O Q EXCHANGE Colloidal Solutions of Copper Sulphide A THESIS PRESENTED TO THE FACULTY OF CHEMISTRY OF STANFORD UNIVERSITY IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY By ROLAND NEAL Colloidal Solutions of Copper Sulphide A THESIS PRESENTED TO THE FACULTY OF CHEMISTRY OF STANFORD UNIVERSITY IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY By ROLAND NEAL COLLOIDAL SOLUTIONS OF COPPER SULPHIDE BY ROLAND NEAI, With the discovery that certain natural copper sulphides yield colloidal copper sulphide on treatment with hydrogen sulphide, 1 and the probability that colloidal solutions of copper sulphide may play a not unimportant part in the pro- cesses involved in the secondary enrichment of copper sul- phide ore beds, a knowledge of the properties of such solu- tions from the chemical point of view becomes a matter of considerable interest. In the following paper are given the results of a preliminary investigation of such solutions, and while the researches are by no means to be considered as com- plete, considerable material of interest is already in hand and given here. The composition of the sulphide content of a colloidal solution of copper sulphide is at the present time a very un- certain matter. Older investigators maintain that by the action of hydrogen sulphide on cupric compounds in the presence of water, a mixture of cupric and cuprous sulphides is formed, while Posnjak, Allen and Merwin 2 claim that only cupric sulphide is thus formed. On the other hand, it has been observed in this laboratory that sulphide precipitated from cupric sulphate by hydrogen sulphide with exclusion of air and sealed up in an atmosphere of hydrogen sulphide shows the presence of distinct crystals of sulphur after some months' standing. There seems to be at least some uncertainty as to the nature of the sulphide in the solutions used in this work, all of which were prepared from cupric compounds. The scope of the present investigations cover: (a) the preparation of colloidal copper sulphide solutions (copper 1 Clark: Bull. University of New Mexico, 1914, Chem. Series, Vol. i, No. 2. 2 Jour. Econ. Geol. 10, 528 (1915). 365231 4 Roland Neal sulphide sols); (b) influence of electrolytes on the dispersion of amorphous copper sulphide; (c) flocculation by electrolytes; (d) influence of hydrogen sulphide on the flocculation by electrolytes; (e) migration velocity in the electric field; (/) influence of electrolytes and of hydrogen sulphide on the migration velocity. Experimental Methods of Preparation of Copper Sulphide Sols. For the most part the methods of preparation used in this work are those which have already been used, and are in the main de- scribed in Svedberg's "Herstellung kolloider Losungen anorganischer Stoffe." Solutions were prepared in a variety of ways in order that it might be determined what influence the mode of preparation had on the properties of the product. Colloidal solutions were therefore prepared by the following three methods : 1. Agitation of well-washed copper hydroxide with hydrogen sulphide water. 2. Agitation of well-washed copper carbonate with hydrogen sulphide water. 3. Agitation of well- washed freshly precipitated copper sulphide with hydrogen sulphide. In all cases the solution was freshly saturated with hydro- gen sulphide from time to time and the containers were kept carefully stoppered to exclude air. They were also kept out of direct light which accelerates the oxidizing action of any air that may gain admission. The following sols were pre- pared as specified and in the latter part of this work will be referred to by the numbers given them here. Sol i. Two grams of copper sulphate were dissolved in 300 cc of distilled water, and the solution brought to faint permanent alkalinity with sodium hydroxide. The pre- cipitated hydroxide was filtered off and thoroughly washed with distilled water. It was then placed in a large volume of freshly distilled water and allowed to stand for ten months, after which the water was decanted off and 800 cc of fresh water added. On now passing hydrogen sulphide into the Colloidal Solutions of Copper Sulphide 5 liquid, the conversion of the hydroxide into finely dispersed sulphide began almost immediately. For a period of ten days, hydrogen sulphide was passed through the solution at frequent intervals, and the solution was subjected to gentle agitation for a period of an hour after each charging with hydrogen sulphide. This sol proved to be perfectly stable. After standing for two and one-half months there was no evidence of any tendency to settle. The color was very dark even when the solution was diluted i to 5 with water. Sol 2. This was prepared similarly to the above, except that the copper was precipitated as carbonate (by sodium carbonate) instead of as hydroxide. The carbonate was, however, washed less thoroughly than the hydroxide, the whole preparation in this case being completed within two weeks. In general appearance the sol was similar to Sol i. Kept saturated with hydrogen sulphide and stoppered, it showed no deterioration and no settlement after many months. Another preparation, identical with Sol 2, but in which the copper carbonate was washed by repeated boiling with water, was not so successful, the dispersion with hydrogen sulphide being both less rapid and less complete. This sol was not used in any of the work to be described. Sol j. This was made by dissolving copper sulphate in water (amounts used as above), precipitating by means of hydrogen sulphide, repeated washing by decantation for a month, and finally dispersing by repeated treatment with hydrogen sulphide accompanied by agitation. This sol main- tained itself in good condition, showing no settling after one year. Sol 4. This was prepared in the same way as Sol 3, except that the copper sulphate was strongly acidified with sul- phuric acid before the precipitation of the sulphide. Upon attempting to disperse the sulphide thus formed after wash- ing as for Sol 3, the result was unsatisfactory, the dispersion not forming at all readily. The precipitate was therefore dialyzed for two weeks, after which it yielded to dispersion with the same ease as that used in preparing Sol 3. Roland Neal Sol 4 (a). This was in all respects a duplicate of Sol 4. Sol 5. This was prepared by adding sufficient ammonia to the copper sulphate solution to redissolve $11 copper hy- droxide, before precipitation with hydrogen sulphide. Other- wise the preparation was identical with Sol 4. Effect of Electrolytes on the Rate and Degree of Dispersion of Copper Sulphide by Hydrogen Sulphide. While the floccula- tion of colloidal solutions by electrolytes has been extensively^ studied, little is known of the influence of electrolytes in in- hibiting dispersion. Considerable time was devoted to an attempt to get some light on this point, but it was soon evident that the experiments would demand an elaborateness of technique and apparatus that would make an extensive investigation in this direction impracticable in the time at disposal. Before abandoning this work, which it is the in- tention to take up later in more thorough manner, some results sufficiently interesting to be recorded here were ob- tained. The method employed was to place in each of several test- tubes the same small amount of copper sulphide (actually the carbonate, thoroughly washed, was used). To the first were then added 5 cc of pure water, to each of the others were added 5 cc of the electrolytes to be studied each at its desired con- centration, this being, of course, always smaller than that necessary to produce flocculation. Hydrogen sulphide was then bubbled slowly through and the dispersion determined by observation from time to time. In all cases where dis- persion took place it seemed to be complete after twenty-four hours. The results were, in general, that in like concentra- tions the inhibitions to dispersion were in the same order as the flocculating powers of the electrolytes, as is, of course, to be expected. One marked peculiarity was however found in the unexpected action of the alkaline hydroxides. These are usually held to stabilize the sulphide sols up to a certain point, but in our work they seemed to show the same inhibition of dispersion as the equivalent solutions of alkaline chlorides. Curiously enough, however, they did show a marked initial Colloidal Solutions of Copper Sulphide 7 acceleration in the rate of dispersion, and probably a marked initial increase in the degree. If three tubes were taken, as above, all containing a small amount of copper carbonate, one containing also pure water, the other two containing N/iooo KC1 and N/iooo KOH, respectively, and hydrogen sulphide bubbled through all three simultaneously, the one containing the potassium hydroxide would show a far more rapid and extensive dispersion than the others, while after a few hours this would have reflocculated to a great extent and the condition of the tube after 24 hours would be about the same as that of the one containing potas- sium chloride. In the mean time, the dispersion in the tube without electrolyte will have far outdistanced the other two. While there seems to be, at present, no satisfactory ex- planation of this phenomenon, it is believed that there may be found in this sort of experiments the nucleus for an interesting method of studying the reverse reaction to flocculation. Effect of Removal of Hydrogen Sulphide from Copper Sulphide Sols. Some experiments were carried out to de- termine to what extent the stability of a copper sulphide sol is dependent on the continued presence of hydrogen sulphide. In this work the apparatus described by Young and Goddard 1 was used. In the parchment bag was placed a one-thousandth normal solution of copper acetate and hydrogen sulphide was bubbled continuously through the surrounding dialyzing water. It slowly diffused into the acetate solution forming the sul- phide, while the acetic acid formed gradually dialyzed out- ward. While there was at first some flocculated sulphide formed, this gradually dispersed under the influence of the stirring and of the hydrogen sulphide, until after two days of treatment at intervals, the sol showed no tendency to settle. When this condition had been reached, the hydrogen sulphide water in the outer compartment was replaced by pure water, and this replacement was repeated every day for ten days, thus ensuring a practically complete removal of free hydrogen 1 Jour. Phys. Chem., 21, 3 (1917). 8 Roland Neal sulphide from the sol. There was no sign of settlement at the end of the ten days, nor was there after an additional twelve days of standing. After this, the apparatus was opened and a sample removed, which was placed in a tightly stoppered bottle. There was no odor of hydrogen sulphide. After standing for five months this sample showed consider- able flocculation and settlement, which in no case happened with sols kept saturated with hydrogen sulphide, even after a period of a year. To the sol remaining in the dialyzing bag there was added a fresh amount of copper acetate solution, sufficient to make the whole about three thousandths normal. This was sub- jected to the same treatment as the first solution. After com- plete dispersion had taken place, and the hydrogen sulphide had been completely dialyzed out, the sol was bottled and stoppered and set away for observation. After two months it was seemingly unchanged, but after four months it was almost wholly flocculated. It is thus quite certain that copper sulphide sols depend for their stability upon the presence of hydrogen sulphide, which agrees with the results of Young and Goddard 1 on the other sulphides. They may, however, maintain themselves in the form of the unstable dispersion for long periods of time. Flocculation by Electrolytes. The literature concerning the flocculation of sulphide sols is so extensive and so well known that it need not be recited here. As particularly concerns copper sulphide sols, Spring and du Boeck 2 carried out floccula- tion tests on this substance, the method being to add 10 drops of the sol to 10 cc of solutions of various concentrations of the electrolytes to be investigated. In the stronger electrolyte solutions distinct flocculation or turbidity appears, in the weaker ones it does not, a tube to which 10 drops of the sol had been added to 10 cc of pure water being used as a com- parison tube. The electrolytes arranged themselves in the 1 Loc. cit. 2 Bull. Soc. roy. belg. (2) 47, 165 (1887). Colloidal Solutions of Copper Sulphide 9 same general order of flocculating power as was found later by Freundlich 1 in his study of arsenic sulphide sols, and as had been found previously by Schultze 2 for antimony sulphide sols even though the experimental methods employed by the different investigators varied considerably. Flocculation experiments were carried out in this work, chiefly to determine whether or not the sols from different sources conducted themselves materially differently from one another. The method used was to prepare a solution of a given electrolyte of a concentration double that which was desired in the final solution. This was saturated with hy- drogen sulphide, after which 2 cc of it were mixed with 2 cc of the sol to be investigated. After mixing, hydrogen sul- phide was again passed through the mixture for a few moments, after which the tube was tightly stoppered and set away for observation. The period of standing which was chosen was twenty-four hours. If at the end of that time the liquid was still colored and turbid, it was considered to be not completely flocculated. In all cases in the following results, flocculation means that the settlement was complete, the liquid water- white and not opalescent. The period of twenty-four hours was chosen instead of a shorter one (Freundlich allowed one- half hour) because a set of preliminary tests showed that the results were more consistent after the longer period. The presence of hydrogen sulphide in the solution and an atmos- phere of hydrogen sulphide above it, were found to greatly increase the regularity of the results obtained. Potassium, calcium and aluminum chlorides were investigated with regard to their flocculating effects. Sols i and 2 (see ante) were used, two sets of experiments being made with Sol 2, and Sol i being investigated at full concentration and in dilutions of 1-5, i -10, 1-15 and 1-20, the dilution being in all cases made with water saturated with hydrogen sulphide. The results are shown in Table I. ( + ) indicates complete flocculation and ( ) indicates incomplete flocculation. 1 Zeit. phys. Chem., 44, 129 (1903). 2 Jour, prakt. Chem., (2) 25, 431 (1882). IO Roland Neal TABLE I Concentrations of Electrolytes Necessary to Completely Flocculate Copper Sulphide Sols Sol K Cl Ca Cl, Al( 2i (+) ( ) (+) (-) (+) ( ) 3N 2N 8AT 7 N 3N 2N 2 100 4 N IOO 3 N IO OOO loN IO OOO 9 N IOO 000 3JV IOO OOO 2N 100 3N IOO 2N 10 000 loN 10 000 9 N IOO OOO 5^V IOO OOO 4 N M T _, 100 IOO IO OOO 9 AT IO OOO 8N IOO OOO IOO 000 I 5 ~p\ii T T/ ~. IO OOO 9 N IO OOO SN Ull. I IO Mr T r 10 000 9 N 10 000 SN T T 5 T^lil T or IO OOO gN IO OOO SN LJii. I 2O 10 000 10 000 When compared with the results of Spring and du Boeck 1 the results show good agreement so far as potassium chloride is concerned. The results for calcium chloride are about 25 percent lower than theirs for barium chloride (they give none for the calcium salt) while the results for aluminum chloride are 50 percent or more lower than theirs for aluminum sulphate. Since the anion has but little influence on the flocculating power of an electrolyte for sulphide sols, it might reasonably be expected that the results for aluminum chloride and sulphate should show closer agreement. The discrepancy is, however, probably to be explained by the longer time given for reaching equilibrium in this work, as well as by the rather fundamentally different manner in which the experiments were carried out. The results of the flocculation experiments show that the amount of electrolyte required does not vary appreciably with 1 Loc. cit. Colloidal Solutions of Copper Sulphide n the source and manner of preparation of the sol, and that the amount of electrolyte required is independent of the dilution of the sol within wide limits. This latter was found to be true within rather close limits by Freundlich 1 for arsenic sulphide sols. Flocculation tests were also made upon sols from which the hydrogen sulphide had been removed by dialysis. The results were the same as those obtained when the sols were saturated with hydrogen sulphide, for which reason it is not considered necessary to give numerical data concerning them. From the results obtained in this work the relative flocculating powers of potassium, calcium, and aluminum for copper sulphide sols are: 1:39:875 Rate of Migration in the Electric Field. The rate of migration of the copper sulphide in the sols under the in- fluence of the electric field was studied quite extensively in order to determine the following factors : (a) Reproducibility of results in successive experiments using the same sol under the same conditions. (b) The influence of the origin of the sol on its rate of migration. (c) The influence of dilution. (d) The influence of electrolytes. (e) The influence of hydrogen sulphide. (/) The combined influence of hydrogen sulphide and electrolytes. (g) The influence of oxygen. The apparatus used in these experiments is shown in Fig. i and is practically self-explanatory. Small platinum wires were used as electrodes and these were allowed to dip about 2-3 mm below the surface. Readings were not begun until a well-defined flat migration surface had established itself. In order to avoid the effect of light, the measurements were made in a darkened chamber. The voltage employed 1 Loc cit. 12 Roland Neal was as near as possible to 100. With the equipment at hand, it was not always possible to maintain this with exactness, and many of the results are calculated to this voltage on the assumption that the migration rate is proportional to the potential fall. It has been the experience in this laboratory that attempts to have the electrodes bathed in pure water containing none of the colloidal material lead to more fluctuating results than are ob- tained when such device is not used, espe- cially when the sols contain electrolytes or other easily diffusible materials. This is probably due to the fact that the conductor is no longer homogeneous as well as to the fact that electrolytes diffuse out into the pure water and materially alter the con- ditions at the surface of migration. While dipping the electrodes directly in the sol means increased electrolysis, it is neverthe- less true that, judging from reproducibility and consistency of results, it is the lesser evil which is encountered in this way. All the results given below were obtained with electrodes dipping directly in the sol. (a) The Reproducibility of Results. For the purpose of determining the repro- ducibility of measurements on one and the same sol in different determinations, a por- tion of Sol 4 was diluted with 2 1 /z vols. of hydrogen sulphide water, and the whole thoroughly saturated with hydrogen sulphide. Five inde- pendent determinations of the migration rate were made on the same day and under identical conditions. Others were made at various intervals of time, often months apart. The results showed that the data obtained on the same day were in very close agreement, but that the rate of migration of the same sol might change quite materially with time. In the following, therefore, all measurements intended for com- Fig. i Colloidal Solutions of Copper Sulphide 13 parison were made on the same day, as far as possible. In considering the whole set of results to be given, it is necessary to bear in mind that results obtained with the same sol at widely separated times, are not in general strictly comparable. (6) Influence of the Origin of the Sol on the Migration Rate. A considerable number of comparisons of migration rates of sols of different origin were made, with the result that in general the rates found were in no case in agreement, the magnitude of the disagreement often being very great. For illustration there are shown in Chart I the values obtained for Sols 3, 4 and 5. It will be recalled that Sol 4 was made by dispersing copper sulphide precipitated from acid copper sulphate solution, while Sol 5 was made by dispersing copper sulphide precipitated from an ammonia solution of copper sulphate. Sol 3 was similarly prepared from neutral copper sulphate solution. The ordinates represent distances traveled in centimeters and the abscissae represent times. As will be seen, Sol 5 migrates about three times as rapidly as Sol 4. This phase of the work was not carried further, the present purpose being rather to determine what factors could cause variations in the conduct of the sols, than to make a detailed study of them. It is the intention to extend this work in the near future. (c) The Influence of Dilution. A number of measure- ments of the rate of migration of a sol in its full concentration 14 Roland Neal and when diluted were made. In order to alter the condition of the sol as little as possible the dilution was always carried out with water saturated with hydrogen sulphide, and the gas bubbled through the diluted sol immediately after dilution, to ensure complete saturation. It was found that the effect of dilution was invariably to increase the rate of migration. The results of two such sets of measurements are shown in Chart II. Here, as throughout the rest of this paper, it is not considered necessary to print the figures for the actual measurements obtained, for the actual values may be read off from the graphs with suffi- cient accuracy. The data in this case are for Sol 4 in full strength and diluted i to i. The migration rate of the diluted sol is nearly double that of the undiluted. (d) The Influence of Electrolytes. The influence of elec- trolytes upon the migration rate of copper sulphide sols was subjected to a rather extensive study. Potassium and calcium chlorides were investigated at less than their flocculating concentrations, and a considerable number of concentrations of each was investigated. In this way there were obtained results from which the relative effects of the two electrolytes in equivalent concentrations could be obtained. All of the different sols were investigated and the influencing factor seemed to be fairly independent of the original migration rate of the sol. That is, whatever the initial value for the migration of the sol, this seemed to be multiplied to about the same extent by the same addition of electrolyte. Variations of the rates for the different sols with time and other causes Colloidal Solutions of Copper Sulphide have made it impossible up to the present to obtain a wholly consistent set of results. For the most part the results on a given chart were obtained in a very short period of time, after a laborious series of preliminary measurements had made evident the danger of comparing results obtained at very considerable intervals of time. A sufficient number of data to indicate clearly the main results obtained have been compiled and are shown in the following charts: Chart III shows the effect of successive additions of potassium chloride to Sol 4. The effect is to increase the migration rate. Chart IV shows the effect of CHART ScL4H 2 S dialled out MINVTL5 10 30 potassium chloride on the same sol after it had been previously dialyzed for a long period so as to remove all hydrogen sul- phide. It will be noticed that the migration rate of the sol itself as well as mixtures of it with potassium chloride solu- i6 Roland Neal tions are all higher than the rates for corresponding solutions saturated with hydrogen sulphide, a matter to which reference will be made later. The effect of the potassium chloride is, how- ever, the same in both cases, namely, to increase the migration rate. Chart V shows the effect of successive additions of potassium chloride to Sol 5, a sol of a high migration rate. The chart shows that the increase in migration rate is ap- proximately proportional to the amount of potassium chloride added. Comparing Chart V with Chart III it is seen that the increase of migration rate produced in the two cases by the same addition of potassium chloride is quite closely pro- portional to the original migration rate of the pure sol. Thus at the end of fifteen minutes, the distance migrated by Sol 4 Colloidal Solutions of Copper Sulphide is i cm, that by Sol 5 is 3 cm and the ratio l /$ = -333 + - the distances migrated at the end of the same period by the same sols to which potassium chloride has been added in I I OOO normality, we have, respectively, 1.5 cm and 4.5 cm, and the ratio = 0.333 + . This is a fortunately chosen case, but in general the rule seems to hold fairly closely. The influence of the potassium chloride on the sol from which hydrogen sulphide had been removed is seen to be much smaller. The effect of calcium chloride is shown in Chart VI. It is entirely similar to potassium chloride in its conduct. One thousandth normal is about the concentration of calcium chloride that will flocculate the sol in twenty-four hours; in fact, it is a trifle above it. It was nevertheless possible to carry out the migration measurements at this concentra- tion, although flocculation occasionally occurred during the experiment. Chart VII shows the relative influence on Sol 5 of potas- 10 15 20 25 30 sium and calcium chlorides at one-thousandth normal con- centrations, the graph for the pure sol being added for com- parison. As will be seen, the effect of the potassium chloride somewhat exceeds that of the calcium chloride, although but little. There is thus no apparent relation between the flocculating power and the migration influence. One seems, however, reasonably justified in surmising that the migration influence stands in a fairly direct relation to some function i8 Roland Neal of the concentrations of the electrolyte added, very possibly the conductivity. In this event the acceleration of the migration rate would be represented by S.K.X where K is the original rate of the uninfluenced sol and X the conductivity induced by the added electrolyte while S is a factor which varies with the condition of the particular sol, and is at present indeterminate. Before this principle can be established with any degree of definiteness, a far more exhaustive series of experiments is necessary. These are planned for the near future. As a matter of fact results of similar experiments per- formed some years ago by R. C. Pollock and F. S. Pratt on arsenious sulphide sols, the results of which are in preparation for publication, show a quite different relationship. Effect of Hydrogen Sulphide. -It has been pointed out above that the effect of hydrogen sulphide is to reduce the rate of migration. Repeated measurements were made of this effect which was always found to occur, whether or not elec- trolyte were present. The results of some measurements of the effect are given in Chart VIII. The reduction of the 4 ISat. w.H^S no Electrolyte 2-" " " n&j 3-No M 2 S no EUctrolijIc, MINWTtS '0 10 15 ?0 25 30 migration rate by hydrogen sulphide is found to be very large. The action in this case was found to be reversible, the migra- tion rate increasing when hydrogen sulphide was dialyzed out and diminishing when it was resaturated with hydrogen sulphide. Colloidal Solutions of Copper Sulphide 19 Here again is found evidence of the rule that the in- fluence of the electrolyte on the migration rate is closely proportional to the value of this quantity for the sol without electrolyte. The distances on Curve 2 are to those on Curve i, taken at the same time as i : 1.35. Those on Curve 4 are to those on Curve 3 as i : 1.34. The different value for this ratio from that obtained from the same sol saturated with hydrogen sulphide illustrates the character of S in the formula suggested above. Influence of Oxygen.- Curing the course of the investiga- tion it was considered wise to determine whether or not free oxygen played any part in influencing the migration rate. The results were negative. Air or oxygen bubbled through a sol freed from hydrogen sulphide by dialysis showed not the least effect. If a sol saturated with hydrogen sulphide were used there occurred always an increase of migration rate which was restored to its original value by resaturation with hydrogen sulphide. If air or oxygen were passed through for a long time the migration rate took on a value very close to that of the dialyzed sol, from which it was concluded that the influence of the air or oxygen was merely that due to displacement of hydrogen sulphide. Summary Copper sulphide sols were prepared in a number of different ways and their conduct with respect to flocculation by elec- trolytes as well as their rates of migration in the electric field were studied. The following conclusions are indicated by the data obtained : (1) The concentration of electrolyte necessary to floccu- late in twenty-four hours is independent of the origin of the sol. (2) It is within wide limits independent of the dilution of the sol. (3) It is independent of the presence or absence of free hydrogen sulphide in the sol. (4) The relative flocculating powers of the chlorides of potassium, calcium and aluminum are as i : 39 : 875. 2O Roland Ncal (5) Evidence is found that sols which have been de- prived of hydrogen sulphide are unstable, and will ultimately flocculate spontaneously. The time required is, however, very long, five or more months. Electrolyte additions show no difference in the conduct of sols saturated and those unsaturated with hydrogen sulphide (see 3). (6) The rate of migration of these sols in the electric field is largely influenced by the origin of the sol. (7) It is accelerated by the addition of electrolytes, but in a way which seems to bear no relation to the flocculating value of the electrolyte used. (8) The acceleration of the migration rate depends upon the migration rate of the original sol, and seems fairly well represented by an expression of the form S.K.X, where K is the migration rate of the original sol and X a function of the concentration (perhaps of the conductivity) of the electrolyte added and S is a factor depending on some specific property of the particular sol used. (9) Other conditions being kept constant, the migration rate is increased by dilution. (10) The migration rate is greatly reduced by hydrogen sulphide, and increased by its removal, the effect being readily reversible. (n) Bubbling air or oxygen through a sol containing hydrogen sulphide affects the migration rate merely by dis- placing hydrogen sulphide. Resaturation with hydrogen sulphide restores the rate to its original value. A considerable portion of the experimental work involved in this paper was carried out at the laboratory of the College of the Pacific, College Park, Cal., the remainder at Stanford University. The above work was carried out at the suggestion, and under the direction of Professor S. W. Young, Depart- ment of Physical Chemistry, Stanford University. Laboratory of Physical Chemistry Stanford University, California Syracuse, N. Y 3C5231 UNIVERSITY OF CAUFORNIA LIBRARY