barium, and lead. The following values, based on the values of a determined by X-ray measurements, and on those of the ratios a: c determined by crystallographic methods, are calculated for the elementary cells of these substances. CaMoO4, a=3.67, c= 5-69 A.; SrMoO1, a=3.79, c=5-97 A.; BaMoO4, a=3.96, c=6.43 A.; PbMoO1, a=3.80, c=6.00 A. The densities, calculated from these data, are, respectively, 4-28, 4-73, 4-84, 6-95, and agree with experimental values. For the alkaline-earth molybdates, the values of a and c are approximately linear functions of the atomic weights, or atomic numbers, of the cation metals. Since a similar relationship holds for the nitrates of these metals, the possibility is suggested of calculating the dimensions of the elementary cells of isomorphous phases, which are too unstable to be experimentally investigated.

F. G. TRYHORN.

Isomorphism between samarium compounds and the corresponding compounds of calcium, strontium, barium, and lead. G. CAROBBI (Rend. Accad. Sci. Fis. Mat. Napoli, 1925, [iii], 31, 83—95).— The results of the author's investigations on the molybdates and phosphates, together with Zambonini's results (A., 1923, ii, 499, 691), show that certain samarium compounds are capable of forming homogeneous mixed crystals with the corresponding compounds of calcium, strontium, and lead, and that the isomorphism of samarium towards the metals of the isomorphogenic calcium group is more pronounced than that of the other metals of the cerium group. T. H. POPE.

New magnetic state of the cobalt ion. (MLLE.) A. SERRES (Compt. rend., 1925, 181, 714-715).Anhydrous amorphous cobalt sulphate has been obtained containing a cobalt ion corresponding with 26 magnetons and possessing a powerful negative molecular field. The magnetic susceptibilities of the salt are recorded over the temperature range 7422°. (Cf. Foex, A., 1925, ii, 345; Théodoridès, J. Phys. Radium, 1922, [vi], 3, 1.)

S. K. TWEEDY.

Effect of temperature on the paramagnetism of cobalt salts in solution. M. CHATILLON (Compt. rend., 1925, 181, 778-780).—The magnetic properties of solutions of cobaltous chloride, sulphate, and nitrate have been examined at temperatures from 10° to 140°. When the reciprocal of the coefficient of magnetisation, corrected for the diamagnetism of the molecule, is plotted against the absolute temperature, straight lines are obtained which do not pass through the absolute zero, but correspond with a Curie point of -12°, and with a moment of 25.0 magnetons, in contradiction to the value 24.5 found by Brant (Physical Rev., 1921, [ii], 17, 768). The anhydrous sulphate according to Serres (cf. preceding abstract)) has a moment of 26.0 magnetons, and this value is retained when the anhydrous salt is dissolved in water, the solution giving a Curie point of -33°. But if the solution is crystallised and the crystals are redissolved, the value of 25.0 magnetons is found. W. HUME-ROTHERY.

Magnetic susceptibility of rubidium bromide, cæsium iodide, krypton, xenon, [potassium, rubidium, and cæsium]. L. CROW (Trans. Roy. Soc. Canada, 1925, [iii], 19, III, 63-70).-The magnetic susceptibilities of krypton and xenon have been determined from observations on rubidium

bromide and cæsium iodide, the experimental method of Meyer (A., 1925, ii, 173) being used. The molecular magnetic susceptibilities of rubidium bromide and cæsium iodide are -62.9 × 106 and -77.6 × 106, respectively. The derived values for the volume susceptibilities at 20° and 760 mm. of krypton and xenon are -13.1 × 10-10 and -16.2 × 10-10. The atomic magnetic susceptibilities of potassium, rubidium, and cæsium are, respectively, -1.9x10-6, -7.0 × 106, and -8.0 × 106. J. S. CARTER.

Magnetic rotatory power of some paramagnetic minerals at very low temperatures. H. K. ONNES, J. BECQUEREL, and W. J. DE HAAS (Compt. rend., 1925, 181, 838-841).—The magnetic rotatory power of certain didymium minerals increases with decreasing temperature down to 4.2° Abs. (A., 1908, ii, 338). The results for tysonite, parisite, and bastnaesite, after a small correction, indicate that Curie's law is followed. Complex results are obtained with xenotime, due to the superposition of two distinct magnetic polarisations; one, generally positive, is due to the Zeeman effect for the absorbed radiations, whilst the other, negative, is peculiar to paramagnetic substances and is due either to orientation phenomena or to distortion in the S. K. TWEEDY. magnetic field.

Absorption spectra of rare-earth crystals in a magnetic field at the temperature of liquid helium. J. BECQUEREL, H. K. ONNES, and W. J. DE HAAS (Compt. rend., 1925, 181, 758—760).—Uniaxial crystals of xenotime (spectrum due to erbium), tysonite, parisite, and bastnaesite (spectrum due to didymium) have been examined. The absorption spectra, which are complex at 80° Abs., become more simple at 14° Abs. At 4-2° Abs., spectra are on the whole even more simple, some bands having disappeared, but others observed only at low temperatures have increased in intensity. The crystals were also examined in a field of 26,170 gauss, the optical axis and beam of light being in the direction of the lines of force. The change of period produced by the magnetic field is independent of the temperature. Great dissymmetry is shown by the intensity of the components corresponding with the absorption of inverse circular vibration, one component sometimes completely disappearing. At 14° reinforced, but this is not always the case at 4.2° Abs. Abs., the component displaced towards the violet is

W. HUME-ROTHERY.

Effect of the addition of dyes on the crystallisation of calcium carbonate. V. KOHLSCHÜTTER and C. EGG (Helv. Chim. Acta, 1925, 8, 697–703; cf. A., 1925, ii, 1035, 1036).—The work on somatoid forms has been extended by an investigation of the effect, on crystallisation of calcium carbonate from solutions of the hydrogen carbonate, of the following dyes Congo-red, gallacetophenone, ponceau, alizarin

Bismarck-brown, eosin, crystal-violet, fuchsin, tropaolin, and chrysoidine. Congo-red is adsorbed principally by calcite with some deformation of the crystals, whilst aragonite exhibits a smaller tendency to adsorb the dye. The formation of aragonite crystalnuclei is hindered and sometimes totally inhibited. The importance of colloidal phenomena in the work earlier stressed is now confirmed by the observation that the most effective dyes exhibit an ultramicroscopic effect either alone or in solutions containing the calcium ion. Thus a fresh solution of Congo-red and calcium hydrogen carbonate exhibits an intense light-cone. The colloids are adsorbed on the crystal surfaces and disturbances of crystal-habit and growth ensue which are peculiar to each colloidal dye. Small additions of Congo-red cause the formation of rhombohedra with red cores, suggesting that particles of the calcium salt of the dye serve as a nucleus for crystallisation of calcite. Higher dye concentrations render the crystal surfaces "passive," resulting in deformation or inhibition in crystal-growth.

R. A. MORTON.

Formation of somatoid forms. V. KOHLSCHÜTTER, M. BOBTELSKY, and C. EGG (Helv. Chim. Acta, 1925, 8, 703-723; cf. preceding abstract).An attempt is made to define more clearly (cf. A., 1925, ii, 1035) the conditions for the appearance of a whole range of somatoid forms. One type of disclike bodies only appears in the presence of a colloid. A whole series of somatoid forms designated a is intimately related with the formation of calcite rhombohedra, and the precipitation of these forms occurs only when the ratio of calcium carbonate to colloidal substance comes within a definite range. The particular form is for the a-series largely independent of the chemical nature of the added colloid. By varying the amount of added material, the whole series of a-somatoid forms can be obtained with one and the same substance. More specific effects occur in the formation of B-somatoid forms when aggregations of calcium carbonate and colloidal substance play a dominating part. The investigation concerns crystallisation under conditions approaching those existing in the living organism. R. A. MORTON.

Electrical conductivity of phosphorus pentachloride. G. W. F. HOLROYD, H. CHADWICK, and J. E. H. MITCHELL (J.C.S., 1925, 127, 2492-2493) -A preliminary paper on the mode of ionisation of phosphorus pentachloride in solution.

R. CUTHILL.

Light scattering of "abnormal" liquids. O. C. H. KITCHING (Trans. Roy. Soc. Canada, 1925, [iii], 19, III, 30-31). The light scattering of aniline, phenol, and acetaldoxime at various temperatures between 20° and 100° has been determined relatively to benzene. With aniline and phenol (normal liquids), the observed temperature coefficients and light scatterings are in fairly good agreement with the calculated values, whilst in acetaldoxime the scattering is more than double and the temperature coefficient less than half the calculated value.

J. S. CARTER.

Light scattering of salt solutions. C. W. SWEITZER (Trans. Roy. Soc. Canada, 1925, [iii], 19, III, 31).-A revision of the measurements previously reported (A., 1925, ii, 512) has led to values which approach more closely to those calculated. Methyl alcohol was used in preference to water as the standard of comparison. J. S. CARTER.

Scattering of light by anisotropic liquids. 19, III, 36-38; cf. A., 1925, ii, 90). The depolarW. H. MARTIN (Trans. Roy. Soc. Canada, 1925, [iii], isation factor of the light scattered by p-azoxyanisole, rendered dust-free by filtration through a Mandler diatomaceous earth filter, is 0.85. For a Mandler diatomaceous earth filter, is 0.85. For a liquid composed of molecules in which vibration can occur in only one direction, the calculated value is 0-86. The experimental result hence supports the hypothesis of needle-shaped molecules in this type of liquid crystals.

J. S. CARTER.

Dependence of absorption coefficient on the thickness of the absorbing colour layer. A. PREDVODITELEV and W. BLINOV (Z. Physik, 1925, 35, 38-49). The absorption coefficients of thin layers of collodion solutions of fuchsin and crystalviolet, of different concentrations, have been determined in a König-Martens spectrometer. The thickness of the films was of the order of 1 μ. The absorptive coefficient was found to be a periodic function of the density distribution of the dye molecules at the surface. It is suggested that this behaviour is due to the transformation into heat of only part of the light energy absorbed, the remainder going to increase the internal energy of the atom or molecule. On this assumption, formulæ have been deduced, and theoretical curves constructed, which are in close agreement with the experimental results.

M. S. BURR.

dynamic action of gases and their equations of Specific heat variations in relation to the 1244-1260). Some preliminary experiments on state. W. J. WALKER (Phil. Mag., 1925, [vi], 50, the flow of highly superheated steam through a nozzle are described. The nozzle plate was bolted between two cast-iron boxes of square section and 6 in. internal dimension. The box to which the steam was first admitted contained an electric heater, and the temperature of the steam was measured by an ironconstantan couple placed just behind the nozzle plate and opposite the centre of the nozzle orifice. The pressure was measured by a Bourdon gauge. From the second box, the steam was conducted to a surface condenser, where it was collected and weighed.

ain.

Results are tabulated giving the observed discharge through a 1 in. nozzle of stream-lined form, for of 480-700° Abs. The results suggest an adiabatic pressures of 26-56 lb. per sq. in. and temperatures relation of the form pumef=constant, in which the factor eft involves the variation of specific heat with temperature. This relation is shown to lead to a general equation of state of the Dieterici and Jeans type.

A. B. MANNING.

Physical properties of nitromethane. J. W. WILLIAMS (J. Amer. Chem. Soc., 1925, 47, 26442652).-Nitromethane consists of an equilibrium

mixture of the two forms CH, NO, and CH,.NO OH, and its physical properties may, therefore, be expected to vary irregularly with the temperature. The specific heat, determined as previously described (A., 1924, ii, 589), depends greatly on the method of drying the nitromethane. When dried by distillation from phosphorus pentoxide, the results do not fall on a smooth curve, but there is a minimum at about 30°. Vapour pressures were determined by the method of Ramsay and Young, but no irregularity exists, the logarithm of the vapour pressure varying linearly with the reciprocal of the absolute temperature. The density-temperature curve is also normal, and between 20° and 101° is represented by the empirical equation d=1.1668-0-001358t0-0000005512. Phosphorus pentoxide forms a stable gel with nitromethane in presence of a trace of water. A. GEAKE.

Absolute critical temperatures. M. PRUDHOMME (Bull. Soc. chim., 1925, [iv], 37, 13301340). Relations between the absolute critical temperatures (T) of substances are pointed out. With simple substances which combine among themselves, the Te difference of suitably chosen pairs is often equal; thus for carbon dioxide and disulphide, carbon monoxide, and carbonyl sulphide T.(ČS2) — Te(CO2)=T(COS)-T(CO). For many non-metallic substances, the Te values are simple multiples or fractions of those for the components. For metallic compounds or carbon compounds containing only one atom of carbon, the metal or carbon atom does not affect the T, values, which are simple multiples

or fractions of the value for the non-metallic element.

These multiples are quite arbitrary, and may differ even for analogous substances as cyanogen chloride or bromide. If the linkings between a carbon and a hydrogen atom are considered as one valency, and single, double, and triple linkings between carbon atoms as 2, 4, and 6 valencies, respectively, the absolute critical temperature of aliphatic hydrocarbons is an exponential function of the number of valencies, namely, T.-cVn, where n=0·489 and log c=2-03911. With the sole exception of ethylene, this rule applies for both saturated and unsaturated hydrocarbons up to hexane. Similar relations exist for saturated alcohols, whilst for esters, ethers, and nitriles the Te values are additive properties of the different groups present, acetic acid and acetonitrile being exceptions.

W. HUME-ROTHERY.

Properties of simple compounds and types of the solid state. E. FRIEDERICH (Z. Physik, 1925, 34, 637).-Corrections of minor points in an earlier E. B. LUDLAM. paper (A., 1925, ii, 374).

Velocity of sound in mixtures of gases. H. B. DIXON and G. GREENWOOD (Proc. Roy. Soc., 1925, A, 109, 561-569).-The velocity of sound in a number of mixtures of simple gases has been measured, using Kundt's double-tube type of apparatus, the first tube containing dry air, the second the experimental gas or mixture. The method of measuring the dust figures was that used by Partington and Cant (A., 1922, ii, 191). The following were investigated: (1) mixtures with y practically constant

but with varying density (e.g., hydrogen and oxygen); (2) mixtures with both y and the density variable (e.g., hydrogen and carbon dioxide); (3) a mixture of carbon dioxide, hydrogen, and air; (4) mixtures with both y and density constant (e.g., carbon dioxide and nitrogen); (5) a single gas (ammonia) with a gas mixture of approximately the same density (e.g., equal volumes of hydrogen and oxygen). In the ammonia experiments, it was necessary to use silica ignited and sieved in place of lycopodium powder, as the latter is attacked by ammonia. In each case, the values found showed satisfactory agreement with those derived from Laplace's formula, y= (Cpa+C'pb)/(Ca+C'‚b), where a and b are the volumes of the gases in the mixture. L. L. BIRCUMSHAW.

Compression and decomposition of nitric oxide. E. BRINER, H. BIEDERMANN, and A. ROTHEN (Helv. Chim. Acta, 1925, 8, 923—928).— The compressibility of nitric oxide has been measured at pressures up to 160 atm. over the temperature range -80° to 10°, and pv values have been plotted against p. The isotherms indicate that nitric oxide is more compressible than the perfect gas, but that when comparison is made with the corresponding data for the non-polymerising gases, nitrogen, oxygen, ethylene, and carbon dioxide, there is no evidence for polymerisation of nitric oxide. The gas exhibits an anomalous temperature-coefficient in the peroxidation reaction which has been explained by polymerisation. Langmuir's octet theory also requires polymerisation of nitric oxide to account for the free valency forces. Both ideas are undermined by the proof that the gas exhibits a normal density.

Nitric oxide under prolonged compression decomposes, yielding a blue liquid which consists of a mixture of nitrous oxide and nitrous anhydride. Nitric oxide decomposes to give nitrous oxide and oxygen, the oxygen reacts with unchanged nitric oxide, yielding the peroxide, which in turn may react with nitric oxide to form the trioxide. The decomposition is reaction is retarded by corrosion of the glass vessels very rapid at 700 atm. At lower pressures, the by nitrogen peroxide. Freshly prepared and condensed nitric oxide is only very faintly blue in colour. Multiple liquefaction does not effect purification, since the blue colour gradually increases in intensity. Liquid nitric oxide is probably colourless and the blue tint is due to nitrous anhydride.

R. A. MORTON.

Assemblies of imperfect gases by the method of partition functions. R. H. FOWLER (Proc. Camb. Phil. Soc., 1925, 22, 861-885).-A mathematical paper giving a general development of the theory of imperfect gases by means of partition functions. The method allows all the usual results of the theory of mixtures of imperfect gases to be deduced simply. The effects of the inverse square law are discussed in detail and the validity of the combined use of the equations of Boltzmann and of Poisson has been examined with special reference to the theory of strong electrolytes advanced by Debye and Hückel (A., 1923, ii, 459, 724). It is concluded

that this theory is established for small ionic concentrations only. A. E. MITCHELL. Measurement of vapour tension of "gasoline "' and other liquids. Ĥ. S. DAVIS (Ind. Eng. Chem., 1925, 17, 1136-1138).-See B., 1925, 980.

the

Vapour-composition relationships in systems phenol-water and phenol-cresol. F. H. RHODES, J. H. WELLES, and G. W. MURRAY (Ind. Eng. Chem., 1925, 17, 1199-1201).-By distillation of the binary systems phenol-water, phenol-o-cresol, phenol-m-cresol, and phenol-p-cresol, and analysis of the original liquid and distillate in each case, the vapour-composition relationships of the systems were determined at various pressures. The curves for the system phenol-water show that azeotropic mixtures are formed with minimum b. p. The percentage of phenol in the constant-boiling mixture decreases with the pressure, being 9.4 at 760 mm. and 4.4 at 260 mm. The concentration of phenol in the distillate remains nearly constant so long as the concentration in the liquid is below 85%. With each of the three cresols, phenol forms ideal" liquid mixtures. The results obtained supply the information necessary for the calculation of the theoretical minimum reflux ratio and the minimum

66

number of sections required in the fractionating columns of stills for the separation of phenol from water and from cresol.

L. L. BIRCUMSHAW.

Liquid mixed crystals. R. WALTER (Ber., 1925, 58, [B], 2303-2310).-The formation of crystalline liquids from mixtures of substances which themselves do not exhibit crystalline-liquid properties (cf. Vorländer and Gahren, A., 1907, ii, 441) has generally caused the crystalline-liquid phases to be regarded as emulsions. The phenomenon is, however, more readily explained in the following manner. The compounds which yield such mixtures are generally very well crystallised (usually aromatic carboxylic acids) and are not readily supercooled. Consequently, a monotropic, crystalline-liquid phase which may be anticipated for them by reason of their chemical constitution remains latent even when the m. p. is not partly below that of the solid phase. If two chemically similar substances of this type are mixed, it may readily happen that the m. p. of the solid phase, which is generally lower than the m. p. of the individual components, falls below the mixed m. p. of the crystalline-liquid phase which with chemically related substances lies between the two latent crystalline-liquid m. p. of the pure components. Such a mixture exhibits enantiotropic crystallineliquid properties. The case is realised experimentally with a mixture of anisic and anisylidenepropionic acids. The m. p. of the latent crystalline-liquid phase of these compounds is determined from the behaviour of their mixtures with p-methoxycinnamic acid and use of the expression, t1=t2-[100(t2t): C1], where t1 is the latent m. p. of a substance, to the known m. p. of an admixed, chemically similar compound, tm the observed m. p. of the mixture, and c1 the molecular concentration of the substance under investigation in mol.%. The observed and calculated values are in good agreement. Many

other pairs of chemically similar compounds have been investigated. H. WREN. Viscosity of heated alloys. J. COURNOT and K. SASAGAWA (Compt. rend., 1925, 181, 661–662).— See B., 1925, 994.

Correlation of the remanent magnetism and specific resistance of some pure iron-carbon alloys. E. D. CAMPBELL (Trans. Faraday Soc., Nov. 1925, advance proof).—The specific resistance and magnetic properties of pure electrolytic iron and a series of iron-carbon alloys containing up to 1.5% of carbon and total elements other than carbon and iron 0.008% have been measured. The magnetic properties were measured using the material as bars with open poles and utilising a special magnetic balance which is described, in which electromagnetic forces were balanced against gravity. For electrolytic iron at 25° slowly annealed in a vacuum the specific resistance is 10-02 micro-ohms/c.c. The increase in specific resistance is almost proportional to the carbide concentration, whilst the magnetic resistance (gilberts required to give 1000 dynes/cm.2) of a bar, magnetised under standard conditions, increases more slowly than the specific resistance up to the eutectoid composition and then more rapidly, indicating that low magnetic resistance is connected with structurally free ferrite. The internal magnetomotive potential (gilberts required to double the dynes) increases to a constant value at 0.6% carbon, whilst the magneto-motive potential (gilberts required to demagnetise) increases continuously, showing that the internal magneto-motive potential of the solvent iron is small compared with that due to dissolved carbides. The force fields of solutes in metallic solution are regarded as transforming electricity into heat, thereby increasing the resistance, whilst the converse transformation is indicated by the thermo-electric potential at solution-solvent junctions. F. G. SOPER.

Absorption of gaseous hydrogen chloride by sulphuric acid. V. CUPR (Publ. Fac. Sci. Univ. Masaryk, 1925, 63, 1-18; cf. A., 1925, ii, 655).Further measurements of the absorption of gaseous hydrogen chloride by sulphuric acid solutions of concentrations corresponding approximately with the hydrates H2SO4,2H2O and H2SO4,4H2O at the temperatures 0, -15.8°, and -25° show that the quantity absorbed diminishes with increasing concentration of sulphuric acid, as was found by Coppadaro (A., 1910, ii, 197) at 17°, 40°, and 70°.

N. H. HARTSHORNE.

Solubility of hydrogen selenide. A. J. MCAMIS and W. A. FELSING (J. Amer. Chem. Soc., 1925, 47, 2633-2637).—The solubility of hydrogen selenide in water is 0-09611, 0-08415, and 0-07317 mol. per litre at 15°, 25°, and 35°, respectively, and 760 mm. In aqueous hydrogen iodide the solubility is greater, and is 0.11012 mol. per litre in the 2.73M acid at 25° and 760 mm. The molal heat of solution in water is calculated to be 2431 cal. at 25°, and the free energy decrease for the reaction H,Se(1M)= H2Se (760 mm.) is 1463 cal. for an aqueous solution containing 0-08017 mol. per litre at 25°. A. GEAKE.

Supersaturation of gases in liquids. N. P. CLARE (Trans. Roy. Soc. Canada, 1925, [iii], 19, III, 32-33; cf. Wyatt, A., 1925, ii, 504).-Aqueous solutions of oxygen up to about 250 atm. saturation pressure are being investigated. On the reduction of pressure of such solutions, a cloud of fine bubbles sometimes appears, suggesting the spontaneous formation of bubbles in the liquid away from the walls. J. S. CARTER.

Initiation of bubbles in supersaturated solutions of gases. F. J. FARNCOMB (Trans. Roy. Soc. Canada, 1925, [iii], 19, III, 32).-Colloidal arsenious sulphide, colloidal ferric hydroxide, starch, platinum wire, lead chloride crystals, phenol globules, and freshly precipitated silver iodide have no apparent effect on the formation of bubbles in highly supersaturated solutions of oxygen. Paraffin-wax globules, benzoic acid, salicylic acid, phthalide, and iron wire have a small but definite effect. Precipitated silver iodide (24 hrs.' settling), precipitated calcium sulphate, and boric acid have a pronounced effect. It appears that bubbles are initiated by substances which are wetted with difficulty. J. S. CARTER.

Relation between solubility and retention of water of crystallisation in the salts of aromatic sulphonic acids. F. EPHRAIM and E. SEGER (Helv. Chim. Acta, 1925, 8, 724-739).-Benzenesulphonic acid and naphthalene-1- and -2-sulphonic acids yield a large number of isomorphous, hexahydrated salts of bivalent metals. The loss in weight sustained by each salt in specified times and at definite temperatures in the presence of phosphorus pentoxide has been determined. The solubility at temperatures from 17° to 82° has been determined for each salt of benzenesulphonic acid and naphthalene-2-sulphonic acid, and at 17° only for each salt of naphthalene1-sulphonic acid. For the first two acids, the solubilities of the salts (expressed by normality) decrease in the order: copper, cadmium, zinc, manganese, nickel, magnesium, cobalt, whilst for the third acid the order is magnesium, cobalt, nickel, copper, cadmium, zinc, manganese. The parallel experiments indicate that the strength of the linkings uniting the salt and its water of crystallisation decreases in the order Co, Ni, Mg, Zn, Mn, Cu, Cd, for the salts of benzenesulphonic acid, and in the order Ni, Co, Mg, Zn, Mn, Cu, Cd for the salts of naphthalene2-sulphonic acid. Thus in these series it is not the most easily, but the least soluble salts which retain their water of crystallisation most strongly. In the case of the salts of naphthalene-1-sulphonic acid, the water is retained in the order Ni, Co, Mg, Zn, Mn, Cd, Cu. The parallelism therefore breaks down to some extent here, and even when it holds, the connexion between solubility and retention of water is in the sense opposed to that expected. R. A. MORTON.

Vanadium oxytrichloride as a solvent. F. E. BROWN and J. E. SNYDER (J. Amer. Chem. Soc., 1925, 47, 2671-2675).-The solvent power of vanadium oxytrichloride has been examined. In general the metals are unaffected and inorganic compounds are insoluble, but most of the non-metals

are soluble. All the organic substances tested dissolve in or react with the solvent. Organic liquids are miscible in all proportions. The liquid aldehydes react vigorously. The solubility of sulphur increases from 3.3 g. per 100 g. of vanadium oxytrichloride at 0° to 30.7 g. at 65°. A. GEAKE.

Influence of very slight traces of water on solubility equilibrium. I. E. COHEN and W. D. J. VAN DOBBENBURGH (Z. physikal. Chem., 1925, 118, 37-48).-Determinations of the solubilities of different specimens of salicylic acid in benzene at 30-5° gave values which varied between the limits 1.03-1.36 g. per 100 g. of solution. It is shown that these irregularities are due to the presence of minute quantities of water occluded in the crystals, the solubility of a carefully dehydrated sample of acid being 100 g. per 100 g. of solution at this temperature. Data are recorded showing the influence of varying small quantities of water (0.004-0·3%) on the solubility of anhydrous salicylic acid in anhydrous benzene at 30.5°. J. S. CARTER.

Solubility influences. I. Effect of some salts, sugars, and temperature on the solubility of ethyl acetate in water. S. GLASSTONE and A. POUND (J.C.S., 1925, 127, 2660-2667; cf. Thorne, ibid., 1921, 119, 262). The solubility of ethyl acetate has been determined at 25° and 50° in aqueous solutions of varying concentrations of the chlorides, bromides, and iodides of the alkali metals and ammonium, and of dextrose, lævulose, sucrose, and lactose, and the values are used to calculate the hydration of these substances (cf. Philip, A., 1907, ii, 935). The lithium salts and nearly all the iodides give uncertain hydration values, apparently due to the solubility of these salts in the ethyl acetate itself. Since the non-electrolytes reduce the solubility considerably, it appears that molecules as well as ions have considerable salting-out power. No single equation will fit the results over the whole range of concentration. The solubility of ethyl acetate in water at 0°, 10°, 25°, 37°, and 50° has been measured.

R. CUTHILL.

Relationship of salts in dilute aqueous solution as determined by their influence on the critical solution temperature of the system phenolwater. J. H. CARRINGTON, L. R. HICKSON, and W. H. PATTERSON (J.C.S., 1925, 127, 2544-2549).— A continuation of the work of Duckett and Patterson (A., 1925, ii, 381). The occurrence of hydrolysis, complex formation, or chemical action causes abnormal results. Otherwise it is found that e=aC, where e is the elevation of the critical solution temperature caused by C g.-mol. of salt per 1000 g. of total mixture, b is a constant characteristic of the phenol-water system, and a is a specific constant for each salt. From the resemblance between this equation and the Freundlich adsorption isotherm, and between the Hofmeister series and the series representing the order of effectiveness of the anions in raising the critical solution temperature, it seems likely that the lastnamed phenomenon is associated with adsorption.

R. CUTHILL.