INORGANIC CHEMISTRY. Inorganic Chemistry. ii. 27 Anodic Formation of Hydrogen Peroxide. FRANZ RICHARZ (Bey. 1909 42 4674-4675. Compare Riesenfeld and Reinhold Abstr. 1909 ii 879).-The secondary formation of hydrogen peroxide at the anode has been observed by the author (Abstr. 1885 624; 1888 12 769). C. H. D. Diseociation of Eydrogen Bromide and Hydrogen Iodide at High Temperatures. KURT VOCEL VON FALCKENSTEIN (Zeitsch. physikal. Chenz. 1909 68 270-280).-The measurements were made by determining the partial pressure of the hydrogen by Lowenstein's method (compare Abstr. 1906 ii 272). The dissociation of hydrogen bromide is 0.50% a t 1024O 0.73% a t llOSa and 1.08% at 1222'; that of hydrogen iodide is 32.9% at 1022' and 37.55% at 1217O. The results and also the fact that at the high temperatures used the ha1 ogen molecules are partly dissociated according to the equations Br2=2Br' and 12=21 will be discussed in a later communication G.S. Electric Conductivity and Density of Solutions of Hydrogen Fluoride. ERNEST G. HILL and ANNODA P. SIRKAB (Proc. Roy. Xoc. 1909 A 83 130-148).-Measurements have been made of the electrical conductivity and density of mixtures of water and hydrogen fluoride ranging from the one pure substance to the other. A satisfactory conductivity cell mas constructed from paraffin wax of low melting point (40.4'). The density measurements were made by weighing a lead cylinder coated with paraffin wax in the various liquids Curves are plotted which show the variation of the specific con- ductivity and of the density with the percentage composition of the solutions.These are of approximately the same type and exhibit a well-defined maximum for a solution containing 75% of hydrogen fluoride. The curve which is obtained when the molecular con- ductivity of the hydrogen fluoride is plotted as a function of the concentration shows two distinct breaks one at a b w t 91% the other between 51% and 55% of hydrogen fluoride. The former corresponds with the composition 9HF,H,0 the latter with HF,H,O and theii. 28 ABSTRACTS OF CHEMICAL PAPERS authors draw the conclusion that definite hydrates of this composition ar0 present in the solutions. From values obtained for the conductivity at 0' and lSo the tern- perature-coefficient has been calculated. This is constant and equal to 0.0125 for solutions containing less than '7-7yA of hydrogen fluoride.By applying this temperature correction molecular conductivity data €or 25' are obtained. On comparison with Ostwald's numbers for dilute solutions aqsatisfactory agreement is obtained a t the dilution v = 4 but for more dilute solutions the authors' numbers are much smaller than Ostwald's the difference increasing with the dilution. H. M. D. Production of Ozone by Ultra-violet Light. EDMOND VAN AUBEL (Compt. rend. 1909,149 IlS3-985).-Bordierand Nogier (Conzpt. rend. 1908 14'7 354) could not find that ozone was produced from the oxygen of the air-by the action of ultra-violet light although previous observers (compare Abstr. 1906 ii 224 ; 1909 ii 657) had observed its formation. Using a quartz mercury lamp as the source of ultra- violet light the author definitely proves that ozone is produced.Instead of using water to absorb the ozone olive oil and light petroleurn were employed in the first experiments being placed in porcelain dishes inside the large glass globe surrounding the quartz tube. After several hours' exposure starch iodide showed the presence of ozone in these liquids. When distilled water was submitted to the action of the ultra-violet light for fourteen hours the presence of ozone mas detected by its action on a photographic plate. Starch iodide paper was turned blue in two minutes except where i t was covered with a piece of quartz and thus prevented from coming into direct contact with the ozonised air. T. S. P. Boiling Point of Sulphur Corrected by Reference t o New Observations on the Absolute Expansion of Mercury.HUGH L. CALLENDAR and HEEBERT Moss (I'roc. Roy. Xoc. 1909 A 83 106-108).-As a result of new measurements of the expansion of mercury between 0' and 300° it has been found necessary to add a correction factor to the result obtained for the boiling point of sulphur by Eumorfopoulos (Abstr. 1908 ii 1029). This raises the tempera- ture from 443.58' to 444*55' which is in practically perfect agreement with the value previously assumed (444.53'). H. M. D. The Dynamic Allotropy of Selenium. HUGO R. ERUYT (Xeitsch. anorg. Chern. 1909 64 305-326).-The study of the electrical con- ductivity is uncertain as a means of investigating the dynamic allotropy of selenium conductivity not being an additive property.The author has therefore used the specific gravity the reciprocal of which is a strictiy additive property of solid solutions. The material was purified as described by Marc in his study of selenium (Abstr. 1904 ii 105 ; 1906 ii 226 280 742 ; 1907 ii 453). Grey crystalline selenium which has not been heated above 120° has D 4-5-4.6 whilst sublimed selenium has D 4.80. The difference is not due to the presence of amorphous selenium After fusion passageINORGANIC CHEMISTRY ii. 29 of the amorphous into the crystalline form and heating to 125O the preparation has D 4.77. The denser form passes into a modification with D 4.77 when heated to 125'. There is thus an equilibrium Se Z Se high temperatures favouring the denser form. Exposure to the light of an arc lamp or even t o bright diffused light the selenium being enclosed in a dilatometer with alcohol produces an increase of volume; the quantity of Se is therefore increased by illumination.The changes in the electrical conductivity are explained on the assumption that Se and Se form solid solutions and also form a compound giving a diagram of the same form as is obtained from mixtures of bromine and iodine (Terwogt Abstr. 1906 ii 15) and alloys of magnesium and cadmium (Grube Abstr. 1906 ii 355). The conductivity has then two minima with a sharp cusp on the curve corresponding with the compound. The anomalous behaviour of specimens cooled from 200" is explained as being due to the combined action of light and heat in altering the equilibrium and also in changing the velocity with which equilibrium is attained.The conclusion is drawn that all phases of selenium contain Se and Se whilst red crystalline selenium is labile and may perhaps also occur in two modifications. C. H. D. Coefficients of Absorption of Nitrogen and Oxygen in Distilled Water and Sea-Water and of Atmospheric Carbon Dioxide in Sea-Water. CHARLES J. J. Fox (Trans. Faraday Soc. 1909 5,68-86).-The apparatus used for determining the coefficients of absorption of nitrogen and oxygen was a modified form of Estreicher's adaptation (Abstr 1900 ii 205) of Ostwald's apparatus. The pressure used could be varied at will whereas Estreicher's had to make the measurements a t exactly atmospheric pressure. Special precautions mere taken to free the water from dissolved air and it is suggested that the method used by Estreicher was not satisfactory so that his values for the absorption coefficients for argon and helium may be respectively 0.2 to 5% and 0.5 to 10% too low.Two series of measurements of the solubility of atmospheric nitrogen in distilled water between 0' and 50' mere made. A correction had to be applied due to the different solubilities of nitrogen and argon and the variation of the partial pressures with temperature. The relation between the absorption coefficient (aN) and the temperature is given by the formula 1000aN = 22.998 - 05298t + 0-0091962P - 0.00006779t3. In the case OF oxygen the formula obtained was 10OOa= 49939 - 1,3440t + 0*28752t2 - 0*0003024t3. These formuls give the following values for gas absorbed in C.C.per litre. 0" 10" 20" 30" 40" 50" Nitrogen ... 23.00 18'54 1 5 5 4 13.55 12.15 11'02 Oxygen ... 49'24 38.37 31.44 26-65 23'30 20.95 Determinations were also made on sea-waters of four different salinities. These values were then combined with those obtained for distilled water and formule obtained connecting the absorption coefficient temperature and salinity (chlorine-content) of the sea-water for nitrogen and oxygen at the partial pressures they have in air.ii. 30 ABSTRACTS OF CHEMICAL PAPERS. Tables are then given showing the amount of nitrogen (containing argon) and oxygen (in c.c.) absorbed by 1000 C.C. of sea-water from a free dry atmosphere of 760 mm. pressure at temperatures from 0’ t o 2S0 and the chlorine content 0-20 per mille.In the case of carbon dioxide the pressure in sea-water varies probably between 1 and 7 parts per 10,000; the total alkalinity also variw and for the open ocean is equivalent to about 40 mg. OH per litre. The carbon dioxide pressure was determined by a modification of Pettenkofer’s method. It was first of all established that the ratio of total CO to total OH is rectilinear and as a result of the law of mass action no C0,”-ions are formed; all the combined carbon dioxide must be in the form of HCO,‘. The effect of varying pressure on only one alkalinity that selected being 40 mg. OH per litre was then deter- mined. From these results an interpolation formula connecting the carbon dioxide per litre of sea-water the salinity the alkalinity the temperature and the pressure was obtained and tables of the calculated results are given.From these tables a calculatiop can be made of the power of the sea to conserve the constancy of carbon dioxide in the air (compare Arrhenius Pld. Mag. 1896 [v] 41 273). The conclusion IS drawn that for the open ocean (OH = 40 mg. per litre and t = 12”) it requires 17.6 times as much carbon dioxide to raise the carbon dioxide partial pressure as for the same volume of air; or if a quantity of carbon dioxide is poured into the air as for example by volcanic action two- thirds will eventually be absorbed by the sea. The final equilibrium will not be attained however until either the insoluble carbonates on the sea-bottom or the carbon dioxide in the air have completely gone into solution in the form of HC03’.The time necessary for this is not known but it must be very important geologically. The concentrations of the free ions HCO,’ and He for ocean-water are calculated to be respectively 2.24 x and 1.6 - 2.6 x 10-6 gram- equivalents per litre ; thus sea-water is just slightly more acid that distilled water. The paper also contains a description of the apparatus used for extracting and analysing gases dissolved in liquids. T. S. P. Red Phosphorus and the so-called ‘‘ Hittorf’s Phosphorus.” ALFRED STOCK and FRANZ GOMOLKA (BeT. 1909 42 4510-4527). -Hittorf’s phosphorus having the high density 2.3 is the only variety other than yellow phosphorus which is at all well defined the ordinary red variety being a mixture. Hittorf’s phosphorus is best prepared by heating 3 grams of pure phosphorus with 200 grams of lead in a sealed Jena glass tube packed in sand to 800’ for forty-eight hours.The glass is broken and removed in a freezing mixture and the lead cleaned by brushing and by washing with hydrofluoric acid. As nitric acid attacks the phosphorus the lead is best removed by electrolysis in acetic acid containing lead. The cathode is placed at the bottom of the vessel a clock-glass being fixed below the rod. The residue thus obtained contains some lead mechanicdly dislodged from the anode and is purified by boiling withINORGANIC CHEMISTRY ii. 31 hydrochloric acid in an atmoaphere of carbon dioxide followed by treatment with hydrofluoric acid. The purest product still contains 1.5% of lead. Bismuth may be used in place of lead but it dissolves only one-fifth as much phosphorus and the crystals obtained are less pure.The metals appear to be held in solid solution Only very minute quantities of Hittorf’s phosphorus are obtained by sublimation. Ordinary red phosphorus melts at 605-610° and Hittorf’s phosphorus a t 620’. The product obtained by fusion in a closed tube is heterogeneous Hittorf’s phosphorus corrected for the dissolved lead has D 2 31-2 33. It is much less chemically reactive than ordinary red phosphorus a t high temperatures but both oxidise slowly in the air to deliquescent products. The statement often made that red phosphorus is stable in air is incorrect. Oxidation of Phosphorus. WILLEM P. JORISSEN and N. H. SIEWERTSZ VAN REESEMA (Chem. Weekblud 1909 6 931-938.Compare Schmidt Abstr. 1902 ii 237 ; Schenck Mihr and Banthien Abstr. 1906 ii 326 ; de Broglie and Brizard Abstr. 1909 ii 535 ; Riboul ibid. ii 718; Elster and Geitel Physikal. Zeitsch. 1902 3 475 ; 1903 4 111 293 436 457 ; Ber. deut. physikccl. Ges. 1906 640).-!I!he work of Elster and Geitel and of Schmidt on the oxida- tion of phosphorus has bean repented. The results obtained support the conclusions of Elster and Geitel but indicate that Schmidt’s statements are erroneous. Preparation of Hypophosphoric Acid. JACQUES CAVALIER and E. CORNEC (Bull. SOC. chim. 1909 [iv] 5,1058-1060).-A number of glass rods are placed across the bottom of an ordinary photographic washing dish and on these are laid transversely sticks of phoephorue separated from each other by glass rods.Water is then poured into the dish until the phosphorus is half submerged. The whole is covered by a glass plate resting on a layer of cotton wadding so that air is slowly b u t continuously admitted. The acid formed can be neutral- ised by sodium carbonate to form sodium hydrogen hypophosphate or the latter salt can be obtained at once by substituting a solution of sodium acetate for water in the dish. From this sodium salt the acid is best recovered by making the lead salt by double decomposition and regenerating the acid with hydrogen sulphide. By this process and using a dish 13 x 18 cm. from 30 to 50 grams of sodium hydrogen hypophosphate may be obtained in five to six days. Decomposition of Water by Hypophosphites in Presence of Palladium aa a Catalyst.ALEXIS BACH (Ber. 1909 42 4463-4470).-The formation of copper hydride from sodium hypo- phosphite and copper sulphate takes place at the ordinary temperature and is an example of the decomposition of water by an oxidisable substance in presence of a substance capable of combining with hydrogen. The same reaction takes place in presence of a catalyst hydrogen being set free An improvement of this method is being studied. C. H. D. A. J. W. T. A. H.ii. 32 ABSTRACTS OF CHEMICAL PAPERS Solutions of sodium hypophosphite and palladium chloride are used the course of the reaction being followed by measuring the volume of hydrogen evolved. The velocity of decomposition increases much more rapidly than the quantity of palladium added the con- centration of the hypophosphite being kept constant.The velocity decreases rapidly with time and is not increased by further additions of palladium but rises to the original value on further addition of bypophosphite. With constant palladium the velocity increases less rapidly than the quantity of added hypophosphite. Hydrocyanic acid prevents the decomposition or if palladium sponge is used instead of the chloride slowly brings it to a standstill. The analogy of the behaviour of palladium in this case to that of a peroxydase is pointed out both departing from the laws of chemical kinetics. The intermediate formation of an unstable palladium perhydridc is assumed. C. H. D. Phosphorescent Oxidation of Arsenic. L ~ O N BLOCH (Compt. rend. 1909 149 775-777. Compare Abstr.1909,; ii 395).-The phosphorescence of arsenic at 200° like that of sulphur and phosphorus is accompanied by oxidation by which arsenic trioxide is produced. As in the case of sulphur no ionisation occurs and there is also no formation of ozone which is produced by both sulphur and phos- phorus. The product of oxidation of arsenic either phosphorescently or with flame always contains arsenic oxide just as some phosphoric oxide and sulphur trioxide are always produced in the parallel cases. The arsenic oxide which may amount to 1/30th of the whole seems to be formed directly from arsenic since arsenious oxide cannot be oxidised under the conditions. The author assumes conversely that arsenic oxide is the sole original product and this is mainly decomposed into arsenious oxide by a secondary change.The synthesis of sulphur chloride and arsenic chloride without incandescence is unaccompanied by ionisation. R. J. C. Formation of Silicon Sulphide in the Desulphurisation of Iron. W. FIELDING (lkans. Faraday Xoc. 1909 5 110-1 1 l).-The object was to find the conditions under which ferrosilicon can react with ferrous sulphide and liberate a sulphide of silicon. The reaction was investigated by heating mixtures of the two compounds in a vaciium at known temperatures. The heating was effected in a crucible in the form of a hollow graphite rod heated electrically. With commercial ferrous sulphide the mass fused at about 930° and a vigorous reaction set in accompanied by a rise in temperature. With pure ferrous sulphide no reaction was observed up to about 1300° so that the readion noted with the impure compound was probably due to reduction of oxide of iron present by the ferro- silicon.I n all the experiments a yellow sublimate appeared on the walls of the tube at 1500'. This was found to consist of approximately 50% of silicon sulphide (assuming the formula to be SiS,) the remainder being iron sulphide which had volatilised silica resulting from the action of moisture in the air on the silicon sulphide and a small amount ofINORGANIC CHEMISTRY. ii. 33 finely-divided ferric oxide. I n different experiments products of variable composition were obtained and it has not been possible completely to identify the silicon sulphide present in the product. T. S. P. Thermal Analysis of the System K2S0,-KF.16. KABANDEEFF (Centr. %in. 1909 728-733).-0n the basis of observations on the rate of cooling of fused mixtures of potassium sulphate and potassium fluoride a temperature concentration diagram has been constructed. The freezing-point curve consists of three branches corresponding with the two components and with a double salt KF,K,SO respectively. The eutectic temperatures 883' and 788O correspond respectively with 41 and 83 mold.% of potassium fluoride. The double salt melts a t 88'7'. It is only stable at temperatures above 5 7 8 O and on cooling below this breaks up into its components. Potassium sulphate undergoes a change in crystalline form at 599'. H. M. D. Existence of Real Percarbonates and their Differentia- tion from Carbonates with Hydrogen Peroxide of Crystal- lisation.E. H. RIESENFELD and B. REINHOLD (Ber. 1909 42 4377-4383).-According to the authors Constam and von Hansen (Abstr. 1897 ii 550) did not definitely prove that the substance which formed at the anode in the electrolysis of concentrated solutions of potassium carbonate was potassium percarbonate. The substance was never obtained pure the hydrogen carbonate and water being always present so that it was possible that it was a hydrogen peroxide additive product of potassium carbonate since the authors have recently proved the anodic formation of hydrogen peroxide in strong solutions of potassium hydroxide at low temperatures (compare Abstr. 1909 ii 879). The authors have now succeeded in preparing a pure potassium percarbonate of the formula K,U,O and containing neither water nor hydrogen carbonate so that it could not be an additive product of hydrogen peroxide.It was made by the electrolysis a t - 30' to - 40' of a very istrong solution of potassium carbonate contained in a U-tube. was a platinum wire the cathode beiag. of platinum Foil and surrounded by parchment paper. The current used was 0.5 ampere. After six hours the salt formed at the anode was collected washed with cold water alcohol and ether and dried and then proved to have the composition given above. When added to a neutral solution of potassium iodide at the ordinary temperature it immediately liberates iodine according to the equation no oxygen being evolved whereas hydrogen peroxide reacts very slowly under such conditions. This reaction is characteristic of a real percarbonate and can be used t o distinguish them from additive products of carbonates and hydrogen peroxide.It is found that the percarbonates described by Tanatar (Abstr. 1903 ii ZOS) when added t o a neutral solution of potassium iodide cause a brisk evolution of oxygen and the solution remains colourless. All these compounds The anode c,o," + 21' = 2c0," + I VOL. XCVIII. ii. 3ii. 34 ABSTRACTS OF CHEMICAL PAPERS. must therefore be classed as additive products of hydrogen peroxide and carbonates. The percarbonates obtained by Wolff enstein and Peltner (Abstr. 1908 ii 180 183 ; 1909 ii 574) are also doubtful. It was not found possible to prepare solid percarbonates of lithium sodium rubidium and cssium although the anolyte reacted towards potassium iodide as if it contained percarbonate.T. S. P. Reduction of Sodium Sulphate by Carbon. ALBERT COLSON (Compt. rend. 1909 149 1076-1078).-When an intimate mixture of sodium sulphate and animal charcoal is heated in an iron pipe practically DO interaction takes place below 950° but at this tempera- ture the reduction is rapid and uniform especially if charcoal is used in the proportion required by the equation Na,SO + 4C = Na,S + 4CO. With increase of temperature the rate of action augments rapidly. T. A. H. Action of Safety Explosives containing Ammonium Nitrate in the Presence of Carbon Paper and Paraffin. H. DAUTRICHE (Compt. rend. 1909 149 926-928) -Safety explosives containing ammocium nitrate are used in coal mines and since oxygen is set free in the explosion it is important to investigate their action on charcoal powder which would be equivalent to coal dust in the mine and on paper and paraan the latter being used in making the cartridges.Experiments carried out under conditions very similar t o those obtaining in a mine with an explosive consisting of 90 parts of ammonium nitrate to 7 parts of trinitrotoluene show that charcoal powder surrounding the cartridges is burnt arid that all the oxygen is used up with the formation of carbon monoxide. The combustion of the paper and paraffin depends on the diameter of the cartridges. When these are 30 mm. in diameter the combustion is feeble but when 40 mm. in diameter it is very marked. T. S. P. True Atomic Weights. Stas’ Determinations.111. Lou19 DUBREUIL (Bull. SOC. chim. 1909 [iv] 5 1049-1053 1053-1055 1055-105S).-A series of three papers continuing (this 1 ol. ii S86) the critical revision of Stas’ determinations the cases now considered being (1) silver iodate (2) silver bromate (3) silver chlorate. The apparent atomic weights found for the three cases are (1) silver 107-9990 and 107-9991 iodine 126.9790 and 126-9991 oxygen 16.0149 and 16.0128 (2) silver 107-9995 bromine 79 9995 and oxygen 16.0061 (3) silver 107.9996 chlorine 35 4996 and oxygen 16-0036. T. A. H. So-called Electrolytic Peroxide of Silver. MARGRETE BOSE (Zeitsch. physikal. Chem. 1909 68 383-384).-The author’s work on the same subject (compare Abstr. 1905 ii 299) has not been mentioned by Baborovskp and Kuzma (compare Abstr.1909 ii 666). G. S. Silver and Thallium Iridichlorides and Iridochlorides. MARCEL DELEPINE (Compt. rend. 1999 149 1072-1 074. Compare Abstr. 1908 ii 702).-The fugitive blue precipitate formed when potassiumINORGANIC CHEMISTRY ii. 35 iridichloride is added to silver nitrate consists of silver iridichloride Ag21rC16. This when kept either alone or in presence of excess of silver nitrate passes into silver iridochloride Ag3T.~-Cl which is yellow and on treatment with ammonia changes into the greenish- yellow argentodiammonium iridochloride ( Ag2NH,)31rC16 which loses part of its ammonia on exposure to air and the whole of it on warming (compare Claus J. pr. Chena. 1847 i 42 348). Thallium iridichloride T121rC16 forms small opaque greenish-blue cubes and is more stable than the corresponding silver salt.Boiling hydrochloric acid decomposes it forming the iridochloride T131rC16 which crystallises out on cooling the liquid in bronze-tinted lamellae. Nitric acid re-converts it into the iridichloride. T. A. H. The Calcium Silicides and their Absorptive Power for Nitrogen. AUALBERT KOLB (Zeitsch. m o r g . Chem. 1909 64 342-367).-By heating together calcium and silicon two silicides are obtained according to the component in excess. The products contain 53.5% and 36.68% Si respectively corresponding approximately with the formulae Ca,Si and Ca,,Si (compare Hackspill Abstr. 1908 ii 589 ; Tamaru Abstr. 1909 ii 400). Both silicides are crystalline evolve hydrogen with acetic acid and evolve spontaneously inflammable hydrogen with dilute hydrochloric acid.Silicones are obtained with concentrated hydrochloric acid. The first silicide yields an orange or yellow crystalline silicone the second yields a silicone containing less silicon. Both silicides absorb nitrogen actively near 1000° the first com- pound being the mme energetic. The products have the respective compositions CaSizNs and Ca,,Sil,N,,~ Structural formulae are pro- posed for the silicides silicones and silico-nitrides. C. H. D. Calcium Ferrites. SIEGFRIED HILPERT and ERNST KOHLMEYER (Bur. 1909 42 4581-4594).-Mixtures of pure ferric oxide and lime are heated in a platinum crucible protected by an outer crucible of spinel mass in an electric furnace of which a hollowed-out carbon rod forms one pole the other being an outer iron cone the space between being filled with granulated carbon.Mixtures containing only small percentages of iron oxide are not completely fusible under these conditions being on y pasty. The last portion solidifies at 1410° when they become completely solid. The lime forms well-developed crystals. When the lime falls below 75 mol.% the primary crystallisation is that of the orthoferrite. The first complete fusion is obtained with 69 mol.% CaO freezing beginning at 1550O. Calcium orthoferrite 3Ca0,Fe203 melts at 1 410° and disintegrates like calcium orthosilicate on f u t ther cooling. The next compound 3Ca0,2Fe,03 melts a t 1450O. At 1220O a reaction occurs in the solid state the compound formed having a formula near to 5Ca0,3Fe203 (compare the aluminate 5Ca0,3A1,03 Shepherd Rankin and Wright Abstr.1909 ii 1015). There IS a eutectic point at 1200O and 50 mol.% CaO and a second maximum at 1400° corresponding with 3-2ii. 36 ABSTRACTS OF CHEMICAL PAPERS. the compound 2Ca0,3Fe20,. Calcium metafeerrite CaO,Fe,O appears to be formed below the eutectic point. The behaviour of mixtures rich in ferric oxide is complicated partly owing to the escape of oxygen and the formation of crystals of magnetite. Ferric oxide melts at 1565' and the magnetic oxide at 1527'. The colour of the mixtures becomes darker with increasing iron. The specific volume curve shows a sharp change of direction at the formula CaO,Fe,O and another less distinct near 20 mol.% CaO. The magnetic properties diminish with the proportion of iron and become insignificant at 67 mo1.s CaO.The electrical conductivity is very small throughout even that of fused ferric oxide being only that of iron. Solid solutions are not formed and the low conductivity is unexplained. The calcium ferrites are much less readily attacked by reagents than the silicates hence the advantage of their presence in cements exposed to sea-water. The mixtures containing 60-70 mol.% CaO are hydraulic. Calcium ferrites are less easily reducible than ferric oxide. C. H. D. Compounds containing Iron Peroxide FeO,. LUDWIG MOESER and H. BORCK (Ber. 1909 42 4279-4283. Compare Abstr. 1903 ii 546).-When a mixture of concentrated solutions of ferric and strontium nitratm (1 mol. 1-2 mols.) is evaporated to dryness and the finely-powdered residue heated a t a temperature of not more than GOO3 in a stream of oxygen until oxides of nitrogen are no longer evolved a compound is obtained which seems to have the formula SrO,FeO,.Determination of the proportion of active oxygen t o ferric oxide showed that i t could not be a compound of strontium peroxide and ferric oxide. A similar compound was indicated in the residue obtained by heat- ing a mixture of barium hydroxide and ferric hydroxide a t 400' in a current of oxygen. These compounds are black substances which are stable below 650° but decompose gradually above that temperature. They are slowly acted on by water with evolution of oxygen. Hydrogen per- oxide reacts violently with liberation of oxygen ; acids act in a similar manner. Hydrochloric acid gives both chlorine and oxygen ; oxalic acid oxygen and carbon dioxide.Mixtures of air and alcohol or other inflammable vapours when led over the heated compounds are oxidised to water and carbon dioxide and when once the combustion has started it proceeds of its own accord. It is possible that a compound of iron peroxide with lithium oxide has been obtained but it was not found possible to prepare siich compounds with other bases. In an appendix H. Borck describes a compound of strontium and ferric oxides which is obtained in a hydrated condition as a yellowish- brown precipitate by warming a suspension of freshly-precipitated ferric hydroxide in a concentrated solution of strontium hydroxide for some time on the water-bath. On heating it loses water and becomes brown and at temperatures above 300' it absorbs oxygen forming the compound SrO,FeO,. T.S. P.INORGANIC CHEMISTRY. ii. 37 Hexahydrated Glucinum Sulphate. MARIO LEVI-MALVANO (Gaxzetta 1909 39 ii 438-440. Compare Abstr. 1906 ii 165).- The author gives a new example of crystallisation of one salt under the catalytic influence of another dissolved salt. When prepared in the ordinary way hexahydrated glucinum sulphate forms a mass of crystals incapable of being measured ; but when a solution containing equimolecular proportions of gluciuum and potassium sulphates is concentrated on the water-bath and allowed to cool it deposits large crystals of hexahydrated glucinum sulphate belonging to the cubic system [ZAMBONINI] and melting partly at 78-80° and completely at 95-96'.T. H. P. Bmic Magnesium Chlorides. WILLIAM 0. ROBINSON and W. H. WAGGAMAN (J. Physical Chem. 1909 13 673-678).-Solutions con- taining from 2.36 to 34.23% of magnesium chloride were shaken with small amounts of magnesium oxide for six months at a temperature of 25O. A t the end of this time the residues had become homogeneous and the solutions constant in composition. From an examination of the diagram on which the solubility data are plotted the conclusion is drawn that the solid substance in equilibrium with solutions contain- ing less than about 10% of magnesium chloride is an indefinite solid solution whereas the solid residue in contact with more concentrated solutions is a basic salt of the composition 3MgO,MgC1,,1 OH,O. Microscopic examination showed that this consists of very small acicular crystals.H. M. D. Zinc Amalgams. ERNST COHEN and KATSUJI INOUYE (Chem. Weekblad 1909 S 921-930. Compare Roozeboom and Byl Abstr. 1901 ii 507; Kerp and Bottger Abstr. 1900 ii 656; Pushin Abstr. 1903 ii 212 ; Hulett Abstr. 1900 ii 543).-The authors have investigated the solubility of zinc in mercury at temperatures between 0" and looo and found that it increases. Kerp and Bottger's method gives erroneous results especially between 20' and 100'. Pushin's results are also incorrect. A. J. W. Double Fluorides of Univalent Thallium. FRITZ EPHRAIM and LEON~D HEYMANN (Ber. 1909 42 4456-4463).-That?lous manganosomunganic Juoride 5TlF 21SlnF3,MnF2 obtained by pre- cipitating manganous acetate with ammonia and hydrogen peroxide dissolving the well-washed precipitate in hydrofluoric acid and adding thallous fluoride forms Clare t-coloured prisms decomposed by water but soluble in cold concentrated sulphuric acid or in dilute oxalic or tartaric acids to violet soliztions decolorised on heating. Attempts to prepare the manganic compound result in the formation of this salt.The three antimony compounds TIF,SbF,; T1F,2SbF3 and T1F,3SbF3 are all crystalline. Thallous fluoride and an excess of asolution of vanadism pentoxide in hydrofluoric acid yield minute insoluble crystals of the salt 3T1F,2V02F. By previously reducing the vanadium solution with sulphur dioxide,ii. 38 ABSTRACTS OF CHEMICAL PAPERS. small green crystals of the salt 2’1’1F,VOF are obtained whilst reduction of the pentoxide with hydrogen followed by solution in bydrofluoric acid and addition of thallous fluoride leads to the formation of green crystals of T1F,VF3,2H,O or of 2T1F,VF,,H20 according t o the proportions taken.The thallous tcmatulum compound 2T1F,TaF5 forms bright glisten- ing crystals. Three tungsten CompoundF 2T1F WO,F 3TlF,2 WO,F and TlF,WO,F are obtained by adding different proportions -of thallous fluoride to a solution of tungstic acid in hydrofluoric acid. C. H. D. Conditions which Determine the Composition of Electro deposited Alloys. Part I. Copper-Zinc Alloys. SAMUEL FIELD (Trans. Paruday SOC. 1909 5 172-194).-To trace the effect of varying conditions on the composition of electro-deposited alloys the composition of electro-deposited brass obtained from a cyanide solution under different conditions of (a) cornposition of solution ( b ) strength of solution (c) temperature (d) current density and (e) presence of free cyanide was determined.The conditions for quantitative deposition from cyanide solutions of the metals separately mere first found and the solutions were then mixed. It was found that with a solutisn containing about equal quantities of the two salts in the absence of any notable amount of free cyanide ( a ) Copper is the more readily deposited. (6) The percentage of zinc increases with the current density and also as the amount of zinc compound is increased. (c) Even with a large excess ofs zinc i n the electrolyte deposits containing a fair proportion of copper are readily obtained ( d ) Dilution raises the percentage of zinc because of the higher E.M.F. necessary to maintain the same current density. (6) Rise in tempera- ture increases the proportion of copper deposited. (f) With appreciable amounts of free cyanide the percentage of copper is always high even with high current density. Free cyanide does not increase the conductivity of the solution to any great extent but it prevents the formation of insoluble single cyanides at the anode. Uniform deposition is not maintained in cold stationary solutions and the insoluble cyanides formed at the anode may completely insulate the plates. With a warm and moving solution uniform deposits may be obtained without the presence of much free cyanide the anodes dissolving freely. These cyanide solutions are subject to continual changes of composition which soon prevent the exact repetition of similar conditions.These changes are due to (2) differing proportions of copper and zinc dissolved at the anodes and precipitated at the cathodes and (2) to the different amounts OF cyanide absorbed or set free by the metals. The zinc compound used in the above experiments corresponded with the composition Zn(CN)2,KCN whilst the copper compound had the composition CuCN,KCN. This latter compound was obtained by saturating a hot solution of Kahlbaum’s potassium cuprocyanide of the composition 3CuCN,KCN + 5ECN with freshly precipitated copper carbouate and then allowing it to crystallise T. S. P,INORGANIC CHEMISTRY. ii. 39 The Corrosion of Iron. J. NEWT~N FRIEND (J.Iron and &eel Inst. 1909 Reprint 3-7. Compare Abstr. 1908 ii 698; Proc. 1909 25 90).-A criticism of recent experiments on the rusting of iron. The presence of an acid is necessary to the initiation of rusting. C. H. D. The Action of Air and Steam on Pure Iron. J. NEWTON FRIEND (J. Iron and Steel Inst. 1909 Reprint 2-11. Compare preceding abstract).-At temperatures above 200' pure iron oxidises in air dried by means of phosphoric oxide. Steam carefully freed from carbon dioxide is without action on iron a t 100-250'. Slight action begins at 330° and is rapid at 445'. Traces of air may have been present. ExperimeEts in silica tubes indicate that the tem- perature at which action begins is near 350'. The action of steam probably takes place in two stages the steam being first dissociated and the oxygen thus set free acting on the iron.An estimate is made of the dissociation pressure of iron oxide. 0. H. D. Reduction of Iron Oxide by Hydrogen and Carbon Monoxide. SIEGFRIED HILPERT (Bey. 1909 42 4575-4581).- The temperature at which the reduction of ferric oxide by hydrogen begins depends on the temperature t o which the'oxide has been previously heated. A sudden fall in the reducibility of oxide pre- pared From the hydroxide occurs at 900° at which point there are indications of a polymorphic change. Oxide prepared from the oxalate only shows this change at 1000'. In both cases the final product behaves like natural hEmatite and is only reduced from 330' onwards. It is impossible to obtain a product containing a deter- mined quantity of oxygen by these means as the reaction takes place unequally in different parts of the mass.Carbon monoxide is active even at 240° but the resulting oxide is impure containing carbon. C. H. D. Chromi-aquo-triammines. E. H. RIESENFELD and F. SEEMANN (Ber. 1909 42 4222-4232. Compare Abstr. 1906 ii 760)-If chromtetroxide-triarnmine is added in small quantities at a time to a cooled dilute solution of hydrochloric acid a violent reaction takes place with evolution of oxygen and chlorine and formation of a red solution. Addition of concentrated hydrochloric acid to this solution produces after a long time a precipitate of dichloroaquotriammine- chromichloride [Cl,Cr OH C1 in the form of reddish-violet di- chroitic crystals which are soluble in water to a blue solution. If concentrated hydrochloric acid is used instead of the dilute acid a light green solution is obtained from which grey needle-shaped crystals of a second modification of the above chloride separate which is scarcely soluble in cold water but dissolves in hot water to a red solution.A third modification is obtained by heating a hydrochloric acid solution of the first chloride for some time at 60'. The solution gradually becomes green in colour and on evaporation in a desiccator dark green crystals are obtained which dissolve in water to a green (NH3)3 Iii. 40 ABSTRACTS OF CHEMICAL PAPERS. solution. is therefore probably contained in the complex in all three salts. all probability they are stereoisomerides Neither of these three chlorides readily loses water which I u N H NH3 C1 I NH C1 ~ I NH C1 I NH NH3 1- / Cr / C1 I H,O I / I br 1 Crr 1 R,O 1 NH H,O I C1 C1 NH NH3 A t present it is impossible to say which formula corresponds with which isomeride.If the aqueous solution of either of the three isomeric chlorides is heated to the boiling point the colour changes to a violet-red and on careful evaporation violet -red crystals of the chlorodiaquot riammine- chromichloride [ (NH,),Cr(OC~2)2]C12 are obtained. These crystals are readily soluble in water to a red solution from which by the addition of concentrated sulphuric acid and precipitation with alcohol pale violet hygroscopic crystals of the chlorodiaquotriamminechromi- sulphate (NHJ3Cr(OCE2)2 SO are obtained. Chromtetroxide-triammine dissolves in concentrated nitric acid to a red solution from which rose-red crystals of nitro-diaquotriammine- chrominitrate [(N H3)3Cr(*H,),] (NO,) separate.Triaquotriamminechromichloride [(NH,),Cr(H,O),]CI is obtained by dissolving chromtetroxy triammine in dilute hydrochloric acid and passing hydrogen chloride into the resulting solution. After a time bright red needle-shaped crystals of the triaquochloride separate. I f the hydrogen chloride is passed in for too long a time the solution becomes green in colour and crystals either of the dichroitic or of the grey monoaquochloride are obtained. If the aqueous solution of the triaquochloride is treated with concentrated nitric acid deep red crystals of the triaquotriammine-chromichloride nitrate - I / I / 1 --_I- [ 1 NO3 [ (NH,)@(OH,),] :& are obtained. Each member of the series of the chromi-aquo-triammines with the exception of the first [Cl,Cr(NH,),] has thus been obtained.The separate membecs are characterised by their different colours not only as solids but also in aqueous solution and it is noteworthy that the colours are exactly analogous to those of the corresponding cobalt compounds as shown by the following table M = CO. [(NH,\,M,H,O] ...... Bluish-green solution ,..... Bluish-green solution. [( NH3),M,2H,0] . . . . . . Reddish-blue . . . . . Violet-red [(NH,),M,3H20] . . . . . . Purple . . . . . . Piirple $ 2 This similarity ‘in colour is all the more remarkable since the ordinary cobalt and chromium salts are so different in colour. The conclusion is therefore drawn that the colour of a salt depends more M = Cr.INORGANIC CHEMISTRY.ii. 41 on its constitution than on the metal which forms part of the cation. Assuming that the atoms in the molecule are bound together by electrical forces the constitution of the molecule will depend on its electron-content and on the distribution of the electrons. It follows that absorption (of light) electron-content and the distribution of electrons are all closely connected a conclusion which has been arrived at in quite another way from optical considerations. T. S. P. The Isomerism of the Stannic Acids. WERNER MECKLEXBURGH (Zeitsch. anorg. Chem 1909 64 368-374).-The two modifications of stannic acid are best regarded as colloidal substances differing in the size of their particles as suggested by van Bemmelen (Abstr.1888 1160; 1905 ii 461). The greater adsorptive power and reactivity of the a-acid indicates that it has the finer structure. The absence of any direct relation of the /3-acid to the crystalloid compounds of tin is evidence of its coarser structure. C. H. D. Double Halogenides of Ter- Quadri- and Quinque-valent Antimony. FRITZ EPHRAIM and S. WEINBERG (Ber. 1909 42 4447-4456).-Derivatives of quadrivalent antimony have n great tendency to decompose into mixtures of compounds of ter- and quin- que-valent antimony with the exception of the thallium salt T1C1,,TlC1,2SbC14 which is stable (Ephraim and Barteczko Abstr. 1909 ii 236). The equilibrium SbCl,+SbCl =1 2SbC1 is greatly dependent on the temperature and on the possibility of ionisation the addition of sulphuric acid or of salts favouring the tetrachloride.On adding solid ammonium chloride to a fused mixture of antimony tri- and penta-chlorides the liquid becomes black and solidifies on cooling to a violet mass which slowly loses its colour at the ordinary temperature ultimately becoming white. The dark metastable salt is best obtained by pouring the hot mixture into chloroform; it then remains for some hours without change. The ammonium compound of antimony tetrahromide (NH,),SbBr prepared by adding the requisite quantity of bromine followed by ammonium bromide to a solution of antimony tribromide in [con- centrated hydrobromic acid forms black octahedra stable in air. The free acid was obtained in a state of doubtful purity.A ferric ammonium andimony chloride 9NH,C1,2FeC1,,3SbC14 is obtained by mixing the chlorides in concentrated hydrochloric acid and forms black octahedra. It may be regarded as 3(NH,),SbCl,+ 3NH,Cl 2FeC1,. Compounds of antimony tri- and penta-halogenides with salts of alkylamines are also described. SbCl,,NH,MeCl and SbCI,,NH,MeCl form very large colourless prisms and micro- scopic crystals respectively. The bromide 2Sbsr,,3NH3MeBr,3H20 forms lemon-yellow six-sided leaflets and the iodidey 2Sb13,3 N H,MeI 5H,O forms yellowish-red leaflets with golden reflex. The methylamine compounds,ii. 42 ABSTRACTS OF CHEMICAL PAPERS The ethylanaine compounds 2SbBr3,3NH3EtBr a yellow salt containing water ; SbBr,,NH,EtBr dark red leaflets and are described. 25 bI 3NH,EtI The following diethylamine compounds are described SbCl,,NHEt HCI long transparent neodles ; S bCl,,NHEt HC1; SbBr,,NHEt,,HBr ; SbEr5,2NHEt,,HBr black crystals violet in thin fragments ; and SbBr,,NHEt,,HBr garnet-red rectangular crystals.C. H. D. Effect of Ferric and Cupric Salt Solutions on Gold. W. J. BICCAUGHEY (J. Amer. Chem. Soc. 1909 31 1261-1270).-In the analysis of gold bullion containing a large proportion of tin it was found that the gold could be most conveniently separated by pre- cipitation with ferrous sulphate. The precipitate however still contained tin and if sufficient hydrochloric acid was present to keep all the tin in solution the gold was not completely precipitated. Preliminary experiments having shown that gold is soluble in solutions of both ferric and stannic salts in presence of hydrochloric acid an investigation was carried out to ascertain the effect of the concentration of the salt and acid and the presence of ferrous sulpbate on the solubility.The solvent action of cupric chloride was also studied. The experiments were conducted a t 38-43' and at 98-100'. The results are tabulated and plotted as curves. It has been found that gold is soluble in solutions of iron alum and cupric chloride containing hydrochloric acid the solubility increasing with the concentration of the acid or salt. An increase in the con- centration of the acid has a greater effect on the solubility than an increase in the concentration of the salt and particularly in the case of the copper salt. The rate of solubility in the presence of iron alum is eleven times greater at 98-100' than a t 38-43' and in the presence of cupric chloride is 32 times greater at the higher tern- perature.By doubling the concentration of the acid the solvent action of cupric chlofide solution is increased seven times at 38-43' and five times at 98-looo. Iron alum is capable of dissolving gold even in presence of a ferrous salt but the solvent action decreases as the concentration of the ferrous salt increases. In the precipitation of gold by ferrous sulphate it is therefore advisable to use considerable excess of the reagent and to allow the solution to cool before filtering it since the solvent action of ferric salts is much greater at the higher temperatures. The solubility of gold in solutions of ferric salts gradually decreases with time and approaches a limit but in the case of cupric chloride the amount of gold dissolved is directly proportional to the time.E. G. Gold Hydrosols. CARL THOMAE (J. pr. Chim. 1909 [ii] 80 518-520).-Zsigmondy's red-gold hydrosol which can be kept for a year is prepared with water distilled repeatedly through a silver condenser. However a hydrosol which is tenable for four months can be obtained with ordinary distilled water as follows. The water 120 c.c. is brought to the boil in a Jena qask and 2.5 C.C. of goldINORGANIC CLZEMISTRY. ii. 43 chloride solution (1 gram of crystallised salt 167) and 3.2 C.C. of potassium carbonate (2.5 100) are added successively. Whilst the water is boiling vigorously and the flask is being violently shaken 5.0 C.C.of formaldehyde solution (3 grams of commercial approxi- mately 36% formalin 100) are added drop by drop. If the first drop causes a separation of gold the addition is stopped and the boiling is continued until the gold has disappeared. Gold should not separate until many drops of the formaldehyde have been added; then the formation of a slight gold mirror is immaterial. The liquid is blue at first and finally dark red. It is decanted into another flask. The formation of a filminess or turbidity indicates that the reducing agent has been added too quickly or in too large quantities. c. s. Magnesium Aurides. G. G. URAZOFF (Zeitsch. anorg. Chem. 1909 64 375-396. Compare Vogel Abstr. 1909 ii 896).-Alloys of gold and magnesium are best prepared by heating the components in the proportions required to form the compound Aulslg in a graphite crucible enclosed in an iron cylinder with screw cap.Combination takes place readily at 700° and the compound may be fused quietly with further quantities of gold or magnesium under a layer of alkali chloride. Four compounds are indicated on the freezing-point curve AuMg with a maximum at 1150'; AuMgP with a maximum at 788'; Au,Mg at a break in the curve at 798O and AuMg with a third maximum at 818'. Gold foras solid solutions with nearly 30 atomic % Mg and the compound AuMg forms solid solutions ranging from 43 to 66 atomic % Au. The compound Au2Mg undergoes a transformation at 721'. Alloys rich in gold may be etched with hydrochloric acid and bromine ; those containing little gold are etched sufficiently by polishing while wet.The microscopical examination confirms the indications of the cooling curves. C. H. D. Atomic Weight of Platinum. EBENEZER H. ARCHIBALD (Proc. Boy. SOC. Edin. 1909,29 721-747).-From a consideration of earlier determinations the author draws the conclusion that the platinum salts analysed must have contained appreciable amounts of impurities for very divergent results are obtained when the weight of original salt is used in the calculation of the atomic weight. To obtain pure platinum the metal was precipitated in the form of ammonium platinichloride the precipitate being thoroughly washed and dried and then reduced in a current of pure hydrogen. After removal of ammonium chloride the platinum-black was boiled with successive portions of concentrated hydrochloric acid t o dissolve out traces of iron. The platinum was then redissolved and the above processes repeated several times. After three operations all indications of iridium bad disappeared.To avoid the difiiculty of removing the last traces of nitric acid from a solution prepared by dissolving platinum in aqua regia the metal was brought into solution by making it tbe anode in an electrolytic cell containing hydrochloric or hydrobromic acid. From such solutions $he potassium and affimonium salts of chloro- and bromo-platinic acidii. 44 ABSTRACTS OF CHEMICAL PAPERS. were prepared With these four salts four series of determinations were made by reduction in a current of pure dry hydrogen.Before reduction the potassium salts were heated at 400° and the ammonium salts at 175O for the purpose of expelling absorbed and occluded moisture. After reduction the platinum metal the halogen salt and the halogen acid given off were estimated. These measurements give several ratios from which the atomic weight of platinum can be deduced. Rejecting those in which the weight of original salt is concerned twelve values are obtained for the atomic weight the lowest of which is 195.21 and the highest 195.24. The mean value adopted is 195.23. Incidentally it is shown that the same results are obtained for the weight of reduced platinum whether this is heated and cooled in hydrogen and weighed at atmospheric pressure or heated cooled and weighed in a vacuum.(0 = 16.) H. M. D. Metallic Iridium Disulphates. - MARCEL DEL$PINE (Compt. rend. 1909 149 785-788; Bull. SOC. chim. 1909 [iv] 5 1084-1088 1126-1 133. Compare Abstr. 1909 ii 408).-The action of sulphuric acid on potassium iridochloride gives a blue solution which does not contain a salt of the same type as the ammonium iridium disulphates already described. The ammonium in the latter can however be wholly or partly replaced by potassium sodium thallium and barium. The iridium disulphates fall into‘ two series namely the green salts which are generally acidic derived from the acid HJJd SO4),(OH)H,O] designated H,A” and the reddish- brown salts which are basic derived from the acid H,[Ir(SO,),(OH),] designated H,A’“. These tautomeric acids are easily converted into one another by addition of excess of base and acid respectively. The following basic salts are described K7H2(AN),,6H20 crystsl- lising in needles varying in colour from old rose to blackish-brown according to thickness. This sait is prepared by pouring a cold solution of NH,*H,(A”) into excess of potassium carbonate solution.Na,H2(A’”),,6H2O and 9H,O reddish-brown clusters of needles or rectangular plates is obtained by precipitation with alcohol. TI,1(KH,)7H,(A’) and T17H,(A”’),,4$H,0 dark brown crystals are very slightly soluble. Ba7H4( A’”)s,Aq is a greenish-brown amorphous precipitate turning pure green on exposure to air and slowly depositing barium sulphate. The salts of the acid H,A” are all soluble in water but less soluble in presenceof alcohol ether or another salt of the same metal The following were prepared by a variety of methods from the corre- sponding ammonium.salts K,NH4H,(A”),,3H20 green to black needles according to size. K,H,(A) 1 &H,O needles. I.,H ( A”’),,3H20 hexagonal or octahedral crystals. K,(A”),H,O opaque square crystals very dark green ; their solution is alkaline and of an impure green colour. K2H,(A”),,6H20 tetrahedra stable only in strong acids. Na8*NH,*H,(A”) 1 8H,O very soluble black crystals almostMlNERALOGICAL CHEMISTRY ii. 46 rectangular. sisting of small green needles only slightly soluble in water. crystallising in long green needles or black rectangular crystals small opaque black tetrahedra. T17(NH,),H,(A”) to T17H,(A”),,6H20 a mixture cone Ba2H2(A”)3,1 2H,O Ba(~H4)2(A”)p q o R. J. C. Oamium. ALEXANDER GUTBLER and IE(. MAISCH (Bey. 1809,424 4339-4243).-In preparation for a revision of the atomic weight of osmium,,the authors have submitted the osmichlorides to a systematic investigation. Sodium osmichloride mas first prepared by heating a mixture of sodium chloride and osmium in a current of chlorine. If the osmium is in the form of granules the reaction takes place extremely slowly b u t when the metal is in a finely divided condition the reaction is complete in about half-an-hour a t a red heat. The sintered mass which is thus obtained was dissolved in cold dilute hydrochloric acid the solution filtered and the filtrate saturated with hydrogen chloride to precipitate any sodium chloride present. After further filtration the sodium osmichloride was obtained by careful evaporation of the filtrate; this salt could not however be prepared in a quite pure condition. It was therefore used to prepare the oamichlorides of ammonium potassium rubidium and caesium by double decomposition with the chlorides of these metals. All these compounds are much less soluble than sodium osmichloride and after recry stallisation from dilute hydrochloric acid they are obtained as dark-coloured well-defined octahedral crystals which are stable in dry a i r ; on being powdered they each give a bright red powder. They are soluble in cold water but the solutions decompose on exposure to the air after a time depositing a black powder. They are easily soluble in dilute hydrochloric acid giving stablo solutions. With the exception of the sodium salt all these compounds are anhydrous. T. S. P.