INORGANIC CHEMISTRY. Inorganic Chemistry. ii. 39 The Behaviour of Water as a Liquid and in Five Solid Modifications Under Pressure. P. W. BRIDGMAN (Zeitsch. anorg. Chem 1912 77 377-455).-The behaviour of water has been examined under higher pressures than in previous experiments liquid water being studied between - 20' and + 80° up to 12,800 kilograms per sq. cm. and solid water between -80' and 97s' up to 20,500 kilograms per sq. cm. Two new modifications of ice one of which is capable of existing above O' have been discovered. Five of the six stable triple points have been found and ten of the eleven trans- formation lines have been followed. In addition to the five stable solid modifications two exist which have no range of stability. All with the exception of ordinary ice are denser than water.The temperature of maximum density of water is independent of the pressure but the maximum disappears at high pressures ; a new anomaly appears however between - 1 5 O and - 20° a t a pressure of 5000 kilo- grams. The increased compressibility a t high pressures suggests a possible compressibility of the atoms as well as of the molecules but the data are still insufficient to construct an equation for the behaviour of liquid water. The results with the solid phases confirm and extend those of Tammann The oquilibrium curve between ice I1 and 111 which was not found by Tammann has now been determined. The water isii. 40 ABSTRACTS OF CHEMICAL PAPERS. enclosed in a vessel filled with light petroleum and the pressure is measured by the change in resistance of a manganin wire.I n order to determine the curve 11-111 along which the volume changes are very small a different mothod has t o be adopted the temperature being varied a t constant volume and the changes of pressure ob- served. Ice VI which is formed at temperatures above 0' under high pressure has the greatest range of existence as regards both tempera- ture and pressure and the form of the curve is such as to indicate that further modifications are not formed. The existence of two modifications having no range of stability is inferred from the occurrence of volume changes which do not lie on any of the equilibrium curves. The observations confirm the impossibility of superheating a solid phase with respect to the liquid whilst undercooling readily takes place and many of the curves have been prolonged into the region of instability but an unstable triple point has not yet been realised and indications of a metastable limit have not betn found.Nuclei of a previously existing phase appear to persist even in a solid phase. The velocity of tiansformation even in solid phases is extrerriely high in the neighbourhood of a triple point but diminibhes rapidly with falling temperature. Thebearing of these results on the theory of the solid state is discussed. A New Modification of Sulphur. A. H. W. ATEN (PYOC. K. Wetensch. Amstsrdum 1912 15 572-583 ; Zeitsch. physikul. Chem. 1912 81 257-280).-1f a solution of rhombic sulphur in sulphur chloride saturated a t the ordinary temperature is heated a t about 150' and then cooled it is found that a further and considerable quantity of sulphur may be dissolved in the liquid.It is shown that this cannot be due to the formation of 8 from S' for the solubility of 8 in sulphur chloride is very small at the ordinary temperature and if solutions saturated with S a t a higher temperature are cooled down this separates out quite readily. I n order to ascertain whether this phenomenon is due to the forma- tion of a new modification of sulphur or to the combination of sulphur with the sulphur chloride systematic determinations have been made of the solubility of rhombic sulphur in sulphur chloride which has been heated with varying proportions of sulphur a t a definite tem- perature and then cooled to either 25" O' or - 60'.The data thus obtained point to the formation of a new variety of sulphur. If this conclusion is correct i t should be possible t o obtain some indication of its formation in the absence of sulphur chloride. This is actually the case for if sulphur is heated alone to 125' and rapidly cooled and its solubility in sulphur chloride examined it is found that the solubility is distinctly greater than that of unheated rhombic sulphur. C. H. D. H. M. D. Tho Relation between the Sulphur Modifications. H. I;. DE LEEUW (Yroc. IT. Akad. Wetensch. Amslerdcm 191 2 15 584-593). -If the sulphur contained in a dilatometer tube is heated t.0 its boiling point and then rapidly cooled so that a large proportion of theINORCANIC CHEMISTRY. ii. 41 S formed remains intact it is found that the temperature of conver- sion of rhombic into monoclinic sulphur is lowered from 95-45' to about 71".As the proportion of S present decreases the transition temperature rises until it reaches 95*45O which must be regarded as the true unary transition point. Kruyt's work (A. 1911 ii 879 ; 191 2 ii 1051) on the dynamic allotropy of sulphur is criticised adversely and in particular it is shown that there is no evidence for the assump- tion of a region of partial miscibility in tbe liquid phase. The occurrence of two liquid layers is due to differences in tempera- ture and when the thermal conductivity is increased by the introduction of platinum wire or gauze the phenomenon in question is only observed very indistinctly or not at all. Alleged Complexity of Tellurium.WILLIAM C. MORGAN (J. Amw. Chenh. SOC. 1912 34,1669-1675).-Flint (A. 1910 ii S45 ; 1912 ii 1051) has stated that by fractional hydrolysis of tellurium tetrachloride he obtained two fractions one giving the atomic weight 124.3 and the other which was still impure the atomic weight 128.85. This work has now been repeated with a large quantity of material biit the results do not show progressive diminution of the atomic weight and thus fail to confirm Flint's conclusion. I n view of these experiments and those of other investigators the author considers that the evidence for the homogeneity of tellurium is quite convincing. [Compounds of Tellurium and Iodine.] FRANS M. JAEGER and J. B. MENKE (Zeitsch. anorg. Chem. 1912 77 320. Compare A. 1912 ii 344)-It is not possible to determine the freezing point of mixtures containing more than 70% of tellurium in an open vessel and mixtures rich in iodine boil.The compound TeI dissociates when heated in carbon dioxide. The Ternary System Tellurium Tetra-iodide-Hydrogen Iodide-Iodine and the Estimation of Tellurium by means of Hydrazine Hydrate. J. €3. MENKE (Zeitsch. cmorg. Chem. 1912 77 282-28s. Compare Jaeger and Menke A. 1912 ii 344).- Tellurium tetra-iodide prepared by melting together its components is heated wit,h hydriodic acid and iodine in a sealed glass tube ten days being allowed for the attainment of equilibrium in each experi- ment. The only solid phases found are TeI4,HI,8H,O and iodine. A hexa-iodide is thus not obtained either by fusion or from solution.'l'he double compound resembles iodine in appearance. For analysis the free iodine is titrated with thiosulphato after neutralisation with sodium hydrogen carbonate ; the total iodine is estimated as silver iodide after removal of most of the tellurium by alkali sulphite. The method of Gutbier and Flury (A 1902 ii 653) is not applicable to tellurium tetra-iodide and the reduction of tellurium by means of sulphur dioxide is not exact as the fine amorphous tellurium readily oxidises. Precipitation by means of hydrazine hydrate is usually incomplete as the filtrate contains colloidal tellurium. In acid solution the precipitation is almost complete but the excess of acid must be noutralised before filtration otherwise the precipitate partly redissolves. H. M. D.E. G. C. H. D. C. H. D.ii. 42 ABSTRACTS OF CHEMICAL PAPERS. The Validity of Werner's Theory of Subsidiary Valencies for Ammoniates. WALTER PETERS (Zeitsch. anorg. Chern. 1912 77 137-190).-The addition of ammonia to inorganic salts has been studied by means of the apparatus of Ley and Wiegaer (A. 1905 i 749). After saturating the solid with ammonia it is weighed trans- ferred to a vacuum until the weight is constant and again saturated. Addition takes place more readily the second time and in some cases a further quantity of ammonia is added. The number of molecules added depends on the degree of saturation of the salt molecule and on the residual affinity of its ions and when salts with the same anion but different cations are compared analo- gies corresponding with the periodic system are observed.The number of molecules of ammonia is most frequently R ~ X or a multiple of six which is in accordance with Werner's hypothesis. The following new compounds have been prepared CuC1,3NH3 ; CuBr,3NH3 ; CuI,3NH3 ; CuCN,$NH ; Cu(NO3),,5NH ; AgCl,SNH ; AgBr,Z$NH ; AgI,+NH ; AgCNO,NH,; AgCNS,ZNH,; AgNO3,3NH,; KAuU1,,3NH3; IT Au(CNS)~,~NH ; ZnS04,H20,6NH,; K2Hg14,2NH3 ; BaHg14,8NH,; Cu,HgI,,GNH,; CdHgl4,6NH,; CrCl,,GNH,; Hg(CBs),,4NH3 ; Hg(C,H302)2,4NH ; Hg(C,H,02)p4NH3 ; UC1,,3NH3 ; UG2C1,,2NH3 ; Mn6I296NH3 ; Mn12,6NH ; MnSO,,fiNH ; iSlnS0,,H20,5NH ; FeS04,H,0,5NH ; Zn2Fe(CN),,7NH3 ; Mn2Fe(CN),,2NH3 ; Cd,Fe(CN),,7NH3 ; Co,Fe(CN),,3NH3 ; Ni2Fe(CN),,7NH3 ; Mn,[Fe(CN),],,GNH ; CLI,[F~(CN)~],,~ 6NH ; Cd,[Fe(CN),]2,1SH20 ; Co,LPe(CN),],,SNH ; Na,FeNO(CN),,ZNH ; Ni(SCN),,GNH ; NiS,O6,6NH3 ; Co(CN),,2NH3 ; Co(SCN),,6NH3 ; RuC1,3NH3 ; RuBr,,3NH3 ; RhC1,,4NH3 ; PdCj2,5NH ; PdI,,6NH3 ; Na2PdCI,,5NH3 ; ( NH4),PdC1,,5NH ; PtC1,,5NH3; Pt12,6NH ; Na2PtC1,,4NH ; Ag,PtCl,,GNH,(7~NH3 at low temperatures) ; K,Pt(CK)),,NH ; KNaPt(CN),,2N& ; CuPt(CN )".4NH0 Zn3[Fe('N),]2,12NH3 ; Ag,Pt(CN),,5NH3 ; MgPt(CN)i,4N'$ ; E;Pt(CN),,GNH ; ZnPt(CN),,4NH ; CdPt(CN),,CiNH ; MnPt(CN)4,4N€I ; FePt (CN),.4NH0 NiPt(CN),,GNH ; CoPt(CN),,6NH;'; Cal%(CN),,3NH3 ; K2Pt( S(?N),,2NH ; CuPt(SCN),,8XH3 ; Ag2Pt(SCN),,6NH ; PtCl,,GNH ; PtBr4,6NH ; Na2PtC1,,6NH3 ; Ag2PtU1,,8NH (lONH a t - 10') ; CuPtC16,18NH3 ; CaPtC1,,12NH3 ; BaPtCI,,GNH ; ZnPtCI6,11NH ; CdPtCl 182JM ; RtnPtC1,,1 lNH,; NiPtC1,,12NH3 ; CoPtC16,12NH,( 18NH a t - 20' ; Na,PtBr6,6NH ; K,Pt(SCN),,lBNH ; (NH,),Pt(SCN),,15NH3 ; CuPt(SCN),,15NH3 ; Ag2Pt(SCN),,14NH ; hu2[Pt(S(11N),],,30NH3 ; BaPt(SCN)6,1 2NH ; ZnPt(SCN)6,15NH ; CdPt(SCN),,18NH3 ; TI,Pt(SCN),,gNH ;INORGANIC CHEMISTRY.ii. 43 MnPt(SCN),,lSNH ; NiPt(SCN),,lSNH ; CoPL(SCN),,14NH3 ; K,Pt( SeCN),,GNH,. Hydrogen chloride and hydrogen phosphide are not absorbed by any inorganic salts under similar conditions. The absorption spectra of aqueous solutions of the compounds CuCI,,GNH and CuPtC1,,18NH3 show that both the long visible waves and the ultra-violet are absorbed the transmitted band being narrower in the case of the double salt. Determinations of the vapour pressure of several of the compounds have been made by means of a tensimeter. C. H. D. Revision of the Atomic Weight of Phosphorus.111. Analysis of Phosphorus Trichloride. GREGORY P. BAXTER and CHARLES J. MOORE (J. Arner. Chem. Soc. 1912 34 1644-1657).-1n earlier papers (Baxter and Joned A. 1910 ii 288; Baxter Moore and Boylston A 1912 ii 347) accounts have been given of determinations of the atomic weight of phosphorus by the analysis of silver phosphate and phosphorus tribromide. The investigation has now been extended to the analysis of phosphorus trichloride. Phosphorus trichloride was prepared and analysed by methods similar to those employed in the case of the tribromide. The product was freed from the pentachloride by distillation in a vacuum. Two series of fractions were collected and analysed in the usual way after decomposition with water and oxidation of the phosphorous acid produced.From the average of twenty-four analyses the value for the atomic weight of phosphorus was found to be 31.018 (Ag= 107*880 ; C1= 35.457) as compared with 31.04 obtained from the analysis of silver phosphate and 31.027 from that of phosphorus tribromide. The averdge of the results from the tribromide and trichloride is 31.023. E. G. The Density of Phosphorus Vapour. ALFRED STOCK GEORGE E. GIBSON and ERICH STAMM (Ber. 1912 45 3527-3539).-A membrane-manometer which is a special modification of the spiral manometer of Ladenburg and Lehmann (Bsr. Deut. physikal. Ges. 1906 4 20) and is to be described in another paper was used to determine the vapour density of phosphorus; the manometer is sensitive to 0.5 mm. difference in pressure.Purified red phosphorus which can be obtained in a purer condition than colourless phosphorus was used in the experiments at the t'emperature of which i t was completely transformed in to vapour. At temperatures between 500" and 700" the vapour density corre- sponds exactly with the forrnult P,. At higher temperatures dissociation takes place but even at 1200" it is only 61% under a pressure of 175 mm. Calculations made on the assumption that the dissociation which occurs is in accordance with the equation P 2P give a very good agreement between theory arid experi- ment. Under atmospheric pressure the degrees of dissociation at 800" 1000° and 1200" are respectively 0.01 0.1 and 0133 ; under aii. 44 ABSTRACTS OF CHEMICAL PAPERS pressure of 0.25 atmos.the degree of dissociation is almost 0.66 at 1200O. Calculation of the heat of dissociation gives values varying from -49,200 to -60,300 cals. The above results do not agree with previous measurements but various measurements were so concordant that there is every probability of their accuracy. The experiments of Preuner and Brockmolier (A 1912 ii 1146) are probably untrustworthy because they used ordinary commercial phosphorus. T. S. P. Boron Hydrides. ALFRED STOCK and CARL MASSENEZ (Be?.. 1912 45 3539-3568).-The boron hydrides obtained by the authors have been prepared by the action of hydrochloric acid on magnesium boride. Full details are given of the conditions needful for the preparation of a satisfactory sample of magnesium boride from magnesium and boron trioxide.The decomposition of the boride with acid was carried out in an apparatus similar to that used in the preparation of antimony hydride (A. 1904 ii 246) the boride being dropped continuously into t h e acid whilst a slow stream of hydrogen is passed through the apparatus. The decomposition is carried out at 50° using 4N-hydro- chloric acid. The evolved gases are passed through U-tubes im- mersed in liquid air a snow-like deposit forming whilst hydrogen passes on. The first portions of gas escaping from the U-tube smell strongly of boron hydride but as soon as the solid forms i t prevents the escape of any more hydride. The deposit consists of boron hydrides together with silicon hydrides and carbon dioxide from impurities in the materials used It is fractionally distilled at varying pressures and temperatures in order to remove the impurities.Hydrogen and traces of silicon hydride are removed by evacuating at the temperature of liquid air. The temperature is then raised to -SOo when the solid Melts to a colourless strong refracting liquid having a pressure of 250 mm. The pressure can be rnpidly diminished t o a few millimetres whereby silicon hydridw carbon dioxide and any traces of hydrogen phosphide or sulphide which may be present are removed. As soon as the gas is free from silicon hydride it is collected separately a t a pressure of less than 3 mm. the temperature of the room in which the distillation is being carried out being maintained above 20° otherwise liquid hydride would be coudensed in the mercury pump used.The gas collected a t this pressure and temperature consists of the boron hydride B,H,,. The residue is then warmed to -40' and dietilled until the pressure falls to 14 mm. the temperature further raised to Oo and distillation proceeded with until the pressure falls to 9-10 mm. in order to remove all the hydride B4Hlo. The remaining iiquici on Turther distillation gives the boron hyiricle B,Hl the pressure falling below 5 mm. Two hundred grams of magnesium boride give about; 100 C.C. of the pure hydride B4H10 and 60 mg. of the hydride B,H,,. Owing to the fact that these hjdrides possess compositians veryINORGANIC CHEMISTRY. ii. 45 different from what was to be expected t h e j were i~nalysed by a numbor of different methods. The hydride B4HIo was analysed as follows (1) a known volume was decomposed by passing slowly through a weighed quartz tube at 500' and the hydrogen collected. The boron was weighed directly and also oxidised to boric acid and estimated volumetrically; (2) atl room temperature the gas is slowly decomposed by water with liberation of hydrogen and formation of boric acid.The results are in accordance with the equation B,HIo + 12H,O 4B(OH) + 1 IH,. (3) Decomposition also takes place in the presence of sodium hydroxide the results according with the equation B,Hlo + 4NaOH + 4H,O = 4NaB0 + 1 lH,. The vapour density agrees with the formula B,H1,. The boron bydride B,H is a colourless liquid b. p. 16-17'/ 760 mm. m. p. ca - 1 1 2 O ot a peculiar and most disagreeable odouv. A few bubbles of the gas affect respiration and cause headache.It is a very unstable substance decomposing at the ordinary tem- perature after a few hours and quicker a t higher temperatures giving rise t o a whole series of new hydrides which await investi- gation. It takes fire spontmeously in the air or in oxygen burning with a green flame. Water and dilute hydrochloric acid decompoEe it and it is oxidised by concentrated nitric acid with explosive violence. It is rapidly absorbed by aqueous sodium hydroxide the solution slowly evolving hydrogen in accordance with the equation already given. It is probable that an intermediate hypoborate corresponding with the oxide B,O (compare Travers and Ray A. 1912 ii 938) is formed since when brought into contact with solid potassium hydroxide the latter becomes coated with a thick layer of a whito salt in the form of a loose powder.Ammonia gives a solid light brown substance insoluble in water whilst alcohol decomposes the hydride with evolution of hydrogen. The solution in benzene is very stable towards oxygen. The hydride B6H12 was analysed similarly t o the hydride B4HIo by decomposing it by heat. The vapour density corresponds with the forruula B,H,,. It is a colourless liquid b. p. 10'115 mm. ca 100°/760 mm. possessiug a highly disagreeable odour and taking fire spontaneously in the air. It is more sensitive towards water and moisture than the hydride B,HIo. With aqueous alkalis hydrogen is immediately evolved. The composition of the above hydrides is not in agreement with that of the boron alkyls nor with the results obtained by Jones and Taylor (T.1881 39 213) Sabatier (A. 1891 979) and Ramsay and Xatfield (P. 1901 1'7 152). A critical examination of the experi- ments of these various investigators shows that no trustworthy conclusion can be drawn from them as t o the composition of boron It is similarly decomposed by electric sparks. hy dride. T. s. P. Condition Diagram of Carbon Dioxide. GUSTAV TAMMANN (Chem. Zentr. 1912 ii 598; from Nuchi*. K. Ges. Wiss. Gatt. 1912 446-452).-The author had observed previously that liquid carbonii. 46 ABSTRACTS OF CHEMICAL PAPERS. dioxide a t temperatures ranging from - 56.7" to - 7.5" and pressures between 5.1 atm. and 2800 kg. per sq. cm. deposited n different crystalline species from that obtained at higher pressures.Repetition of the experiments up to 4000 kg. pressure showed that a triple point does not occur in the melting curve (compare A. 1912 ii 29). T. A. H Carbon Subsulphide C,S,. ALFRED STOCK and PAUL PRAETORIUS (Bey. 1912 45 3568-3578).-1n the preparation of carbon telluride (A. 191 1 ii 722) a substance possessing a very penetrating odour was always an accompanying product. This has proved itself to be the tricarbon disulphide C,S first obtained by Lengyel (A 1894 ii 90) which the authors propose to call carbon subsulphide from analogy to carbon suboxide. Lengyel's method of preparation (Zoc. cit.) is not a convenient one and the authors have sought for a better. The compound is produced when an arc is formed between carbon electrodes under liquid carbon disulphide but i t is accompanied by many by-products.It is also produced when carbon disulphide is passed through a heated quartz tube the most favourable temperature being 1000-~lOOo ; carbon is deposited in the tube at the same time and diminishes the yield of subsulphide. If the quartz tube is filled with a spiral of iron wire the formation of subsulphide takes place below 800° but i t soon ceases owing to the formation of iron sulphide. The action of metals led the authors to investigate the formation of subsulphide by the disintegration OF metals by means of an arc under liquid carbon disulphide the disintegration of the anode exposing fresh hot surfaces continually to the carbon disulphide. Iron copper and aluminium electrodes gave no eubsulphide whilst electrodes of arsenic bismuth lead tin antimony cadmium or zinc yielded the required compound large quantities of the metallic sulphides being formed a t the same time. The best results were obtained with a carbon cathode and an anode consisting of antimony mixed with 7% of carbon.The brownish- red solution obtained is filtered and then shaken with mercury and phosphoric oxide to remove sulphur and polysulphides and water. A solution of carbon subsulphide in carbon disulphide is thus obtained from which by distillation finally under a high vacuum the pure substance is obtained in a receiver cooled to - 40" as a yellowish-red solid. At the ordinary temperature i t forms a bright red strongly refracting liquid the vapour density of which corresponds with the formula C,S,; the m. p.is -0-5' and the substance possesses normal molecular weight in carbon disulphide solution. The alcoholic solution decomposes after a short time. Dilute solutions in carbon disulphide are salmon-coloured stronger solutions yellowish-brown ; above 1% strength the solutions gradually deposit black polymerisation products even in the dark ; solutions below 1% strength are stable in the dark but not in sunlight. Carbon disul- phide solutions have no action on copper or mercury but are decom- posed by the sulphides of these metals owing probably to the formation of additive products. The polymerisation at 90' is a reaction of the second order andINOKGANIC CHEMISTRY. ii 47 can be readily measured by observing the diminution in pressure ; the reaction takes place at the surface of the containing vessel.Carbon subsulphide can be considered as the anhydride of thio- malonic acid (compare carbon suboxide) since with aniline it readily gives thiomalonanilide. Its constitution may therefore be expressed as S:C:C:C:S. The black polymerisation product is not affected by water sodium hydroxide hydrochloric acid or chlorine water. Heated in a vacuum it gives carbon disulphide and a black residue which still contains 39% of sulphur. H. VON WARTENBERG (Zeitsclh. ccno~g. Chem. 191 2 79 71-87).-Platinum is often attacked when heated in a silica vessel in a reducing atmosphere even if not in contact with the silica indicating the formation of either silicon vapour or silicon hydride. The disociation of silicon hydride (prepared by the action of alcohol on purified silicon-chloroform followed by decomposition of the ester by sodium) has been studied.It is necessary to use a catalyst and the nickel used in Sabatier's method of reduction has proved to be suitable. Without a catalyst decomposition is hardly measurable below 350° even after several days. Measurements up to 380' give the result SiH,=Si+2H2 - 8700 cal. From 100' to 150' the pressures observed are higher than would be expected probably owing to the occurrence at the lower temperature of the reaction BSiH = Si,?? + 2H,. The volatilisation of crystallised silicon in a stream of hydrogen has been measured up t o 1315". Silicon hydride is not to be detected in the issuing gas. The experiment has been repeated with argon io place of hydrogen with similar results the diminished loss of silicon being accounted for by the formation of a thin protecting layer of the nitride Si,N as nitrogen is not completely removed from argon even by passing over heated likhium.The value SiH,= Si (vapour) + ZH - 44,000 cal. is calculated. The reduction of silica by hydrogen has been studied at 1350-1400" a t which temperature silica is not volatile. The observed sublimation at lower temperatures is therefore due to reduction followed by oxidation. Carbon monoxide and graphite are also capable of reducing silica. The Velocity-of Crystallisation and Crystallieing Power of Silicates. ERWIN KITTL (Zeitech. ccnorg. Chem. 1912 77 335-360). -The crystal!isation of a number of silicates has been examined in relation to the viscosity.Crystallisation takes place on the heating stage of a microscope and the growth of the crystals which usually form radiating groups of needles is measured by means of a micro- meter eyepiece. The linear velocity of cry stallisation varies with the undercooling in the manner described by Tammann for organic substances. The horizontal maximum of the curve is often very short or diminished to a point and is best marked in silicates of low viscosity. Olivine bronzite and hypersthene show sharp maxima whilst polysilicstes show a low flat curve. The maximum is usually 20-30' below the initial temperature of crystallisation. Metasilicates T. S. P. The Reduction of Silica. C. H. D.ii. 48 ABSTRACTS OF CHEMICAL PAPERS. form the longest needles the larger number of centres in orthosilicates causing in terforence.Orthosilicates of the alkalis form glasses without crystallising with the exception of lithium orthosilicate which cryst.allises if cooled slowly. Simple and mixed magnesium calcium ferrous and manganous silicates crystallise readily. Alkali metasilicatea with the exception of litbium form only glasses. The simple metasilicates of the metals mentioned above crystnllise readily whilst the mixed ruetasilicRtes remain partly glassy. Polydicates e~pecially those of complex constitution tend to form glasses. C. H. D. The Proportion of Potassium Chloride in Potassium Chlorate and the Nephelouetric Control. FRIEDRICH MEYER and ARTHUE STAIILER (Zeitsch. ano7.g. Chem. 1912 77 255-256.Compare A. 1911 ii S d l ; Guye A. 1912 ii 552).-The criticism of Guyeoverlooks the fact that the solution in the nephelorneter is not one of silver chloride in pure water but contains an excess of either silver or chloride considerably diminishing the solubility. I n the case of potassium chlorate the error cannot affect more than the sixth decimal place in the atomic weight of chlorine or potassium. C. H. D. Molten Carnallite. KURT ARNDT and HEINRICH Huao KUNZE (Zeitsch. Elektrochem. 1912 18 994-99S).-Determinations of the density electrical conductivity and freezing point of carnallite and mixtures of carnallite and potassium chloride are given. The carnallite used was prepared artificially by mixing magnesium chloride and potassium chloride in the requisite proportions.In all experiments decomposition of the magnesium chloride was prevented by the constant and careful addition of small quantities of ammonium chloride. The freezing points were deduced from cooling curves obtained by the use of a thermocouple and are as follows pure carnallite 496'; carnallite + 6% KC1 493' ; carnallite + 12% KC1 484' ; carnullite + lS% KCl 4'70' ; carnallite + 24% KCL 460' and carnallite + 30% KCl 436'. The density determinations were made by weighing a platinum sinker in the molten mixture (compare Brunner A 1904 ii 244). Tables of results are given which show that at 550° 650' and 750° the addition of potassium chloride up to 30% causes a slight decrease in the density of the mixture. Pure carnallite has the density a t 570° 1.711 ; 600° 1.698 ; 650° 1.678 ; TOO' 1.658; 7509 1.638 and 780' 1.625.It is shown from the density determinations that carnallite and potassium chloride form no definite compounds. Tho conductivity measurements were made by the method described by Arndt (A. 1906 ii 418) and the following specific conductivities obtained %-Excess of potassium chloride. - + Temp. 'i0.3-20.5. 10.4-10'7. 0*8-1'0. 550" 600 650 700 750 0 '96 0.92 0 *87 1 '07 1.03 0.9s 1'18 1 *14 1-09 1 '29 1-25 1'20 1'40 1-36 1.31INORGANIC CHEMISTRY. ii. 49 By extrapolation from the above figures the specific conductivity of pure carnallite was deduced. The values are at 570° 0.92; 600° 0.97 ; 650° 1.08 ; 700° 1-20 and 780° 1-37. J. F. S. Thermal Analysis of Binary Mixtures of Alkali Nitrites with Other Salts of the Same Metals.D. MENEGHINI (Gazxettcc 1912 42 ii 47%-479).-'I!he paper deals with the thermal analysis of the syatems KN0,-KNO NaN0,-NaCI and NaN0,-NaBr. Fused mixtures of potassium nitrite and nitrate in the interval from 10% to 100% of nitrate give an uninterrupted series of mixed crystals whilst at the ordinary temperature between the same limits of concen- tration there is no miscibility. I t is imposaible t o obtain nitrite free from nitrate. The experiments with sodium nitrite were carried out with a nitrite containing 15% of nitrate The diagrams of the systems NaN02-NaC1 and NaN0,-NaBr belong to Roomboom's type 111. R. V. S. The Action of Ozone on Alkali Hydroxides. WILHELM TRAUBE (Ber. 1912,45 3319-3320. Compare A. 1912 ii 844).- The previous work of Manchot and Kampschulte (A 1908 ii 101) on the same subject is acknowledged.T. S . P. The Electrolytic Preparation of Sodium Hypoohlorite. PAUL H. PRAUSNITZ (Zeitsch. Xkltrochem. 19 12 18 1025 -1 080).- After careful consideration of the theory of the electrolytic production of sodium hypochlorite from sodium chloride the paper deals first with an experimental examination under laboratory conditions with enclosed cells of the various factors which influence the reaction. I n the second part the results are applied to the technical conditions the author using in this connexion working models of the best known technical electrolysers. The experiments were all made except where the contrary is stated with 5N-solutions of sodium chloride t o which a 0.2N-solution of potassium chromate had been added The reactions were studied with regard to the maximum concentration of available bleaching chlorine obtainable and the current efficiency. The first section includes (a) The influence of the anode current density (Dn) and the anode shape The best results were obtained with a short bright platinum anode in the form of a wire in which case a maximum concentration of 88 grams per litre of available bleaching chlorine was obtained with a current density D = 14.3 amperes per sq.cm. Massive platinum anodes (sheets or wires) are far more efficient than gauze electrodes. Platinised platinum anodes showed a slight superiority over bright anodes for current densities up to Da=0*44 ampere per sq. cm. but abovo this value there was no advantage gained by platinising the electrodes.With D =2 0.44 ampere per sq. cm. using platinised electrodes a maximum concentration of 48 grams per litre available bleaching chlorine was obtained with a current efficiency of 91%. The platinised electrodes showed their superority .especially when they had been previously cathodically polarised. Grey platinum electrodes behaved similarly to the platinum- black electrodes. VOL. civ. ii. 4ii. 50 ABSTRACTS OF CHEMICAL PAPERS. ( 6 ) Influence of temperature. The best results were obtained a t the ordinary temperature. A t 30' the concentration of hypochlorite mas diminished whilst at 0' and 15' the formation of chlorine hydrate (CI,,lOH,O) on the anode interfered with the process. ( c ) Influence of stirring the electrolyte.The effect of stirring is to reduce the maximum concentration of the hypochlorite. (d) Influence of dilution of the electrolyte. The salt concentration was changed from 5N to 2N and l.1N; in each case there was a decrease in the maximum concentra- tion of the hypochlorite and of the current efficiency. (e) Influence of the addition of various substances to the electrolyte. The object of the addition was to prevent cathodic reduction. The experiments were carried out with carbon anodes in addition to platinum which had hitherto been used. The results are (I) Potassium chromate pre- vented reduction almost entirely ;. (11) the addition of calcium chloride and Turkey-red oil (sulphonated castor oil) allowed a reduction of 20% although a 30% higher maximum coneen tration of hypochlorite could be obtained.The action of the Turkey-red oil is probably the forma- tion of an anode coating which causes a great increase of the current density. The best results are obtained by combining the chromate and Turkey-red oil additions; in this way for example 62 grams per litre of available bleaching chlorine was obtained with a 62% current efficiency. By using Portland cement to make the cell malls of from which something is dissolved an improvement of both anode and cathode actions was effected. Wheat starch and soluble starch gave much inferior results. The author experimented with models of the apparatus used in the old Kellner process the new Kellner process the Schuckert process and the Weichert process.Working with the old Eellner process it was found that the highest Concentration of hypo- chlorite obtainable was 12-20% below that given in closed vessels. This is attributed t o the influence of the carbon dioxide from the air and to the circulation of the electrolyte over the anode. I n the new Kellner process it was found that the maximum hypochlorite concen- tratiou was independent of the current density and the rate of flow of the electrolyte. The anode in this process must be placed beneath the cathode for the best results. The yield is increased by the addition of potassium chromate and Turkey-red oil to the electrolyte and in this case a concentration of 68 grams per litre available bleaching chlorine was obtained from a 5N-solution of sodium chioride with a 41% current efficiency.I n the Schuckert process the current density has but little iufluence on the concentration of the product; the best results were achieved by adding potassium chromate to the electrolyte. The addi- tion of calcium chloride and sodium resinate gave less favourable results. I n the Weichert process using a dilute sodium chloride solution ( 5 5 O Be) with the addition of 0.2% potassium chromate a yield of 8-10 grams per litre of available bleaching chlorine was obtaiued when working with a current of 14 amperes at 95 volts and with the electrolyte floLving through the cell a t 40 litres per hour. J. F. S. Crystallographic Study of Lithium Chlorate. (Mlle.) H. BKUHL (BUZZ. SOC. frang. Min. 1912,35 155-165).-Lithium chlorateii. 51 INORaANlC CHEMISTRY.was prepared by mixing solutions of lihhium sulphate and bsriam chloritte. It is extremely deliquescent and is very soluble in alcohol the latter property affording a means of purifying the salt. Under the microscope the anhydrous salt LiCIO (m. p. 125-127*5") was observed in three different crystalline modifications. There is also the hydrate Li,UlO,,&H,O (m. p. 63-65') which forms plates with secondary twin-lamellE. L. J. S. REZSO BALL^ and EMIL UITTLER (Zeitsch. nnorg. Chem. 1912 77 456. Compare A. 1912 ii 758).-The biaxial modification of lithium metltsilicnte previouily described is already known and is probably monoclinic. Its pro- The Binary System Zi2SiO3-Al,(SiO,),. perties persist in -its solid solution - with aluminium metasilicnte. C.H. D. The Carbonic Acid of Natural Waters which is Able to Attack Calcium Carbonate. J. TILLMANS and 0. HEUBLEIN ( C / L R ) ~ . Zen.fr. 1912 ii 1355-1396 ; from Gasundheits-lny. 19 12 35 669-677).-A certain amount of carbonic acid is necessary to pro- tect calcium hydrogen carbonate from decomposition so that although " free," it does not attack calcium carbonate. Thus a soft water mill attack limestone f a r more than a hard water containing the same amount of free carbon dioxide. A curve has been made connecting calcium hydrogen carbonate with the carbon dioxide which is necessary for its protection. Having found the amount of combined carbon dioxide and the total free carbon dioxide in a water i t is thus possible to say how much of the latter is capable of attacking calcium carbonate.Small amounts of magnesium carbonate do not seriously affect the results. J. C. W. Physical Action of Neutral Reagents on the Tribasic Alkaline Earth Phosphates. Investigation of the Colloidal State. WILLIAM OECHSNER DE CONINCK (Rev. Gen. Chim. Y u w App!. 1912 15 285-886).-1f the normal phosphates of calcium strontium and barium are shaken up with water methyl alcohol or ethyl alcohol for several days colloidal solutions are obtained which can be s2parsted from the excess of phosphate by filtration through filter paper. I n this way the author has prepared colloidal solutions of calcium phosphate in water and methyl alcohol and of strontium phosphate in methyl aud ethyl alcohol. The solutions are unstable and on this account the solutions of barium phosphate which appear t o be formed in methyl and ethyl alcohole are coagulated during the process of filtra- tion.H. M. D. The Ternary Alloys of Magnesium Zinc and Cadmium. GIUSEPPE BRUNI and CARLO SANDONNINI (Zeitsch. anorg. Chem. 1912 78 273-297):-The portion of this system which includes the con- stituents Zn-Cd-MgZn2 has been described previously (A . 19 10 ii 954). Cadmium retains up to 3 atomic % of zinc in solid solution. The crystallisation-interval i s very small throughout the magnesium- cadmium series. A transformation curve for the compound DlgCd is 4-2ii 52 ABSTRACTS OF CHEMICAL PAPERS drawn showing a eutectoid point; near 80 atom. % Cd traces of a eutectoid structura being recognisable in slowly-cooled alloys having about this composition. The compounds MgZn and MgCd form a simple eutectiferous series.The entire triangular diagram is divided into two parts by a continuous eutectic curve connecting the two eutectic points of the magnesium- zinc system and separating the MgZn area from that of solid solutions and from that of zinc. This curve passes through amaximum. The curves of secondary and eutectic crystallisation have been completely traced. It appears that the P-solid solutions retain zinc in solution at high temperatures but that its solubility disappears at about 250'. The alloys are best etched by means of nitric acid in alloy1 alcohol. The photo-micrographs illustrate the whole series of ternary alloys and serve t o fix the limits of stability of the solid solutions. C.H D Action of Potassium Amide on Thallium Nitrate in Liquid Ammonia Solutions. EDWARD CURTIS FRANKLIN (J. P l y i c a l Chem. 19 12 16 682-703).-The author describes in detail the preparation and properties of thallium nitride and potassium ammonio- thallite which have already been briefly mentioned (A. 1912 ii 452). The interaction in liquid ammonia solution of thallium nitrate and potassamide according to the equation 3TlN0 + 3KNH2 = TI,N + SKNO + 2NH yields a a dense black precipitate of thallium nitride ihich is readily soluble in excess of potassamide and also in liquid ammonia solutions of ammonium nitrate. I n the latter case the thallium nitride is reconverted into nitrate thus Tl,N + 3NH,N03 = 3TlN0 + 4NH From the solution in excess of potassamide well formed golden- yellow crystals are deposited.These consist apparently of iso- morphous mixtures of potassamide and dipotassium ammoniothallite with ammonia of crystallisation. The potassamide may be eliminated by contact with excess of thallium nitride or by repeated crystallisation. Dipotassiurn ammoniothallite when freed from excess of ammonia in a vacuum a t - 40° has the formula TlNK2,4NH3 or TlNH,,2KN H2,2N H 3 At + 20° in a vacuum ammonia is given off and the yellow substance TlNK,,2NH3 or T1NH,,2KNH2 results. On further deammoniation a t + 100' the substance turns black and approximate3 to the composi- tion TINK l$NH8 alternatively written Tl,N,6KNH2 or as a simplo mixture (T1,N + GKNH,). TINH 2 KNH,,2NH3 is more reasonable in view of the isomorphous mixed crystals with potassamide but on the other hand thallium amide is unknown in the free state.The extreme composition observed in the mixed crystals was T1NK2,4NH 3*6KNH and the higher the proportion .of potassamide the greater the solubility. The ammonia of crystallisation was always proportional to the thallium content the proportions being T1 4NH and TI 2NH in preparations dried at - 33' and + 20' respectively. Thallium nitride and potassium ammoniothallite and its mixtures The alternative formula for dipotassium ammoniothallite,INORGANlC CHEMISTRY. ii. 53 explode with great violence when subjected to heat shock or the action of water or dilute acids. They were hydrolysed for analysis by means of water vapour. No evidence could be obtained of the formation of an.amide or imide of thallium even at - 33". R. J. C The Quaternary System KCI-CuC12-BaC12-H20. FRANS A. H. SCHREINEMAKERS and (Miss) W. 0. DE BAAT (Proc. Lz. Ahad. Tetensch. Amsterdam 191 2 15 467-472).-The equilibria occurring in this quaternary system have been examined at 40" and 60". The choice of these temperatures is determined by the fact that the ternary system KCI-CuC12-H20 is characterised by the formation of only one double salt 2KC1,CuCl2,2H2O below 57O whereas a further double salt KCl,CuCl may occur between 57" and 92". The experimental data may be conveniently represented by means of a tetrahedron the four apices of which correspond with the fow components. The spacial model for 40" shows saturation surfaces corresponding with KCI BaC12,2H,0 CuC12,2H20 and 2KCl,CuCl2!2 H20 whereas that for 60" show an additional surface corresponding with KC1,CuC I 2.The temperature corresponding with the disappearance of this surface has been found to be 5 5 * 7 O and this temperature corresponds with the reversibility of the change represented by BaC1,,2H20 + CuC1,,2H20 + 2KCI,CuCJ2,2H2O KCI,CuCl + solution. H. M. D. Mutual Solubilities of Cuprous Chloride and Ferrous Chloride as also of Cuprous Chloride and Sodium Chloride and the Transformation Point of FeC1,,4H20 FeC1,,2H20 + 2H20. ROBERT KREMANN and F. Noss (Monutsh. 1912 33 1205-1215).'At 21.5" the solubility of cuprous chloride is increased by tho addition of ferrous chloride as also that of ferrous chloride by the addition of cuprous chloride the solid phases in equilibrium with the solution being respectively cuprous chloride and ferrous chloride ; there is no formation of a double salt.The solution saturated with respect t o both the salts contains 73.20 grams of ferrous chloride (anhydrous) and 23-20 grams of cuprous chloride in 100 grams of water. The system cuprous chloride-sodium chloride-water a t 26.5' behaves similarly to the system ferrous chloride-cuprous chloride- water. At the point where it is simultaneously saturated with respect to the individual salts i t contains 5'7.21 and 44.14 grams of cuprous chloride and sodium chloride respectively per 100 grams of water. FeC1,,4H20 as determined by cooling curves was found to be 65.2'. T. S. P. The transformation point of the system FeCI2,2H2O + 2H20 Ammoniacal Sodium Cuprous Trithionate.KSHITIBEIUSHAN BHADURI (Zeitsch. anorg. Chem. 1912 78 327).-Copper sulphate is precipitated with ammonia and the precipitate is washed and dissolved in ammonia. Sodium thiosulpbate is added and after remaining over- night blue crystals are obtained the solution becoming colourlees.ii. 54 ABSTRACTS OF CHEMICAL PAPERS. The product after washing with water and alcohol and drying forms sky-blue crystals soluble in water or more readily in ammonia and giviag the reactions of a trithionate. The salt has the composition 3Cu2S,06,2Na2S30,,9N H,. C. H D. The System HgC12-CuC12-€€20. FRANS A. H. SCHREINEMAKERS and J. C. THONUS (Proc. K. Akad. JVetensch. Amsterdam 1912 15 472-474).-1n order to ascertain whether mercuric chloride and cupric chloride form a double salt the form of the solubility isotherm has been determined by a series of solubility deasurements at 35".The composition of the residual '' solid phase" as well as that of the solution was determined and the numbers so obtained show that no double salt is formed at this temperature. A solution containing 21.5% of cupric chloride and 50.5% of mercuric chloride is saturated both with respect t o cupric chloride (CuC1,,2H20) and to mercuric chloride. The middle portion of the isotherm where two solid phases coexist is of very limited dimensions. The data show that the solubility of mercuric chloride is greatly increased by the addition of cupric chloride. Whereas in pure water the solubility amounts t o 8*5% it increases to 52.8% in presence of 18.06% of cupric chloride.When the quantity of cupric chloride present is further increased a slight diminution occura in the percentage amount of dissolved mercuric chloride. H. M. D. The Acid Dissociation C o n s t a n t of Aluminium and Zinc Hydroxides. ROLAND EDGAR SLADE (Zeitsch. anoyg. Chern. 19 12 77 457-460. Compare P. 1910 26 236; Klein A. 1912 ii 351; Hantzsch ibid. 644).-An approximate calculation shows that aluminium liydroxide must have an acid dissociation constant of a t least 10-10 and probably higher and that zinc hydroxide in its most stable form must have a corresponding constant of.not less than 0.5 x 10-12. C. H. D. Ultramarine Compounds. L. WUNDEB (Zeitsch. anorg. Chem. 191 2 '77 209-23S).-The replacement of sodium in blue ultramarine by other metals has been studied by Heumann (A.1877 ii 572 ; 1878 113; 1879 437 692 ; 2880 217-367) and has now been further developed. The material is heated with the [reacting solution to 120-180" in a closed glass tube enclosed in sand in a steel cylinder. Mercurous nitrate yields a greyish-blue mercuro-ultramarine which loses mercury on heating and leaves a pure blue residue which becomes yellow on heating and again blue on cooling this change being capable of being repeated indefinitely. The product is very stable towards acids. Mercuric salts are reduced by ultramarine. Cadmium yields :in olive-green product whilst lead replaces the sodium completely. Lithium ammonium and calcium ultramarines have also been prepared. The decomposition of ultramarine by alum solutions is not due to the acidity of tho latter but to the intermediate formation of an aluminium derivut ive.Chromium has the same effect. Ferrous ultramarine is dark greyish-blue and is readily decomposed by acids.INORGANIC CHEMISTRY. ii. 55 Yellow phosphorus reduces ultramarine to a leuco-compound which again becomes blue if heated in air. Red ultramarine is decomposed by acids without evolution of hydrogen sulphide and is not changed by alum solution. Red ultramarine is best prepared by the action of nitric oxide on the blue compound sulphur being removed in the process. Violet ultramarine is decomposed by alum and also by acids sulphur dioxide being evolved. Dry hydrogen reduces it sulphur being removed as hydrogen sulphide. Red ultramarine is regarded as an acid the sodium salt of which is the violet compound The blue compound contains the group A1-S-Na which is absent from the red.Formulae are proposed for all these compounds. C. H. D. The Heusler Ferromagnetic Alloys of Manganese. EDGAR WEDEKIND (Zeitsch. angew. Chem. 1912 25 2524-2526).-A claim for priority against Heusler (ibid. 2253) in respect to manganese borides and the magnetic properties of binary compounds of manganese. C. H. D. Ternary Alloys of Iron-Manganese-Copper. NICOLA PARRAVANO (Gaxxetta 1912 42 ii 513-531).-0f the three binary systems comprised in the tornary system Fe-Mn-Cu the system Fe-Cu has been studied by Sahmen (A.? 1908 ii 186) and the others by the author (A 1912 ii 1175 and this vol. ii 58). The system Fe-Mn-Cu belongs to type I of the cases of solubility gaps in ternary systems discussed by Parravano and Sirovich (A 19 1 1 ii 705) and in a theoretical discussion the author applies the con- clusions there set forth to the construction of diagrams from the experimental results in the present instance.I n addition to the diagrams photographs are given of a number of the ternary alloys which are thus shown to have the structure to be expected from their position in the space diagram. R. V. S. Revision of t h e Atomic Weight of Iron. V. Analysis of Ferric Oxide. GREGORY P. BAXTER and CHARLES RUGLAS HOOVER (J. Amer. Chern. SOC. 1912 34 1657-1669).-From analyses of ferric oxide by reduction with hydrogen at 900° Richards and Baxter (A. 1900 ii 407) obtained the value 55.88 for the atomic weight of iron.Baxter (A 1904 ii 177) analysed ferrous bromide and found the value 55.845; on repeating this work with purer material Baxter Thorvaldson and Cobb (A. 1911 ii 287) obtained the result 55.838. Baxter and Thorvaldson (A. 1911 ii 288) further proved t h a t meteoric and terrestrial iron give identical values. I n the present investigation ferric oxide was prepared by igniting carefully purified ferric nitrate in a current of air. A weighed quantity of the oxide was reduced to the metal by heating it in a current of hydrogen at 1050-1 100' for several hours. Two specimens of ferric nitrate were employed one containing iron of terrestrial origin and the other meteoric iron. The average of twelve analyses yielded 55,847 as the atomic weight of iron (0 = 16.000).This when combined with the result obtainedii. 56 ABSTRACTS OF CHEMICAL PAPERS. from the analysis of ferrous bromide (Baxter Thorvaldson and Cobb Zoc. cit.) gives the value 55.84. The identity of terrestrial and meteoric iron was confirmed. E. G. The System Iron-Carbon. 111. ANDREAS SMITS (Zeitsch. Elektrochern. 1912 18 1081-1086. Compare A 1912 ii 165 769).- An answer to Ruff (A 1912 ii 917; compare also A. 1911 ii 897). The remainder of the paper is a theoretical consideration of the results of Wittorf (A 1912 ii 259) who shows that the system iron-carbon undergoes two inverse separations by raising the temperature namely C -+ Fe,C -+ FeC -+ Fe,C -+ C. In this connexion the author discusses the relations of the separations of the substances C FeC FeC C which he shows are successively deposited on raising the temperature.PZ' curves of the system are given in which two quadruple points are indicated at both of which graphite FeC solution and vapour are in equilibrium. On raising the temperature a t the lower point FeC is formed at the expense of the graphite whilst at the higher point graphite is produced at the expense of the FeC. The thermal relationships of the various changes are considered. J. F. S. The System Iron-Iron Sulphide. RICEARD LOEBE and E. BECKER (Zeitsch. anorg. Chern. 1912 7'7 301-319. Compare Treitschke and Tammann A. 1906 ii 547).-Ordinary ferrous sulphide always contains free iron and also oxide. A product contain- ing 98.72% FeS is obtained by repeatedly melting natural pyrites.The mixtures are made by fusing this product with Swedish iron in porcelain tubes using a kryptol furnace. The equilibrium diagram shows a freezing-point curve of simple form the two branches meeting in a eutectic point at 985O and 85% FeS. The solid components are practically immiscible on the iron side whilst solid ferrous sulphide retains about 1%Fe in solution. The formation of two liquid layers and the occurrence of a crystallisation interval are not observed. The transformations in the solid state have been examined by the differential method and it is found that the sulphide is without influence on the transformation points of iron. A polymorphic trans- formation of ferrous sulphide occurs at 298O and has been studied by the dilatometric method using mercury as the filling liquid.The transformation a t 1 3 8 O is also observed in all mixtures containing more than 7% Fe (compare Rinne and Boeke A. 1907 ii 471). The brittle- ness of iron containing sulphide is due to this transformation. Photo-micrographs of the alloys are given. The eutectic tends to become segregated in mixtures rich in sulphide. Alloys which have been in contact with oxygen contain a new eutectic. Fusion of iron sulphide with iron oxide however leads to the elimination of sulphur and formation of ferrite and the structure is better developed by fusing the sulphide in contact with air. The composition of the eutectic is uncertain. The brittleness of iron containing sulphur at a red heat is due to the absorption of oxygen in this form.C. H. D.INORGANIC CHEMISTRY. ii. 57 Stability of the Oxides of Cobalt in the Interval from Co,O to COO. SIEGFRIED BURGSTALLER (Chern. Zentr. 1912 ii 1525-1526 ; from Abhandl. Beut. naturwiss.-med. Ver. Bohrnen 1912 3 83-143).- The stability of the oxide of cobalt has been followed dynamically in view of the numerous oxides mentioned in the literature and of the statical measurements of the dissociation pressures made by Richards and Baxter (A 1900 ii 78) and by Smith and Foote (A. 1908 ii 847). By means of baths of diphenylamine b. p. 301*9'/760 mm. anthra- cene b. p. 339*8O/760 mm. and anthraquinone b. p. 377*2'/760 mm. the dissociation temperature of cobaltic oxide has been narrowed down to 372-373O/760 mm. Above this temperature the reduction of the oxide has been followed in a sulphur bath and in an electric oven and after allowing for the possible sources of error (such as the time taken in warming up and in cooling down) the alteration of the oxygen con- centration of the oxide with time by rising or falling temperature has been graphically portrayed. From the dissociation temperature of cobalic oxide to a point between '705.5" and 869O cobalto-cobaltic oxide is stable above which the region for cobaltous oxide occurs.The transition points cannot be determined however because regions of solid solution of one oxide in the otber must be passed through. From the reactions curve of the oxidation of cobalt in air i t is also seen that between 300' and 431° cobalto-cobaltic oxide is formed and that at 869' the region of solid solution between Co,O and COO is entered.The existence of solid solutions and the fact that the speed of the dissociation becomes very slow after a short time explain the numerous references to improbable oxides of cobalt. The Replacement of Metals from Aqueous Solutions of their Salts by Hydrogen at High Temperatures and Pressures. IV. VLADIMIR IPATIEV and B. ZRJAGTN (Ber. 1912 45 3226-3229. Compare A. 1912 ii 50).-In solutions of 2N- and 0.2N-cobalt sulphate and initial pressures of 100 atmos. of hydrogen no formation of a precipitate takes place within twenty-four hours at 103". A t 145-1 50' a rosy-violet coloured crystalline precipitate is formed having the composition CoS0,,H20 ; metallic cobalt also separates. If nitrogen or air is used instead of hydrogen the 2N-solution gives only the salt CoSO,,H,O.When a quartz tube is used instead of aglass one the reaction is not complete at 150" within four days cobalt sulphate still remaining in solution. A glass tube could not be used for so long a time without being attacked. A t 150° 2"- and 0.2N-cobalt nitrate solutions bebave similarly to the sulphate solutions a t 103'. At 205-210' a crystalline brownish- black precipitate- of cobaltic oxide which is contaminated with silicic acid from the glass tube is formed within a day (compare nickel nitrate A. 1912 ii 51). A similar result is obtained with N-solu- tions both in glass and quartz tubes whereas N-solutions oE nickel nitrate give a small quantity only of a green crystalline precipitate containing 52*90-53.36% of nickel. J.C. W.ii. 58 ABSTRACTS OF CHEMICAL PAPERS Solutions of cobalt chloride behave similarly to those of nickel chloride (Zoc. cit.) the reaction being reversible iarnely CoCI + H 3 Co + 2HC1. T. s. P. The Origin of the Colour Produced by Cobalt Solutions in Certain Mineral Colours. SIEGFRIED BURGSTALLER (Chem. Zentr. 1912 ii 1523-1525; from Abhandl. Deut. natuywiss.-med. Ver. Bollmen. 191 2 3 57-80).-The coloured residues obtained by moistening certain inorganic oxides with cobalt solutions and then igniting are shown to be solid solutions in which a white substratum such as alumina does not modify the blue or red colour of the cobalt ion whereas a yellow oxide such as hot zinc oxide produces a mixed colour. Rinmann’s green prepared at 750-760° from zinc oxide moistened with cobalt nitrate was extracted with ammonium carbonate in which a solid solution of 7.05 mol.COO to 100 mol. ZnO dissolved whereas free cobaltous oxide was insoluble. Thhard’s blue obtained by evaporating aluminium nitrate with cobalt nitrate solution and igniting the residue also contained cobaltous oxide. No solvents could be found which would separate the solid solution from the free oxide and other methods for investigating its constitution are being pursued. J. C. W. Ternary Alloys of Nickel-Manganese-Copper NICOLA PARRAVANO (Gazzettcc 19 12 42 ii 385-394).-The paper contains an account of this ternary system derived from the results of its thermal analysis and from the thermal analysis of the three con- stituent binary systems.Of these the system Ni-Mn has been previously described by the author (A. 1912 ii 1175) and the remaining two by other writers. The results of some new experi- ments with the system Cu-Mn are also given. The nature of the ternary system is exhibited in diagrams and photographs are reproduced of a number of the ternary alloys prepared. The Constitution of the Halogen Compounds of Bivalent Molybdenum. IVAN KOPPEL (Zeitsch ccnorg. ClLem. 19 12 7’7 289-300).-The compounds having the empirical composition AIoC1 and MoBr are known to react in n manner which indicates a more complex constitution. The properties are now reviewed and con- stitutional formulae are proposed. The most important properties are the existence and stability of the radicles (Mo3R?)” the formation of ;t crystalline hydroxide the resistance to oxidation the difference in solubility between the anhydrides and the hydrates and the arnphoteric character of the radicles.The formulae given assume a co-ordination number 4 the insoluble hydrates for example being [Mo3CI,R,,2H20]H,0 where It is a halogen. The amphoteric hydroxy-compound is . represented as [Mo,C1,(OH),,2H20] and the formation of salts from it is necessarily Molybdates. GEORG WEMPE (Zeitsch. ccnoyg. C l m ~ . 1912 78 29S-326).-Ammonium tetramolybdate i s best obtained by the R. V. S. accompanied by isomeric change. c. H. n.INORGANIC CHEMISTRY. ii. 59 action of warm hydrochloric acid on the ordinary molybdate and forms transparent triclinic crystalp (NH,),0,4M00,,2H20 of which 3.67 grams dissolve in 100 C.C.of water at 18". Further decom- position gives the trimolybdate ( NH,),0,3Mo0,,H20. Sodium tetra- molybdate Na20,4Mo03,6H,0 obtained by the action of nitric acid on the ordinary salt is much more soluble than the ammonium salt. The trimolybdate is obtained with 6 or 9 H,O from the mother liquor. Potassium tetramolybdate has not been obtained the corresponding solutions yielding instead the trimolybdate K,0,3M00,,3H20. The potassium hydrogen tetramolybdate has the composition K,O H20,8Mo03 12 H,O. Lithium tetramolybdate obtained from Rammelgberg's salt and hydrochloric acid forms triclinic crystals Li,O 3H,O 16M00,,6~5H20 from which a trimolybdate is obtained with 2 or 4 H,O. A second tetramolybdate is Li,0,H20,8M00,,1 OH20. The mbidium tetra- molybates obtained are Rb20,4Mo03 ; Rh20,H20,8M00,,3H,0 ; Rb,0,Mo0,,Rb20,3Mo0,,5H,0 ; and Rb,0,4Mo03,4H20 whilst the trimolybdate has been obtained with 3H,O.The caestum salts are Cs,0,4Mo0,,3H20 ; 5Cs20,12Mo03,1 lH,O ; and Cs,0,Mo03,Cs2~,3Mo03,4~5H20. The calcium Ca0,2H20,1:!Mo0,,~1H,0 and Ca0,H,0,8Mo03,16H20 ; barium Bit0,H20,8RIo0,,1 4H20 and qBa0,4Mo03),7 H,O ; strontium 2Sr0,3H20,20Mo03,21H20 ; thallous 'l'1,0,41!Co0,,H20 ; silver Ag,0,4Mo03,6 H ,O ; cadmium Cd0,1T20,8M00,,6H20 ; magnesium Mg0,H,0,8M00,,20H20 ; and zinc Zu0,H20,8M00 1 4H20 salts have also been obtained. C. H. D. Iso- and Hetero-poly-acids. VI. Hydrates of Same Hetero-poly-acids. ARTHUR ROSENHEIM arid JOHANNES JAENICKE (Zeih3cl~ anwg. Chem. 1912 77 239-251. Compare A. 1911 i 109 265 ; ii 116 612).-The following new hydrates of 12-phospho- tungstic acid have been prepared H,[P( W207),],28H,0 which is metastable at the ordinary temperature and passes into the stable 22-hydrate.The transformation point must be below 0". The addi- tion of concentrated nitric acid t o the solutions of ptecipitates the 1 %hydrate. 12-Silicotungstic acid also yields two new hydrates H@( W,0,),]33H,O which is stable below 29" and the 22-hydrate which is stable between 29' and 53'. 12-Borotung~tic acid forms H,[B( W2(/7)6],28H,0,. which does not pass into a 32-hydrate when heated. Nitric acid precipitates a 10-hydrate. The 22-hydrate previously described is hexagonal and is derived from an acid isomeric with that of the 28-hydrate thu two hydrates crj-stallising simultaneously from solution.Metatungstic acid which is also H 12-hetero-poly-acid yields only a single hydrate HI,[ H,( W20,),],22H,O. 12-Phosphoruolybdic acid yields a 28-hydrate H7[P(Mo,0,)6],28H20 which apparently yields a 12-hydrate on dehydration but the individual character of this hydrate is uncertain. A 22-hydrate is obtained by precipitation with nitric acid. 13-Silicomolybdic acid also forms a H,[%( h~020,),],28H20 fromii. 60 ABSTRACTS OF CHEMICAL PAPERS. which nitric acid precipitates a 12-hydrate but a 22-hydrate has not been obtained. C. El. D. Action of Acids on Uranous Oxide. A. COLANI (Compt. rend. 1912 155 1249-1251).-Usin,a uranous oxide prepared by reduction of the green oxide with hydrogen the anthor bas obtained results widely different from those of Raynaud (A.1912 ii 166 948). Whilst but a small amount of the uranous oxide goes into solution with sulphuric acid a considerable amount is converted into the sulphate and as such remains in the insoluble portion. Uranous oxide only dissolves slowly in hydrochloric acid the amouut dissolved in a given time varying widely with the mode of preparation of the oxide. Solution is continuous over a very long period of time. Asserted Explosibility of Uranyl Nitrate. LAUNCELOT W. ANDREWS (J. Amer. Chent. ~ o c . 1912 34 1686-1687 ; Chem. Zed. 1912 36 1463).-With reference to the statement of Ivanov (A. 1912 ii 455) on the explosibility of uranyl nitrate the author records the explosion of a quantity of uranyl nitrate in the preparation of which ether bad been employed.On adding strong solution of potassium hydroxide t o some of the salt which had escaped decom- position a marked odour of ether was produced whilst another specimen when disFolved in water gave a deep blue colour to Congo- red paper indicating the presence of nitric acid. It is considered that in this case the explosion was due to the water of crystallisation of the uranyl nitrate CJr02(N03)2 6H20 having been replaced by ether and nitric acid and it is regarded as probable that Ivanov's explosion was due to the Fame cause. VC'. G. E. G. The S y s t e m Tin-Iodine. WILLEM REINDERS and S. DE LANGE (PYOC. K. Akad. Wetensch. Amsterdam 1912 15 474-481*C).-The freezing-point curve for mixtures of iodine and stannic iodide consists of two branches meeting a t a point corresponding with 19.6' and 60% by weight of stannic iodide.The corresponding boiling-point curve which has been determined for atmospheric pressure shows neither maximum nor minimum. When stannic and stannous iodide are heated together a t 350° two liquid layers are formed one of which consists of almost pure stannic iodide whilst the other layer represents stannous iodide together with at most 6% of stannic iodide. When iodine is melted in contact with a small excebs of tin the product firbt formed is almost exclusively stannic iodide. The reaction SnI + Sn = 2Sn12 takes place extremely slowly and even a t 350' the velocity of t h i s change is comparatively small. Pure stanriic iodide melts at 143.5' and boils at 340° the corres- ponding temperatures for the stannous salt being 320' and 720".H. M. D. The Volatisation of Vanadic Acid by Halogens. ERNST 13. AUERBACH and K. LANCE (Zeitsch. angew. Chew%. 1912,25 2532-2523). -The observation of Ephraim ( A . 1903 ii 48'7) that vanadic acid * and Zcitsch. anorg. Chenh. 1912 79 230-238.INORGANIC CHEMISTRY. ii. 61 volatilises with hydrochloric acid is confirmed. The loss when vanadic acid is heated with sodium chloride is also appreciable. When barium chloride is used i t is found that the greater part of the chlorine may be driven off with very little loss of vanadium the chloride first formed being decomposed by the atmospheric moisture. If heated in a stream of dry carbon dioxide red vapours are obtained which condense to an oily liquid of unknown composition containing vanadium. It is necessary in the estimation of vanadium by means of mercurous nitrate that chlorides should be absent.C . H. D. Hetero-poly-acids Containing Vanadic Acid. WILHELM PRANDTL (Zeitsch. ano3.g. Chcm. 1912 79 97-124. Compare A. 1912 ii 167).-The author's view that compounds of this kind are derived from a hexa-vanadic acid is contrasted with the formulation due to Miolati (A. 1908 ii 595) and Rosenheim (A. 1911 i 109 265; ii 116 612 ; this vol. ii 59). Most oE the well-defined COM- pounds of this class may be represented as additive compounds of 4 or 6 mols. of molybdate and 1 mol. of hexavsnadate. Ammonium paramolybdate 3( NH4)20,7M003,4H20 is regarded as a compound of 1 mol. of diammonium trimolybdate with 4 mols. of ammonium hydrogen moly bdate ( NH4),E4M03012 4NH4H MOO,.Acids remove ammonia from the monomolybdate molecule which then polymerises to the trimoly bdate from which the hexamolybdate is obtained by elimination of water. [With SIGJIUND P ~ a ~ o w s ~ ~ . ] - T h e following new salts have been obtained KVO3,K,MoO4,4H?0 or K3V,0,,3K2Mo04,1 2H2O,. the normal salt of the series which is pale yellow and very sensitive to acids. Na4V,Ol7,6NaHMoO4,l 2H20 also yellow. Na6V60 s 6 Na,H,Mo30i 36H20 pale yellow. B~,V,O,,~GN~,H,MO~~~~ 2OH,O or 74 H,O according to the conditions ot precipltation. Na H,V,017 Nn2H,Mo3012,4NaHMo0 1 6H,O ; ~3HV60,7,2KHMo04,2K,H4Mo,0,,,6H,0 which are more deeply coloured. (NH,),v,o, 2NH4HMo04,2 ( NH,),H4Mo,0,,,8H,0 ; (NH4),HV60,,,2 ( NH4),H,Mo3O1,,4N H4H Mo04,6H20.These salts are tabulated and correlated with others of the same series described in the literature. Chemical Reactions of @Gold and Crystallised Gold. MAURICE HANRIOT and FRAN~OIS RAOULT (Conapt. rend. 1912 155 1085-1088. Compare A. 1911 ii 791).-Contrary to general belief ordinary gold as well as the brown (p) modification is attacked by nitric acid the amount dissolved whilst small varying with the concentration of the acid the @modification being readily acted on. Fuming hydrochloric acid in the presence of oxygen also has a marked solvent effect. The best solvent for /3-gold is a hot solution of auric chloride containing hydrochloric acid. On cooling the solution the dissolved gold crystallises out in the metallic state and a determination of its magnetkation coefficient shows that in this form it consists almost entirely of the @-variety.This solvent h j s a much more marked action on the brown than the yellow gold and it is the P-variety which passes into solution thus leaving the insoluble residue C. H. D.ii. 62 ABSTRACTS OF CEEMICAL PAPERS. enriched with the a-variety. This fact is borne out by comparison of the magnetisation coeficients of the origins1 brown gold the crystallised gold and the insoluble residue. 'CV. G. The Preparation and Application of Colloidal Platinum Metals ALADAR SKITA and W. A. MEYER (Ber. 1912 45 3579-3589).-See this vol. i 53. Osmium Tetroxide as an Oxygen-carrier and the Activation 3329-3336).-1n the presence of about 1 mg. of osmium tetroxide per 500 c.c. 1% solutions of p-leucaniline and leucomalachite-green become strongly coloured after a few hours on shakiog in the presence of air.Under an oxygen pressure of 10 atmospheres and temperatures of 50-100° the following oxidations take place in the presence of 0.01 gram of osmium tetroxide per 200 C.C. of liquid 40% alcohol is oxidised in three to four hours to acetic acid and aldehyde or acetal traces of crotonaldebyde also being formed ; 50% methyl alcohol gives formaldehyde and formic acid. The action on the aldehydes takes place very slowly or not a t all; €or example paracetaldehyde and acetal shows no signs of oxidation within eight days at 20'; vanillin only after twenty-four hours. The acids formic acetic and oxalic are very stable towards osmium tetroxide. Aniline sulphate is readily oxidised and anthracene gives anthraquinone in acetic acid solution. The following oxidising actions of potassium chlorate in neutral or slightly acid solutions take place very readily in the presence of osmium tetroxide but not at all or only extremely slowly in its absence arEenic to arsenic acid hypophosphorous acid to phosphoric acid ; hydrazine sulphate quantitatively t o nitrogen indigo-carmin to indigo- white quinol to quinhydrone aniline to emeraldin; mono- and di- methylaniline are oxidised ; benzidine to diphenoquinone-di-imide p - leucaniline and leucomalachite-green t o the dyes ; catechol resorcinol and tannin are oxidised anthracene to anthraquinone ethyl alcohol to acetal no chlorination taking place ; slightly acid potassium formate to carbon dioxide. Substances containing the ethylene linking such as ethylene propylene amylene indene r-pinene and dichloroethylene CHCl:CHCl are readily oxidised ; trichloroethylene CHCKCCI is only slowly oxidised and perchloroethylene not a t all. Ally1 alcohol is much more readily oxidised than ethyl alcohol ; isoeugenol is readily oxidised whilst vanillin is not. Benzene containing thiophen gives a reaction due to the presence of the thiophen. Benzoic acid bromobenzoic acid naphthalene and saturated ketones O f ChlOrate Solutions. KARL A. HOFBIANN (i?e!r. 1912 45 Acetylene is oxidised. T. s. P. are not affected and benzaldehyde only very slowly.