Distribution of pyridine between water and benzene. R. M. WOODMAN and A. S. CORBET (J.C.S., 1925, 127, 2461-2463).-The ratio of the weight percentages of pyridine in the two layers at 25° is constant at about 2-7 over the range 0-6% pyridine in the total system, then decreases with increasing pyridine concentration to a minimum of 0.8, finally rising again, and reaching unity at the critical point. The results are supported by density data, and may be explained by compound formation between the solute and one of the solvent liquids. R. CUTHILL.

Adsorption of vapours by alumina gel. J. H. PERRY (J. Physical Chem., 1925, 29, 1462-1468).— The adsorption of the vapours of ten organic substances by alumina gel, prepared from aluminium sulphate, has been studied (cf. Munro and Johnson, A., 1925, ii, 191). The hard, glass-like gel, obtained at 60° and containing 29.98% of water, was activated by passing dry air, free from carbon dioxide and heated at 200°, through it at the same temperature. Air saturated with vapour was then passed over the gel at 25°, and adsorption followed by weighing. The adsorption decreases in the series methylene chloride, ethylene dichloride, carbon tetrachloride, and chloroform. Chlorine derivatives are more or less oxidised in air to acid, which, on a large scale, would destroy the activity of the gel. Methyl ethyl ketone and ethyl formate are readily adsorbed, but a reaction with the gel prevents the recovery of the former; toluene is more efficiently adsorbed than hexane. With methyl acetate the adsorption is nearly complete in the early stages, but decreases rapidly above a certain limit. This substance can only be removed from the gel with great difficulty. Carbon disulphide is satisfactorily adsorbed, but the gel becomes yellow when recovery is attempted. The saturation values for the above-mentioned substances vary between 2.46 g. for methyl ethyl ketone and 3.8 g. for chloroform per 10 g. of gel.

Preliminary experiments show that benzene or ether, adsorbed by alumina gel from mixtures with water vapour, is slowly displaced by the latter on continued passage of the mixed vapour through the gel. L. S. THEOBALD.

Adsorption of dissolved substances. J. BANCELIN (J. Chim. physique, 1925, 22, 518-555).Measurements of the amounts of various dyes and salts adsorbed from aqueous solution by glass and silica surfaces at different temperatures gave, in general, results of the order of 10-8g./cm.2of adsorbent. The adsorbed layer is thus not unimolecular, but is of the order of one-tenth saturated. The relationship between the concentration of the solution and the quantity of solute adsorbed is represented, within wide limits, by Perrin's equation; within certain narrow limits of concentration, Freundlich's equation is, however, satisfied. The quantities adsorbed on the free surface of the solution approach a limiting value indicating saturation of the surface and conformity with Perrin's equation. Measurements of the surface tensions of the solutions have been made and, within wide limits, the adsorption at the free surface of the solution and on the surface of mercury

is in agreement with Gibbs' equation. When a surface concentrating substance such as saponin is added to the solution, the amount of solute adsorbed at the surface is of the same order as that on the solid surfaces. The degree of precision of the measurements is low. A. E. MITCHELL.

bromine by wood charcoal and their removal by Active charcoal. IV. Binding of chlorine and ammonia and water vapour. O. RUFF, E. RIMROTT, and H. ZEUMER (Kolloid-Z., 1925, 37, 270-274). Carefully de-tarred alderwood charcoal takes up chlorine and bromine at a red heat and subsequent evacuation to 1 mm. pressure at the same temperature does not remove any appreciable quantity of these substances. They replace hydrogen which is still present in the charcoal, either (i) on its surface, or (ii) in impurities. Evidence for (i) is given by the fact that a part of the halogen may be made to react hydrobromic acid, whilst the variation in the amount with ammonia to give nitrogen and hydrochloric or and the somewhat easier introduction of chlorine of halogen taken up with the experimental conditions, compared with bromine, point to (ii). The principal reactions of the halogenated charcoals with water vapour and with ammonia lead to the elimination of carbon dioxide and hydrogen cyanide, respectively. The results afford no explanation of the cause of the activity of a charcoal, but rather make this problem more obscure, since it is found that tarry substances in the interior and foreign atoms on the surface do not affect the power of being activated. Further, there is no difference in this respect between the de-tarred charcoals used and a very carefully purified specimen. N. H. HARTSHORNE.

Adsorption. I. Adsorption by coconut char- · coal from alcohol-benzene and acetone-benzene mixtures. F. G. TRYHORN and W. F. WYATT (Trans. Faraday Soc., Nov. 1925, advance proof).—The partial vapour pressures of the components of alcoholbenzene and acetone-benzene mixtures have been determined by evaporation of about 5-8% of the binary mixture at 20° and ascertaining the change in composition by measurement of the refractive index of the mixture. The values derived in this way were reproducible to within 2%. The adsorption of the components from these mixtures by specially prepared charcoal was determined by measuring the increase in weight of the charcoal and the change in refractive index of the liquid. The changes of composition of the liquid mixtures when in direct contact with the charcoal are similar to those which are found when adsorption takes place from the vapour phase, and the results support the conclusion that an adsorbed film which is in equilibrium with a saturated vapour must also be in equilibrium with the liquid in contact with that vapour. Alcohol is selectively adsorbed from all mixtures of alcohol and benzene; acetone is selectively adsorbed from mixtures containing up to 72 mol.% of acetone in the vapour phase. These results indicate that the surface of the charcoal is to some extent polar in its action.

F. G. SOPER.

Negative adsorption. Surface tensions and activities of some aqueous salt solutions. A. K. GOARD (J.C.S., 1925, 127, 2451-2458; cf. Goard and Rideal, A., 1925, ii, 510). The surface tensions of aqueous solutions of varying concentrations of the chlorides of lithium, sodium, potassium, and cadmium,

and of silver nitrate have been determined at 20°, and the results used, with the aid of activity data from various sources, to calculate the (negative) surface adsorption of the solutions. The results do not agree with the view that the surface layer of salt solutions consists of a single layer of oriented water molecules. Solutions of cadmium chloride behave anomalously, probably due to the undissociated molecules as well as the ions playing a part in the adsorption.

R. CUTHILL.

Adhesive forces in solution. V. Adsorption of complex__compounds. N. SCHILOV and B. NEKRASSOV (Z. physikal. Chem., 1925, 118, 79-88). -The adsorption phenomena observed when charcoal is brought into contact with aqueous solutions of substances of the type [(NH4)2SO4,MSO,6H2O], cuprammonium compounds, and complex cobalt compounds have been investigated. In all cases, it is found that a disruption of the complex compounds occurs. Although ammonium sulphate is adsorbed to approximately the same extent from all the double sulphates, the order of increasing adsorbabilities of the bivalent metal sulphates is: nickel, zinc, magnesium, manganese, cadmium, cobalt, iron, copper. The complex cobalt ions are reduced to the cobaltous state, the ion [Co(NH3)2]" being adsorbed. In the case of the cuprammonium compounds the ion. [Cu(NH3)2]" is adsorbed. Sodium is adsorbed much less than cobalt from solutions of sodium salts containing cobalt as part of the complex anion. Unsymmetrical cis-1: 2-[Co(NH3), (NO2)2]Cl is adsorbed much more readily than the symmetrical transcompound. Maximum adsorption is obtained with the non-ionised compound, [Co(NH3)3(NO2)3].

J. S. CARTER.

Nature of the interfacial layer between an aqueous and a non-aqueous phase. F. L. USHER (Trans. Faraday Soc., Nov. 1925, advance proof). The total charge on the particles of an aqueous gamboge suspension has been found by measurement of the quantity of ferric ions associated with the gamboge after coagulation by these ions. The charge per particle, which is 1-45 × 10-3 e. s. unit, is 100 times that calculated from the observed migration velocity on the basis of a Helmholtz double layer of molecular dimensions, and 7000 times that calculated assuming the double layer to be diffuse. It is also found that the flocculating power of various cations on a suspension of cupric ferricyanide is proportional to their effect in reducing the concentration of ferricyanide ions by the formation of insoluble ferricyanides. These results support the theory that the surface charge is due to ionisation of molecules furnished by the material itself or adsorbed from the surrounding liquid. The increase of ionisation with dilution explains the increased stability of certain suspensions and their greater adsorption of the flocculating ion in

diluted systems. The variation in migration speed with particle size is also satisfactorily explained.

F. G. SOPER.

Attack of hydrogen chloride and ammonium halide on metals. K. A. HOFMANN and F. HARTMANN. See this vol., 37.

Production of oxide films on copper at the ordinary temperature. U. R. EVANS (J.C.S., 1925, 127, 2484-2491; cf. A., 1925, ii, 688).Cathodic treatment of copper in 0.1N-sodium hydroxide at the ordinary temperature results in the production of colours similar to those obtained by exposure to air at high temperatures. If the current density is sufficient to cause evolution of hydrogen at the cathode, the colours appear only after the termination of the electrolysis, and when the electrode is allowed to remain in the solution. The effect is destroyed by rubbing or vigorous stirring, but is independent of the current density. Compact copper free from oxide and from porous copper (produced by reduction of oxide) is not easily rendered colourable by cathodic treatment. The phenomenon is probably due to hydrogen stored in the copper, which permits of alternate oxidation and reduction.

If copper is anodically treated in 0.1N-sodium hydroxide, it becomes passive at high current density, but at low current densities becomes coated with a black deposit, and if the solution is stirred, bright colours appear on the anode.

Copper is much less readily attacked by silver nitrate solution after considerable exposure to air than when freshly ground; this is presumably due to the presence of an oxide film in the former case.

R. CUTHILL.

Rhythmic reactions. K. SEIDEL (Kolloid-Z., 1925, 37, 298-299).-The damp root hairs of oat, wheat, rice, and other seedlings develop periodic bands when placed in contact with silver nitrate crystals. Periodic plasmolysis is also observed when glycerol is allowed to diffuse into the roots. The results are discussed. N. H. HARTSHORNE.

Influence of light on lead chromate bands. E. HATSCHEK (Kolloid-Z., 1925, 37, 297-298).— Polemical, in reply to Dhar and Chatterji (cf. A., 1925, ii, 959). The author's paper on lead chromate bands (cf. Proc. Roy. Soc., 1921, A, 96, 496) has been misunderstood. The reaction between lead acetate and potassium chromate or dichromate in agar gives explanation of the screw-like bands was attempted. bands only in the light and not in the dark. No that the precipitate is initially of a colloidal nature. The author doubts the statement of Dhar and Chatterji

N. H. HARTSHORNE.

Growth of structures formed by reactions at the boundary between solutions of electrolytes in water and those in gels. V. MORÁVEK (Pub. Fac. Sci. Univ. Masaryk, 1925, No. 29).-When a layer of 0.1N-lead nitrate solution is placed on a 5% gelatin gel, 0.1N with respect to potassium dichromate, fibrous structures, 0·1-0.3 mm. wide, grow from the gel into the nitrate solution. The walls of these are formed by a gelatin membrane coated with

precipitated lead chromate, and growth usually occurs vertically or horizontally, although occasionally in spiral formation. The gelatin membrane is first formed and the lead chromate deposited on it discontinuously. The velocity of growth has a maximum value of 0.16 mm./min. The dimensions, shape, and velocity of growth of the fibres are influenced by the concentration of the gelatin, by substitution of chromate for dichromate ions, addition of foreign ions, and by change of temperature. The nature of the growth of these fibres, and the influence of ions and of temperature on it, resemble in many respects the phenomena of cell-formation. F. G. TRYHORN.

Diffusion experiments in solutions. R. FÜRTH (Physikal. Z., 1925, 26, 719-722).-The diffusion of a coloured solute in a solution contained in a small cell, about 10×5×1 mm., is followed colorimetrically by comparison with a solution of known concentration in a similar neighbouring cell. The solutions are viewed through a low-power microscope fitted with a micrometer eyepiece, the field being limited to a narrow horizontal strip. The displacement of a band of definite concentration is thus obtained as a function of the time. The method requires only a relatively short time (1-1 hrs.) for one series of observations, and is capable of considerable accuracy, Fick's law is verified for solutions of picric acid and Congo-red, the respective diffusion coefficients being 6.4×10% to 6.5x 106 and 1.8x106 to 1-6x106. A modification of the method, using the refractive index as a measure of the concentration, has been applied to solutions of sucrose. The results are less accurate, but show that here Fick's law is not applicable, the diffusion coefficient passing through a minimum with increasing concentration. A. B. MANNING.

Specific gravity of concentrated solutions of phosphoric acid. W. H. Ross and R. M. JONES (Ind. Eng. Chem., 1925, 17, 1170-1171).-See B., 1925, 987.

Temperature of steam evolved from a [boiling aqueous] solution. E. REISSMANN [with SCHREBER] (Z. angew. Chem., 1925, 38, 1040-1044; cf. A., 1925, ii, 105).—Experiments were performed using an apparatus similar to that previously described but in which the heating of the steam by radiation from the hot oil-bath is minimised. It is also shown that superheated steam blown into a boiling solution is in equilibrium with the latter if its temperature before entry is the same, but that if cooler it absorbs heat from, and if warmer gives up heat to, the solution. The results are taken by the author [Schreber dissenting], to confirm his previous conclusions.

W. T. K. BRAUNHOLTZ.

Viscosity and chemical analogy with reference to the viscosity of aqueous metallic acetate solutions. L.-J. SIMON (Compt. rend., 1925, 181, 862-864). The methods previously outlined for determining neutralisation points viscosimetrically are now applied to the mixing of N-solutions of some fatty acids with N-solutions of sodium and potassium hydroxide. The viscosity-alkali concentration curve is a straight line for acetic acid and is nearly so for propionic acid. The curves for chloro- and trichloro

acetic and formic acids show at the neutralisation point the discontinuity previously observed with mineral acids. Solutions of trichloroacetic acid are much more viscous than equivalent solutions of chloroacetic acid; the latter are similarly more viscous than solutions of acetic acid, although the viscosities of their sodium salts in equivalent solution are fairly close. The viscosities of the alkali salts of trichloroacetic acid are very near those of the corresponding propionates in equivalent solution. Comparison of the viscosities of equivalent solutions of the acetates or nitrates of the metals reveals the well-known similarities between, for example, lithium and calcium etc. (cf. A., 1925, ii, 512). S. K. TWEEDY.

of salts. G. S. ADAIR (Proc. Roy. Soc., 1925, A, Osmotic pressure of hæmoglobin in absence 109, 292-300).-Osmotic pressures of hæmoglobin determined by previous workers varied from 3.5 to 12.1 mm. per 1% of protein. The high value was accepted as representing hæmoglobin of mol. wt. low values were attributed to aggregation. Using 16,700 (i.e., the same as the equivalent), whilst the the conductivity of the solutions as a test of their freedom from impurities, it has been found that the pressure of 3.2 mm. per 1% of protein as the isosolution of lowest conductivity gives an osmotic electric point. If the low values were due to aggregation, they should decrease still further at higher concentrations of the protein, but actually the figures indicate a slight increase. It is concluded that the mol. wt. of hæmoglobin is about four times the equivalent.

S. BARRATT.

Colloid chemistry of bismuth and its compounds. A. KUHN and H. PIRSCH (Koll. Chem. Beihefte, 1925, 21, 78-96).-The preparation of sols of bismuth and of some of its compounds is described. Bismuth sols containing 3.5 mg. per c.c. may be prepared by the reduction of bismuth tartrate solutions with sodium hyposulphite in weakly alkaline solution. This shows an improvement on the methods of Lottermoser and Vanino, which with stannic acid as a protective colloid give sols containing only 1.5 mg. per c.c. Bismuth sulphide sols are stable only in the presence of protective colloids. With 1% of gum arabic a sol containing 10 mg. per c.c. may be obtained. Bismuth sulphoiodide sols cannot be prepared without protective colloids. Sols containing 0.3 mg. of bismuth per c.c. with 0.59% of gum arabic, 1.6 mg. with 0.5% of hæmoglobin, and 0.3 mg. with 0-25% of gelatin may be obtained. The gum arabic sol is reversible. By mixing bismuth sulphide, sulphoiodide, or hydroxide with wool fat, subsequent intensive pulverisation of the mixture at the temperature of liquid air, and dissolution of the product in sesamé oil, fairly stable sols result containing 12-16 mg. of bismuth per c.c. for 15-20% of protective colloid. By Bredig's method stable sols of the metal may be prepared containing 8 mg, per c.c. with 5% of wool fat and 0-6 mg. per c.c. with 1% of caoutchouc-ether sol. By Svedberg's protected arc method (Medd. Nobel-Inst., 1910, 5, No. 10, 4) metal sols cannot be obtained in water even with a protective colloid, but in sesamé oil with at least

Negative adsorption. Surface tensions and activities of some aqueous salt solutions. A. K. GOARD (J.C.S., 1925, 127, 2451-2458; cf. Goard and Rideal, A., 1925, ii, 510).-The surface tensions of aqueous solutions of varying concentrations of the chlorides of lithium, sodium, potassium, and cadmium, and of silver nitrate have been determined at 20°, and the results used, with the aid of activity data from various sources, to calculate the (negative) surface adsorption of the solutions. The results do not agree with the view that the surface layer of salt solutions consists of a single layer of oriented water molecules. Solutions of cadmium chloride behave anomalously, probably due to the undissociated molecules as well as the ions playing a part in the adsorption.

R. CUTHILL.

Adhesive forces in solution. V. Adsorption of complex compounds. N. SCHILOV and B. NEKRASSOV (Z. physikal. Chem., 1925, 118, 79-88). -The adsorption phenomena observed when charcoal is brought into contact with aqueous solutions of substances of the type [(NH4)2SO4,MSO4,6H20], cuprammonium compounds, and complex cobalt compounds have been investigated. In all cases, it is found that a disruption of the complex compounds occurs. Although ammonium sulphate is adsorbed to approximately the same extent from all the double sulphates, the order of increasing adsorbabilities of the bivalent metal sulphates is: nickel, zinc, magnesium, manganese, cadmium, cobalt, iron, copper. The complex cobalt ions are reduced to the cobaltous state, the ion [Co(NH3)2]" being adsorbed. In the case of the cuprammonium compounds the ion [Cu(NH3)2]" is adsorbed. Sodium is adsorbed much less than cobalt from solutions of sodium salts containing cobalt as part of the complex anion. Unsymmetrical cis-1:2-[Co(NH3)(NO),]Cl is adsorbed much more readily than the symmetrical transcompound. Maximum adsorption is obtained with the non-ionised compound, [Co(NH3)3(NO2)3].

J. S. CARTER.

Nature of the interfacial layer between an aqueous and a non-aqueous phase. F. L. USHER (Trans. Faraday Soc., Nov. 1925, advance proof). The total charge on the particles of an aqueous gamboge suspension has been found by measurement of the quantity of ferric ions associated with the gamboge after coagulation by these ions. The charge per particle, which is 1-45 × 10-3 e. s. unit, is 100 times that calculated from the observed migration velocity on the basis of a Helmholtz double layer of molecular dimensions, and 7000 times that calculated assuming the double layer to be diffuse. It is also found that the flocculating power of various cations on a suspension of cupric ferricyanide is proportional to their effect in reducing the concentration of ferricyanide ions by the formation of insoluble ferricyanides. These results support the theory that the surface charge is due to ionisation of molecules furnished by the material itself or adsorbed from the surrounding liquid. The increase of ionisation with dilution explains the increased stability of certain suspensions and their greater adsorption of the flocculating ion in

diluted systems. The variation in migration speed with particle size is also satisfactorily explained.

F. G. SOPER.

Attack of hydrogen chloride and ammonium halide on metals. K. A. HOFMANN and F. HARTMANN. See this vol., 37.

Production of oxide films on copper at the ordinary temperature. U. R. EVANS (J.C.S., 1925, 127, 2484-2491; cf. A., 1925, ii, 688).Cathodic treatment of copper in 0.1N-sodium hydroxide at the ordinary temperature results in the production of colours similar to those obtained by exposure to air at high temperatures. If the current density is sufficient to cause evolution of hydrogen at the cathode, the colours appear only after the termination of the electrolysis, and when the electrode is allowed to remain in the solution. The effect is destroyed by rubbing or vigorous stirring, but is independent of the current density. Compact copper free from oxide and from porous copper (produced by reduction of oxide) is not easily rendered colourable by cathodic treatment. The phenomenon is probably due to hydrogen stored in the copper, which permits of alternate oxidation and reduction.

If copper is anodically treated in 0.1N-sodium hydroxide, it becomes passive at high current density, but at low current densities becomes coated with a black deposit, and if the solution is stirred, bright colours appear on the anode.

Copper is much less readily attacked by silver nitrate solution after considerable exposure to air than when freshly ground; this is presumably due to the presence of an oxide film in the former case.

R. CUTHILL.

Rhythmic reactions. K. SEIDEL (Kolloid-Z., 1925, 37, 298-299).-The damp root hairs of oat, wheat, rice, and other seedlings develop periodic bands when placed in contact with silver nitrate crystals. Periodic plasmolysis is also observed when glycerol is allowed to diffuse into the roots. The results are discussed. N. H. HARTSHORNE.

Influence of light on lead chromate bands. E. HATSCHEK (Kolloid-Z., 1925, 37, 297-298).Polemical, in reply to Dhar and Chatterji (cf. A., 1925, ii, 959). The author's paper on lead chromate bands (cf. Proc. Roy. Soc., 1921, A, 96, 496) has been misunderstood. The reaction between lead acetate and potassium chromate or dichromate in agar gives explanation of the screw-like bands was attempted. bands only in the light and not in the dark. No The author doubts the statement of Dhar and Chatterji that the precipitate is initially of a colloidal nature.

N. H. HARTSHORNE.

Growth of structures formed by reactions at the boundary between solutions of electrolytes in water and those in gels. V. MORÁVEK (Pub. Fac. Sci. Univ. Masaryk, 1925, No. 29).-When a layer of 0.1N-lead nitrate solution is placed on a 5% gelatin gel, 0.1V with respect to potassium dichromate, fibrous structures, 0.1-0.3 mm. wide, grow from the gel into the nitrate solution. The walls of these are formed by a gelatin membrane coated with

precipitated lead chromate, and growth usually occurs vertically or horizontally, although occasionally in spiral formation. The gelatin membrane is first formed and the lead chromate deposited on it discontinuously. The velocity of growth has a maximum value of 0.16 mm./min. The dimensions, shape, and velocity of growth of the fibres are influenced by the concentration of the gelatin, by substitution of chromate for dichromate ions, addition of foreign ions, and by change of temperature. The nature of the growth of these fibres, and the influence of ions and of temperature on it, resemble in many respects the phenomena of cell-formation. F. G. TRYHORN.

Diffusion experiments in solutions. R. FÜRTH (Physikal. Z., 1925, 26, 719–722).—The diffusion of a coloured solute in a solution contained in a small cell, about 10×5×1 mm., is followed colorimetrically by comparison with a solution of known concentration in a similar neighbouring cell. The solutions are viewed through a low-power microscope fitted with a micrometer eyepiece, the field being limited to a narrow horizontal strip. The displacement of a band of definite concentration is thus obtained as a function of the time. The method requires only a relatively short time (1-1 hrs.) for one series of observations, and is capable of considerable accuracy. Fick's law is verified for solutions of picric acid and Congo-red, the respective diffusion coefficients being 6.4 × 10% to 6.5 × 106 and 1.8 × 10-6 to 1.6 × 10-6. A modification of the method, using the refractive index as a measure of the concentration, has been applied to solutions of sucrose. The results are less accurate, but show that here Fick's law is not applicable, the diffusion coefficient passing through a minimum with increasing concentration. A. B. MANNING.

Specific gravity of concentrated solutions of phosphoric acid. W. H. Ross and R. M. JONES (Ind. Eng. Chem., 1925, 17, 1170-1171).-See B., 1925, 987.

Temperature of steam evolved from a [boiling aqueous] solution. E. REISSMANN [with SCHREBER] (Z. angew. Chem., 1925, 38, 1040-1044; cf. A., 1925, ii, 105).-Experiments were performed using an apparatus similar to that previously described but in which the heating of the steam by radiation from the hot oil-bath is minimised. It is also shown that superheated steam blown into a boiling solution is in equilibrium with the latter if its temperature before entry is the same, but that if cooler it absorbs heat from, and if warmer gives up heat to, the solution. The results are taken by the author [Schreber dissenting], to confirm his previous conclusions. W. T. K. BRAUNHOLTZ.

Viscosity and chemical analogy with reference to the viscosity of aqueous metallic acetate solutions. L.-J. SIMON (Compt. rend., 1925, 181, 862-864). The methods previously outlined for determining neutralisation points viscosimetrically are now applied to the mixing of N-solutions of some fatty acids with N-solutions of sodium and potassium hydroxide. The viscosity-alkali concentration curve is a straight line for acetic acid and is nearly so for propionic acid. The curves for chloro- and trichloro

acetic and formic acids show at the neutralisation point the discontinuity previously observed with mineral acids. Solutions of trichloroacetic acid are much more viscous than equivalent solutions of chloroacetic acid; the latter are similarly more viscous than solutions of acetic acid, although the viscosities of their sodium salts in equivalent solution are fairly close. The viscosities of the alkali salts of trichloroacetic acid are very near those of the corresponding propionates in equivalent solution. Comparison of the viscosities of equivalent solutions of the acetates or nitrates of the metals reveals the well-known similarities between, for example, lithium and calcium etc. (cf. A., 1925, ii, 512). S. K. TWEEDY.

Osmotic pressure of hæmoglobin in absence of salts. G. S. ADAIR (Proc. Roy. Soc., 1925, A, 109, 292-300).-Osmotic pressures of hæmoglobin determined by previous workers varied from 3.5 to 12.1 mm. per 1% of protein. The high value was accepted as representing hæmoglobin of mol. wt. 16,700 (i.e., the same as the equivalent), whilst the low values were attributed to aggregation. Using the conductivity of the solutions as a test of their freedom from impurities, it has been found that the solution of lowest conductivity gives an osmotic pressure of 3.2 mm. per 1% of protein as the isogation, they should decrease still further at higher electric point. If the low values were due to aggreconcentrations of the protein, but actually the figures indicate a slight increase. It is concluded that the mol. wt. of hæmoglobin is about four times the equivalent.

S. BARRATT.

Colloid chemistry of bismuth and its compounds. A. KUHN and H. PIRSCH (Koll. Chem. Beihefte, 1925, 21, 78-96).-The preparation of sols of bismuth and of some of its compounds is described. Bismuth sols containing 3.5 mg. per c.c. may be prepared by the reduction of bismuth tartrate solutions with sodium hyposulphite in weakly alkaline solution. This shows an improvement on the methods of Lottermoser and Vanino, which with stannic acid as a protective colloid give sols containing only 1.5 mg. per c.c. Bismuth sulphide sols are stable only in the presence of protective colloids. With 1% of gum arabic a sol containing 10 mg. per c.c. may be obtained. Bismuth sulphoiodide sols cannot be prepared without protective colloids. Sols containing 0.3 mg. of bismuth per c.c. with 0.59% of gum arabic, 1.6 mg. with 0.5%, of hæmoglobin, and 0.3 mg. with 0-25% of gelatin may be obtained. The gum arabic sol is reversible. By mixing bismuth sulphide, sulphoiodide, or hydroxide with wool fat, subsequent intensive pulverisation of the mixture at the temperature of liquid air, and dissolution of the product in sesamé oil, fairly stable sols result containing 12-16 mg. of bismuth per c.c. for 15-20% of protective colloid. By Bredig's method stable sols of the metal may be prepared containing 8 mg, per c.c. with 5% of wool fat and 0.6 mg. per c.c. with 1% of caoutchouc-ether sol. By Svedberg's protected arc method (Medd. Nobel-Inst., 1910, 5, No. 10, 4) metal sols cannot be obtained in water even with a protective colloid, but in sesamé oil with at least