gliadin-KC1 sol and KI sol come up to each other. (4) In a medium of alcohol about 50 per cent. by weight the viscosities of the KCl and KI sol coincide within the experimental error. (5) This phenomenon takes place in the same alcohol concentration wherein the gliadin alcohol sol without electrolyte has its maximum of viscosity. It does not follow that protein solutions of the same concentration: with the same viscosity in water will show the same percentage rise in viscosity in 60 per cent. alcohol by volume. We may consider two aqueous solutions of the same protein concentration with the same viscosity the first solution obtained by adding a certain quantity of KCl the second one by adding a different quantity of KI.The colloid particles in these solutions do not possess the same charge for in aqueous medium the decrease in viscosity by adding electrolytes to mos 38 COLLOID CHEMISTRY OF GLIADIN SEPARATION KCI KI KC1 KI emulsoids is not onlv the result of the discharging effects of the ion, but a second factor the lyotropy plays a leading part. As the decrease in viscosity due to the lvotropic effect of the ion is less for the C1 ion than for the I ion at least for positive sols and with small electrolyte concentrations a larger quantity of KCl than of KI must be added to the same protein solution to get the same lowering in viscosity. Therefore in the system KC1-gliadin the particles are less charged than in the system KI-gliadin.On adding alcohol to these two solutions the separated particles will be completely dispersed and clear solutions will result. The vis-cosities of these alcoholic solutions must be different because the charge and the solvation of the particles are different. Consequently the percentage rises in viscosity for these KC1 and KI gliadin solutions must be different in a medium of the same alcohol percentage. On the other hand it is to be expected that solutions separated with equivalent quantities of KCl and KI (i.e. solutions in which the particles have the same charge but different solvation) will also show different percentage rises in viscosity in a medium of the same alcoho! concentration.This is in accordance with the experimental facts of Table 111. (d) We next sought to ascertain whether the phenomena found with KC1 and KI in a concentration of 12 mil. mol. in 50 per cent. alcohol by weight were the same with all concentrations of these electrolytes, provided that the electrolytes were equimolar in solution. Table IV. gives viscosity measurements carried out upon an acid gliadin sol in 50 per cent. alcohol by weight by adding to this sol 0.8 2 and 6 mil. mol. KC1 and KI. 1.076 1.062 1'045 1-07 I 1.053 1'029 } qs 3'147 3.102 3.050 3'150 3.102 3'048 } qs+a TABLE TV.-INFLUENCE OF CHANGING CONCENTRATION OF KCl AND KI ON THE RELATIVE VISCOSITY OF AN ACID GLIADIN SOL IN ALCOHOL 48 PER CENT. BY WEIGHT. I 1 0.8. 2. 6. 1 Mil. Mol. Electrolyte.I I From these data i t is evident that the effects found in 50 per cent. alcohol by weight are independent of the electrolyte concentration, except that by increasing the concentration of electrolyte the viscosity of the gliadin sol in 50 per cent. alcohol by weight decreases. Summing up our investigations we can say that in about 50 per cent. alcohol by weight the effects of KI and KC1 on the positive gliadin particle are the same. Before discussing these measurements the figures of the measurements in acetone medium will be given in order to have more facts a t our disposal H. L. BUNGENBERG DE JONG AND W. J. KLAAR 39 Acetone Medium. -4 large number of measurements were carried out in varying acetone-water media and with different m i - and divalent electrolytes.This medium was preferred to the alcoholic medium because the latter must be absolutely free from oxidising agents to get accurate determinations. Traces of free iodine-caused by the oxidising actions of impurities in the alcohol on KI-have a marked lowering influence on the viscosities of gliadin sols. Acetone was free from these substances. A pure prepara-tion was used which was moreover rectified at its boiling-point. The investigations in acetone media can be divided into three sections : ( I ) The influence of acetone on acid gliadin sols in the presence of dif-ferent univalent electrolytes (ie. negative ions). (2) The influence of acetone on acid gliadin sols in the presence of di or polyvalent electrolytes. (3) The influence of changing p on the effects mentioned under the first and second sections.(4) The phenomena mentioned under sections I and 3 for higher protein concentrations. For the determinations in this part of the paper the electrolytes were added only in very small concentrations. The influence of higher elec-trolyte concentrations will be dealt with separately. I. Univalent Electrolytes. From the previously mentioned analogous behaviour of acid gliadin sols in alcohol and acetone media without electrolytes i t was to be expected that the effects of electrolytes on sols in acetone would be analogous to those in alcohol. Aqueous sols separated by addition of electrolytes became water-clear again when acetone was added. In higher acetone concentrations for a second time a turbidity of the sol appears.If a sufficient quantity of electrolyte is added and the acetone concentration in the sol is high enough flocculation re-sults owing to the discharging of the practically desolvated particle. If the concentration of 3.05-acetone in the sol is too low but if there is a sufficient quantity of electrolyte there results on the other hand sediments which intermediate between separation I ,oo- I I and flocculation. Relative vis- 5 10 15 20 cosity curves from the figures in the following tables were ob-concentration in a gliadin sol separated by electrolytes; these curves had a form analogous with that of a sol without electrolyte. The maximum in the relative viscosity curve was found at the same acetone concentration with or without electrolytes.It was found that the relative viscosities of gliadin sols with equivalent concentrations of KC1 and KI approached one another with increasing acetone concentration. In acetone 44 per cent. by volume (at the maximum) the relative viscosities of the sols were iden tical. At higher acetone concentrations the relative viscosity curves separate again. Table V. (Fig. 4) This was actually the c x e . /-'* e. \B r (microscopically observed) are - *A C.C. Acetone per 25c.c. FIG. 4. -See Table V. Curve A = KCl B = KJ. tained by varying the acetone We shall refer to this later 40 COLLOID CHEMISTRY OF GLIADIN SEPARATION gives a series of measurements carried out on sols containing 4 mil. mol. KI and KCI and varying acetone concentration. The purpose of the following experiments was to determine whether tpis coincidence of the relative viscosities of gliadin sols with 4 mil.mol, KI and KC1 in acetone of about 44 per cent. was the case for all con-centrations of these electrolytes. TABLE V.-INFLUENCE OF VARYING ACETONE CONCENTRATION ON THE RELATIVE VISCOSITY OF GLIADIN SOLS WITH 4 MIL. MOL. KCI AND KI. 6. g. 12. 15. 18. 20. -~ 1'233 1.443 1.579 1'600 1'479 1'221 0.987 1.308 1.547 1'705 1'730 1'590 1.286 1'005 1.047 I*&I 1-072 1.079 1.081 1.075 1.053 1.013 sl. op. E l . JIOCC. 1.29s 1-540 1.702 1.729 1.591 1'292 1-016 1.029 1,053 1.067 1.078 1-081 1.075 1.058 1.024 OP* cc Acetone. lac %+ac %c appearance 778+aC -qs+ac rlae appearance 7sSa;c For this purpose 12 C.C.acetone? different equivalent quantities of KI and KCl and 2.5 C.C. of an acid gliadin sol were put into volumetric flasks of 25 C.C. capacity. After making up to the mark at 25' C. the final concentrations of the electrolytes were respectively 0 2 4 6 8, 10 12 14 20 28 and 40 mil. mol. per litre. At the same time the TABLE VI.-INFLUENCE OF VARYING CONCENTRATIONS OF KI AND KCI ON THE RELATIVE VISCOSITY OF AN ACID 44 PER CENT. ACETONE GLIADIN SOL. ~~ Mit Mol. %+Ac+ KCI ~ A C + W+ KCI T A ~ + w + KI 7)s +Ac+KI l s + A c + KI I TAc %+AC+KC~ II qAc 0. 2. 4. 6. 8. 10. 1% 20. 28. 40. 80. 1805 1737 1722 1716 1713 1703 1702 1697 1692 1695 1694 1 6 0 2 - - - - - 1605 1600 1597 1601 1602 1805 1742 1722 1714 1711 1702 169s 1694 1688 1685 1677 1602 - - - - - 1600 I597 I593 I594 I587 1127 1088 1075 1070 1069 1063 1062 1061 1060 1057 1057 1127 1086 1075 1071 1070 1064 1061 1060 1060 1059 1058 I corresponding acetone-water-electrolyte mixtures were made.The viscosity determinations were carried out upon sols which had been left for twelve hours. Table VI. gives the results of these measurements. If we examine these data the following facts can be observed (a) The relative viscosities of sols with equivalent concentrations of KI and KCl are the same within experimental error in about 44 per cent. aceton H. L. BUNGENBERG DE JONG AND W. J. KLAAR 41 medium. (b) By increasing the electrolytic concentration in the sols a. gradual decrease in the relative viscosity takes place. In acetone concentrations of more than 44 per cent.by volume the viscosity curves of KI and KC1 separate again (see Table VI.). In order. to get a clearer insight into this phenomenon the influence of other rnono-valent ions on an acid gliadin solution in higher acetone concentrations was studied. For this purpose 20 C.C. acetone and 3.5 C.C. of electrolyte solutions of varying concentration were added to 2 C.C. of an acid aqueous gliadin sol in steamed flocculation tubes so that the final concentrations of the used electrolytes were respectively 4 5 6 and 7 mil. mol. and the final acetone concentration was about 80 per cent. by volume. The electrolytes used were KCl KBr KNO and KI. The appearance of the liquids in the tubes shows a t once that the same concentrations of TABLE VII.-INFLUENCE OF VARYING CONCENTRATIONS OF DIFFERENT UNIVALENT ELECTROLYTES ON A GLIADIN SOL OF 80 PER CENT.2 ACETONE BY VOLVME. Nil. Mol. KI I \ 4. 5. 6. 7. slightly opal- opalescent slightly cloudy cloudy escent cloudy cloudy cloudy cloudy trace sediment opalescent slightly cloudy cloudy cloudy slightly cloudy slightly cloudy slightly cloudy opalescent trace sediment sediment sediment slightly cloudy cloudy cloudy cloudy slightly cloudy slightly cloudy opalescent clear trace sediment sediment sediment sediment very cloudy very cloudy very cloudy very cloudy slightly cloudy opalescent clear clear sediment sediment sediment sediment after I hr. after 36 hrs. after I hr. after 36 hrs. after I hr. after 36 hrs. after I hr. after 36 hrs. different univalent electrolytes in this acetone medium have different influences on the partly desolvated gliadin ; e.g.after thirty-six hours 2 slight trace of sediment appeared on the bottom of the tube containing the highest KI concentration (7 mil. mol.) while the tubes containing respectively 6 5 and 4 mil. mol. were only turbid. On the other hand, with a KC1 concentration of 4 mil. mol. the protein has practically settled down and with higher KCl concentrations the supernating liquid was practicrtlly clear. KNO and KRr were found to be between these two extremes. Fig. 5 shows the different influences of the monovalent ions in acetone of about 80 per cent. From this we may conclude that in higher acetone concentrations the lyotropic influence of the ions reappears but in a series reversed from that in aqueous medium.In other words in higher acetone concentrations KC1 ha; the strongest desolvating influence on th 42 COLLOID CHEMISTRY OF GLIADIN SEPARATION particle while KI influences the solvation of the particles least. In the last section of this paper we shall deal with this phenomenon; for the present it suffices to say that in acid gliadin solutions the lyotropic series of the monovalent ions is dependent on the nature of the medium. We thought it desirable more closely to study viscosimetrically the influence of the other monovalent electrolytes on an acid gliadin sol in varying acetone media. As the relative viscosity curves with varying acetone concentration of a KI and KCl gliadin sol intersect in acetone of about 44 per cent.by volume there were two possible ways in which the other monovalent electrolytes might behave ( I ) The respective viscosity curves of a gliadin sol with the same concentration of mono-valent electrolytes in varying acetone medium might run through the same point of intersection (it?. about 44 per cent. acetone) ; (2) the respective viscosity curves might intersect a t different acetone concen-trations. If we found that the respective viscosity curves of a gliadin sol with the same concentration of KCl KBr KNO, K I and KCNS ions meet at a point corresponding to the same percentage of acetone (in other words in a special acetone concentration (44 per cent. & acetone) the relative viscosity is independent of the nature of the monovalent electrolyte and we should have thus an indication that the important criterion is not the structure of the ion but only the state of the protein particle in this medium.However if we found that the respective viscosity curves of a gliadin sol with different monovalent ions intersect a t different concentration of acetone such a deviation should be markedly found for KCNS and KNO, because of the complex nature of the negative ions. In the case of KBr the properties of which stand between KC1 and KI a big devia-tion from the behaviour of KC1 and KI was not to be expected. In order to investigate these possibilities the following experiments were carried out. To 2.5 C.C. of an acid gliadin sol in several volu-metric flasks of 25 C.C. that quantity of electrolyte was added which was necessary to ensure that the final concentration was 6 mil.mol. after adding respectively 0 3 6 10 12 1 5 and 18 C.C. acetone and making up to the mark with distilled water. At the same time the correspond-ing acetone gliadin sols without electrolyte were made. The electrolytes used were KCl KBr KI KNO and KCNS. At the same time we sought to ascertain whether an organic electrolyte sodium sulpho-methane * (CH,. S0,Na) fitted in the same series. The viscosities of the aqueous solutions were determined immediately after the solutions were made up those of the acetone sols after twenty-four hours. The vis-cosities of the gliadin sols with the same acetone concentration though with different electrolytes were measured in the same viscosimeters.Table VIII. gives the data of these measurements. We may conclude from this : ( I ) With increasing acetone concentration the relative viscosities of the sols with the same quantity of different univalent electrolyte, increase and approach each other. ( 2 ) In acetone concentration under 44 per cent. the lyotropic series remains the same as in the aqueous medium (CNS>I>NO,>Br>Cl). (3) Within experimental error the relative viscosity of a sol in 44 per cent. acetone is independent of the nature of the added univalent electrolyte provided that the concentration is the same. * We are indebted to Prof. Backer of the University of Groningen who pro-vided us with this and other substituted methane preparations H. L. BUNGENBERG DE JONG AND W. J. KLAAR 43 ELECTROLYTES.Electrolyte Without { { . CH3S03Na KCl { ' KBr { { K 1 { KNO, ' KCNS { (4) In acetone of a concentration of about 44 per cent. by volume, (5) In acetone concentration of more than 44 per cent. the respective the relative viscosity curves intersect. viscosity curves separate. I 0. 3. 6. 10. 12. 15. 18. 1.359 1.616 - 1.823 - 1.351 -rogo 1.102 1-120 - 1.140 - 1'106 - 1.304 1'539 - 1.719 - 1-282 1.040 1'057 1.067 - 1-074 - 1.050 - 1.299 1.538 1-716 1.719 1.582 1.274 1.036 1'053 1.066 1'073 1'074 1.069 1'044 - 1.293 1.535 1.716 1.717 1.582 -1.032 1.049 1'064 1,073 1,073 1~069 -- 1'293 1'534 1.718 1.717 1'583 1.280 1'029 1'049 1.063 1'074 1'073 1.070 1.049 - - - - 1'718 - 1.282 1.018 - - - 1'073 - 1-050 - 1.286 1328 1.716 1.719 1.584 1.289 1'015 1'043 1.059 1'073 1'074 1'071 1'056 i 1 - 09108 15932 20404 20404 16994 08664 I cc.Acetone. %+aC 78+UC 'lac %+UC %+ac qac %+ac 3ac q8+UC c (6) This second separation gives a lyotropic order the reverse of that shown in aqueous medium (Cl>Br>NO,>I>CNS). (7) Sodium sulphomethane has a weaker lyotropic effect in aqueous medium than KC1 therefore it takes the place in the lyotropic series, which the SO,"-ion would occupy if i t could act as a univalent instead of as a bivalent ion 44 COLLOID CHEMISTRY OF GLIADIN SEPARATION (8) In acetone concentration of more than 44 per cent. sodium suI-phomethane behaves in an abnormal way. I t was to be expected that i t would have a desolvating effect stronger than KC1.I t is possible that the anomaly in this organic medium is due to the organic character of the anion. From a large number of viscosimetric determinations carried out on acid gliadin sols of practically the same H-ion concentration and con-taining varying quantities of KC1 and KI as well as of acetone it was possible to construct the three-dimensional graph given in Fig. 6. On the axes are plotted the electrolyte concentration the relative viscosity of the sols and the acetone concentration. In the interests of brevity, we refrain from giving the data in a table. As the sols were-made on different days slight fluctua-tions in the values used are possible. a This is however not I important for our I purpose was merely to get a general im-pression of the be-haviour of gliadin in I these media.I t is to be ex-pected that while the X viscosimetric curves, obtained by the ad-dition of varying quantities of dif-ferent electrolytes te different proteins will have the same shape in aqueous medium, this will probably not X Axis Concentration of acetone. be the case in binary mixtures of solvents Y Axis Relative viscosity. Z Axis Concentration of electrolyte. - KI. - .- KCl. so that possible dif-ferences between dif-ferent proteins will be emphasised under these circumstances. From the graph obtained in this way it appears that the shape of the curve of the relative viscosity plotted against the acetone concen-tration is not dependent upon the quantity of electrolyte added i.e.the charge of the sol. By increasing the electrolyte concentration the-curves merely lie on a lower level. The acetone concentration at which the maximum occurs does not change with the charge of the sol. FIG. 6. 2. Polyvalent Ions. In order to study the influence of bivalent ions the same method was followed as in the case of monovalent ions. Here too we studied the change in the relative viscosity of gliadin sols with a constant quan-tity of K,SO, but with varying acetone concentrations H. L. BUNGENBERG DE JONG AND W. J. KLAAR 45 CH2(SoaK)z [ t The shape of the curve was similar to that of gliadin sols with a univalent ion with a maximum a t an acetone concentration of about 44 per cent. At higher acetone concentrations flocculation occurs as a result of the discharging influence of the bivalent ion and of the partial desolvation of the gliadin particle.By comparing the curves of gliadin sols with equivalent concentration of KC1 and K,SO, i t appeared that the maximum in the relative viscosity curve for the sol containing sulphate was lower than that for the sol with KCl. This was to be expected; since the sulphate ion being bivalent has a stronger discharging influence. We next investigated the influence of two different bivalent elec-trolytes namely K2S04 and CH,(SO,K),. When these electrolytes were added to aqueous gliadin sols the viscosity determinations showed that the lyotropic influence is not the same. This difference is not so striking 1'824 1-772 1.701 1-695 1.693 1.692 1'140 1'071 1'063 1.060 1'058 1.058 TABLE IX.-INFLUENCE OF VARYING K,SO AND CH2(S0,K) CONCENTRATION ON THE RELATIVE VISCOSITY OF A POSITIVE AQUBOUS AND 44 PER CENT.ACETONE GLIADIN SOL. I. 2. 4. 6. 8. I ** 1.089 1.043 1.033 1.027 1.022 1.018 v. sl. op. sl. cld. sl. cld. cld. cld. 1.089 1.035 1'027 1.021 1.017 -op. sl. cld. cld. v. cld. flocc. 1.824 1.711 1'699 1.697 1-691 1.691 K2SO* 1.140 1.070 1-062 1.060 1.057 1'057 Mil. eq. q s + w appearance 75+w I appearance 9s + Ac 75 + Ac - - I qAc as is found in the case of different univalent ions though it is clearly perceptible. In aqueous media CH2(S03K) has a stronger dehydrating influence than K,SO,. The differences in viscosity in aqueous solutions found for equivalent concentration of these electrolytes disappear in 4.4 per cent.acetone In this acetone medium the relative viscosities of gliadin sols with equivalent Concentration of these two electrolytes co-incide within experimental error. Table 1X. gives a review of viscosity .data. In higher acetone concentration the same phenomenon is displayed as in the case of univalent ions namely the reversal of the lyotropic series. -4s the lyotropic difference between the SO ion and the CH,(SO,) ion is v e r y small this reversal of the lyotropic series in higher acetone con-centration is not so marked as for the univalent ions. I t appeared that with gliadin solutions containing about 75 per cent. acetone and respec-.tively 2 4 6 and 8 mil. eq. of the afore-mentioned electrolytes thos 46 COLLOID CHEMISTRY OF GLIADIN SEPARATION solutions containing 4 6 and 8 mil.eq. of either electrolytes were com-pletely flocculated after twelve hours. The solution containing 2 mil. eq. K,SO was only partly precipitated while the solution of 2 mil. eq. CH,(SO,K) was merely cloudy and did not show any precipitate. Thus ail the effects produced by addition of bivalent ions to a 44 per cent. acetone gliadin sol are the same as for the monovalent ions except that the discharging effect is not the same. We next studied the influence of tri- and tetra-valent ions on an acid gliadin sol of 44 per cent. acetone. The trivalent electrolyte used was CH(SO,K),. As this substance is only slightly soluble in the ace-tone medium it was only possible to reach a final concentration of I mil. eq. When more electrolyte was added not only the salt but part of the protein was precipitated (after the addition of acetone).1 I 1 I 5 10 15 20 mil. quiv. electrolyte per litrc FIG. 7 . 4 e e Table X. , B = K2S0, , C = K,Fe(CN),. Curve A = KCI By addition of K,Fe(CN) to acid 44 per cent. acetone gliadin sols, another phenomenon was observed. The first traces of K,Fe(CN) have no influence on the appearance of the solution but there is a sharp drop in the viscosity. After this at a concentration of K,Fe(CN) of about I to 4 mil. eq. the solutions are more or less cloudy. The viscosities of these solutions were determined after thirty minutes instead of twenty-four hours because after a day the protein had settled down. By addi-tion of more K,Fe(CN) the solutions are again absolutely clear and there is a gradual small rise in viscosity.Table X. gives the results (Fig. 7). When the data of Table XI. for KC1 K,SO, CH(SO,K) and K,FeCy, are plotted we observe the following facts ( I ) On discharging of acid 44 per cent. acetone gliadin sols there is a distinct divergence of the relative viscosity curves for ions of different valency ; ( 2 ) the viscosities of sols discharged with ions of different valency do not reach the same level. Znfluence of Changing H-Ion Concentration on 4 Per Cent. Acetone Ciliadin Solutions containing Electrolyte. The foregoing conclusions are valid only for an acid-acetone sol of In order to obtain a more complete insight into maximal viscosity H. L. BUNGENBERG DE JONG AND W. J. KLAAR 47 the influence of electrolytes of different valency it was thought desirable to study the effect a t different p ~ .With this end in view the following questions were studied ( I ) What is the effect of uni- and polyvalent ions on the relative viscosity of an uncharged 44 per cent. acetone gliadin sol? ( 2 ) What is the effect of changing p H on the relative vis-cosity of a sol containing varying quantities of KCl? (3) What is the influence of changing p H on the relative viscosity of a sol containing a definite quantity of uni- or polyvalent electrolyte ? (In our experiments we used 10 mil. eq. of electrolytes because the viscosity of an acid sol practically did not decrease when still more KCI was added.) For the experiments in this part the following method of preparing the sol was used.In a thermostat a t 25' C. 1.3636 gr. gliadin was shaken with a mixture of acetone and water. This mixture was made a t 25" C. by diluting 200 C.C. acetone with distilled water to 500 C.C. in a volumetric flask. The protein solution thus obtained was not filtered in order to avoid evaporation of acetone but was allowed to stand for thirty minutes so that small particles of filter paper could settle down. Two and a half cubic centimetres of this gliadin solution were pipetted into 25 C.C. volumetric flasks with 11 C.C. acetone and the necessary quantity of electrolyte and made up to the mark a t 25' C. The advantageof I I 2.5 580 7.5 10 u.5 mil. tquiv. cktrdyte pcr litre FIG. &-See Table XI. Curve A = K,Fe(CN), 9 7 B = &SO , C = KCI.preparing the solutions by this method was that the protein as well as the acetone concentration (44 per cent.) were practically the same as in those used in our former experiments. Thus it is possible to compare these experiments with the previous measurements. From the Table XI. (Fig. S) i t will be evident that addition of varying quantities of uni- and polyvalent electrolytes has wholly different effects on the relative viscosity of a practically uncharged gliadin sol. Prac-tically no lowering in the relative viscosity of these sols took place by addition of varying quantities of KC1 ( I The values of these measurements plotted against the KCl concentration give a horizontal line. On the other hand the relative viscosity curve, obtained by addition of varying quantities of the bivalent K,SO, shows another shape.The first additions of K,SO give a slight lowering in the relative vis-cosity reaching a minimum at a concentration of about 1.5 mil. eq. By further increasing the K,SO concentration the relative viscosity rises gradually even above its initial value. With K4Fe(CN)6 the same phenomenon (but even more pronounced) can be observed. Very small quantities cause a decrease in the relative viscosity. The minimum in the relative viscosity is found a t a concentration of about 0.25 mil. eq. By increasing the K4Fe(CN)6 concentration a rise in the relative 8 10 and 13 mil. eq.) COLLOID CHEMISTRY OF GLIADIN SEPARATION I Y i' N P I I Y U P % t' z f h h H s f' 00 u, !-I d op I s U LD H Q U I I 3 ? f In h W 0 m ? W m H OI OI ? m I 0 * F I I I I I I I I I I u n U 9 I * N Dp U I I I I I I I I I I 3 I m n 0 d P viscosity occurs and this reaches a value distinctly above the relative viscosity of the sol without elec-trolyte at a concen-tration of 3 mil.eq. We may conclude from these experi-ments that the mini-mum in relative vis-cosity shifts with the valency of the nega-tive ion. The higher the valency the lower will be the concentra-tion of electrolyte at the minimum and the sharper will be the rise in relative viscosity afterwards by adding more elec-trolyte. We next investi-gated the influence of changing H-ion concentration on an acetone gliadin sol with varying quan-tities of KCl.Table XII. gives a series of measu remen t s, showing the final concentration of elec-trolyte and acid in the sols. It will be apparent (I) that w i t h increasing charge of the gliadin sol the discharging effect of KCl will be more prominent ( 2 ) that the level of the relative viscosity (ob-tained by discharging different charged sol with 10 to 14 mil. eq. KCL) rises with in-creasing charge of the initial sol (3) that this rise in level does not stop a t that PH at which a sol with-out electrolyte has it H. L. BUNGENBERG DE JONG AND W. J. KLAAR 49 TABLE XI.-INFLUENCE OF VARYING EQUIVALENT CONCENTRATIONS OF K C I K,SO AND K F e ( C N ) ON THE RELATIVE VISCOSITY OF A GLIADIN SOL WITHOUT ACID I N ACETONE 44 PER CENT.BY WEIGHT. Mil. Eq. per Litre. 0. 0.125. o'q. o.5. I. 1-5 2. 3. 4. 8. 12. 1 Electr. 1 1.690 - - - 1.689 - - 1.687 1.688 1.689 1'689 TJs+Ac 7s + Ac 1.057 - - - 1.056 - - 1'055 1'055 1.056 1'056 1 1 1.690 - 1'688 1-686 1.683 1'683 1'683 1.686 1'689 1.690 1.695 1.057 - 1.055 1'054 1.052 1.052 1.052 1.054 1.056 1.056 1.060 1 I- I TABLE XI1 -INFLUENCE OF CHANGING ACID CONCENTRATION ON THE RELATIVE VISCOSITY OF A GLIADIN SOL WITH VARYING KC1 CONCENTRATION IN ACETONE 44 PER CENT. BY WEIGHT. Mil. Eq. HCl Per Litre. 0 { { 0'2 0'4 { 0.6 4'0 { Milli-equivalents of Potassium Chloride Per Litre. a 3. 6. 8. 10. 14. 1.689 1.687' 1.688 1.688 1,692 1.689 1.056 1.055 1'056 1.055 1.058 1'056 1.730 1.699 1-697 1.692 1.692 1.692 1.081 1.062 1.061 1.058 1.058 1'058 1.775 1.713 1'705 1.705 1'704 1.701 I'IIO 1-071 1'066 1.065 1-065 1.063 ~ 1'812 1'724 1.714 1.712 1'709 1'703 1.113 1-077 1.072 1.070 1.068 1'065 1.826 1'727 1'718 1,715 1.710 1'710 1-141 1.080 1.074 1.072 1'069 1'06g 1.808 1.740 1.726 1.723 1.718 1.716 1'130 1.088 1'079 1.077 1'074 1'073 1'794 1.735 1.728 - 1.717 1.715 1-122 1.085 1.080 - 1.074 1'072 ~-1'748 1.727 1.718 - - 1,712 1-092 1.080 1'074 - - 1.070 5 0 COLLOID CHEMISTRY OF GLIADIN SEPARATION maximal viscosity but that its highest value is reached a t a slightly higher H-ion concentration ; in other words the relative viscosity maxima of a sol with and without electrolyte are found a t different H-ion concentrations.By increasing the acid concentration of a sol beyond its viscosity maximum the discharging effect of the Cl' ions will cause a decrease in viscosity. Therefore addition of KCl as a discharging agent will have a smaller effect because of the smaller charge of the particles in the sol. At length a H-ion concentration will be reached at which addition of KCl will have practically no effect on the relative viscosity of the sol, in which case the relative viscosity curve will be practically a horizontal line but a t a level considerably higher than that of the sol without acid and varying quantities of KCl. We may now discuss briefly the influence of varying H-ion concen-trations on 44 per cent. acetone gliadin sols with and without 10 mil.eq. KCl K2S04 and K4FeCy (Table XIII.) (Fig. 9). 5 10 mil. mol. HC1 per litn 20 FIG. 9.-See Table XIII. Curve A = Without electrolyte , B = KCl , C = K,SO, , D = K,Fe(CN),. It is evident that increase of the H-ion concentration of a sol con-taining 10 mil. eq. KC1 gives a t first an increase in the relative viscosity until at a concentration of about 1-2 mil. eq. HC.1 the relative viscosity becomes practically independent of further increase of acidity in the sol. As was to be expected the maximal relative viscosities of a sol with and without 10 mil. eq. KCl are found a t different acid concentrations. Changing the p~ of sols containing 10 mil. eq. K2S0 gives a completely different effect. The initial small additions of acid cause a lowering in the relative viscosity.A minimum is reached a t about 1-2 to 1.6 mil. eq. HCl. By further increasing the acid concentration the relative viscosity curve rises again to reach a value of 8 units above the initial viscosity at an acid concentration of 20 mil. eq. HCl. By changing the p of sols containing 10 mil. eq. K,Fe(CN) an analogous effect occurs except that whereas the sols with 10 mil. eq. K,SO are all perfectly clear a t the acid concentrations used those con-taining 10 mil. eq. K,Fe(CN) are not clear. The first traces of acid alter the appearance of the originally clear sol but by adding 1.6 to 4 mil. eq. HC1 the sols become more or less cloudy and in some cases part o H. L. BUNGENBERG DE JONG AND W. J. KLAAR 51 the protein settles down after twenty-four hours.(The viscosities of these cloudy solutions were determined after one hour instead of after twen ty-four hours.) Here we have a separation phenomenon in 44 per cent. acetone sol, probably caused by the discharging effect of a polyvalent complex ion. The sediment when examined under the microscope showed a nurnher of small drops. By further increasing the acid concentration the sols become per-fectly clear again. Coincident with these phenomenon there is a change in the intensity of the yellow colour of the sols. TABLE XIII.-INFLUENCE OF CHANGING HC1 CONCENTRATION ON THE RELATIVE VISCOSITY OF GLIADIN SOLS WITH 10 M. EQ. KCl K,SO AND K,FeCy6 IN ACETONE 44 PER CENT. BY WEIGHT. WLhout electrolyte 10 m. eq. KCl { 10 m. eq. K ! W , 10 m.eq. K,FeCy, 0. 0'4. 0-8. 1.2. 1%. 2-5 4. 10. 20. 1.690 i'S12 1.S26 1.808 1.794 - 1.748 1-724 1.718 1'056 1-110 1-141 1.130 1'122 - 1'092 1.078 1'074 1.689 1.694 1.708 1-714 1-715 1.716 1'716 1.714 1.714 1'056 1-059 1.06s 1'072 1'072 1.073 1.073 1'072 1.072 1.695 1'694 1.691 1.689 1'689 1'694 1.696 1'705 1.703 1.060 1'059 1.057 1.056 1'056 1.059 1'061 1'066 1.065 1.706 1'702 1.702 1.687 1'679 1.680 1.682 1.685 1'6go 1.067 1'064 1.064 1.055 1'050 1.050 1-051 1.054 1'056 sl. cld. sl. cld. cld. Gliadin Sols of Double and Triple Concentration. In view of the above-mentioned investigations on the influence of equimolar quantities of KC1 and KI on an acid gliadin sol i t was evident that the differences in viscosities found in aqueous solution disappeared in 44 per cent.acetone. It was desirable to investigate whether the concentration of the gliadin would have any influence on this phenomenon the more so as we had observed by studying a large number of data that there was always a. tendency for the KI sol to have a slightly smaller relative viscosity in 44 per cent. acetone than the KCl sol especially a t higher electrolyte concentration (8 m. mol.). These differences always were however, practically within the region of the experimental error. Acid sols were made with double and triple protein concentration, and the influence of changing acetone concentration on the relative viscosity of these sols with 8 m. mol. KCl and KI was investigated. Table XIV. gives the results 5 2 COLLOID CHEMISTRY OF GLIADIN SEPARATION I t is evident that with increasing protein concentration the differ-ences found with KCl and KI in 44 per cent.acetone are practically pro-portional to the gliadin concentration. However it struck us that by using sols of higher H-ion concentration these differences decreased with increasing acid concentration. For example a sol with 20 mil. eq. HC1 did not show any difference in viscosity. It was possible that these differences in viscosity in such an acid 44 per cent. acetone sol did not occur because there is no lyotropic effect between KC1 and KI sols of such high H-ion concentration. To test this possibility i t was necessary to study the effect of KC1 and KI in aqueous sols of this H-ion concen-tration As it was impossible to make these sols containing 8 mil.mol. KC1 or KI (because the separated phase settled down) we decided to make sols containing the same quantity of electrolyte and acid but with the addition of varying quantities of acetone. These sols would be partly stabilised by the presence of acetone but nevertheless in the smaller acetone concentrations lyotropic effects ought to be found if this effect did exist a t all. %+ac 1 - 1 tac TABLE XIV.-INFLUENCE OF 8 MIL. MOL. KCl AND KI ON ACID GLIADIN SOLS OF DOUBLE AND TRIPLE CONCENTRATION IN VARYING ACETONE CONCENTRATIONS. a. 6. 10. 12. IS. 18. KC1 1.084 1.131 1.148 1,349 1.139 1.0go KI 1'044 1.119 1'144 1.146 I'X39 1'104 Double concentration Acetone. c*c. I I For this purpose 4.0908 gr. gliadin were dissolved in 50 C.C. of an acetone water mixture (2 3) ; 2-5 C.C.of this sol were pipetted into a volumetric flask the necessary quantities of acetone acid and electrolyte were added and the whole was made up to the mark (25 c.c.) with dis-tilled water. The quantities of acid and electrolyte are given in Table XV. In this way it was possible to make sols of 3 5 7 10 12 and 16 C.C. acetone per 25 C.C. The sol containing 3 C.C. acetone was slightly opales-cent with KCl and opalescent with KI. When still less acetone was added flocculation occurred. With high H-ion concentration a distinct difference between KCI and KI can be observed in low acetone concentration. This difference decreases towards the maximum. In higher acetone concentrations the lyotropic order is reversed. From this must follow that there still exists a certain lyotropic effect in aqueous medium.The differences for KCI and KI in acetone medium disappeared not only a t higher H-ion concentrations but also when sols of triple concen-tration were made without any acid. In this case only sols with 10 and 12 C.C. acetone per 25 C.C. could be made. It was impossible t H. L. BUNGENBERG DE JONG AND W. J KLAAR 53 Without Acid. 7. 10. 12. 16. prepare sols containing 8 mil. mol. KCl and KI in smaller acetone con-centrations so that it is not certain that there is a lyotropic difference between KCl and KI in aqueous medium in such an unchanged sol.* C.C Acetone. TABLE XV.-INFLUENC& OF VARYING ACETONE CONCENTRATIONS ON THE RELATIVE VXSCOSITY OF GLIADIN SOLS OF TRIPLE PROTEIN CONCENTRATION CONTAINXNG 8 MIL.MOL. KCl AND KI. 1.856 1-86, 1.161 T-T~IL. pocc . flocc. 9s + Ac 7 s + Ac 1.860 1.865 %+ Ac b + Ac 1.162 1.166 jlocc. I qAc - -.I 6. 10. 12. 15. 18. 1 C.C. Acetone. HCl Conc. a 7 s + Ac %+Ac - 1-697 1.928 1.931 1'761 1.385 1.070 1'176 1.2~5 1'207 I - I ~ I 1.134 clr. ~ A C xa 12. 16. 1% 5. 7. HCI Conc. o. 3. 20 m. eq. 1'410 1-607 1'764 1.908 1.912 1'639 KCl { frocc. j¶occ. 1'143 1.164 1.182 1.193 1'195 1.166 1. op. 1'390 1.592 1.754 1.904 1'910 1.648 K I { pocc. pocc. 1.127 1.153 1.176 1'190 1.194 1,172 C.C. Aeetone. %+AC 9s + Ac VAc Ts+Ac 7s + Ac 9Ac I As a lyotropic difference was found at some points of the p H region, whereas this difference disappeared completely a t other p ~ we decided to investigate this influence of the p~ more fully.At the same time the *From later determinations we have concluded that in any case in small electrolyte concentrations no lyotropic effects exist in practically iso-electric sols. This was to be expected because the order of the lyotropic series is reversed at the transition from a positive to a negative sol 54 COLLOID CHEMISTRY OF GLIADIN SEPARATION TABLE XVI.-INFLUENCE OF CHANGING H-ION CONCENTRATION ON A 44 PER CENT. ACETONE GLIADIN SOL OF TRIPLE CONCENTRATION WITH VARYING QUANTITIES OF KI AND KCL M. Eq. HCI 1*6 { 2'2 1 3'2 0. 2. 4. 8. 12. I 6. 1.867 - 1.865 1.865 1.866 -1,167 - 1.166 1.166 1.167 -KCl { - - 1.863 1.864 1-866 -- 1'164 1.165 1.16s - K1 { 2'001 - 1.908 1.893 1.S85 1.881 1.251 - 1.193 1.183 1.178 1.176 KCI { - 1.g08 1-893 1.882 1.881 - 1.192 1.183 1.177 1.176 2.151 2'031 1.989 1.948 1.930 1.921 1.344 1.270 1,243 1.218 1.206 1.202 KCI { 2-03' 1'984 1'942 1.924 1.913 KI { - - 1.270 1'240 1.214 1.203 1.197 2.127 - 1.994 1'960 1'940 1-928 1.330 - 1,246 1-225 1'213 1.206 KCI { M.Eq. Electr. 7s + Ac 7s + Ac 7Ac 7 s + Ac 7 s + Ac VAc 7 s + Ac 7 s + Ac 7Ac 7 s + Ac 7s + Ac 7 Ac 7 s + Ac 7s + Ac 7AC % + Ac 7 s + Ac VAc 7 s + Ac 7 s + Ac ~ A c 7s + Ac 7 s + Ac ~ A c H. L. BUNGENBERG DE JONG AND W. J. KLAAR 5 5 20 4 I. Eq. HCl. I 4 I0 I TABLE XVI. (continued). 0. 2. 4. 8. 12. I 6. 2'091 - 1.986 1.953 1.935 1.924 1.301 - 1-241 1.221 1-210 1'20t KCl { 1.968 - 1.943 1.932 1'924 1.913 1230 - 1215 1 .2 ~ 8 1.203 1.197 KCl { 1.921 - 1.918 1.914 I * ~ I O z-grr 1'201 - 1.199 1'196 1.194 1.195 KC1 { 4. Eq. Electr. 7s + Ac 7s + Ac 9Ac -7 s + Ac 7 s + Ac -7 Ac 9s + Ac 7s + Ac qAc 9s + Ac 9 s f Ac qAc -t +- Ac 9s f Ac 7 Ac 7s + Ac TAc 7s + Ac influence of the quantity of electrolyte added was determined. In the interests of brevity only the data for sols of triple concentration are given but analogous determinations were made with sols of double concentration (Table XV'I. Fig. 10). From this table it is evident that the difference between KCl and KI sols of equimolar concentration changes with varying H-ion concentration. In a sol without acid as well as in a sol with high H-ion concentration the difference disappears.The maximum divergence of the KC1 and KI curves is found a t about the maximum of viscosity of the sols with electrolyte. The maximum in the relative viscosity curves shifts to higher H-ion concentrations with the quantity of electrolyte added as was found before with the dilute gliadin solutions. For one and the same H-ion concentration the difference in relative viscosity between the KC1 and KI sols increases with the electrolyte concentration and seems to reach a constant final value 56 COLLOID CHEMISTRY OF GLIADIN SEPARATION 2:2 m k HCI per litre 10 I2 o:a 1.6 2 3*2 4 FIG. x o . C e e Table XVL. Curve A = o millimol. 99 B = 4 9 9 # ? c = 8 99 9 9 D = 12 9 9 $ 9 E = 16 9 9 Determination of the Composition of the Solvation Layer of Uncharged Ciliadin in Varying Alcohol Concentrations.For these determinations we did not use the same gliadin as for our other experiments but a preparation made from fresh gluten ; 3 or 6 gr. gliadin were dissolved in flasks a t 60" C. in varying quantities of alcohol of 65 per cent. by volume. After cooling down to 25" C. in a thermostat, the solutions were diluted either with water or alcohol of a temperature slightly lower than 25" C. so that the final temperature after mixing was 25" C. The flasks were stoppered with corks and silverfoil and held for two weeks in the thermostat. From preliminary experiments it was evident that in this time equilibrium was reached. After this period the original liquid had separated into two distinct layers.The nature of the protein-rich layer was in a high degree dependent on the alcohol percentage of the original solution. Starting from very dilute alcohol, the sediments became more and more liquid with increasing alcohol concentrations in solution. For instance in the lowest alcohol concen-tration used in our experiments the layer was a thick liquid whereas in alcohol of about 27 per cent. it was very mobile. In alcohol concentra-tions greater than 60 per cent. the layer was with increasing alcohol concentration in solution successively liquid a perfect gel tough and finally in alcohol of about go per cent. granulated. After a fortnight the supernatant liquid was completely decanted if possible and the weight of the remaining sediment was determined.I H. L. BUNGENBERG DE JONG AND W. J. KLAAR 5 7 complete decantation was impossible on account of the nature of the layer part of the sediment was transferred into and weighed in another flask. The protein and the alcohol were then determined in the supernatant liquid as well as in the sediment. For the determination of alcohol, 10 C.C. of the supernatant liquid were pipetted and weighed in a flask and at the same time 10 C.C. were evaporated and dried at 104' C. to constant weight (protein determination). The same method was used for the protein-rich layer but in this case alcohol and protein were determined in the same sample of known weight. The alcohol concen-tration of both the layers was determined by the distillation method. Preliminary determinations had proved that this method could very TABLE XVII.(a). Grams Specific Grams Alcohol Total Weight in Material Alcohol and ~ ~ ~ { ~ f ~ ~ t . Alcohol in Per Cent. Appear-Grams. 1 3o:'C. I Water. I 15,r5. ! Dist. 1 W$ht. I ance' Analysis of the protein-rich layer. 6.0606 2'4637 3'5969 0.99818 x 0-23S5 6-70 liquid 5.4404 2'1104 3'3300 0.995~0 x 0'6240 18.74 9 9 2'7770 0'7947 1.9823 0.99603 x 0'535s 27-03 9 9 3'3321 I.2272 2'1049 0'99920 t 0'2120 IO'O7 9 9 5'8738 2.3493 3'5245 0'99643 x 0'4798 13-61 *. I 11 I11 IV V Specific Gram Alcohol Weight of Protein Grams loc*c. So'. I 1:::. I Water. I Alcohol and Azi:&ft. Alcohol in Per Cent' Weight. 1 Dist 1 by I Analysis of the supernatant liquid. 9.8331 o 0379 9'7952 0.99509 x 0.6674 6.81 -9'7829 0'0533 9.7296 0'99630 t 0.9960 10'24 -9.7154 0.06fj 9,6479 0-99058 x 1.3360 13-85 -9.6612 0.1103 9'5509 0.95789 x 1'7659 18.49 -*5'7020 *0'2016 5'5004 0.9-740 x 1'45;o 26.49 -t = in 50c.c.x = in 25 C.C. dist. = distillate. I I1 111 IV v well be used but much care was necessary on account of foaming. The determinations of the upper layers gave an experimental error of about 0.2 per cent. but even with the protein-rich layer this accuracy could be reached after some practice. Before the distillation of the protein layers some water was added and after complete evaporation in a bath of saturated CaCI solution the distillation was repeated twice with newly added water. The protein remainder was then dried at 104O C. and weighed. The alcohol distillate was received in a volumetric flask of 25 or 50 C.C.containing some water and constantly cooled with running water. Afterwards the flask was made up to mark with distilled water In this determination thz quantity of sol was only weighed and not pipetted so the protein content was determined in 5'7020 gr. sol 58 COLLOID CHEMISTRY OF GLIADIN SEPARATION Specific Weight of Alcohol Dist. 15/15. at 25' C. and the specific gravity of the distillate was determined in an 0 s twald pyknometer. From these determinations all the necessary data are known for the calculation of the alcohol and water concentration in the protein-rich layer and the supernatant liquid. For this purpose the dry material (protein) was subtracted from the total weight and the percentage of alcohol in the remaining liquid calculated.* Table XVIII.and Fig. 1 1 give the results. Grams $coho1 In Dist. TABLE XVII (b). Protein Per IOCC I ~ O C. Grams Alcohol and Water. Analysis of the proteinlrich layer. 5'3461 6-7798 11.6912 6.0670 4'5247 8.9929 14'3303 9.5715 1.1435 4'2026 1-62 12 5'1586 2.1883 3-8787 1~6805 2-8442 3'5557 8.1355 3'2942 5.6987 4'6853 9.6450 1'7410 7's305 0'98203 x o'g67go x o'gg226-f 0.97342 x 0'95703 x 0.97879 x 0.9 8975 t O'g5040 X Weight of I0 cc Sol. Specific Weight of Alcohol Dist. 15/15. Analysis of the supernatant liquid. 8'7470 8'7847 8.5652 8.3052 8.2976 8-2386 8'0055 8'1544 0'5404 0.6581 0'2200 0'0202 0.0 I 75 0.0088 0'0062 0'0130 8.2066 8'1266 8.2850 8.2801 8'2298 S-1482 7'9925 8.3452 2.7988 3'4415 5'5475 2.930 2.157 4'5095 8'1500 7'2707 Grams Alcohol in Dist.5.6227 5'5737 6'1086 6.7882 6'7526 6.9896 7.1482 7'4535 Alcohol Per Cent. by Weight. 66.60 66'71 68-19 75'54 75'84 79-13 84.50 92-85 Alcohol Per Cent by Weight. 68.46 68-58 73-17 8 1.93 81-55 84'93 87'73 93'25 Appearance of the Layer. liquid perfect gel toughy tough very tough sPoWY granulated 9 9 A B C D E F G H - ~ ~~ t = in 50 C.C. x = in 25 C.C. dist. = distillate. On the horizontal axis of Fig. 1 1 is plotted the alcohol percentage in the supernatant liquid (dispersion medium) while on the vertical axis the corresponding value for the protein-rich phase (solvation layer) is found.Curve A represents the equality of the alcohol content of the two layers. At lower concentrations the values of our determinations actu-ally coincide with this curve. Therefore in this case the solvation of the protein particles in the protein-rich layer has the same composition as * When our experiments were completed a publication of the Carlsberg Lab-oratorium l1 appeared wherein the same question of composition of layer and supernatant liquid of alcoholic gliadin solutions was studied. These investiga-tors did not get the same results as ourselves. They made their determinations a t oo C. and used another more indirect method for alcohol determinations as a result of which it is impossible t o check their results with ours.l1 G. Haugaard and A. H. Johnson Compt. Rend. Labovatoire Cnvlsbsrg 18, VOl. 2 1930 H. L. BUNGENBERG DE JONG AND W. J. KLAAR 59 the dispersion medium. Any deviation from curve A represents a dif-ference in alcohol composition between layer and dispersion medium. This effect is found in alcohol concentrations of more than 60 per cent. Curve B represents the values of our determinations in these media. From the first part of curve B a steady increase in this divergence from curve A by increasing alcohol concentration was to be expected but the experiments disproved this. At higher alcohol concentrations a decrease in this divergence takes place. This effect however must be visualised as an unreality caused by the difficulty of separating the two layers.If part of the supernatant liquid remains in the protein layer the differ-FIG. II.-See Table VII. ence in the ratio alcohol-water must naturally diminish. That this is actually the case can be seen theoretically by the increase of the ratio total weight to dry material and. practically by centrifuging the protein layer in alcohol of about 92 per cent. when this layer separates again into two layers. Up to 80 per cent. alcohol a proper separation of the two layers was possible so that the real divergence in composition of the protein-rich layer and supernatant liquid can be expressed by curve C. These curves A and C intersect in an alcohol concentration of about 50 per cent. This is the alcohol concentration a t which the maximum in viscosity is found but we cannot deny that a graphical method based on only three points provides too insecure a basis on which to draw any definite conclusions.Discussion of Results. The following principal facts can be deduced from the foregoing ex-perimental work on dilute protein solutions 60 COLLOID CHEMISTRY OF GLIADIN SEPARATION (a) By increasing the acetone or alcohol concentration in a sol con-taining equivalent quantities of electrolytes of the same valency the differences in relative viscosity found in aqueous medium decrease. (Tables III. V. VIII.) (b) At certain acetone or alcohol concentrations this difference dis-appears within experimental error to reappear in higher concentrations of these media. (c) The value of the relative viscosity in the maximum of the acetone (alcohol) relative viscosity curve is dependent on the quantity of elec-trolytes and on their valency.We will now endeavour t o give an explanation of these effects. The other experimental facts are for the most part more or less com-plicated variations of these cases. The phenomena mentioned under (a) (b) and (c) can be considered as the result of a change in solvation through a change in medium. This change in solvation may take place either with the ions or with the colloid particle or with both. It is however practically certain that the effects are for the most part caused by a change in the solvation !ayer of the particle and that the effects of the change in solvation of the ions are of minor importance.It is well known that for emulsoids the constitution of the solvation layer is of great impor-tance in determining the behaviour of these systems with regard to ex-ternal influences whereas on the other hand for ions i t is very dificult to indicate circumstances a t which the ions alter their solvation. This is the more evident when the magnitude of the solvation layers of ions and emulsoids are compared. For ions the hydration is but small, and is often not more than one layer of closely bound molecules of water. This supposition becomes the more probable when we consider the fact that small quantities of electrolytes have no influence either on the viscosity of water or on the viscosity of water-acetone mixtures. The differences in the viscosities of water-acetone mixtures with and without 10 mil.mol. KI are within the experimental error. Any pronounced influence of acetone-water mixtures on the solvation layer of an electro-lyte would certainly appear in the viscosity. Whereas the foregoing facts only prove that the change in the solva-tion of the ions in mixed media will be very small we can show that there is a marked increase in the solvation layer of the colloid particle in these media. When a drop of a separated sol (prepared by adding a quantity of acid too small for total peptisation) is brought under the microscope and acetone is added it swells and loses its outlines till at last a water-clear solution is obtained. From these data it is sufficiently clear that the increase or change of the solvation layer of the particle by addition of acetone or alcohol must be the primary cause of the dis-appearance of the differences mentioned under (a).As the composition a t which the maximum in relative viscosity occurs is always the same for a sol with or without electrolyte and as at this point it is only the value of the relative viscosity which changes on addi-tion of electrolyte the explanation of the phenomena obtained with an electrolyte sol can in the first instance he referred back to sols without electrolytes. We will therefore first consider this simple case. The increase in relative viscosity of the sol with increasing alcohol concentration must be caused by a gradual increase of the protecting layer. (Tables III. V. VIII.) (Tables IV. VI. IX. X.) The reasons for these may now be discussed H.L. BUNGENBERG DE JONG AND W. J. KLAAR 61 From our determinations of the composition of the solvation layer and the dispersion medium it is evident that at least at small alcohol concentrations (up to 27 per cent.) the alcohol concentration of both layer and medium are the same although we had expected a preferential adsorption of alcohol in the form of alcohol hydrate. The solvation layer of the particle which in water consists purely of water molecules, will gradually by increasing the alcohol percentage in the medium, change into a layer of alcohol hydrate molecules. The increase of the solvation layer by this effect will result in an increase in relative viscosity of the sol. This increase in volume of the layer will reach a maximum a t which the maximal concentration of hydrate in the layer as well as in the medium is reached.In higher alcohol concentrations a completely different effect occurs ; there is a preferential adsorption of water! so that the layer is poorer in alcohol than the external phase. As the gliadin is alcoholophobic, the particle will be practically dehydrated in higher alcohol concentra-tions. If water is added to such a system i t will be preferentially ad-sorbed by the particle. Approaching the maximum (by diluting the alcohol) the quantity of hydrate will increase until a t last the solvation layer and the external phase have the same composition. We can therefore say that the left hand part of the viscosity graph represents a system of increasing solvation with increasing alcohol or acetone concentration whereas the right hand part gives a decrease in solvation with increasing alcohol or acetone in the medium.Before discussing the influence of electrolyte on an alcohol or acetone sol it is necessary to consider for a moment the condition of a sol particle in aqueous medium. In an acid gliadin sol the particles are charged posi-tively by the adsorption of H ions on the surface. The following clues-tion now arises Is the surface of the particles entirely or only partly covered by these ions ? Supposing the particles to be spherical and the specific gravity of the gliadin to be unity a rough calculation can give an insight into this question. Assuming that the radius of an H ion is the same as that of an H atom which is known i t is possible to calculate the area that can be covered by all the H ions of the acid solution.On the other hand a simple calculation will give the diameter of the globules in which a known quantity of gliadin must be divided so that the total surface of these globules will be the same as the area that can be covered by all the H ions. In the case of the gliadin solution used in our experi-ments (1.5 gr. gliadin + 4 C.C. 0-IN HC1) the diameter of the particles must be about 4p if the whole surface were to be covered. Now the particles in such a sol are amicroscopical that is to say they have a diameter of less than 0 . 2 ~ . (Although the particles of an emulsoid sol are invisible on account of their solvation we can deduce their size from the fact that in a partially discharged sol the drops of the separated phase which must consist of several particles desolvated to a consider-able extent are on the verge of visibility.) For the same quantity of protein the total surface of the particles with this diameter will be larger than that of particles with a diameter of 4p.Consequently it is impossible for the H ions of the acid to cover this surface completely. Moreover only a small proportion of the H ions in the medium is ad-sorbed by the protein as appears from the slight differences in H-ion concentration of the acid solutions with and without gliadin. We may conclude that only a very small proportion of the surface of the particles is taken up by the H ions ; these adsorbed H ions are centres of hydration 62 COLLOID CHEMISTRY OF GLIADIN SEPARATION Between these centres are large areas unprotected by this hydration caused by charge ; we must nevertheless assume that these uncharged parts of the surface are nevertheless protected by a hydration layer.This water is attracted as a result of the polar character of the protein molecules. By addition of equivalent quantities of electrolytes of the same valency to a gliadin sol in aqueous medium solutions of different vis-cosity are obtained i.e. sols of different hydration. In the first instance, we may assume that the discharging effect of equivalent concentrations of ions of the same valency is the same. The differences in viscosity found for these concentrations may indicate that the electrolytes are differently adsorbed by the gliadin particles.This difference in adsorption has been pointed out more than once in the literature of colloid chemistry. The quantity of water expelled from the protecting layer by these different adsorptions cannot be the same; therefore these effects must cause a difference in viscosity of these sols. It stands to reason that this type of adsorption must be of a different nature from that causing the discharging of the particles otherwise the state of charge of the particles in solutions containing equivalent quan-tities of electrolytes of the same valency could not be the same. We must therefore distinguish between two kinds of adsorptions ; the one causing the discharge of the particles through the adsorption of ions of opposite charge the other being an adsorption of both positive and negative ions or molecules which causes a change in hydration of the sol without influencing the boundary potential.An indication of the existence of this second kind of adsorption can be found in the investi-gations of Weiser and Schilow,12 who have proved that there is no equivalence between the quantities of electrolytes adsorbed by the sol particles. Weiser and Dhar l3 also have shown that a discharged flocculated sol still adsorbs electrolyte. They suggest that this adsorp-tion will only take place after the particles are completely discharged. In his discussion of the results of cataphoretic measurements Huyzingl4 proves that i t is impossible for these two adsorptions to occur the one after another but that on the contrary they will occur simultaneously.We can agree with this point of view if we take into consideration that a large part of the surface is not protected by charge. Is it possible to give more facts indicating the existence of an ad-sorption other than that which causes the discharge of the particles? We believe that the clarification by heating of a separated sol may be an indication of this supposition. For the time being we visualise this action as an expulsion of the adsorbed electrolyte by the increased molecular movement. At the same time however we see an indication of this kind of adsorption in the analogous shape of the curves indicating the addition of alcohol or acetone to gliadin solutions containing surface-active material or electrolytes.We will now discuss shortly the influence of alcohol and acetone on gliadin sols containing electrolytes. By the addition of alcohol to such a sol there will first occur the same phenomena as that already found 12 Weiser J . Physical Chem. 29 955 1925 ; Schilow 2. physik. Chem., 1s Dhar Koll. Z . 28 457 1924; 35 144 1924. 14 J. J . Huizing Invloed uan Electrolyten op de kataphoretiscb snelheid Diss., 94 25 1920. Utrecht 1928 H. L. BUNGENBERG DE JONG AND W. J. KLAAR 63 with a sol without electrolyte i.e. an adsorption of alcohol-hydrate molecules instead of water molecules. By this increase in solvation the gliadin drops (which in water separated out of solution by the dehydrating action of the electrolytes) are once more dispersed.In our opinion how-ever this fact cannot give a complete explanation of the phenomena observed with gliadin sols separated by electrolytes in varying alcohol or acetone media. * We imagine that coupled with the formation of alcohol hydrate molecules on the surface of the particles the molecularly adsorbed elec-trolyte will decrease and at the maximum the electrolyte will be com-pletely thrown off the surface or will be present in the same quantity. Since in aqueous solutions of equimolar concentrations the quantity of KI adsorbed on the surface of the protein is larger than the quantity of KC1 a larger quantity of KI than of KC1 will be thrown off the surface by the addition of alcohol. Therefore the rise in relative viscosity will be larger for the KI sol than for the KC1 sol and in consequence the relative viscosity curve will be steeper for a KI sol.In the behaviour of a sol containing electrolyte in varying alcohol media we see an analogy with that of a sol containing surface-active material in varying alcohol media. In the last-mentioned case it is supposed that the increase of the relative viscosity of a sol containing for instance small quantities of resorcinol is caused by the disappearance of the adsorbed substance from the surface of the particle. So far we have not been able to test our supposition that in the case of sols containing electrolyte the elec-trolyte will be expelled from the particle by the addition of alcohol. Ultra-filtration of these alcoholic sols was impossible. On the other hand all experiments indicate that a certain sensitivity of the sol for electrolyte still exists at that alcohol or acetone concentra-tion a t which the maximum in the viscosity occurs.The effects are, however the same for equivalent concentrations of electrolytes of the same valency but different for electrolytes of different valency. The higher the valency of the negative ion the more pronounced will be the influence of the electrolyte. According to the above explanation the protecting layer at the maximum is of such a formation that the effects of the molecularly adsorbed electrolytes have disappeared so there still exists in the med-ium an essentially different kind of adsorption. Having regard to the differences in relative viscosity found for sols containing equivalent concentrations of electrolytes with anions of different valency this ad-sorption must be an ionic adsorption which only influences the boundary potential.I t is readily understood that in these solutions there is still an electric effect if we consider the fact that even at a distance it is possible for an ion t o infliience the electric field of a centre of charge, which directs polar molecules. Thus in Tables 111. and XI. the decrease in relative viscosity at the maximum is only caused by discharging effects. This therefore represents the purely electroviscous effect, which in aqueous medium is always coupled with lyotropic influences. Looking a t the right-hand part of the relative viscosity-alcohol (acetone) graph we see firstly that the same phenomena will occur as with a sol without electrolyte i.e.a decrease of the hydrate in the solvation layer and a corresponding increase of water in the layer. But coupled therewith there will be a decrease in solubility of the electrolyte (larger *Especiallv the fact that gliadin solutions of 44 per cent. acetone are per-fectly water-&ear even in the presence of higher electrolyte concentration 64 COLLOID CHEMISTRY OF GLIADIN SEPARATION for KC1 than for KI) with increasing alcohol concentration. Data are to be found in the literature. Since the solvation layer is relatively richer in water than the dispersion medium the KCl will be driven to the water-rich phase and adsorbed on the surface of the particles. KI will do the same but to a less extent. In these media we shall thus get a reversal of the lyotropic series (Table IX.).An analogous explanation can be given for the other monovalent ions. When the lyotropic influence of different electrolytes of the same valency is compared the strongest adsorbed electrolyte shows the most marked lyotropic influence. The composition of the medium alone decides which electrolyte will be most strongly adsorbed. If now we consider the effects found in more highly concentrated gliadin sols we observe at the maximum slight differences in relative viscosity between these sols containing equivalent quantities of KI and KCl. These differences are practically proportional to the protein con-centration and could be explained by assuming that there still exist slight lyotropic influences of different electrolytes of the same udency, i.e.the molecularly adsorbed electrolytes will not be thrown off the particle to the same extent. Consequently the lyotropic effects in diluted protein solutions cannot be zero a t the maximum but must still have a definite value slightly different for every electrolyte. It is also possible to explain these differences in a wholly different manner, still assuming however that the lyotropic influences a t the maximum are zero or the same for different electrolytes of the same valency. The principal facts found in 44 per cent. acetone sols with higher protein concentrations are :-(I) The difference in relative viscosity of sols with equimolar quan-tities of KC1 and KI increases practically proportionally to the protein concentration (Table XV.) ( 2 ) At any one protein concentration but with increasing electrolyte, this difference increases until the sol is practically discharged to remain practically constant with higher electrolyte concentration (Table XVII.).(3) The difference between the KCl and the KI curve is maximal when the sols have their maximurn viscosities (Table XVII.). (4) This difference becomes a minimum a t the iso-electric point as well as in sols with higher H-ion concentrations (Table XVI.). The most striking facts are those mentioned under (3) and (4). From these facts it can be deduced that this difference in viscosity is dependent upon the charge of the sol particle. The higher the charge of the sol the larger will be the difference found by the discharging of the sol with KC1 and KI.In an iso-electric sol as well as in one with high H ion concentration (in which the particles are nearly discharged by the action of the acid) the differences for KCl and KI have disappeared. To explain these facts it is necessary to consider the phenomena of charging and discharging of a sol in aqueous medium from a general point of view. When acid is added to gliadin the protein is peptised by the preferential adsorption of the H ions and the particles will be charged positively. At the same time the ions of opposite charge (in this case the C1 ions) will be adsorbed by the H ions on the surface of the particle, but to a less extent. These will form a double layer with the H ions, and in this way the charging effect of the H ions will be partly neutralised.By increasing the acid concentration the number of positive centres of charge (i.e. the adsorbed H ions) will increase but a t the same time the discharging effects of the C1 ions will increase also and to a large H. L. BUNGENBERG DE JONG AND W. J. KLAAR 65 extent. The adsorption of the H ions will reach its maximum a t a lower H-ion concentration in the sol than that of the C1 ions. Both the maxi-mum and the drop in viscosity curve after the maximum can be explained b y these contradictory forces of the ions. Consideration must however be given to the fact that in aqueous media the electrolytes exercise their lyotropic influence in addition to t h e effects of their charge. In the case under discussion therefore the acid will exercise a lyotropic effect on the gliadin in the same way as KCl and other electrolytes in addition to the effect of its charge.If, however we assume for a moment that the acid does not show a lyo-tropic influence it is to be expected that in higher acid concentrations a final state will be reached a t which the viscosity curve has a hori-zontal direction i.e. the viscosity will be independent of further addi-tion of acid. In this region every adsorbed H ion will be opposed by a firmly-bound negative ion. It is moreover to be expected that such a sol will not be influenced by the addition of electrolyte ; a t any rate if this electrolyte has the same negative ion as the acid used for peptisation and the lyotropic effects of these salts are zero.This is actually the case for sols with such a maximum concentration of alcohol or acetone t h a t the lyotropic effects have disappeared (and in certain other cases n o t mentioned here). Table XIV. It is now easy to understand why the viscosities of an iso-electric sol and of a sol discharged either with an excess of acid or with electrolytes in 4.1 per cent. acetone never reach the same level (Table XIV.) and consequently must have different solvations. In our opinion the situa-tion can be visua!ised as follows. Every adsorbed H ion which is a +centre of charge will attract and direct polar substances (ie. hydrate molecules) in its field of force. When a negative ion which also is in poqsession of a solvation layer approaches this centre this ion will be strongly attracted by the H ion.The result will be the formation of a new dipole by the two ions. Consequently both the ions will lose their solvation. The newly-formed dipole however will also have a field of force and will direct hydrate molecules but to a far less extent. The hydration of such a dipole must be much smaller than the original hydra-tion of the H ion. In the case of a sol discharged by an excess of acid all the centres of charge on the surface of the particle will be changed into HCl dipoles. Every one of these dipoles will direct a certain number of hydrate molecules. On the other hand the iso-electric state of a par-ticle will be quite different. In this case practically no adsorbed H ions are to be found on the surface ; nor will there be dipoles which can direct hydrate molecules.Therefore it follows that there must exist a dif-ference in hydration between these two uncharged states of the particle, and consequently in the viscosities of the sols. In aqueous medium these facts will be complicated by the lyotropic influence of the acid. Applying the foregoing reasoning to the discharging by KC1 and KI of a 44 per cent. acetone sol with maximum viscosity we see that with KCl as discharging agent the negative C1 ions will be firmly bound in the double layer by the H ions and form HCI dipoles while in the case of KI the same effect will occur but of course HI dipoles will be formed. By increasing the KC1 and KI concentrations in the sol the number of dipoles on the surface of the particles will increase until the particles are practically discharged.The addition of more electrolyte after this point will not have any influence on the number of dipoles. As the -water binding capacity of HC1 dipoles will be larger than of HI dipoles, 66 COLLOID CHEMISTRY OF GLIADIN SEPARATION we may expect that the same will be the case for the capacity of binding hydrate molecules. Consequently we may expect an increase in the divergence of the relative viscosity-electrolyte concentration curves with increasing KCl and KI concentration in the sol. This difference will be practically constant when the sol is discharged by these electrolytes. Experimental measurements are in accordance with the above theory, (See (2) above). The other effects found in higher protein concentrations can also be explained by the formation of dipoles.For instance comparing sols of the same H-ion concentration but with different protein concentration, the number of H ions adsorbed on the total surface will be seen to increase proportionally with the protein concentration. On discharging of these sols with KC1 or KI the number of KCl and KI dipoles must increase in direct proportion with the protein Concentration. Since there is a difference in solvation between the KC1 and KI dipoles the difference between the relative viscosity curves of the KC1 and KI sol will increase proportionally with the number of dipoles i.e. proportionally with the protein concentration (see ( I ) above). By increasing the acid concentration of a sol with maximal viscosity the number of HCl dipoles on the surface of the particle will increase, as a result of the discharging effect of the C1 ions of the acid.On the addition of KCl and KI to such a sol only part of the adsorbed H ions will be free to form dipoles with the C1 or I ions of the added electrolyte, since some of the H ions on the surface have already bound C1 ions of the acid i.e. the ratio between the HC1 and the HI dipoles is increased. By still further increasing the acid concentration of an electrolyte-free sol practically all the H ions on the surface will firmly bind the negative C1 ions of the acid. The result of adding KC1 or KI to such a sol will be that practically no HI dipoles can be formed and consequently the factor that gives the differences in viscosity between KC1 and KI sols will have disappeared.We can thus deduce theoretically that with increasing acid concentration in a sol beyond the maximum the divergence in the relative viscosity of such a sol discharged by KCl and KI will decrease as a result of an increase in the ratio HC1 to HI dipoles and that in a sol with high H ion concentration these viscosity curves will run together. This reasoning is therefore in accordance with the experimental facts mentioned under (3) and (4). It is evident that the differences in relative viscosity of KI and KCI sols of higher protein concentrations can be easily explained by the for-mations of HCl and HI dipoles. We pointed out above that in 44 per cent. acetone the lyotropic effects of electrolytes have disappeared and we deduced that the effects of elec-trolytes in that medium were only caused by a change in the boundary potential.We may now ascertain whether the effects caused by acid and electrolytes to 4 per cent. acetone sols are in accordance with this supposition. Considering first the effects obtained by adding varying quantities of KCl K$04 or K,Fe(CN)6 to a 44 per cent. acetone sol without acid (Table XII.). For KCl the relative viscosity curve thus obtained will be practically horizontal while for K,SO the first traces will give a slight sloping down of the curve. By increasing the K,S04 concentration a rise will occur. The minimum for the K,Fe(CN) curve will be more pronounced. The explanation of these curves is as follows The con-centrations of KCl used in these experiments are insufficient to chang H.L. BUNGENBERG DE JONG AND W. 3. KLAAR 67 the state of charge of the sol particles. On the other hand the first additions of K,S04 will take away the charge of the particles and a decrease in viscosity will be the result. This sol though made without acid is not quite iso-electric but has a small positive charge. By further increasing the KzSO4 concentration in the sol the particles will adsorb the bivalent negative SO ion and will be gradually charged negatively, and in consequence a rise in relative viscosity will occur. The same will be the explanation for K4Fe(CN),. This explanation is based on the fact that the sol is charged slightly positively. The following facts are in accordance with this :-(a) A slight decrease in the relative viscosity of this sol occurs by the addition of very dilute sodium hydroxide in other words by adding sodium hydroxide the sol loses its charge and reaches its iso-electric state.We did not find a distinct iso-electric point but a trajectory of p ~ where practically no change in viscosity occurs. (b) By the addition of polyvalent electrolytes to the original sol the minima in viscosity do not occur a t the same electrolyte concentration. The higher the valency of the negative ion the smaller will be the elec-trolyte concentration at the minimum in the relative viscosity curve (for K,SO 1 - 5 mil. eq. ; for K,Fe(CN) 0.25 mil. eq.). (c) The rise in the relative viscosity beyond the minimum will be the larger the higher the valency of the negative ion.(d) The phenomena found by adding acid to sols containing 10 mil. eq. K,SO and K,Fe(Cb?) (Table XIV.). The above facts confirm the proposition that the 44 per cent. acetone sol without acid must be slightly positively charged and that it is possible to reverse the charge of a gliadin particle in 44 per cent. acetone hy adding polyvalent electrolytes. Are the facts observed with a sol containing 10 mil. eq. KC1 K,SO, or K,Fe(CN) and changing p~ in accordance with our point of view that all these effects can be explained by a change of charge of the par-ticles. The differences in viscosity of the KCI K,SO and K,Fe(CN)6 sols without acid must be caused by a difference in charge. The K,SO, and K,Fe(CN) sols will be negative but to a different degree (the charge of the K,Fe(CN) sol will be the higher) while the KCI sol will be charged positively.By adding acid to a KCl sol the charge of the particles will be increased by the adsorption of the positive H ions and a rise in viscosity will be the result. On the contrary the negatively-charged K2S04 and K,Fe(CN), sols will lose their negative charge by addition of small quantities of acid and become electrically neutral. By further increase of the acid concentration the sol will become positively charged for the same reason as with the KC1 sol. The discharging of the negatively charged K2S0, and K,Fe(CN) sols by the positive H ions will be the more rapid for the least charged sol in this case the K,S04 sol. Therefore in this case the viscosity minimum by addition of acid will be found at a lower acid concentration than for the K,Fe(CN) sol and at the same time the rise in viscosity after this minimum will be quicker for the K,SO sol than for the K,Fe(CN) sol because in the first case the bivalent SO ion alone resists the recharging by the positive ions in the latter case the tetra-valent Fe(CN) ion.In this case too therefore all the facts are in accordance with our suppositions and we may conclude that all the phenomena discussed in this part are due to change of charge of the particles 68 COLLOID CHEMISTRY OF GLIADIN SEPARATION There is still another conclusion that can be drawn from the fore-going facts namely that it is possible to get a negatively-charged sol by adding polyvalent electrolytes a t a p at which the original sol with-out electrolyte is charged positively. Thus the real iso-electric point can only be determined in the absence of electrolytes. Similar pheno-menon in aqueous medium was pointed out by Kruyt and Tendeloo,15 working with gelatin sols containing different polyvalent electrolytes. However in their viscosities lyotropic effects are included. From the data of Tables XU. and XIII. i t is evident that the maximum in the relative viscosities for sols without electrolyte and practically discharged with electrolyte do not occur at the same H-ion concentration. The explanation of this effect is that the percentage increase of the free dis-charging ions is larger for a sol without electrolyte than for a sol with electrolyte. Since the maximum is formed through the opposing forces of the positive and the negative ions and the percentage increase of the negative ions (in this case the C1 ions) is smaller for a sol with electrolyte, the maximum will be reached at a higher HC1 concentration than with the sol without electrolyte. In conclusion we should like to emphasise that the discussion of results must be taken merely as an attempt to explain the phenomena observed. For this explanation it was necessary to introduce a few suppositions. We hope to test these in due course especially the ad-sorption of electrolytes in alcohol- and acetone-water mixtures. Summary. The following points were studied :-I. The influence of varying alcohol or acetone concentration on acid gliadin sols containing equimolar quantities of different monovalent elec-trolytes. A reverse of the lyotropic series was observed in high alcohol and acetone concentrations. 2. The influence of increasing concentrations of mono-valent electro-lytes in the viscosity maximum in the alcohol-(acetone) -water diagram. 3. The influence of polyvalent ions in 44 per cent. acetone. 4. The influence of the protein concentration on the viscosities of sols containing electrolyte. 5. The influence of the acid concentration in the sol on the effects mentioned under 2 3 and 4. 6. The composition of the solvation layer of alcohol-gliadin sols. 7. An explanation of the foregoing facts is given based on the principle that there exist two kinds of adsorption ; the one being an ionic-adsorption, influencing the boundary potential the other a molecular adsorption causing the lyotropic effects. 8. In special alcohol or acetone media the last-mentioned effect is the same or zero for electrolytes of the same valency. The differences between sols containing equivalent quantities of electrolytes of different valency can be referred back to different state of charge of the particles. 9. The $JH a t which a sol is in the iso-electric state is dependent upon the electrolyte in solution. The iso-electric point can only be stated in electrolyte-free sols. l6 H. R. Kruyt and Tendeloo. Kon. Akad. Wet. 3 4 4 1925