ii. 44 ABSTRACTS OF CHEMICAL PAPERS. Inorganic Chemistry. Molecular Volumes Physical Properties and Molecular Models of the Halogens. I?. A. HENGLEIN (2. anorg. Chem. 1921 118 165-171).-1t has been shown by Biltz (A. 1921 ii 437 487) that there is a linear relation between the molecular volumes of the halogens and of their compounds. It is now shown that there is likewise a linear relation between the atomic volumes of the halogens at their boiling points and others of their physical properties including melting point boiling point critical point latent heat of fusion or of evaporation and normal potential; also between the molecular volumes of the halogen acids a t the boiling point and many other of their physical constants. Owing to the close similarity in structure of the different halogen atoms their properties are determined principally by the atomic radius.In contrast with most groups of the periodic system the molecule of a halogen element has an especially simple structure; it can be represented diagrammatically by two cubes having a common edge with the valence electrons situated at the corners. E. H. R. The Catalytic Formation of Hydrogen Chloride from Hydrogen and Chlorine without Explosion. BERNHARD NEUMANN [with BERGDAHL BROY and KARWAT] (2. angew. Chem. 1921,34,613-620).-All non-explosion methods for the synthesis of hydrogen chloride gas yield a product contaminated with chlorine. Hoppe who first suggested the use of metallic chlorides as catalysts to this reaction employed the chlorides of aluminium tin and zinc in concentrated solution a t a temperature of 130".A repetition of his work has shown that under the most favourable conditions and with a very slow current of gas not more than 70% conversion is obtained. The reaction is favoured by an increase of concen- tration and of temperature but a practical limit is set to the latter by the point a t which evaporation becomes rapid. The author has used solid chlorides and higher temperatures. The hydrogen and chlorine were generated electrolytically in separate cells in the same electric circuit passing into a mixing flask containing water and thence to a quartz tube in the shape of a pipette filledINORGANIC CHEMISTRY. ii. 45 with granules of quartz impregnated with the selected chloride and heated. The emerging gases were absorbed in a 10-bulb potash tube.The mixed gases were shielded from light and drawn through the apparatus by uniform slight suction fluctuation of pressure being a source of explosions. With a low gas velocity a complete conversion was effected by magnesium chloride a t 300" by calcium chloride a t 305" by aluminium chloride (which a t this temperature is almost entirely decomposed) a t 350" and by quartz unimpregnated by any salt a t 380". Conversion was improved by warming the mixing flask to 50° with the effect of adding 1 mol. of water to each mol. of hydrogen chloride this temperature being an optimum. The dilution of the gases with oxygen had at 380" no influence on the reaction. The reactions Cl2+H,O=C1OH+HC1 ClOH+H,=H,O+HCl are considered to occur removal of the hydrogen chloride generated being facilitated Sulphate-free Sulphites for Standard Sulphur Dioxide Solutions.S. LANTZ SHENEFIELD PRANK C. VILBRANDT and JAMES R. WITHROW (Chem. and Net. Eng. 1921 25 953-955).- Pure sodium sulphite heptahydrate was prepared by passing care- fully purified sulphur dioxide into a solution of sodium carbonate to saturation adding the requisite amount of sodium carbonate to transform the resultant sodium hydrogen sulphite into the normal sulphite and cooling the solution to 0" whereby a crystalline meal of the heptahydrate was obtained. All the operations were carried out in the absence of oxygen and the damp crystals were found to be free from sulphate. They were placed in a desiccator over sulphuric acid and sodium pyrogallate was used as oxygen absorbent.Although every precaution was taken to exclude air leakage the crystals after one week contained 7.52% of sodium sulphate deca- hydrate. Similar oxidation occurred in the preparation of pure dry calcium sulphite and the opinion is expressed that in both cases this is due to autoxidation of the salt. It is concluded that the validity of all investigations in the literature based on the preparation of sulphate-free sulphite for use as a sulphur dioxide standard is doubtful. Polythionic Acids and Polythionates. E. H. RIESENFELD and G. W. FELD (2. anorg. Chem. 1921 119 225-270).-A method has been devised for estimating the proportions of tri- tetra- and penta-thionates together in solution with sulphite thiosulphate and sulphate and the method has been applied to the study of the formation and stability of the polythionates.The hexathionic acid described by Debus (T. 1888 53 278) does not exist. The total polythionic acids can be determined by treat- ment in neutral solution with a mercuric salt when four equivalents of sulphuric acid are produced for each molecule of polythionate for instance 2S,O,"+ 2Hg" + 4H,O =2HgS+ 4SO,"+ 8H'+4S. The acid can then be titrated. Alternatively the polythionates can be oxidised with bromine in alkaline solution; the whole of the sulphur is oxidised to sulphate and is precipitated and weighed by the partly dissociated chlorides. c. I. A. R. P.ii 46 ABSTRACTS OF CHEMICAL PAPERS. as barium sulphate. The trithionate is estimated by boiling the solution with an excess of copper sulphate when the following reaction occurs S306”+CU”+2H20=CUS+2~04”+4Ho The copper sulphide is filtered ignited and weighed as copper oxide.When boiled with excess of alkali the polythionates form thio- sulphate and sulphite thus 2S306”+60H’=~20,”+4~03’’+ 5X20,”+ 3H,O. The sulphite and thiosulphate formed are estim- ated by titration with iodine. By applying three of these reactions data can be obtained from which to calculate the proportions of tri- tetra- and penta-thionate present. The interaction between hydrogen sulphide and sulphur dioxide was studied by leading a current of the former as gas into an aqueous solution of the latter a t O” until a definite ratio of the reacting substances was present in solution. The total poly- thionate was estimated after sixty hours. The optimum ratio for polythionate formation was 2S0 1H2S.On the other hand with the ratio 2H2S SO all the sulphur was precipitated in the elementary form. Evidence was obtained that immediately after the prepar- ation of a solution in the ratio 2SO,:H,S an intermediate com- pound is formed which can be precipitated a t a low temperature as the barium salt; in solution this changes to barium thiosulphate. The proportion of tri- and tetra-thionic acids formed (determined after fourteen days) depends on the sulphur dioxide concentration the formation of tetrathionic acid being favoured by low concentration. The proportion of pentathionic acid is however practically con- stant. These results are contrary to those obtained by Heinze (A.1919 ii 334). The sulphuric acid which is always formed reaches a maximum near the commencement of the reaction; it must therefore he formed from the intermediate compound not by oxidation of the polythionic acids. Of the three polythionic acids the tetrathionic acid is the least stable and decomposes relatively quickly into tri- and penta- thionic acids. The trithionic acid decomposes more slowly with formation of sulphur dioxide whilst pentathionic acid decomposes only in the course of months with separation of sulphur. The order of stability is the same in neutral as in acid solution; in alkaline solution all the polythionates decompose quickly into thiosulphate and sulphite. The phenomena observed are explained on the assumption that the above intermediate compound is a hydrate of the unknown sulphur monoxide SO.This is stable in acid solution for a time but in neutral or alkaline solution quickly forms thiosulphate. In acid solution it slowly polymerises to pentathionic acid. By combination with sulphurous acid it forms tri- and tetra-thionic acid 3SO+H,S0,=.H,S,06 ; S0+2S02+H20=H2S306. By hydrogen sulphide it is reduced to sulphur. 3H20 ; 2S406”+60H’=3S20,”+2S0,”f3H20 ; 2S50,+60H’= E. H. R. The Preparation of Hydrogen Selenide from Metallic Selenides. L. MOSER and E. DOCTOR (2. anorg. Chem. 1921 118 284-292) .-The selenides of magnesium aluminium iron,INORGANIC CHEMISTRY. ii. 47 and zinc were prepared in a similar manner to the corresponding tellurides (this vol. ii 48) by passing selenium vapour over the heated metal in a vacuum.The aluminium and magnesium compounds were also prepared by direct combination of the metal with selenium in a crucible starting the reaction between the mixed components with a burning magnesium wire. All the products were considerably contaminated with metal except aluminium selenide which prepared by the latter process was practically pure. Aluminium selenide Al,Se forms a light brown powder unstable in air and magnesium selenide MgSe is very similar. Zinc selenide ZnSe is citron-yellow and iron selenide FeSe is black and metallic; both are stable in air. Hydrogen selenide was prepared in an apparatus similar to that used for hydrogen telluride by dropping the metallic selenide slowly into acid. The best results were obtained using aluminium or magnesium selenide.The gas was liquefied at the temperature of a solid carbon dioxide-ether mixture and by revaporisation was obtained pure. It is not decomposed by daylight in the liquid or gaseous form but is sensitive to ultra-violet light. Dry oxygen has no action on the dry gas but in presence of moisture decomposition is rapid. E. H. R. Selenious Acid and Heteropolyselenites. ARTHUR ROSEN- HEIM and LEONHARD KRAUSE (2. anorg. Chem. 1921 118 177- 192).-A large number of heteropolyselenites with vanadates and molybdates have been described by Prandtl and others (A. 1907 ii 477; 1912 ii 167; 1916 ii 333) who described numerous well- crystallised salts which however had variable compositions according to the conditions of preparation. These compounds have been re-examined with a view to determine whether they may not have a semi-colloidal constitution similar to that of the periodates (A.1919 ii 508). Experiments were first made to determine the state of aggregation of selenious acid in aqueous solution. Depression of the freezing point of water indicated slight association which has a tendency to decrease with time. The dissociation was determined by the hydrogen-ion concentration method and from the electrical conductivity the results obtained being 4.85. and 3.45. respectively. These results are taken to be consistent with the presence of associated (H,SeO,) mole- cules in solution. The method of estimating selenious acid by heating with potass- ium iodide and hydrochloric acid and distilling the iodine over into potassium iodide was improved by the addition of phosphoric acid to the hydrochloric acid.This prevents the formation of selenium iodide which may be the cause of low results. It was also found that selenious acid may be accurately titrated with sodium hydroxide; the best indicators are for the formation of NaHSeO p-nitrophenol and for complete neutralisation to Na,SeO thymolphthalein. Lithium selenite forms the hydrate 4Li,Se04,3H,0 ; ibs solubility has a negative temperature coefficient.ii. 48 ABSTRACTS OB CHEMICAL PAPERS. MoZybdoseZen~tes.-The composition of molybdoselenites was found to depend on the ratio of molybdate to selenious acid in the solution from which they were precipitated. When less than 1 mol. of selenium dioxide was present to 1 mol.of molybdate the potassium and barium salts corresponded with 2R20,2SeO2,5MoO,xH2O and the ammonium salt with 3(NH,),0,2Se0,,8M003,6H20. With more than 1 mol. of selenium dioxide per mol. of molybdate in solution salts were obtained in which the proportion of base was variable but the ratio SeO MOO was always very nearly 1 1. The ammonium salt 2(NH,),O,5Seb.,,5MoO3,8H2O white microscopic prisms was obtained by adding 15 mols. of selenium dioxide to a saturated solution of ammonium paramolybdate. K20,2Se0,,2M00,,3~5H20 forms microscopic prisms and the barium salt BaO ,2Se02,2Mo03,7H,0 a white crystalline precipitate. VunadioseZenites.-Vanadioselenious acid has the! composition 4SeO2,3V,Os 10H,O. An extensive series of experiments showed that as the concentration of selenious acid in the mother-liquor increased from zero to 5N the ratio of SeO to V,O in the solid phase increased from 4 3 to 5.5 3.Similar behaviour was shown by the ammonium vanadioselenites having the approximate formula 3(NH,),O 12Se0,,8V20,. It is concluded that the variable composition of the vanadic acid compounds is due to the formation of adsorption compounds although in the case of the molybdic acid compounds this is not no clearly demonstrated. The potassium salt E. H. R. The Preparation of Hydrogen Telluride from Metal Tellurides. L. MOSER and K. ERTL (2. anurg. Chem. 1921 118 269-283) .-A new method for preparing metal tellurides was devised which consisted in distilling tellurium at a low pressure (8 mm.) over the hot finely divided metal. In this way the tellurides of magnesium MgTe aluminium Al,Te iron FeTe and zinc ZnTe were prepared.The aluminium compound was obtained in an almost pure state; it is a blackish-brown lustrous amorphous substance decomposing in the air with formation of tellurium hydride. The other tellurides were all more or less contaminated with excess of the respective metal. Magnesium telluride forms a brown sintered mass iron telluride is grey and metallic and zinc telluride is pale brown. The last two are stable in air. For the preparation of hydrogen telluride a special apparatus was designed in which the powdered metallic telluride was dropped very gradually into acid in an atmosphere of nitrogen. Aluminium telluride proved the most suitable substance from which to generate the gas and hydrochloric acid the best acid to use.The yield of gas obtained under the best conditions was more than 80% of the theoretical. The gas was liquefied by passing through a tube immersed in a mixture of solid carbon dioxide and ether. In the liquid state hydrogen telluride is very sensitive both to daylight and ultra-violet light but the dry gas is quite stable in light. The dry gas is however immediately oxidised by oxygen. E. H. R.INORGANIC CHEMISTRY. ii. 49 The Preparation of Telluric Acid. JULIUS MEYER and HANNS MOLDENHAUER (2. anorg. Chern. 1921 119 132-134).- Telluric acid can be prepared in a pure state and in almost theoretical yield by oxidation of tellurium tetrachloride with chloric acid. Tellurium (10 grams) is boiled with 10 C.C.of nitric acid and 3 C.C. of hydrochloric acid until completely dissolved. To the hot solution is added gradually a concentrated solution of 9 grams of chloric acid and the solution is boiled until no more chlorine is evolved. A slight excess of chloric acid is added to avoid forma- tion of any explosible chlorine oxide. The solution is filtered through asbestos and concentrated by distillation in a vacuum on the water-bath thereby removing chlorine. The telluric acid can be crystallised out by the addition of concentrated nitric acid collected and finally freed from chlorine and nitrogen oxides by drying in a vacuum. It is obtained as a crystalline snow- white powder readily soluble in water having the composition H 6TeO 6. R. 0.E. DAVIS L. B. OLMSTEAD and F. 0. LUNDSTRUM ( J . Amer. Chem. SOC. 1921 43 1580-1583; cf. this vol. ii 56).-Vapour pressure curves have been constructed for the following solutions Ca(N0 ) 22*48% NH 19-18y0 H,O 58.34%; NaI 32-34y0 NH 16.062 H20 51.60% ; NH,*CNS 77.84y0 NH 22.16% ; NH,NO 33.7% NH 18+i2~0 H,O 47.48%; CaC1 12-9y0 NH 22.9y0 H,O 64.2%; Ca(N0,)2 55-8y0 NH 25-77y0 H,O 18.43%; NH 28.15% H,O 71.85%; NaI 64.88% NH 26.02% H,O 8.20% over the temperature range -16" to 40". It is shown that solutions of ammonium nitrate in ammonia and ammonium thiocyanate are very corrosive to iron and steel the calcium chloride-ammonia solution is less corrosive and calcium nitrate- ammonia and sodium iodide-ammonia solutions show no immediate corrosive action. Calcium nitrateammonia solutions seem to be the most promising of these solutions for practical use as an absorbent for ammonia in the synthetic ammonia process.The Action of Metals such as Copper and Zinc on an Aqueous Solution of Ammonium Nitrite. N. R. DHAR (2. anorg. Chem. 1921 119 176176).-The action of a solution containing ammonium nitrite on copper was attributed by Hof- mann and Buhk (A. 1921 ii 43) to the hydrolysis of the nitrite with formation of free nitrous acid. The observation that the metal is attacked even in presence of urea however renders this explanation improbable. Further solutions of other nitrites such as zinc nitrite which are hydrolysed as much as ammonium nitrite do not attack copper. The activity of the ammonium nitrite is probably related to its instability and ready decomposition into nitrous oxide and water. The Structure of Pyrophosphoric Acid.111. D. BALAREFF (2. anorg. Chem. 1921,118 123-130; cf. A. 1915 ii 446; 1917 ii 467).-In previous papers it has been shown that there is a good E. H. R. Vapour Pressures of Ammonia-Salt Solutions. J. F. S. E. H. R.ii. 50 ABSTRACTS OF CHEMICAL PAPERS. deal of evidence in favour of the unsymmetrical structure of pyro- phosphoric acid. The synthesis of the pyro-acid by combination of the ortho- and meta-acids in sulphuric acid would give added support to the hypothesis of an unsymmetrical structure but attempts in this direction were not successful even in acid con- taining 15% of free sulphuric anhydride. The change of colour of the salt NaAg,P,O from white to yellow on heating has been attributed to its decomposition into NaPO and Ag,PO but it is now shown that this does not occur the colour change being probably due to some physical change in the salt.No evidence as to the structure of the pyro-acid could be gained from a study of the dehydration of dihydrogen phosphates of alkali metals. The potassium salt heated a t 244" loses water and changes to the acid pyrophosphate K,H,P,O,; the rubidium salt behaves in the same way. The sodium salt loses water very slowly at 180" and in the course of about one hundred and seventy-eight hours becomes completely converted into Na,H,P,O,. The products of further dehydration a t a higher temperature depend on the water vapour pressure. In moist air a t 305" only soluble meta- phosphate is formed whilst in dry air at 330" about 75% of the metaphosphate formed is insoluble.Phosphoryl bromide dehydrates orthophosphoric acid to pyrophosphoric acid but not to the meta-acid. The action is a complex one and depends on the temperature and proportions of the interacting substances. A dilute solution of an alkali pyrophosphate after prolonged boiling shows the presence of orthophosphate proving that hydra- tion occurs slowly. E. H. R. Iso- and Heteropoly-acids. XVII. Polyborates in Aqueous Solution. ARTHUR ROSENHEIM and FELIX LEYSER (2. anorg. Chem. 1921,119 1-38).-An attempt was made to prepare simple and complex polyborates wit,h the object of comparing these with salts of other acids such as telluric antimonic periodic plumbic and stannic acids which show semi-colloidal properties.Methods for the quantitative estimation of boric acid were examined. The polarimetric method depending on the influence of boric acid on the optical rotation of tartaric acid is of limited application on account of the disturbing influence of salts or other substances present in solution. Titration with sodium hydroxide in presence of mannitol using phenolphthalein as indicator gives trustworthy results. Free boric acid in the presence of borate can be detected by boiling a sample of the substance for some minutes with dry acetone filtering evaporating the filtrate on a watch glass moisten- ing with a few drops of methyl alcohol and igniting when the characteristic flame coloration is given if free boric acid is present.An investigation of the equilibrium in the system Na,O-B,O,- H20 a t 0" confirmed the existence of the three salts Na20,B20,,8H,0 Na20,2B20,,10H20 and Na20,5B20,,10H20. Sodium penta- borate can readily be crystallised from solutions containing Na,OINORGANIC CHEMISTRY. ii. 51 and B,O in the ratio 1 5 but sometimes only crystalrJ of borax are obtained probably because the pentaborate is metastable a t ordinary temperatures and borax is the less soluble salt. Potassium pentaborate K20,5B20,,8H,0 is a well-defined characteristic salt separating from solutions in which the ratio B,O, KOH is 3 1 or higher. Its solubility is very low not much greater than that of potassium perchlorate. Potassium monoborate crystallises with 8H20 a t O" with 2.5H20 a t 30". Rubidium pentaborate is very similar to the potassium salt but crystallises with 10H,O.Thallium pentaborate like the potassium salt crystallises with 8H20 but is more soluble than the latter. Guanidine forms a diborate crystallising in elongated prisms with 48,o and a penta- borate with 8H20. Experiments on the dehydration of pentaborates showed that in the general formula R20,5B20p,xH20 two molecules of water are probably constitutively combined. Conductivity experiments indicated that in dilute aqueous solution the pentaborate anion is hydrolysed into the diborate anion and boric acid. I n presence of great excess,of boric acid this hydrolysis is prevented and the specific conductivity of sodium pentaborate a t 0" appears to be 85% of that of sodium diborate. Experiments on the hydrogen- ion concentration of solutions containing varying ratios of NaOH to B2.03 confirm the existence of a pentaborate ion in concentrated solution.The pentaborate ion appears to form complex anions with a number of metals. Whilst borax solution immediately precipitates zinc or cadmium hydroxide from a solution of a salt of the metal sodium pentaborate does not. Cobaltous hydroxide dissolves in sodium pentaborate solution to form a red solution in which although alkalis do not readily precipitate it the cobalt is in the kation. When this solution is oxidised with hydrogen peroxide however some cobaltic oxide is precipitated and a yellow solution is formed containing a complex cobalt anion. Nickel chromium (Cr"') manganese (Mn") and copper also appear to form complex anions. A very small quantity of a copper compound was isolated having approximately the composition 2Na,0,4C~0,12B,0~,50H,O.E. H. R. The Atomic Weight of Carbon. E. MOLES (Anal. Pis. Quim. 1921 19 255-259).-The value 12-005 for the atomic weight of carbon given by Richards and Hoover (A. 1915 ii 96) is held to be based on an erroneous value for the atomic weight of sodium. The value 12.000 is claimed to be more exact. G . W. R. The Oxidising Properties of Carbon Suspensions. F. FEIGL (2. anorg. Chem. 1921 119 305-309).-The oxidising effect of blood charcoal was studied in a qualitative manner by boiling solutions of different oxidisable substances with a sus- pension of the charcoal. I n acid solution hydrogen sulphide was oxidised to sulphuric acid potassium iodide to iodine mercurous salts to mercuric oxalic acid to carbon dioxide.I n alkaline solu-ii. 52 ABSTRAaTS OF UHEMICAL PAPERS. tion potassium iodide was oxidised to iodate alkaline sulphidea and sulphites t o sulphate cuprous and cupric sulphides to copper sulphate cobalt sulphide to sulphate potassium chromite to chromate. Sodium thiosulphate was unacted on in alkaline or neutral solution and sodium nitrite was unaffected in alkaline solution. A quantitative study of the oxidation of tervalent chromium to chromate was made after a method had been devised for removing from the solution a product formed by the interaction of the charcoal and potassium hydroxide which liberates iodine from potassium iodide. This was accomplished by boiling with potassium permanganate and removing the excess with hydrogen peroxide.The experiments showed that the proportion of chromate formed increased with the proportion of charcoal used but that with a constant quantity of charcoal the amount of chromate formed increased with the quantity of chromium salt taken. Different charcoals varied widely in their oxidising power but the differences seemed to bear no relation to the ash content. E. H. R. Aqueous Carbonic Acid Solutions. E. WILRE (2. anorg. Chem. 1921 119 365-379).-The dissociation constant of carbon dioxide solutions was measured by the conductivity method using a solution through which the gas was being continuously circulated. When an ordinary saturated solution was used without circulation the conductivity was found to increase during measurement probably through electrolytic changes caused by the current.Even with the greatest precautions variable results were obtained confirming the observations of earlier workers. It was observed that by contact with the metal electrodes (gold) even without passage of current the conductivity gradually increased. In three hours the dissociation constant K . lo7 increased from 3.07 to 4.5. Light seemed to have an effect in the same direction. E'or measur- ing the hydrogen-ion concentration a special hydrogen electrode was used consisting of a palladium capillary into which hydrogen was forced a t a pressure of 20 atm. The hydrogen-ion concentration was determined in presence of sodium potassium and barium chloride.In these solutions carbon dioxide has the character of a strong acid increasing with the concentration of salt. The hydrogen-ion concentration increases more rapidly than the total carbon dioxide concentration. The observations can be explained on the assumption that a solution of carbon dioxide in water contains orthocarbonic acid H,CO which containing no ketonic oxygen is a very weak acid. In concentrated salt solutions it is dehydrated to form the strong acid CO(OH),. Behaviour of Amorphous Carbon and Sulphur at High Temperatures. Carbon Sulphides. J. P. WIBAUT (Proc. K . A h d . Wetensch. Amsterdam 192 1 24 92-101) .-The action of sulphur on amorphous carbon a t high temperatures has been investigated. Pure sugar charcoal has been heated with sulphur a t temperatures from 400" to 1000" under reduced pressure for E.H. R.INORGIANICl CHEMISTRY. ii. 53 prolonged periods of time. A slow evolution of hydrogen sulphide due to the amall amount of hydrogen present in the carbon is observed and a carbon-like substance containing 1*98y0 of sulphur obtained. This subsfance yields no sulphur to toluene even after prolonged boiling and the residue after this treatment contained 2°03y0 of sulphur. Prolonged heating in a vacuum a t temperatures up to 1010" did not reduce the sulphur content nor was any volatile compound obtained. Prolonged shaking with bromine water oxidised 9% of the sulphur to sulphuric acid and heating in a current of hydrogen at temperatures up to 750" removes 77% of the sulphur as hydrogen sulphide; this reaction is exceedingly slow and must be regarded as an action between a sulphur com- pound and hydrogen and not as an action between hydrogen and sulphur vapour.This was further proved by the fact that heating in nitrogen did not reduce the sulphur content. The author con- siders that a solid carbon sulphide is formed which bears a strong resemblance to coal coke (cf. Stock and Praetorius A. 1913 ii 46). A further sulphide containing 3.5% of sulphur has been obtained by heating carbon purified by chlorine with sulphur. This substance has similar properties to the compound containing 2-0y0 of sulphur. Tho Deviations from the Gas Laws of Carbon Disulphide. ALFRED SCHULZE (2. anorg. Chem. 1921,118,223-230).-A number of observations on the properties of carbon disulphide vapour indicate that it is associated to a small extent.The increase of pressure observed when carbon disulphide and ether vapours are mixed a t constant volume at 80" under atmospheric pressure indi- cates association of the former to the extent of o.14y0 whilst vapour density determinations by Dumas's method give results corresponding with 2 yo association. Compressibility experiments at 80" showed @5y0 more association a t 2 atmospheres than a t 1 atmosphere pressure. The PV curves at 78-82' and 130.48" show that the amount of association decreases with increasing tempera- ture but a t constant temperature increases with increasing pressure. It is probable that in the liquid phase association is more con- siderable. E. H. R. J. F. S. New Theory of the Constitution of Hydroxides particu- larly those of the Basic Metallic Oxides.FR. TIEMANN (Chem. Zeit. 1921 45 1125).-To furnish an explanation for a number of phenomena in organic inorganic and electrolytic pro- cesses which are not in consonance with existing ideas the author proposes a new theory of the constitution of the hydroxides of the pronounced electropositive metals. It is suggested that these compounds do not contain hydroxyl groups but are to be regarded merely as true hydrates of the corresponding oxides that is to say that sodium and calcium hydroxides for example are not correctly represented by the formulze NaOH and Ca(OH)2 but are actually Na20,H20 and CaO,H,O a molecule of water being closely associated with the metallic oxide rJimilarly to the " water ofii 54 ABSTRACTS OF CHEMICAL PAPERS crystallisation " of salts.This applies to all the elements of groups I and I1 of the periodic system whilst the constitution of the hydroxides of those of groups I11 and IV (aluminium zinc) will depend on the electrochemical conditions under which they are produced. Only the hydroxides of the metalloids and non-metals are to be regarded invariably as true hydroxyl derivatives. With increasing basicity of the oxides the associated water molecules become increasingly firmly bound exactly as in the case of the increasingly basic character of salts containing associated water. So the dehydration of the hydroxides of calcium strontium and barium is effected with increasing difficulty in the order named. The sucrosates are cited to illustrate the application of the theory.If calcium hydroxide is regarded as a hydroxyl compound the chemical character of sucrose or dextrose is quite incompatible with the idea of a " neutralisation " of hydroxide looked on as a generator of hydroxyl ions. There can therefore only be a question of the displacement of the associated water by the sugar and the sucrosates must be formulated C1,H,,O 11 ,2CaO C ,H 120.G ,CaO etc. The isomerism of the hydroxides of tin and aluminium is also explained by reference to the theory e. g. AI(OH) and A4120,,H20+2H20 can both exist as individual substances and either one or the other will be produced according to the conditions of the reaction. The non-appearance of hydrogen peroxide de- rived from the union of two hydroxyl groups during electrolysis of a metallic hydroxide is explained by the new theory as due to the absence of hydroxyl groups.The electrolysis actually is that of say Na,O,H,O the associated water taking no part in thc process. There takes place simply a direct fission into metlal and oxygen which are liberated a t their respective poles. An assump- tion of the appearance of hydroxyl ions in any electrolytic process is quite unjust'ified. G. F. M. Preparation of Alcoholic Potassium Hydroxide Volumetric Solution. S. 3'. MCCALLUM ( J . Ind. E72g. Chein. 1921 13 943).-A solution which does not darken in colour when kept is prepared by dissolving potassium hydroxide in methyl alcohol (purified mood spirit) ; the solution must be filtered through glass- wool to remove insoluble potassium carbonatc etc.before it is used. 1 . P. s. Existence of Tetra-hydrated Sodium Sulphate in Mixed Crystals with Sodium Chromate. THEODORE W. RICHARDS and W. RUELL MELDRUM ( J . Amer. C h m . Xoc. 1921 43 1543- 1545).-lt is shown that crystals of the tetrahydrate of sodium chromate Na2Cr0,,4€I,O dissolve sodium sulphate as Na2S0,,4H,0 a form of sodium sulphate otherwise unknown to the extent of somewhat less than half the quantity corresponding with the same weight of sodium chromate in the supernatant liquid. When sodium sulphate was in largc excess no crystallisation could be induced by " seeding " the saturated mixture with the crystals of the mixed tetrahydrate above the transition temperature oE sodiumINORGANIC CHEMISTRY.ii. 55 sulphate and below this temperature only crystals of the deca- hydrate could be formed. Thus under these conditions the tetra- hydrate is so much more soluble than the phases containing more sodium sulphate as to be incapable even of metastable existence. J. F. S. Ammonium Radicles. 111. Ammonium. HANS HETNRICH SCHLUBACH and FRITZ BALLAUF (Bey. 1921 54 [B] 2825-2834; cf. A. 1920 i 822 and this vol. i 16).-The authors' experience with tetraethylammonium leads them to expect that the ammonium radicle would be extremely sensitive to rise in temperature and that there is no hope of isolating it by the electrolyses of solutions of ammonium salts in liquid ammonia on account of the impos- sibility o€ avoiding the thermal effect of the current and that the only prospect of success lies in displacement reactions effected a t a low temperature.They find that when well-cooled ammonium chloride is added to a solution of potassium in liquid ammonia a t -70" in the apparatus described previously for the preparation of tetraethylammonium decolorisation of the solution takes place before the calculated volume of hydrogen has been evolved the deficit amounting to as much as 65%; according to Moissan the whole of the hydrogen is evolved by the time the solution becomes colourless. The deficit cannot be attributed to the solubility of hydrogen in liquid ammonia since this is shown to be too small to account for the observed effect and it appears therefore that colourless ammonium is actually present in the solution. This conclusion is supported by the observation that the remainder of the hydrogen is evolved rapidly when the solution is cautiously warmed a t about -40".Repetition of Moissan's experiment shows that the non-observation of the production of ammonium is due to operatioil in too concentrated solution and consequent decomposition of the radicle by the heat liberated during the reaction. When a solution of potassium (14%) is added to a solution of ammonium chloride (1%) in liquid ammonia a t -70° a violent reaction is observed and each drop of added solution is immediately decolorised formation of ammonium and its con- version into the colourless form appearing to occur instantaneously ; tetraethylammonium and ammonium therefore stand to one another in the same relationship as triphenylmethyl to methyl.In spite of the violence of the reaction the yields of ammonium by this method are good and readily reach 50% ; the influence of concentration is however again apparent and it is to be expected that an improvement in yield would bc observed with more dilute solutions. The behaviour of ammonium towards the reagents used with tetraethylammonium is described. Corresponding with the rapid isomerisation to the colourless form the equilibrium is here greatly displaced in the direction of the latter and it is probable that dissociation and consequent reaction only occur in close proximity to the temperature of decomposition. A reaction with dimethyl- pyrone could not be observed. Iodine on the other hand appears3. 56 ABSTRACTS OF CHEMICAL PAPERS.to react immediately with ammonium but the quantitative examination of this change could not be completed on account of the experimental difficulties involved. Vapour Pressure of the System Lithium Nitrate Ammonia. R. 0. E. DAVIS L. B. OLMSTEAD and F. 0. LUNDSTRUM ( J . Amer. Chem. Xoc. 1921 43 1575-1580).-The solution of ammonia in lithium nitrate has been studied with the object of finding an absorbent for ammonia in the synthetic production of this gas. The use of ammonium nitrate (Kurilov A. 1898 ii 156) and ammonium thiocyanate (Foote and Hunter A 1920 ii 246) suffers from the serious drawback that the liquids produced when these salts adsorb ammonia attack metals rapidly. A large number of salts have been tested as to their suitability for this purpose and of these lithium nitrate alone forms a liquid with ammonia in the absence of water whilst calcium nitrate tetrahydrate liquefies in the presence of a little water.The ammonia contained in 1 C.C. of the lithium nitrate solution saturated a t 24" is equivalent to 26.0 C.C. of 0.95N sulphuric acid whilst that for the calcium nitrate solution under identical conditions is 18.5 C.C. of 0.95N acid. Vapour pressure measurements have been made for the solution 36.34y0 ammonia 63.66% lithium nitrate and for several other mixtures containing 6.06-58-66~0 of water. The solutions of ammonia in lithium nitrate have no action on machine steel iron wire and nichrome wire after several months' contact but nickel steel shows a slight action after several months. The results show that a solution of lithium nitrate in ammonia with a small percentage of water should be a good absorbent for the removal of ammonia from mixtures of nitrogen hydrogen and ammonia.The absorption could be effected at 0" and a large proportion of the ammonia released either by a small increase of temperature or by reduction of the pressure. J. F. S. GEOFFREY ISHERWOOD HIGSON (T. 1921,119 2048-2055). J. H. REEDY ( J . Amer. Ohem. Xoc. 1921 43 1440-1445).-1n an earlier paper (A. 1915 ii 733) it was shown that the electrode Ag IAgBrO,! O*lNKBrO only reached a steady value (0.631 volt) after being kept for five days if the bromate was prepared by the action of bromine on silver nitrate solution but if it was obtained by double decomposition of silver nitrate and potassium bromate the correct value was a t once obtained.Investigation now shows that silver bromate is dimorphous existing as tetragonal bi-pyramids and as hair-like crystals. The tetragonal crystals are stable at temperatures below 98.5" (the transition point) and the fine hair-like crystals are stable above this temperature. Difference of solubility of the two forms explains the irregular behaviour of the electrode mentioned above. The solubility of silver bromate has been determined a t temperatures from 25" to 90" and the following values have been found 25" 0.196 ; 35" 0.269 ; 45" 0.371 ; 55" 0.497 ; 65" 0.648 ; 75" 0.832 ; 85" 1.055 H. W. The Reaction between Persulphates and Silver. Silver Bromate.INORGANIC CHEMISTRY. ii. 57 and go" 1.325 the solubilities being expressed in grams per 100 grams of water.The solubility curve indicates 98.5" as the transi- tion temperature a value which is confirmed by a dilatometric measurement of this quantity. Dry silver bromate melts a t 308-310" and is stable toward heat and light but in the presence of water it darkens slowly a t the ordinary temperature and rapidly a t high temperatures. If a little impurity such as dust is intro- duced into heated silver bromate decomposition occurs with explosive violence. Silver bromate crystals absorb a considerable quantity of air which is slowly evolved a t high temperatures. It is shown that silver bromate may be used as a standard in iodo- metry. The method of use consists in placing 1 gram of bromate with an excess of potassium iodide in 150 C.C.of water in a 250 C.C. flask ; this is heated on a water-bath to effect double decomposition. The contents of the flask are cooled and made up to 250 C.C. Samples of 25 C.C. are withdrawn acidified with dilute hydrochloric acid and titrated with sodium thiosulphate. Arsenious oxide gives a result about 0.3% higher than silver bromate but after recrystal- lising the arsenious oxide from hydrochloric acid this figure was reduced to 0.05y0. This indicates that whilst silver bromate may have a somewhat higher oxygen equivalent than arsenious oxide this defect is fully compensated by its greater definiteness. J. P. S. Alkali Silver Thiosulphates and their Ability to Add Ammonia. ERIK JONSSON (Ber. 1921 54 [B] 2556-2564).- Additive compounds of alkali silver thiosulphate and ammonia have been described previously by Schwicker (A.1889 942) and by Meyer and Eggeling (A. 1907 ii 347) who however do not record analyses of their products. A repetition of their work has given somewhat different results. The ability to form additive cornpounds seems to depend on the presence of unused subsidiary valencies of the silver atom and is most marked in compounds of the type K,S,O,,Ag,S,O,; it is scarcely noticeable in the case of the salts 2M,S,03,Ag,S,0 and 5M2S,0?,3Ag2S,0,. The existence of colourless and yellow alkali silver thiosulphates (cf. Meyer and Eggeling Zoc. cit.) is confirmed but it appears doubtful whether their isomerism is explicable by assigning the respective formulix AgS*SO,*OK and KS*SO,*OAg since their behaviour towards ethyl iodide indicates that the silver is attached to the sulphur atom in each case. Conversion of the colourless into the yellow modification can be effected frequently by cautious warming with water but too drastic treatment leads to tjhe formation of silver sulphide sulphur dioxide and sulphate.It appears therefore that the yellow compounds are intermediate products in the decomposition of the colourless salts and the transformation is possibly explained by such a scheme as KO*SO,*SAg + MO*S*SO,*.Ag. The following individual substances are described the saZt 2K,S203,Ag,S,03 colourless prisms from silver nitrate and potassium thiosulphate in the presence of ammonia; the salt 5KzSz0,,3Ag,S,0 long colourless prisms ; the compound 3KAgf$0,,NH3,2H,0ii.58 ABSTRACTS OF CHEMICAL PAPERS (cf. Schwicker loc. cit. who regards it as KAgS,O,,NH,) colour- less plates which are converted by warm dilute ammonia and into a yellow salt of the same composition and are transformed by warm water into the compound KAgS203 1.5H20 colourless needles and KAgS203 .yellow hexagonal pyramids ; the salt 2NaAgS,0,,3H20 small irregular plates ; the salt 5(NH4)2%03 3 3&2S@3 long prisms and the compound (NH4)AgS20 prismatic crystals (by the action of ammonium thiosulphate on a solution of silver oxide in ammonia) ; the salt 5Rb,S20p,3Ag,S203 colourless prisms (cf. Meyer and Eggeling Zoc. cit.) which is transformed by warm water into the salt 3RbzS20,,4Ag2S2O3,. yellow prisms ; the salt 3RbAgS20,,NH,,2H,0 (Meyer and Eggehng record the unstable yellow salt Ag,S,03,3NH3,H20. RbzS203,Ag2S203,NH,) ; H.W. Metallic Hydrides. 11. Hydrides of the Alkaline-earth Metals and of Lithium. FRITZ EPHRAIM and EDUARD MICHEL (Helv. Chim. Acta 1921 4 900-924; cf. A. 1921 ii 638).-The preparation of the hydrides and the measurement of their dissociation tensions is recorded. When attempts are made to compare the tensions of the different hydrides with one another it becomes apparent that all measure- ment of dissociation pressure of the alkali and alkaline-earth hydrides are vitiated by the use of impure material containing a greater or less proportion of dissolved metal which tends to depress the tension. Within each group the effect of the metal increases with its atomic weight and the influence of sublimation lies in the same direction.In the cases of caesium and barium hydrides these actions render the measurement of dissociation pressures a t high temperatures almost impossible. The influence of the atomic weight of the metal on the stability of the alkali hydride cannot be regarded as elucidated completely but the authors consider from their own experiments that a slight diminution of stability with increasing atomic weight of the metal is probable. The tension curves of lithium hydride could not be measured sincc a material which would withstand the chemical action of the hydride and metal could not be found. It is however established that it is the most stable of all the alkali or alkaline-earth hydrides which is in accordance with its great heat of formahion.Calcium hydride appears to be more stable than barium hydride whilst thc strontium compound occupies an intermediate position. The behaviour of the alkali and alkaline-earth metals towards hydrogen is not confined to the formation of hydrides XH and XH but extends also to the production of solutions the phenomenon being more marked with the hydrides of the alkaline earths than with those of the alkalis. The absorption of hydrogen occurs previously to and in part simultaneously with the formation of the hydrides; this occurs to a greater extent with tlhe alkaline earth than with the alkali metals. The formation of hydrides occurs slowly with the alkali metals,MORQANIC CHEMISTRY. ii. 59 rapidly and with incandescence in the cases of the alkaline-earth metals.This appears to be due to the greater solubility of the hydride in the metal. The same explanation applies to the ob- servation that calcium hydride for example can be formed a t a temperature which is certainly considerably higher than the tem- perature of dissociation of the pure hydride. The liquid nature of the alkali hydrides a t the temperature of their formation con- tributes also to the slowness of absorption of hydrogen since the eutectic mass protects the metal from further action. Investigation of the hydrides of lanthanum and cerium (Math- mann and Baur A 1903 ii 213) and of neodymium and praseo- dymium (Mut'hmann and Beck A. 1904 ii 409) have given results similar to those now observed with the alkaline-earth metals except that the displacement of the tension due to the presence of an excess of metal is even more considerable.The increase in the action of an excess of metal with increasing atomic weight is there- fore apparent not only within a group in the periodic system but also from left to right with increasing valency of the metal. H. W. The Discovery of an Equilibrium between Cement and Lime-water. RICHARD LORENZ and GUSTAV HAEGERMANN (2. anorg. Chem. 1921 118 193-201) .-When finely-ground Port- land cement which has been previously treated with water and dried is stirred with a fixed quantity of water in absence of air the quantity of lime taken up by the water eventually reaches a maximum value. This maximuin is much less than the solubility of lime in water and depends on the quantity of cement present in proportion to the water and also to some extent on the fine- ness of the particles.The existence of this maximum is shown to depend on the partition coefficient of lime between the water and the silica-alumina gel formed by the decomposition of the cement constituents such as monocalcium silicate and tricalcium aluminate. This partition coefficient was determined by repeatedly treating the cement with fresh water until the whole of it had decomposed The ratio of lime in the solid phase to lime in the water was then found to be about 7.0. The existence of the partition coefficient shows that no definite compound is formed between the lime and the constituents of the gel. [See also .I. 8oc. Chem. Ind. 1922 1 5 ~ . ] E. H. R.The Solubility of Glucinum Sulphate in Water and Sul- phuric Acid at 25". HTJBERT THOMAS STANLEY BRITTON (T. WILHELM BILTZ and GUSTAV F. HUTTIC (2. anorg. Chem. 1921 119 115-131).- For the investigation of the ammoniates of magnesium haloids special precautions were taken in the preparation of pure mag- nesium chloride bromide and iodide. Ephraim's results (A. 1912 ii 546) which were not confirmed were probably due to the fact that his magnesium chloride contained basic salts. The 1921 119 1967-1971). Ammoniates of Magnesium Haloids.ii. 6Q ABSTRACTS OF CHEMICAL PAPERS. hexammoniate of magnesium chloride is formed in about fifteen hours a t room temperature when pure magnesium chloride is saturated with ammonia and the product is extraordinarily voluminous.Magnesium bromide behaves similarly but in the case of the iodide the increase in volume when the hexammoniate is formed is relatively slight. Observations were made on the time taken for the vapour pressure to become steady with different proportions of ammonia in the solid phase. From these observ- ations definite rules can be formulated regarding heterogeneous equilibrium in a solid-gas system. When two non-miscible sub- stances are present in the solid phase equilibrium is reached with gradually diminishing velocity usually in the course of a few hours depending on the temperature and the nature of the sub- stances. When one constituent just disappears for instance from a mixture containing principally a lower ammoniate and a small quantity of a higher ammoniate equilibrium is attained very rapidly.When unsaturated mixed crystals are present for example of two ammoniates equilibrium is reached very slowly often requiring several days. This case occurs with the magnesium haloids containing 5.5 to 5.8 mols. of ammonia. When the ammonia content is very nearly 6 mols. addition of a very small quantity of ammonia even a few hundredths of a mol. causes a very rapid rise in vapour pressure equilibrium being rapidly attained. On account of this the dissociation pressures of the hexammoniates could not be accurately determined. These ammoniates may be compared with the zeolites which unlike normal hydrates lose water very rapidly. The phenomenon has not been observed with other ammoniates. Magnesium chloride and bromide both form a diarnmoniate and a monoammoniate; the iodide forms only a diarnmoniatc.The table gives the heats of formation in Cals. and the absolute tem- peratures a t which the dissociation pressure is 100 mm. Saturated mixed crystals. 2NH,. lNH,. 17-2 ; 475" 38.7 ; 636" - MgC1 13.3; 367" 17.9; 495" 20.8; 573" MgBr 15.2 ; 420' 20.1 ; 559' 21.7 ; 606" MgI2 E. H. R. The Solidification Diagram of the Zinc-Arsenic Alloys. W. HEIKE (2. u w g . Chem. 1921 118 26&268).-Thermal examination was made of alloys containing from 6.6 to 92% of arsenic and with the aid of the results the equilibrium diagram was constructed. Two compounds both melting without decom- position are formed ZngAs2 m p. 1016" and ZnAs2 m. p. 771"; the former has a transition point a t 671". With excess of zinc pure zinc appears with Zn3As2 no solid solutions being formed.Arsenic dissolves little zinc but on the other hand is soluble to a considerable extent in the compound ZnAs,. Two eutectics are formed at 62% and Sl*5y0 As respectively the corresponding tem- peratures being 730" and 723". Both compounds -are,very brittle,INORGANIC CHEMISTRY. ii. 61 and ZnAs shows a well-marked cleavage. shows super-cooling during crystallisation of the alloys. This compound always E. H. R. The Fusion Diagram of Cd(N0,),,4H20+Ca(N0,),,4H,0 at Pressures of 1 to 3000 kilo. /cm.2. MEINHARD HASSELRLATT (Z. anorg. Chem. 1921 119 313-324).-The fusion diagram of the system Cd(N03),,4H,0+ Ca(NO3),,4H2O a t the normal pressure has been previously determined (A. 1913 ii 484).It was shown that the stable form of calcium nitrate forms a eutectic with cad- mium nitrate at 91% Ca(N0,),,4H20 and 40.6" whilst the unstable form of the calcium salt forms a continuous series of mixed crystals with the cadmium salt. The effect of increased pressure up to 3000 kilos. per sq. cm. on the diagram has now been investigated. The general form of the diagram is unchanged. The p-t curve for mixed crystals containing a high proportion of the calcium salt could not be followed at higher pressures on account of the rapid change of the calcium salt into the stable form. Excess of the cadmium salt inhibits this change but as the pressure increases more cadmium salt is needed t o produce this effect. With in- creasing pressure the m. p. of the stable calcium salt rises much more rapidly than that of the unstable.The latter does not form mixed crystals with cadmium nitrate. The lowering of the m. p. of the stable calcium salt by the cadmium salt is independent of the pressure. With increasing pressure the eutectic point moves towards the cadmium side; at 1000 kilo./cm.2 it is at 79% calcium nitrate 47%" ; at 2000 kilo./cm.2 74% and 55" and at 3000 kilo./cm.2 71% calcium nitrate and 61.5". E. H. R. Light Reactions of the Oxides of Titanium Cerium and the Earth Acids. CARL RENZ (Ndv. Chim. Ada 1921 4 961- 968; cf. A. 1921 ii 316).-!I'itanium dioxide cerium dioxide niobium pentoxide and tantalum pentoxide are in themselves stable towards light but become markedly photosensitive in the presence of suitable media. Reaction is due to reduction; this is the more remarkable since the oxides are reducible by purely chemical means with considerable difficulty.Titanium dioxide niobium pentoxide and to a less degree cerium dioxide undergo reduction when exposed to light in the presence of certain organic liquids and reducing solutions par- ticularly glycerol. A lower oxide appears to be formed (zirconium dioxide is not photosensitive and does not form a lower oxide) which on exposure to air or on being heated regenerates the original material. In the case of niobium pentoxide the process depends to some extent on the presence of impurities notably stannic and tungstic acids zirconium compounds and titanic acid or its anhydrides. Brown vanadium pentoxide becomes black with greater or less rapidity when exposed to light beneath glycerol benzaldehyde cinnamaldehyde cuminol or aqueous mannitol solu- tion; a lower oxide initially vanadium tetroxide is produced.Solutions of citric or tartaric acid in absolute alcohol become greenii. 62 ABSTRACTS OF CHEMICAL PAPERS. and ultimately blue when illuminated in the presence of vanadium pentoxide ; carbon dioxide is evolved freely. Similar decomposition is observed with mandelic acid but in this instance the vanadium pentoxide is blackened. Brown neodymium oxide containing praseodymium does not alter in appearance when illuminated under glycerol or phenylhydrazine ; it becomes bluish-grey when warmed with the latter owing to conversion of the brown to the yellow oxide of praseodymium. When exposed to sunlight in the presence of glycerol benzaldehyde or tartaric acid dissolved in alcohol bismuth oxide becomes grey and ultimately dark black.Reduction to the lower oxide and possibly to the metal Cakes place. I n similar circumstances antimony trioxide is also photosensitive. H. W. Concentration of the Erbium Earths. PAUL H. M-P. BRINTON and C. JAMES ( J . Amer. Chem. SOC. 1921 43 1397- 1401).-Four methods for the concentration of t'he less basic of the rare earths have been investigated ; the methods examined are (1) formation of basic nitrates (2) crystallisation of chlorides from 1 1-hydrochloric acid (3) formation of basic chlorides and (4) formation of basic thiosulphates. The authors highly recom- mend the first method for the separation of erbium holmium dysprosium and the less basic earths from yttrium and the second method for the separation of holmium and dysprosium from yttrium.The basic nitrate formation was carried out with (a) a solution of yttrium nitrate containing a little erbium and ( 6 ) a yttrium-erbium-holmium mixture. I n the former case the solution was boiled and treated with a fairly concentrated solution of sodium hydroxide and thoroughly boiled. The yttrium hydroxide which first precipitated soon dissolved. The addition of sodium hydroxide was continued until minute crystals of the basic nitrate were observed swirling through the liquid. The whole was then set aside to cool when a mass of needle-like crystals of the basic nitrate was obtained. These were collected dis- solved in the least amount of nitric acid and treated with sodium hydroxide solution as before.The basic nitrate crystals from this when dissolved in nitric acid gave a rose-red solut)ion which exhibited intense absorption bands of erbium thus showing that the erbium was rapidly collecting in this frackion. The original filtrate was treated several times with sodium hydroxide and although the concentration of nitrates was kept high the erbium absorption bands soon began t o fade. The results obtained with the second mixture were equally good. The cr~7stallisation of the chlorides was cff ccted with a solution containing yttrium holmium and dysprosium. The solutioii in hydrochloric acid was boiled down unt(i1 acid of constant boiling point was obtained. The solution was then evaporated until a scum appeared on the sur€ace when i t was set aside for fifteen to twenty hours.The crystals were separated by decantation and the crystallisation proceeded with; by the time the tail fraction had become No. 9 and the head fraction No. 4 owing to the combination of smallINORGANIC CHEMISTRY. ii. 63 head fractions it was found that the atomic weights had become 02.5 and 124.0 respectively. After four further fractionations the atomic weight of the tail fraction No. 12 was 91.5 whilst that df the head fraction No. 4 was 133.70. The order of separation in order of decreasing solubility of the chlorides is erbium yttrium holmium dysprosium. J. F. S. Disglomeration and Formation of the Autogenous Lead Tree. A. THIEL (Ber. 1921 54 [B] 2755-2758; cf.A. 1920 ii 622) .-Disglomeration which has been observed previously in thc cases of lead and copper is also exhibited by tin when the latter is preserved for some time under stannous chloride solution in a loosely stoppered bottle. Large uniform crystals of lead become strongly corroded when preserved for some weeks beneath Heller's solution ; a consider- able quantity of lead powder is formed but as expected there is no evidence of disglomeration that is formation of deep fissures a t the boundaries of the crystallites. Unexpectedly in the light of the previous theory the large crystallites readily exhibit the formation oE t'he lead tree when preserved beneath a solution of lead nitrate acidified with nitric acid. The phenomenon is observed only after the formation of a white skin of basic salt on the metal and is explained in the following manner.The presence of the skin inhibits the contact of dissolved lead salt and metal by con- vection and greatly impedes the diffusion of the lead ion. Beneath the skin therefore the solution soon cont'ains practically only lead nitrate and is poor in lead ions whereas the external solution still contains lead nitrate and therefore has a much higher lead ion concentration. The possibility of the formation of a short- circuited concentration cell is thus provided. H. W. The Chemical Behaviour of Crystallised Binary Com- pounds with one Component Nobler than Hydrogen. G. TAMMANN (2. anorg. Chem. 1921 118 93-104).-The author draws a comparison between metallic mixed crystal series and binary compounds.In the former case the members of a mixed crystal series behave chemically and electrically in a similar manner to one or other component according to the proportion of each present with sharply defined limits a t molecular fractions which are generally a simple multiple of 1/8. This behaviour is correlated with the lattice structure of the mixed crystals and may be expected also in crystallised binary compounds which have a similar lattice structure the difference being that in the latter case the proportions of the two kinds of atom are fixed. It is to bc expected that binary compounds will show similarity chemically or electrically to one or other component and when two or more compounds of the two elements are formed some will resemble one compoiicnt and some the other.As an example confirming this view the compounds of lead and palladium are cited. In this series the compounds Pb,Pd PbPd PbPd and PbPd have been identified. I n a solution of lead nitrate againstii. 64 ABSTRACTS OF CHEMICAL PAPERS. lead these all give a potential equal to that of palladium but as soon as any free lead is present the potential drops to zero. Chemically all the palladium-lead compounds are as resistant as palladium. Binary compounds can be conveniently classed as resistant or non-resistant the former showing the properties of the nobler the latter those of the baser component. The above principles are applied to the discussion of a large number of binary compounds principally metallic such as those present in alloys of gold silver copper and platinum besides sulphides silicides and carbides and it is shown that the com- pounds can be classified as resistant or non-resistant.The more base the inactive component is the greater is the number of atoms of the nobler component necessary to protect it. I n general a smaller number of gold than of silver atoms are needed to protect a given atom of a baser metal. These considerations apply to compounds in which one component is nobler than hydrogen. When both are less noble than hydrogen the classification into resistant and non-resistant does not apply since both constituents and their compounds decompose water. Apart from compounds of this type it is possible in a series of binary crystallised con- glomerates to determine from a few measurements of their galvanic potential which of the components they will resemble in their chemical character.E. H. R. The Chemical and Electrical Behaviour of some Series of Alloys WILHELM JENGE (2. anorg. Chem. 1921 118 105- 122).-With a view to test the theory put forward by Tammann (preceding abstract) that crystallised binary compounds when no hffusion of the atoms is possible may be expected to behave chemically and electrically as one or other of the constituent elements a number of series of alloys in which binary compounds are formed were examined. The alloys were used as anodes and subjected to the action of halogen sulphate or nitrate ions and were tested against acids and alkalis. In the cobalt-silicon series in which the compounds formed are Co,Si Co,Si CoSi CoSi CoSi those having less than 32% of silicon were readily attacked by acidic ions but those with higher silicon content were resistant. That is to say CoSi and the higher silicides behave as silicon the other compounds as cobalt.Towards cold acids the same compounds were respectively resistant and non-resistant and towards cold alkalis all were resistant except CoSi the behaviour of which resembled that of silicon. No sharp demarca- tion of properties was found in the behaviour of the alloys towards hot acids and alkalis because the cobalt loses its passivity and decomposes water. In the series of nickel-silicon alloys some- what similar results were obtained but the compound Nisi unlike CoSi was not resistant to halogen ions. I n the manganese- silicon series only Mn,Si was non-resistant to acids and all that is Mn,Si MnzSiy and MnSi were resistant to sodium hydroxide.Alloys of antimony with cadmium and tin and of bismuth with thallium were also examined. The compounds Cd3Sb and Zn3Sb,INORGANIC CHEMISTRY. ii. 65 have the potentials of cadmium and tin respectively whilst CdSb and ZnSb soon approximate to the hydrogen potential. The cadmium alloys precipitate antimony and lead from solution and Zn,Sb precipitates not only lead but also cadmium. A bismuth- thallium alloy with the composition Bi,TI gives the hydrogen potential but after etching with hydrochloric acid the bismuth potential indicating that the surface thallium atoms have been removed leaving only bismuth exposed. Of the lead-thallium alloys those with over 0.49 mol.of lead show the lead potential those with 0 to 0.475 mol. of lead show t'he thallium potential. Great difficulty was experienced in measuring the potentials of the alloys of magnesium with copper lead cadmium and tin but i t appears that at the moment of contact with the electrolyte they ha;; the magnesium potential which however rapidly fa&. E. H. R. Physical Chemistry of the Oxides of Lead 11. The Supposed Enantiotropy of Lead Monoxide. SAMUEL GLAS- STONE (T. 1921,119 1914-1927). Binary Systems of the Sulphates Chromates Molybdates and Tungstates of Lead. F. &!I. JAEGER and H. C. GERMS (2. anorg. Chem. 1921,119 145-173).-The paper comprises an account of thermal investigations of the binary systems of PbO with SO CrO MOO and WO respectively and of the different systems formed by pairs of the compounds PbSO PbCrO PbMoO and PbWO,.By an optical method the transition tem- perature of PbO from the red tetragonal low temperature form to the yellow rhombic high temperature modification was found to be 587". The melting point of pure lead oxide is 879". The following melting points were also freshly determined and differ slightly from accepted values chromium trioxide 198" ; molybdenum trioxide 795"; tungsten trioxide 1473". The examination of the binary systems formed by lead monoxide with the acid anhydrides was limited in each case to the partial system PbO-PbM"'0,. In the system PbO-PbSO the existence of the following compounds was recognised Bb4SO7 Pb,SO Pb,SO PbSO,. The f i s t has no real melting point but decomposes at 897" forming Pb,SO m.p. 961" which has a transition point at 450". Pb,SO (lanarkrite) has m. p. 977". PbSO decomposes markedly at 1135" and its m. p. is estimated by extrapolation to be 1170". The transition point of PbSO from the p to the low temperature a form is 864&1". Three eutectics are formed a t 89 mols. % PbO and 835" 60 mols. yo PbO and 950" and 34 mols. yo PbO and 960". I n the system PbO-PbCrO the compounds Pb,CrO Pb7Cr,01 Pb,CrO and PbCrO were recognised. The first has no real melting point and exists only below 815'. Pb,Cr201 m. p. 854" has a reversible transition point at 744" and forms with Pb,CrO m. p. 920" a eutectic at 68 mols. % PbO and 841'. Lead chromate PbCrO is found to be trimorphous; the a-form VOL.CXXII. ii. 3ii. 66 ABSTRACTS OF CHEMICAL PAPERS. is stable below 707"; the P-form between 707" and 783" and the y-form above 783" melting at about 844" with evolution of oxygen. The heat effect of the change CI p is small and is sharper in binary mixtures with lead oxide than in the pure substance. The eutectic between PbCrO and Pb,CrO occurs a t 820" but this part of the equilibrium diagram could not be determined accurately on account of decomposition. The system PbO-PbMoO shows only two compounds Pb,MoO m. p. 951" and PbMoO m. p. 1065". There are two eutectics a t 87.5 mols. % PbO 762" and a t 40 mols. yo PbO 933". The tungstates correspond with the molybdates Pb,WO m. p. 899" and PbWO m. p. 1123" with a transition point a t 877". The equilibrium diagram for the system PbCr0,-PbS04 is largely conjectural owing to the considerable amount of decompo- sition occurring a t higher temperatures.There is a gap in thc mixed crystal series between about 40% and 30% PbCrO,. Thc mixed crys\als have transition points a t 934" 874" and 748". In the PbS0,-PbMoO system mixed crystals are formed con- taining up to 6 mols. yo sulphate or 2 mols. yo molybdate. There is a eutectic a t 57 mols. yo molybdate and 962". Ah 879" the mixed crystals have a transition point. In the PbS0,-PbWO system the mixed crystals of the P-type separating a t the eutectic temperature contain respectively 37 mols. yo sulphate and 7 mols. Yo tungstate. The eutectic is at 51 mols. yo tungstate and 995". At 875" occurs the ~ZCI-sulphate transformation and a t 859" the corresponding tungstate change.The PbCr0,-PbMoO diagram is largely hypothetical. The composition of the limiting mixed crystals on the molybdate side is 48 mols. % PbCrO a t 838" the eutectic temperature. Transitions occur a t 799" of y-+P-chromate mixed crystals and a t 697" @-+a. In the PbCr0,-PbWO system the eutectic temperature is 837" and the limiting mixed crystals on t'he tungstate side contain 41 mols. :& PbCrO,. Lead molybdate and lead tungstate form an isodimor- phous mixed crystal series with a transition temperature a t 1082". A mixture containing 75 mols. yo PbMoO is in equilibrium a t this temperature with bot!h kinds of mixed crystal. Photochemistry of Thallous Chloride. 11. CARL Rmz (HeZv. Chirn. Acta 1921 4 950-960).-A continuation of previous work (A.1920 ii 71). Thoroughly illuminated blackish-brown thallous chloride in consequence of .photolysis contains as primary product more or less grey to slate-grey photothallous chlorides in addition to yellow intermediate thallous-thallic chlorides formed in accordance with the scheme 6TlC1-t- Light = photochloride+ TlC13,3T1C1. Thallic hydroxide formed by subsidiary actions is also present. These phases characterise the photo-processes in those cases in which the change of colour extends over the scale greyish-brown dark greyish-brown blackish-brown and hence occur when t'tlallous chloride is illuminated in the dry condition under water and in the presence of solutions of many neutral salts. In the presence of reducing agents or of organic hydroxy- E.H. R.INORGANIC CHEMISTRY. ii. 67 acids the action of light on thallous chloride only leads to the production of the photochloride. The formation of thallous- thallic chlorides is not observed in the presence of alkalis or alkali carbonates which decompose these compounds immediately. The production of photothallous chlorides and of thallic hydroxide does not occur in the presence of hydrochloric acid even without the addition of organic substances. Photothallous chloride can be prepared by purely chemical methods if ferrous sulphate is added to a boiling saturated aqueous solution of thallous chloride and the mixture is treated with an excess of ammonia. The black precipitate of photothallous chloride and iron hydroxides is allowed to settle and is subse quently washed with hydrochloric acid until the iron compounds are dissolved ; the slate-grey photothallous chloride so obtained behaves in exactly the same manner as the photosynthetic product. H.W. Ammoniates of Cupro- and Thallo-haloids. WILHELM BILTZ and WILHELM STOLLENWERK (2. anorg. Chem. 1921 119 97-1 14).-The formation and vapour pressures at different temperatures of ammoniates of cuprous and thallous chloride bromide and iodide were investigated using apparatus similar to that employed in experiments on the ammoniates of silver haloids (A. 1921 ii 201). When saturated with ammonia gas cuprous chloride first shrinks to a yellow mass then swells and becomes greyish-white. Saturation at -70" to -30" requires a t least a day. When the excess of ammonia is allowed to evaporate at room temperature and atmospheric pressure cuprous chloride triammoniate remains.I n damp air it quickly turns green. The pressure isotherms also indicate the existence of a sesqui- ammoniate and a monammoniate. Cuprous bromide behaves similarly forming a white triammoniate a sesquiammoniate and a monammoniate. Cuprous iodide absorbs ammonia quickly a t room temperature. It forms four compounds containing re- spectively 3 2 1 and Q mol. of ammonia. In the following table are given the heats of formation Q in Cals. and the temperatures in absolute degrees a t which the dissociation pressures of all these compounds are equal to 100 mm. 3NH3. 2NH,. 1 &NH3. lNH,. WH3. CuCl 9.48; 283" - 12-61; 326" 16.i3; 417.5' - CuBr 9.50; 283" - 13.15; 339" 14-64; 369.0' - CUI 10.37; 286-5" 11.30; 298" - 14.70; 371.0" 15.22; 390" Thallous haloids do not absorb ammonia a t the ordinary tem- perature but in liquid ammonia they all form triammoniates.The vapour pressures are all very close to those of ammonia itself. The triammoniates are soluble to a certain extent in liquid ammonia the solubility increasing with rising temperature and with the atomic weight of the halogen. The heat of formation is about 7.1 Cal. for the ammonia compound of each of the three haloids. No lower a,mmoniates are formed. E. H. R. 3-2ii. 68 ABSTRACTS OF CHEMICAL PAPERS. The Action of Molten Alkali Chlorides on Copper Oxide. J. ARVID HEDVALL and GUNNAR BOOBERG (2. anorg. Chem. 1921 119 213-216).-1t was shown in a former paper (Hedvall and Heuberger A.1921 ii 508) that potassium chloride could not be used as a flux in the fusion of cupric oxide with aluminium oxide on account of a reaction taking place between the potassium chloride and copper aluminate. It is now shown that when copper oxide is heated with potassium chloride cuprous oxide is formed and oxygen evolved. This is best demonstrated by adding cupric oxide in small quantities to a mixture of potassium and sodium chlorides a t 1000" and continuing the heating for one and a half hours. At the same time a basic cupric chloride is formed which by prolonged heating with sodium or potassium chloride solution is obtained as the compound 3Cu0,CuC12,4H,0. Phenomena of Diffusion in Metals in the Solid State and Cementation of Non-ferrous Metals.I. Cementation of Copper by means of Ferro-manganese. G. SIROVICH and A. CARTOCETI (Gaxxetta 1921 51 ii 245-261).-A bar of copper was arranged centrally in a porcelain tube glazed internally and the tube then packed with ferro-manganese containing 5% of wood charcoal both these materials being capable of passing through a sieve with 64 meshes per sq. em. and of being retained by one of 324 meshes per sq. em. The tube was closed by mcaiis of rubber stoppers luted with sodium silicate one of the stoppers having two holes to admit a thermo-couple for measuring the temperature and a glass tube bent a t right angles and with its end dipping into mercury. After the tube had been heated for some hours a t 900" in a Heraeus furnace considerable proportions of the manganese were found to have penetrated the copper (cf.J . SOC. Chem. Ind. 1922 1 7 ~ ) . T. H. P. E. H. R. Tervalent Copper. G. SCAGLIARINI and G. TORELLI (Gmzetta 1921 51 ii 225-228).-Contrary to Moser's statement (A. 1907 ii 549) the action of potassium persulphate on cupric hydroxide in presence of barium hydroxide a t temperatures obtained by cooling with ice and salt results in various changes in the colour of the solution and in the deposition of a tenuous amaranth-red precipitate which may be purified by repeated washing with ice- water by decantation. The compound thus obtained yields oxygen when treated with sulphuric acid oxidises hydrochloric acid with liberation of chlorine oxidises ammonia in the cold with production of nitrogen nitrous acid and traces of nitric acid decolorises per- manganate and decomposes potassium iodide with liberation of iodine in quantity greater than that corresponding with the pro- portion of copper present.Since i t does not yield hydrogen peroxide when treated with dilute acid the compound lacks the grouping characteristic of peroxides and is thus different from the orange- yellow copper peroxide obtained by means of hydrogen peroxide. The ratio between the percentages of copper and active oxygen present is in agreement with the formula Cu,03. T. H. P.INORGANIC CHEMISTRY. ii. 69 Production of Single Crystals of Aluminium and their Ten- sile Properties. H. C . H. CARPENTER and CONSTANCE F. ELAM (Proc. Roy. Soc. 1921 [A] 100 329-353; cf. A. 1921 ii 641).- A continuation of work previously published (Eoc.cit.) on the production of large crystals of aluminium. The metal used in the present work had a purity of 99.6y0 the impurity being 0.19% silicon and 0.14y0 iron. The test-pieces used were 70 mm. with a parallel portion 103 mm. long 26 mm. broad and 3 mm. thick and were estimated to contain 1,687,000 small crystals in the parallel portion (103 x 26 x 3 mm.). The authors first describe the treatment necessary to convert the whole of the crystals into a single crystal. Three separate processes are shown to be necessary (i) the aluminium strip is heated at 550" for six hours (ii) the strip after cooling is subjected to a stress which is equivalent to 378 kilos. per sq. cm. and gives an average elongation of 1.6% on 76 mm.(iii) the test- piece is finally placed in a furnace a t 450" and the temperature raised 15-20' per day up to 550' and then for 1 hour a t 600". Applying this treatment to thirty-eight test pieces showed that nine pieces consisted of a single crystal fourteen of two crystals nine of three crystals four of four crystals and two of six crystals. The tensile strength of aluminium strips consisting of known num- bers of crystals has been determined. It is shown that for strips consisting of 150 crystals per 25 mm. it is 708-740 kilos. per sq. cm. and these give an elongation of 36-38% on 76 mm. The tensile strength of strips consisting of a single crystal varies between 598 and 642 kilos. per sq. cm. and these strips suffer an elongation of 34-86% on 76 mm.The varying tensile strength and elongation was accompanied by differences in the type of stretching and fracture. Strips consist,ing of two crystals have a tensile strength of 441-550 kilos. per sq. cm. and suffer an elongation of 29-70y0 on 76 mrn. whilst strips consisting of three crystals have a tensile st'rength of 456-567 kilos. per sq. cm. and suffer an elongation of 36-55y0 on 76 mm. A further series of experiments on the pro- duction of single crystals in bars is described. The Thermal Treatment of certain Complex Aluminium Alloys. LBON GUILLET (Compt. rend. 1921 173 979-982).- In order to determine the effect of each constituent on the behaviour of duralumin under thermal treatment (cf. ibid. 1919 169 508) the author has studied alloys of aluminium and copper aluminium and silicon aluminium silicon and copper aluminium magnesium and silicon and quaternary alloys containing all four elements.Measurements of hardness have been made on annealed samples and on samples tempered a t different temperatures the measure- ments being made in the latter case immediately after tempering and also after the alloy had been kept for forty-eight hours a t 20". Prom the resulfs of these measurements and from micrographic examinations of the alloys it is shown that the simultaneous presence of silicon magnesium and copper is indispensable to obtain the interesting results given by tempering high resistance aluminium alloys. W. G. J. F. S.ii. 70 ABSTRACTS OF CHEMICAL PAPERS. Solubility Limits of Carbon in Ternary Steels.I. The System Chromiun-Iron-Carbon. KARL DAEVES (2. anorg. Chem. 1921 118 55-66).-Experiments were made to determine the influence of chromium on the solubility of carbon in iron and to determine the position of the corresponding solubility line in the ternary chromium-iron-carbon diagram. The solubility falls off rapidly at first as the chromium content increases then more slowly the general form of the curve being hyperbolic. Points on the curve were determined by observing what chromium content was necessary with a given carbon content to cause the appear- ance of a eutectic in the structure of the metal. To make the hard alloys workable for the preparation of polished surfaces it was necessary to heat for several hours a t SOO" just below the Ac point by which treatment the solid solution was broken up and the metal softened.Etching was accomplished by electrolysis in ammonium persulphate solution. In eutectoid alloys the cementite is practically unattacked by hot sodium picrate solution. Cold alkaline potassium ferricyanide turns the hard constituent of the eutectic brown to yellow leaving the mixed crystals untouched. The solubility curve explains many of the known .properties of chromium steels. The melting point of steel and the arrest points are little affected by chromium up to lo?&. [Cf. J . SOC. Chern. Ind. 1922 1 6 ~ . ] Solubility Limits of Carbon in Ternary Steels. 11. The System Tungsten-Iron-Carbon. KARL DAEVES (2. anorg. Chem. 1921 118 67-74).-The effect of tungsten on the solu- bility of carbon in iron was studied in the same way as that of chromium (preceding abstract) and a solubility curve of similar form was obtained separating eutectic from non-eutectic steels in the ternary diagram.Sudden changes in the physical properties of tungsten steels are correlated with changes of composition involving the passage from one side to the other of this limiting curve. The appearance of so-called double carbides of iron and chromium or of iron and tungsten observed by different workers is attributed to the same cause Small amounts of tungsten in E. H. R. steel raise the melting point but larger amounts depress it E. H. R. The Colour of Iron Alum. JANE BONNELL and EDGAR PHILIP PERMAN (T. 1921 119 1994-1997). Complex Selenates. JULIUS MEYER (2. anorg. Chem. 1921 118 147).-A large number of new complex selenates and incidentally some simpler compounds which have not hitherto been described were prepared for comparison with the corre- sponding sulphates.The new selenates described belong to the chromi- and cobalti-series and show the closest resemblance to the sulphates differing from these occasionally only in their water of crystallisation. On account of the ease with which selenic acid is reduced difficulties were a t times encountered in the preparation of certain of the compounds.lTORGANIC CHEMISTRY. ii. 71 [With LEONHARD S~~~c~.]-Chromiselenates. Violet chromic selenate [Cr(H20),],(Se0,),,3(or 4)H,O forms a crystalline powder readily soluble in water from which it is precipitated by alcohol or acetic acid.Its aqueous solution dissolves chromic hydroxide with formation of green basic salts. When the violet salt is heated in solution or in the solid state a t 90" it changes irreversibly into a green chromiselenate. The green salt prepared in the solid state has the composition Cr,(SeO,) 10H,O and dissolves very slowly in water probably only after addition of water. The green salt may have a constitution of the type [Cr(SeO,)(H,O),],SeO,. When a solution of the violet salt is boiled for some time a green compound is formed which is precipitated by alcohol as a green oil and dries to an amorphous green solid. It is very soluble in water and gives no precipitate with barium salts or with ammonia. It is probably a triselenatochromic acid [Cr( SeO,),]H,. ChloropentaquochromiseZenate [CrC1(H20),]Se0,,3H,0 was pre- pared from chloropentaquochromichloride and sodium selenate ; it forms a bright green powder very soluble in water and alcohol. Attempts to obtain other chloro-selenates corresponding with known chloro-sulphates were not successful.forms a green crystalline powder readdy soluble in water slightly so in alcohol. An attempt to prepare a corresponding double chromi-aluminium selenate failed although sulphates of the type [~C1,(H,o),](so,),[M(H2~)6] where M=Cr Fe Al or V were prepared by Werner and Huber (A. 1906 ii 170). Hexamminechromiselenate [Cr(NH,),],(seO,) was prepared from the corresponding nitrate and selenic acid. It is precipitated from aqueous solution by alcohol as a heavy yellow finely crystalline powder. The salt is amorphous whilst the corresponding sulphate has 5H20.Chloropentamminechromiselenate [CrCl(NH,),]SeO mas pre- pared from purpureochromichloride and silver selenate. It forms a heavy red amorphous powder sparingly soluble in water. The corresponding sulphate is much more soluble and crystallises with 2H,O. HexacarbamidechromiseZenate [Cr(NH2*CO*NH,)6]2( SeO,) pre- pared from hexacarbamidechromichloride and silver selenate was obtained as a bright green finely crystalline powder moderately soluble in water from which alcohol precipitates it. Triethylenediaminechromiselenate [Cr en,],(SeO,) from the corre- sponding chloride and silver selenate is a reddish-yellow heavy crystalline powder soluble in water and precipitated by alcohol. When the dry salt is heated a t loo" the colour changes to reddish- violet.Aluminium selemte which has not before been described forms a white crystalline powder easily soluble in water and precipitated by alcohol. It appears to contain less than 18H,O but the analysis did not distinguish between 15 and 17H,O. [With HANIT$ MOLRENHAUER.]-Comp~ex cobaltiselenates. The Dichlorotetraquochromihexaquochromiselenute [c~c~2(fI,o),l(seo,),~~~~~2o),l,ii. 72 ABSTRACTS OF CHEMICAL PAPERS. complex cobaltiselenates prepared were confined to those containing only one cobalt complex and to those with 6 5 or 4 molecules of ammonia or 4 molecules of pyridine. Hexammine ( lut eo) co balt isel enat e [ Co (NH ) ( S e 0,) 5H2 0 c orre - sponds in every respect with luteocobaltisulphate. Aquopentamminecobaltiselenate [CO.(H,Q)(NH,),]~(S~O,)~,~H~O was prepared both from the corresponding cobaltichloride and from selenatopentamminecobaltiselenate.The salt is similar in physical and chemical properties to roseocobaltisulphate. Diaquotetramminecobaltiselenate [ C!( H20)2(NI~3)4]2( Se04),,3H20 was prepared from carbonatotetramminecobaltiselenate and selenic acid. It dissolves in water to a deep red solution from which alcohol precipitates it as a bright red crystalline powder. It loses its water of crystallisation on exposure to air. Chloropentamminecobaltiselenate [CoC1(NH3),]SeQ4 was prepared from purpureocobaltichloride and silver selenate ; it corresponds in its properties with purpureocobaltisulphate. Chloroaquotetramminecobaltichloride selenate {[CoC1(H20) (NH3)41C112Se04 was obtained when dichlorotetrammincobaltichloride was treated with silver selenate through hydration of one of the nuclear chlorine atoms.It forms a violet crystalline powder giving a violet aqueous solution. Nitropentamminecobaltiselenate [ Co (NO,) (NH,) ,]SeO from the corresponding chloride and silver selenate forms bright yellow microscopic crystals giving a yellowish-brown aqueous solution. It forms a periodide as does the corresponding sulphate. [ Co (SO,) ( NH3),I,Se04,2H,0 was prepared from the corresponding sulphato- bromide and silver selenate. It is precipitated from aqueous solution by alcohol in rose-coloured leaflets consisting of microscopic rhombic tablets. The corresponding sulphatosulphate contains only 1H,O. Sulphatopentammineco baltiselenate Acid selenatopentamminecobaltiselenate [ co( SeO4)(NH3) 5]SeO4H,2H& was prepared by treating chloropeiitamminecobaltichloride with concentrated selenic acid.From the diluted solution the acid salt crystallised in reddish-violet crystal aggregates. It closely resembles the sulphato-sulphate and forms the starting material for the preparation of a series of selenatopentamminecobalti-salts including several of the following. [co(seo4> (NH3) 512(Se04) ,H20 was obtained by treating the above acid selenate with alcohol; it has a brighter red colour than the acid salt. SelenatopentarnminecobaltisuErphate [Co(SeO,) (NH3),],SO4,H2O was obtained from the selenato-bromide and silver selenate. It is precipitated by alcohol from aqueous solution in bright red lustrous tablets.This salt is metameric with the above sulphatopent- amminecobaltiselenate but the two are not isomorphous as the latter crystallises with 2H20. Xelenafopentamminecobaltinitrate [CO(S~O,)(NH,)~]NO~ was pre- Normal sel enat opent ammineco ba 1 t iselenat e ,I N ORGANIC CHEMISTRY. ii. 73 pared from the above acid selenate and ammonium nitrate. It separates in well-formed bright red sparingly soluble crystals. Selenatopentamminecobaltibromide [ Co( SeO,) (NH,),]Br was pre- pared from the above acid selenate and hydrobromic acid. It is thrown down by alcohol from aqueous solution as a bluish-red precipitate. Sel enatopentamminecobaltihexachloroplatinate [ Co ( S e 0,) (NH,) ,I2PfC1 2 H2 0 forms lustrous orange-red tablets sparingly soluble in water.In the tetramminecobalti-series only carbonic acid of the bi- valent acids could be introduced into the complex. With two univalent acid radicles stereoisomerism becomes possible and ,it was found possible t o prepare the 1 2- and 1 6-dinitrotetrammine- co baltiselenates. Carbonatotetramrninecobaltiselenates [ Co( CO,) ( NH,),],Se0,,3H20 is similar to the corresponding sulphate crystallising in dark red leaflets which lose their water of crystallisation on exposure to air. Acid dichlorotetramminewbaltiselenute [CoC12(NH,),]Se04H crys- tallises in dark green well-formed needles but is unstable and readily changes to the chloroaquotetrammine salt described above. Acid dichlorotetrapyridinecobaltiselenute [CoC1,Py4]Se0,H,2Hz0 is more stable than the dichlorotetrammine salt ; it crystallises in lustrous green leaflets. The salt corresponds with the sulphate described by Werner and Feenstra (A. 1906 i 450).1 2-Dinitrotetramminecobaltiselenute [ Co( N02)2(NH,),],Se0 was prepared from ffavocobaltinitrate (Jorgensen A. 1898 ii 592) and ammonium selenate; it forms dark brown crystah. 1 6-Dinitrotetramminecobalti~ele~te stereoisomeric with the last was prepared from croceocobaltichloride (Jorgensen Zoc. cit.) and silver selenate; it is precipitated from aqueous solution by alcohol in the form of minute bright yellow crystals. The electrical conductivities of many of the above salts in aqueous solution were measured a t 25" and their magnitudes were found to agree with the constitutions ascribed to the different salts.E. H. R. The Green Colour of Tungsten Trioxide. J. A. M. VAN LIEMPT (2. anorg. Chem. 1921 119 310-312).-Tungsten trioxide generally has a yellow colour but is sometimes green. A number of explanations of this phenomenon have been offered but it is now shown experimentally that the green colour is due to re- duction at ordinary temperatures by traces of organic matter to lower oxides. Provided the green oxide has not been ignited the yellow colour may be restored by heating it in a current of oxygen. E. H. R. Chlorination by Mixed Carbon Monoxide and Chlorine. F. P. VENABLE and D. H. JACKSON ( J . Elisha Mitchell Sci. Soc. 1920 35 87-89) .-Chlorination is successfully accomplished with a mixture of carbon monoxide and chlorine containing the former in excess in the following cases zirconium dioxide a t 480" stannic 3"ii.74 ABSTRACTS OF CHEMICAL PAPERS. oxide a t 400" magnesium oxide at 475" aluminium oxide a t 45OU ferric oxide a t 460" chromic oxide a t 625" manganese dioxide at 460" uranoso-uranic oxide at 500". With chlorine in excess the requisite temperature for zirconium dioxide is 425" and for ferric oxide 370". CHEMICAL ABSTRACTS. Antimonic Acid and the Use of Sodium Antirnonate in Analysis. E S . TOMULA (2. anorg. Chem. 1921 118 81- 92).-The constitution of antimonic acid and of the salts derived from it has never been satisfactorily settled and an attempt has now been made to solve the problem by the application of physico- chemical methods. The conductivity of the potassium salt was measured a t 25" a t dilutions from V=32 to 1024 and the basicity of the acid by the Oswald-Walclen rule was found to be 1.This rules out the possibility that the salt is a pyroantimonate K2H2Sb20 and since it gives a solution having an acid reaction it cannot be the metantimonate KSbO,. It must therefore be the orthoantimonate KH,SbO,. Hydrogen-ion determinations in a 1 /1024N-solution by the calorimetric method confirmed this view. The dissociation constant a t this dilution was found to be c(=0*957 and the hydrogen-ion concentration CH= 10-6'3. The equivalent conductivity of the sodium salt was on account of its low solu- bility determined only at dilutions V=512 and 1024 and was found to be of the same order as although slightly lower than that of the potassium salt. The hydrogen-ion concentration at V= 1024 was CH=10-6'4 and it is concluded that the two salts have the same constitution.Delacroix ( A . 1898 ii 340; 1900 ii 145) and Senderens (A. 1899 ii 557) both isolated a soluble and an insoluble form of antimonic acid which they called ortho- and pyro-acids but they differed as to which was which. Conductivity experi- ments on the potassium salts show that the soluble acid is the ortho-acid whether prepared by Senderens's or Dclacroix's method. It is concluded however that a concentrated solution of antimonic acid is not a true solution but a supersaturated colloidal pseudo- solution from which the acid soon separates in the insoluble form. Determinations were made a t 18" 25" and 33.5" of the solu- bility of sodium antimonate in water in aqueous sodium acetate and in aqueous methyl and ethyl alcohols.Expressed in mg. of Na,O,Sb,O,,GH,O per 100 C.C. of solution the solubility at 18" is in water 56.4 in equal volumes of water and ethyl alcohol 0.1 and in 2.5y0 sodium acetate 3.1. The following method is recommended for the estimation of antimony as sodium antimonate. The antimony must be in alkaline solution as sodium sulphantimonate Na,SbS4 and must be free from potassium since in presence of potassium salts pre- cipitation is incomplete. The solution is warmed a t 80" and stirred while a solution of 30% hydrogen peroxide is run in drop by drop until vigorous evolution of oxygen commences and it is then boiled until all oxygen evolution ceases. The alkaline solution is neutralised with acetic acid until it is acid to phenolphthalein but still weakly alkaline to litmus.It is stirred a further quarterINORGANIC CHEMISTRY. ii. 75 of an hour and then one-half its volume of 96% alcohol is added after which stirring is continued for ten minutes. After twelve hours the crystalline sodium antimonate is filtered washed on the filter with a solution containing 3 grams of sodium acetate 3 grams of acetic acid and 400 C.C. of ethyl alcohol per litre and finally with 50% alcohol. The dried precipitate is separated from the filter-paper which is burnt separately is ignited for fifteen minutes in a porcelain crucible and weighed as sodium metanti- monate NaSbO,. Special directions are given for procedure when tin is present as it is then. necessary to redissolve and re- precipitate the sodium antimonate.[See also J . Xoc. Chem. Ind. 1922 12A.1 E. H. R. The Reaction Limit of Chemical Agents on Copper-Gold Alloys and their Galvanic Tension. G. TAMMA" (2. anorg. Chem. 1921 118 48-54).-The reaction limit in different copper-gold alloys is reached when the molecular fraction of gold present is 1/8 2j8 or 4/8 according to the chemical reagent used. The reactivity of the mixed crystals may be regarded as due to the loosening of the copper atoms from their lattice combination by the chemical agent or from another point of view to the action of the chemical agent on copper atoms which have become detached from the lattice on account of their solution tension. From the latter point of view it was important to determine how the solution tension of the alloys varied with the composition.Measurements were made against a gold electrode in a number of electrolytes and against silver with silver sulphate as electrolyte. The results showed that the limiting composition beyond which no copper ions appear in the solution and the alloy behaves electrically as pure gold is a t 2/8 mol. fraction of gold. This method does not give such sharp limiting values as the chemical method however owing to the sensitiveness of the galvanic tension to impurities on the surface of the metal. The case of the cell silver I saturated silver sulphate I copper-gold is specially interesting since when the proportion of gold in the alloy does not exceed 0.145 mol. silver is visiblyprecipi- fated and the metal becomes negatively charged whilst the alloys richer in gold do not precipitate silver and assume a weak positive charge. This weak positive charge indicates a superficial deposit of silver so that the surface acts as a silver-gold alloy of corre- sponding composition. It is shown from consideration of the mixed crystal lattice that when 1/8 mol. of gold or less is present in the copper alloy conditions are favourable for the formation of silver crystals. It is suggested that those agents which find their active limit a t 2/8 mol. of gold corresponding with the solution tension limit for copper ions act first on the copper ions in solution but as soon as the osmotic pressure of the copper ions exceeds the solution tension the agent attacks the mixed crystal surface. E. H. R. Ruthenium Tetroxido. F. KRAUSS (2. anorg. Chem. 1921 119 217-220) .-An aqueous solution of ruthenium tetroxide has 3*-2ii. 76 ABSTRACTS OF CHEMICAL PAPERS. apparently a weak acid reaction although this is dificult to demon- strate on account of the rapid decomposition of dyes by the solution. The solution behaves as an electrolyte and is decomposed by the current with formation of a green colour. With alkali hydroxides it forms salts but only the ammonium salt could be obtained in the pure state. It was prepared by adding concentrated ammonia to a concentrated solution of ruthenium tetroxide in water until the colour changed from yellow to greyish-brown. By evaporating a salt of the composition (NH,),RuO was obtained. Under certain conditiona which could not be accurately determined a mono- and a di-hydrate of this salt were obtained. In the preparation of ruthenium tetroxide besides the yellow compound a brownish- red substance was observed less soluble in water than the tetroxide. This has not been identified. Ruthenium tetroxide can be estim- ated by distilling it in a current of dry air at 15" into a specially constructed weighed flask dissolving in a little water reducing with alcohol evaporating with dilute hydrochloric acid igniting in a stream of hydrogen and weighing the ruthenium. E. H. R.