INORGANIC CHEM18TR.Y. Inorganic Chemistry. Quantitative Syntheais of Water, By EDWARD H. KEISER (Amel*. Chem. J., 1898, 20, 733-739. Compare Abstr., 1887, 1078; 1891, 1154).-By the following method, the quantities of hydrogen and oxygen combining to form water can be determined gravimetri- cally without the necessity of gasometric measurement. A light glass cylinder, terminated at the bottom by a narrow tube carrying a glass stopcock, and closed a t the top with a round glass cap with a three- way stopcock, is divided by a slight constriction into two compart- ments, the smailer and lower of which contains purified phosphoric anhydride. After completely exhausting the tube, i t is weighed, and pure spongy palladium, contained in a thin glass tube having a number of small holes near the bottom, is introduced into the larger compartment; the tube is again exhausted and weighed;, pure hydrogen is admitted so as to completely saturate the palladium, and after partially exhausting, the amount of hydrogen absorbed is determined.Pure oxygen is then passed through the apparatus a t such a rate that the water formed is completely absorbed by the anhydride, and the temperature does not rise sensibly. After completely oxidising the hydrogen, the exeess of oxygen is removed by exhausting the apparatus, and the weight of the water formed is determined. In the author’s experiments, the tube used was counterpoised during88 ABSTRACTS OF CHEMICAL PAPERS. all weighings by a duplicate; both the oxygen and hydrogen were prepared electrolytically, and were purified by being passed over heated palladium, and subsequently through tubes of fused caustic potash and phosphoric anhydride.The average of four experiments carried out by this method, ranging from 15.874-15.886, gives 15.880 as the atomic weight of oxygen, W. A. D. Electrolytic Production of Chlorates, Bromates, Iodates, and Hypochlorites. By WILHELY VAUBEL (Chem. Zeit., 1898, 22, 33 l).-The author finds, on employing an electrolytic cell containing a saturated solution of sodium hydrogen carbonate surrounding the anode, and separated by a porous diaphragm from a strong brine solution in contact with the cathode, that the whole of the chlorine migrating into the anodic compartment is oxidised to chlorate, whilst carbonic anhydride and caustic soda are obtained a t the anode and cathode respectively.The temperature is maintained at 60-70° during electrolysis, and a current density of 5-10 amperes per sq. dcm. and an E.M.F. of 4-5 volts are employed, A table showing the yield of chlorate after the expiration of certain ampere-hours is given, and the author compares his results with those of previous investigators, The experimental conditions for the production of potassium chlorate are quite similar, as are also those for the preparation of bromates and iodates. When the above electrolysis is performed at lower temperatures, hypochlorite, and not chlorate, becomes the chief product, but, as in the former experiment, there is no formation of chloride round the anode. G. T. M. Electrolytic Preparation of Perchloric Acid and its Salts.By FRITZ FOERSTER (Zeit. EZektrocliem., 1898, 4, 386) .-Potassium chlorate in neutral or alkaline solution is scarcely reduced by a high cathodic current density when the cathode is made of platinum, lead, copper, zinc, or nickel, but is vigorously reduced when it is of iron and to some extent when i t is of cobalt. I n neutral or acid solutions, chlorates are readily oxidised at the anode to perchlorates. A series of experiments shows that, with high current density, concentrated solutions, and low temperatures, the yield of perchlorate is a very good one. A 50 per cent. solution of sodium chlorate was electrolysed with equal platinum electrodes 1.5 cm. apart and a current density of 8-3 amperes per sq. dcm. The solution was at the ordinary tempera- ture, and 4.5 volts were required.The yield increased from 86 to 98 per cent. of the theoretical amount, falling off only when the greater part of the chlorate was oxidised. With a 5-6 per cent. solution of potassium chlorate and 8 ampbres per sq. dcm. the yield was about 82 per cent., but fell off considerably when about one-half of the chlorate was oxidised. In alkaline solutions, oxidation ia observed at the beginning of the electrolysis, but even with high current densities it soon ceases, so that the direct production of perchlorates from chlorides would appear to be imprac- ticable. T. E.INORGANIC CHEMISTRY. 89 Electrolysis of Solutions of Calcium Chloride. By H. BISCHOFF and FRITZ FOERSTER (Zeit Elektrochem., 1898, 4, 464).- Oettel's statement that a better yield of chlorate is obtained in the electrolysis of calcium chloride than in that of potassium chloride is confirmed.Measurements of the gases evolved during the electrolysis give the percentages of the current employed (a) in the formation of hypochlorite, chlorate, and perchlorate, ( b ) in reducing hypochlorite at the cathode, (c) in decomposing water. When a solution of 74.3 grams of calcium chloride in 500 C.C. of water is electrolysed with 6.7 volts, and current densities of 9.1 amperes per sq. dcm. at the anode and 13 amperes per sq. dcm. a t the cathode, 85.7 to 90.4 per cent. of the current yields oxygen compounds of chlorine, mainly chlorate, 1.4 to 2.9 per cent, reduces hypochlorite, and 7.8 to 12.8 per cent. decom- poses water.With a solution of 100 grams of potassium chloride + 7.5 'grams of potash in 500 C.C. of water, the E.M.F. being 4.8 volts and the current densities the same as before, the corresponding values were 50.9 to 61.9, 12 to 19.9 and 20.8 to 29.6. Both solutions were cooled with ice. A solution of barium chloride gave results very similar to those obtained with potassium chloride, The small reduction observed with calcium chloride is probably due to a layer of calcium hydroxide on the cathode which acts as a diaphragm; the greater part of the calcium hydroxide formed combines with chlorine, but part of i t remains undissolved, free hypochlorous acid existing in the solution and volatilising with the gases evolved. Further experiments show that at 20-25' the best yield of chlorate is obtained from solutions con- taining at least (preferably much more than) 10 per cent.of calcium chloride, and with a current density of 10 amperes per sq. dcm. at the anode and at least double that amount at the cathode; the yield is nearly 90 per cent. The deposit of calcium hydroxide on the cathode considerably increases the E.M.F. required ; at 50°, however, 4.55 volts suffice, and the yield is but slightly reduced (80 to 87 per cent.). Determinations of oxygen in admixture with hydrogen should not be made in the phosphorus pipette, but by means of copper in presence of a solution of ammonia. T. E. Density and Molecular Weight of Ozone. By ALBERT LADENBURG (Ber., 1898, 31, 2830-2831. Compare this vol., ii, 18).- The calculation of the amount of ozone present in the gas employed in the experiments previously described was based on the assumption that 253.06 parts of iodine are liberated by 48 parts of ozone.Since the object of the experiment was t o determine the molecular weight of ozone, the author has recalculated the results on the simpler assump- tion that 1 molecule of iodine is set free by 1 molecule of ozone. He thus finds that the gas employed contained 84.4 per cent. by weight of ozone, and that the density of ozone is 1.469. Colour of Sulphur Vapour. By J. LEWIS HOWE and S. G. HAMNER (J. Anzev. Chem. Xoc., 1898, 20, 757-759).-The authors, after referring to the varying statements with regard to the colour of sulphur vapour, describe experiments which show that this colour varies with the temperature, being orange-yellow just above the boiling point of sulphur, becoming rapidly darker with rise of temperature A.H.90 ABSTRACTS OF CHEMICAL PAPERS. until it becomes deep red comparable to the red of ferric thiocyanate ; this red colour is most intense at about 500". Above this temperature, the colour becomes lighter until at 634O, the limit of temperature at which observationa were made, the colour is straw-yellow. G.. W. P. H. The Change in Sulphur by Heat. By FRIEDRICH W. KUSTER (Zeit. cmorg. Chem., 1898, 18, 365-370).-The author has determined the amount of insoluble sulphur which is formed on heating sulphur at known temperatures for a known time. The sulphur, pTeviously crystaIlised from carbon bisulphide, is sealed up in a vacuum in a glass tube, and, after heating, the insoluble portion is determined by ex- traction with carbon bisulphide (so-called insoluble sulphur is slightly soluble in carbon bisulphide).When heated at 141.7", about 5.2 per cent. of insoluble sulphur is formed, the amount formed being independent of the time ; after 1 hour, this amount is approximately the same as after 16 hours. When heated at 1 8 3 O , the amount of insoluble sulphur is also independent of the time, and also apparently of the temperature, since about 6 per cent. of insoluble sulphur was obtained. When heated at 448" for 15 minutes and then gradually cooled, 1.8 to 3.3 per cent. of insoluble sulphur is formed ; when, how- ever, the molten sulphur is suddenly cooled by plunging into cold water, 30.9 to 34.2 per cent.of insoluble is obtained. The formation of the insoluble modification and the converse formation of soluble sulphur take place, therefore, with extreme rapidity, so that the amount of insoluble sulphur which is present after crystallisation is not dependent on the temperature and time of the heating, but on the rate of the crystallisation and the temperature at which it takes place. Therefore, different samples of sulphur which have been heated for different times above the melting point after remaining some time at a lower temperature (loo'), are practically identical as regards the concentration of the insoluble modification, from which it follows that the difference observed in the velocity of solidification, &c., must be assigned to some other cause than the different concentration of the insoluble modification (Abstr., 1897, ii, 439).A sample of insoluble sulphur which has been kept for 5 months, when treated with carbon bisulphide, gives the same percentage of soluble matter as when freshly prepared. On evaporating the solution, the sulphur separates in solid drops which show no signs of crystal- lisation under the microscope, but give evidence of crystallisation when subjected to polarised light. This sulphur is not completely soluble in carbon bisulphide, so that the '' insoluble " modification, when dissolved, is not completely converted into the soluble modification. The author concludes that the soluble and insoluble modifications have different molecules in solution, that they are not only physical isomerides, but chemical isomerides, having a relation to each other similar to that of ozone to oxygen.By THEODOR CURTIUS and JOHANNES Rrssonl (J. pa. Clbem., 1598, [ii], 58, 261-309. Compare Abstr., 1891, 57 ; 1892, 1 l2).-The azoimide, HN,, was obtained in aqueous solution by dis- tilling with dilute sulphuric acid either the ammonium salt (Abstr., E. C. It. Azoimide.INORGANIC CHEMISTRY. 91 1892, 113) or the lead salt precipitatied from the mother liquor of that salt; excess of acid should be avoided, especially in the latter case. Azoimide is decompoeed but very slowly when boiled with dilute hydrochloric acid ; most of the nitrogen is liberated as the gas ; a little ammonia is formed, but neither hydroxylamine nor hydrazine. In aqueous solution at the ordinary temperature, azoimide is very stable; such loss of strength as does occur is due t o volatilisation.The metallic salts (azoimides, azides, or nitrides) were prepared (1) by precipitation, in the case of AgN,, HgN,, Pb(N,),, TlN,, and Cu(N,),; (2) by dissolving the metal in the dilute acid of 16-17 per cent. strength (applicable in the case of Zn, Fe, Cd, and Mn, but the solutions are decomposed on evaporation, basic azoimides or even the hydroxides of the metal being formed and azoimide given off); (3) by dissolving the freshly precipitated hydroxide or carbonate of the metal in the aqueous acid and evaporating the solution; (4) by double decomposition of the sulphate of the metal with barium azoimide and evaporation of the filtered solution.In the analysis of the salts, the nitrogen was sometimes determined by combustion, the substance being mixed with plenty of powdered lead chromate in a long porcelain boat, but more often by distillation with dilute sul- phuric acid, the azoimide evolved being collected in excess of N/10 potassium hydroxide, of which the excess was estimated with N/10 hydrochloric acid, phenolphthalein serving as the indicator ; in the residue from the distillation, the metal was determined. Crystallographically the azoimides of potassium, rubidium, and thallium appear to form an isomorphous tetragonal group, those of barium, calcium (and stron- tium?) an isomorphous rhombic one; in all cases, the double re- fraction is very marked. As regards solubility in water a t 16O, the azoimides of Na, I(, Rb, Cs, NH, arrange themselves, as compared with the corresponding halogen salts, in the order of increasing solubility F, C1, N,, Br, I (which is also that of increasing formula-weight), those of Ba, Ca, Sr, Li, T1 in the order F, N,, C1, Br, I. When the azoimides of the alkali and alkaline earth metals are heated in a small capillary tube closed a t one end (melting point tube), azoimide is evolved, and the metal is left (in this way, small quantities of Cs, Rb, Ba, Sr, Ca may be prepared) ; none of them are very explosive, only lithium and the alkaline earth azoimides exploding at comparatively Iow temperatures, and only thallium azoimide when hammered ; in aqueous solution, they attack glass, although the acid itself has no such action; in aqueous solution, they are stable, but the'solution has often an alkaline reaction.The azoimides of the heavy metals are often very ex- plosive, perhaps most of all the potassium-platinum derivative, which explodes spontaneously with frightful violence, even in aqueous solution. No azoimide has been found as yet to crystallise with water. Ammonium azoimidg NH,*N,, melts and decomposes violently at 160" ; very volatile ; the vapour density at 100' in Hofmann apparatus gave themolecular weight = 29.3; 30.3 (calculated 60), so that complete dissociation must have taken place ; the spectrum resembles that of am- monium chloride, showing the hydrogen lines and the red-yellow part of the nitrogen spectrum ; crystals apparently rhombic. Hydrazine92 ABSTRACTS OF CHEMICAL PAPERS, azoimide, N2H59N3, begins to melt at 65', and decomposes energetically at 108'.Lithium axoimide, LiN, : crystals anisotropic ; deliquescent ; explodes between 115' and 298'. Sodium azoimide, NaN3 : crystals apparently hexagonal ; unchanged a t 350'. Potassium axoimide, KN, : crystals tetragonal ( a : c = 1 : 0.5810) ; melts and decomposes above 350'. Rubidium azoimide, RbN, : the best crystallised of the salts examined ; crystals tetragonal [a : c = 1 : 0.57851 ; slightly hygro- scopic ; melts at 330-340'. Casium axoirnide, CsN,, crystalline, deliquescent ; melts at 310-318'. Thallium azoimide, TIN, : crystals, tetragonal [a : c = 1 : 0*5SS] ; explodes when struck, also when heated strongly, but is unchanged a t 340'. Calcium axoimide, Ca(N3)2 : crystals rhombic [a : b : c = 0.320'7 : 1 : 0.8815 ( P ) ] ; hygroscopic ; explodes at 144-156'.Strontium azoimide, Sr(NJ2 : crystalline ; hygroscopic ; decomposes violently at 194-196O. Barium axoimide, Ba( N3)2: crystals rhombic [a : b :c = 0-3424 : 1 : 0.84611; hardly hygroscopic ; decomposes violently at 2 1'7-221". Magnesium and beryllium axoimides are readily decomposed by hot water. Basic zinc uxoimide, N,*Zn*OH (a) ; crystals ill-defined, anisotropic. Basic manganese axoimide, N,mMn*OH. Cadmium axoimide, Cd (N3)2 : yellow, biaxial crystals ; forms with pyridine a colourless, crystalline compound, Cd(N,),,SC,NH,. Cupric axoimide, Cu(N,),, from copper sulphate and sodium azoimide, but also by dissolving copper in the aqueous acid ; dark brown and crystalline ; insoluble in water ; very explosive.Aluminium forms no azoimide ; from a solution of the sulphate, sodium azoimide precipitates the hydroxide. Chromium azoimide is formed in solution by dissolving chromium hydroxide in aqueous azoimide ; it decomposes on evaporating the solution, the residue containing only 2N, per 3Cr. Ferrous sulphate gives a colourless solution with cold aqueous azoimide ; on boiling, a yellow solid is precipitated. The solution turns red when shaken in the air; a deep red solution is also obtained when solutions of ferric chloride and azoimide are mixed ; this becomes colourless slowly in t-he cold, rapidly wheu boiled, all the iron being precipitated. Tin is precipitated from a solution of stannous chloride by sodium azoimide, and the precipitate is, in part at any rate, an azoimide.Basic nickel axoirnide, N,*Ni*OH, with some Ni(N3)2(?), from nickel carbonate and aqueous azoimide : green, crystalline; explodes a t 247-271'. Basic cobalt axoimide, N,*Co*OH, with some CO(N,)~ : violet and possibly amorphous ; potassium cobaltoaxoimide, KN,,Co(N;),, precipitated when strong solutions of the two azoimides are mixed, is bright blue (gives a pink solution) and crystalline, and explodes at 225' ; the ammonium analogue, (NH4)Nq(CoN,),, is similar in appearance and properties. An analogous bright green nickel com- pound, KN,,Ni (NJ2(P), also crystalline and explosive, was obtained. By mixing strong solutions of platinochloric acid and potassium azoimide, or aurochloric acid and sodium azoimide, solutions were obtained from which, in the first case, a brownish-red, in the second an orange crystalline, residue, extremely explosive in both cases, remained on evaporation (potassium platinoazoimide and sodium auroazoimide 9).C. P. B.INORCIANIC CHEMISTRY 3 93 Metaphosphimic Acids. 111. By HENRY N. STOKES (Amer. Chew. J., 1898, 20, 740-760 and Zeit. Anorg. Chem., IS, 36-58. Compare Abstr., 1897, ii, 28 and 94).-Although penta- and hexa-phosphonitrilic chlorides (Abstr., 1898, ii, 70), on hydrolysis give rise to the corre- sponding pen ta- and hexa-metaphosphimic acids, from heptaphospho- nitrilic chloride the acid (H,PNO,), + H20 is obtained. The salts of all three new acids differ from those of tri- and tetra-metaphosphimic acid in being amorphous; hence, in many cases, they cannot be obtained satisfactorily pure, Pentametaphosphimic acid, N H PO( OH) *.NH * PO(OH)>NH, PO(OH)<NH.PO(OH).NR.PO(OH) is best prepared by shaking pentaphosphonitrilic chloride (4 parts), dissolved in ether free from alcohol (20 parts), with a solution of sodium hydroxide (5 parts) in water (20 parts) during 60 hours.On adding alcohol, the pentasodium salt, P5N501,H5Na5 + 2H,O, is precipi- tated as a thick syrup, which is mashed repeatedly with 60 per cent. alcohol, dissolved in water, reprecipitated by alcohol and again washed until free from sodium chloride; it is then freed from water by stirring several hours with frequently renewed absolute alcohol. Thus prepared, it is a white, sandy powder, which retains 2H20 after dryingat loo', and dissolves in water with developmentof heat, the solution having an alkaline reaction. When dissolved in 80 per cent.acetic acid, i t is converted into the tetrasodium salt, P5N50,,H6Na4 + 2H20 ; the latter is precipitated on adding alcohol and resembles the pentasodium sdt, but has a neutral reaction, On adding a magnesium salt to an aqueous solution of sodium pentametaphosphimate strongly acidified with acetic acid, the mclgnesium salt, P5N50,,H6Mg2 + 5H#, is preci- pitated ; on dissolving tbis in dilute nitric acid, adding ammonia until a precipitate just forms, and filtering, a solution containing the salt (P,N,0,,H9)2Mg is obtained. The silver salt, P5N501,H,Ag5, precipi- tated by adding the calculated quantities of nitric acid and silver nitrate to II solution of the tetrasodium salt, is a white powder, which is not affected by light or by heating a t 100'; it is decomposed, how- ever, by cold caustic alkalis.Salts containing more than five atoms of silver can be obtained, which are yellow in colour. When silver nitrate is added to an ammoniacal solution of sodium pentametaphos- phimate, a yellow, amorphous salt, P5N5011H3Ag9, is obtained, which i g probably a derivative of amidotetrimidopentaphosphoric acid, OH *PO( NH2)* [NH*PO(OH) J3*NH* PO(OH),. Somewhat impure pen ta- metaphosphimic acid separates as a gelatinous precipitate on adding alcohol to the filtrate obtained after decomposing its silver salt with hydrogen sul phide . When sodium pentametaphosphimate is heated during 8 hours with dilute acetic acid, it is decomposed into a mixture of the sodium salts of tetrametaphosphimic acid (Abstr., 1897, ii, 94), tri-imidotetra- phosphoric di-imidotriphosphoric and orthophosphoric acids, together with other substances.The first of these acids is completely precipi- tated during the reaction as a sparingly soluble acid sodium salt, in spindle-shaped crystals. On concentrating the filtrate, and adding sodium acetate, sodium tri-imidotetraphosphate, probably P4N,01,H5Na,, crystallises out in small rhombic or hexagonal plates ; on dissolving in VOL. LXXVI. ii. 794 ABSTRACTS OF CHEMICAL PAPERS. water and adding silver nitrate, the silver salt, P,N,O,,H, Ag,, separates as an amorphous precipitate which becomes crystalline on washing with water. Sodium ILexameta~hosp~~i~~~~~te, Y,N,O1,HGNa, + 2H,O, prepared from hexaphosphonitrilic chloride, closely resembles sodium pentameta- phosphimate.No definite magnesium salt could be obtained, but the silver salt, P6N60,,H6Ag6, separates on adding the calculated quantity of nitric acid and silver nitrate to a solution of the sodium salt ; i t forms a white, gelatinous precipitate, and is decomposed by cold caustic patash. I n presence of ammonia, a yellow silver salt is pre- cipitated. Free hexametaphosphimic acid cannot be obtained, as on decomposing its silver salt suspended in water by hydrogen sulphide, filtering, and adding alcohol, no precipitate separates ; on attempting to evaporate, decomposition occurs, and a gum-like residue is obtained. On heating sodium hexametaphosphimate with dilute acetic acid, 30 per cent.of the theoretical quantity of tetrametaphosphimic acid is obtained. When the hydrolysis of heptaphosphonifrilic chloride is effected by sodium hydroxide as in the cam of pestaphosphonitrilic chloride, amidoheximidoheptaphosphoric acid, OH*PO(NH,)* [NHePO(OH) J6*NH*PO(OH),, is formed j the sodium salt, P7N70,,H,.57Na7.43 + 2H,O, and the silver salt, P7N70,bH9Ag7, were analysed. Attempts to prepare amides by acting on ethereal solutions of tetra- and penta-phosphonitrilic chloride with aqueous ammonia gave amorphous substances of indefinite composition. In discussing his results, the author points out that tetrametaphos- phimic acid is by far the most stable of the metaphosphimic acids. To explain this, and the formation of tetraphosphimic acid from the penta- and hexa-acids by hydrolysis, a hypothesis as t o the stability of the acid rings is put forward, which is similar to von Baeyer’s tension theory of carbon rings. By ALBERT L&VY and H.HENRIET (Compt. rend., 1898, 127,353-355).--When air that has been completely freed from carbonic anhydride by passing it over potassium hydroxide is allowed to remain for 2 hours in contact with potassium hydroxide solution, a further quantity of carbonic anhy- dride is formed (compare Abstr., 1898, ii, 573). I f the potassium hydroxide solution contains 7 grams per litre, all the carbonic anhydride existing as such in the air is completely absorbed in 10 minutes, and all the organic matter in the air in the form of gas or vapour is com- pletely oxidised in 2 hours.Experiments were made during July, 1898, in which the carbonic anhydride absorbed by the potassium hydroxide was determined after it had been in contact with the air in large flasks for 10 minutes and 2 hours respectively. The difference was often from 3 to 6 litres per 100 cubic metres of air, but sometimes was as high as 11 or 12 litres, and in one exceptional instance it reached 56 litres. It is obvious that this difference gives useful information as to the quantity of gaseous organic matter present in the air. I n all cases, thedifference observed W. A. D. Atmospheric Carbonic Anhydride.INORGANIC CHEMISTRY, 95 in the manner indicated is somewhat less than its true value, because even in 10 minutes the organic matter is appreciably oxidised in presence of the alkali.C. JI. B. The Various Theories relating to the Constitution of the Arnrnonio-metallic Salts. By FRITZ REITZENSTE~ N (Zed. anopg. Chem., 1898, 18, 152--210).-An historical review of the theories that have been put forward on this subject. By LOUIS KAHLENBERCI and AZARIAH T. LINCOLN (J. Physical Chem., 1898, 2, 77-90).-The conclusion arrived at by Kohlrausch (Abstr., 1893, ii, 166) that in solutions of sodium silicates these salts are hydrolytically decomposed into sodium hydroxide and colloidial silicic acid, has been confirmed by investigating the freezing points of such solutions. The freezing points and the electrical conductivity of solutions of the silicates of potassium, lithium, rubidium, and cesium show that these salts are also decomposed by water into colloidal silicic acid and the hydroxide of the alkali metal.The silicates of the alkalis all show an analogous behaviour when dissolved in water. The same solution is obtained whether a silicate is dissolved in water, or whether solu- tions of caustic alkali and colloidal silicic acid in proper propor- tions are mixed, Since colloidal silicic acid has but little effect on the freezing point, the degree of hydrolytic decomposition of the silicates can be calculated from the lowering of the freezing point of their solutions. Silicates of the general formulae M,SiO, and MHSiO, are practically completely hydrolytically dissociated when one gram-molecule is contained in 48 litres. Silicates of the general formula M,Si,O,, are practically completely decomposed by water when one gram-molecule is present in 128 litres.A comparison of the electrical conductivity of silicate solutions with that of solutions of the alkali hydroxides shows that the values of the former approach the latter as the solutions become more dilute, the retarding influence that the silicic acid has on the mobility of the ions gradually becoming less. Fzom the results of the above investigations, it appears safe t o con- clude that, in natural waters, silicic acid always occurs in the colloidal state; only in very rare instances are the solutions of the silicates so concentrated that they are not practically completely hydrolytically decomposed. H. C. Sodium Oxides. By ROBERT DE FORCRAND (Compt. wnd., 1898, 127, 364--366).-When dry air, free from carbonic anhydride, is passed through sodium heated somewhat above its melting point, the first product is a bulky, grey, arborescent mass, consisting of the sub- oxide Na,O, mixed with a small proportion of sodium.If the heating is continued, this substance burns, and is converted into a yellow mixture of the monoxide and dioxide, or completely into the dioxide Na,O?, which, however, generally retains a small proportion of water. No trioxide is formed. The suboxide does not rapidly absorb water vapour, but is gradually oxidised when exposed to air ; if thrown into water, it reacts violently, with liberation of pure hydrogen. E. C. R. Solutions of Silicates of the Alkalis. C. H. B. 7-296 ABSTRACTS OF CHEMICAL PAPERS. Interaction of Sodium Arsenite and Sodium Thiosulphate.By LEROY W. MCCAY (Chem. News, 1898, 78, 209).-When sodium thiosulphate and sodium arsenite in the theoretical quantities are made into a thick paste with caustic soda and rubbed in a mortar, the following change takes place : Na,As03 + Na2S203 = Na,As03S + Na2S03. The sodium orthomonothioxyarsenate is separated by treat- ing the paste with water, filtering, and crystallising. It is generally obtained by similar treatment of a paste of sulphur, caustic soda, and arsenious oxide in stochiometrical amounts. D. A. L. By H. VAN ERP (Rec. Trau. Chim., 1898, 17, 296-299).-As a rule, the efflorescence formed on walls consists almost entirely of sodium sulphate and carbonate, with varying quantities of water of crystallisation ; nitrates, nitrites, phosphates, and ammonium salts can seldom be detected, but calcium carbonate and sand are usually present.In the case of a wall of a corridor in the Harlem Museum, which had been covered for some time with a ciirtain, it was found that, beneath the latter, a deposit of slender needles, several centimetres in length, of pure sodium sulphate (with 10H20) had formed. Ammoniacal Lithium Chlorides. By J. BONNEFOI (Compt , rend., 1898, 127, 367-369).-The compound LiCI,NH3 is formed by the action of dry ammonia on pure and dry lithium chloride at a tem- perature exceeding 85", or by heating the compounds containing a higher proportion of ammonia. Its heat of dissolution at 15" is + 5.385 Cal. and hence, LiCl sol. + NH, gas = LiCl,NH, sol. develops + 11.842 Cal. Saline Efflorescence of Walls.W. A. D. Its vapour pressures are as follows. t 8 8 O 9 6" 109.2" 119" p 256 mm. 367 mm. 646 mm. 975 mm. and the heat of formation calculated by Clapeyron's formula agrees closely with the number directly determined. The compound LiCI,BNH, is obtained by the action of ammonia on lithium chloride between 60" and 85", or by heating the higher com- pounds between these limits. Its heat of dissolution is +2.668 Cal. hence, LiCl sol, + 2NH3 gas = LiCl,BNH, sol. develops + 23.355 Cal Its vapour pressures are i? 68.5" 7 7" 83' 89.2" p 373mm. 658 mm. 739 mm. 980mm. The compound LiCl,SNH, is formed between 20" and 60". Its vapour pressures are t 43" 5 0" 60" 62.2" 65' p 320 mm. 473 mm. 790 mm. 882 mm. 1011 mm. The oompound LiC1,4NH3 is formed below 13"; its heat of dissolution is + 0.292 Cal., hence LiCl sol.+ 4NH, gas = LiC1,4NH3 sol. develops + 43.335 Cal., LiCl,SNH, sol. + NH, gas = LiC14NH3 sol. develops + 8.879 Cal. Its vapour pressures are t O0 9" 14.5" P 384 mm. 640 mm. 850 mm.INORGANIC CHEMISTRY. 97 The heat developed by the combination of the four successive molecules of ammonia with the lithium chloride is + 11.843, + 11.517, + 11 -097, and + 8.879 Gals. respectively. In all cases, Ulapeyron's formula gives results that agree with the direct determinations (compare Abstr., 1897, ii, 371). Dicalcium Phosphate. By A. BAR ILL^ (Chew. Cents*., 1898, i, 434-435 ; from Rgp. Pharm., 1897, 529).-When ammonia is added to a solution of monocalcium phosphate, half the phosphoric acid remains in solution as ammonium phosphate, and dicalcium phosphate is precipitated.By adding calcium chloride solution to the filtrate, the rest of the phosphoric acid is precipitated as monocalcium phos- phate. According to the author, the latter reaction takes place in three phases. Ammonium chloride, tricalcium phosphate, and tetra- hydrogen calcium phosphate are first formed, then by the action of water the last compound forms tricalcium phosphate and phosphoric acid, and finally the whole of the tricalcium phosphate acts on the phosphoric acid t o form monocalcium phosphate. To prepare dicalcium phosphate, 1000 grams of well burnt bone ash are stirred with hot water and then gradually mixed with 1454 grams of commercial hydrochloric acid of sp.gr. = 1.17. When the action is completed, 3 litres of hot water are added, the solution of monocalcium phosphate and calcium chloride is filtered, the filtrate diluted to 10 litres and a mixture of 442 grams of ammonia of sp. gr. = 0.925 with 20 times its weight of water is slowly added. The precipitated dicalcium phosphate, after collecting, washing with water until free from acid and drying at 60°, is a light, snow-white, lustrous powder, which consists of flat, hexagonal crystals. Dicalcium phosphate crystallises with 4H,O. I t s solution is not decomposed by boiling or evaporating, and its saturated solution in water containing carbonic anhydride contains 0.069 per cent. of the salt, and might be used medicinally instead of Crystallisation of Anhydrous Calcium and Strontium Sul- phides.By A. MOURLOT (Compt. rend., 1898, 127, 408-410).- Calcium or strontium sulphide can be obtained in a crystalline condi- tion by heating a moderately large quantity of a mixture of the cor- responding sulphate with carbon for a few minutes only in an electric furnace with a current of 1000 amperes and 60 volts. If the amorphous sulphides are heated in a carbon dish in an electric tube furnace until they are completely melted, they crystallise on cooling. Both sulphides crystallise in the cubic system; the sp. gr. of the calcium compound = 2.8 a t 15O, and of the strontium compound = 3.7 at 15O. I n chemical properties, they closely resemble anhydrous barium sulphide (Abstr., 1898, ii, 376). Composition of Phosphorescent Strontium Sulphides.By Josh R. MOURELO (Compt. rend., 1898, 127, 229-231).-Strontium sulphide, prepared from minerals, contains strontium sulphate, in quantity depending on the degree of exposure to air, and also small quantities of sodium chloride ; it may also contain calcium, barium, iron, and aluminium compounds. When prepared from pure strontium C . H. B. tricalcium phosphate. E. w. w. C. H. B.9s ABSTRACTS OF CHEMICAL PAPERS. sulphate, carbonate or oxide, it usually contains some sulphate, but if special precautions are taken and it is obtained quite pure, it is non- phosphorescent. 0. H..B. Phosphorement Strontium Sulphide. By Josk R. MOUIZELO (Compt. rend., 1898, 127, 372--374).-Strontium thiosulphate, pre- pared by the action of sodium t,hiosulphate on strontium chloride, retains with great tenacity small quantities of sodium chloride, and consequently this salt is present in any strontium sulphide prepared from it by the action of heat ; the phosphorescence is yellowish-green, and somewhat intense.An intensely phosphorescent sulphide is obtained by a method very similar to that adopted by Verneuil for the preparation of calcium sulphide. For 100 grams of strontium carbonate, 2 grams of bismuth subnitrate, 2 grams of sodium carbonate, and 0.12 gram of sodium chloride is taken ; after moistening the carbonate with the sodium salts, it is heated to redness, and then mixed with 21 grams of sulphur and the bismuth salt, and heated to bright redness for 4 hours. The pro- duct is relatively very stable. C. H. B. Double Compounds of Cerium Tetrachloride.By IVAN KOPPEL (Zeit. anorg. Chem., l898,18,305--31l),-A solution of cerium tetra- chloride is obtained by treating hydrated cerium dioxide, suspended in methylic or ethylic alcohol, with dry hydrogen chloride ; it forms a yellowish-brown solution, which deposits yellow, crystalline crusts when evaporated in a vacuum, and decomposes at once, with evolution of chlorine, when removed from the mother liquor or when brought into contact with water. This solution forms stable, double compounds with pyridine, quinoline, and the triethylamine chlorides, but not with mono- or di-ethylamine hydrochloride. I n the preparation of these compounds, the author employed commercial cerous oxalate, and he points out that by precipitating the cerium in the form of one of these double salts, a very easy method is obtained for the preparation of pure cerium compounds.Dipyrzne cerium hexachloride, ( U,NH,),,H2CeC16, obtained by adding pyridine hydrochloride to the above solution of cerium tetra- chloride and then precipitating with ether, crystallises in small, lustrous, quadratic leaflets, is easily soluble in methylic alcohol, less so in ethylic alcohol, and is at once decomposed by water, with evolu- tion of chlorine. Diquirzolim cer6m-z hexachlode, (C,NHy),,H2CeC1,, obtained in a similar manner to the preceding salt, is a yellow, crystalline precipi- tate, less soluble than the preceding salt, and at once decomposes when treated with water. Di-triethylarnirye cerium hexachloride, (NEt,),,H,CeCI,, crystallises well from alcohol in octahedra.Yellow, crystalline salts, which are probably the double compounds K,CeCl, and (NH,),CeCI,, are obtained by adding finely powdered potassium or ammonium carbonate to the solution of cerium tetra- chloride containing strong hydrocbloric acid, and then shaking with a large quantity of ether ; when separated from the solution, they a tINORGANIC CHEMISTRY. 99 once decompose, with evolution of chlorine. to obtain double salts with calcium, mercury, or ferric chloride. The author was unable E. C. R. Praseodymium. By CARL VON SCHEELE (Zeit. cmorg. Chem., 1898, 18, 352-364. Compare Abstr., 1898, ii, 519).-Praseodymium chloride, PrCl, + 7H20, crystallises in large, deliquescent crystals when the concentrated solution is allowed to remain over sulphuric acid ; sp.gr. at 16' = 2.251. The bromide, with 6H20, is obtained by dissolving the oxide in hydrobromic acid. The ptutinochloride, PrCl,,PtCI, + 1 2H20, crystallises in large, yellow crystals and slowly gives off water when allowed to remain over sulphuric acid. The platinobromide, with 10H20, separates in large, dark red crystals and effloresces when allowed to remain over sulphuric acid. The uuri- chloride, with 10H,O, crystallises in beautiful, deliquescent crystals and is very soluble in water. The auribi-omide crystallises in long needles which deliquesce extremely rapidly on exposure to the air. Praseodymium pZatinocyanide, 2Pr(CN),,3Pt(CN), + 1 8H20, obtained by adding the theoretical quantity of praseodymium sulphate to a solution of barium cyanide and platinum cyanide, crystallises in black, fluorescent prisms ; i t effloresces over sulphuric acid, becoming red, and giving off 4H20 ; sp.gr. at 1 6 O = 2,663. PrcGseodymium nitrate, Pr(N03)3,+ 6H20, cry stallises in long, deli- quescent needles. The double sodzum salt, Pr(N03),,2NaN0, + H30, crystallises in small, deliquescent needles. The ammmium salt, with 4H20, crystallises in large, deliquescent crystals ; sp.. gr. = 2.155. Praseodymium sulphute, Pr,3S04 + 158H20, is obtained by allowing a dilute solution to crystallise very slowly ; it crystallises with 8H20 at the ordinary temperature; sp. gr.=2-822; and with 5H20 when a saturated solution is concentrated on the water-bath. The an- hydrous salt has the sp. gr.=3*720; 23.64 parts are soluble in 100 parts H,O at O', and 17.7 parts are soluble in 100 parts H20 at 20'.The double potassium salt, Pr2(S0,),,3K2S0, + H20, is a heavy, crystalline precipitate, very sparingly soluble in water and easily soluble in hydrochloric and nitric acids ; sp. gr. = 3.275. The double ammonium salt, with 8H20, separates in large crystals, is not altered by exposure to the air or by remaining over sulphuric acid, is sparingly soluble in water, and becomes anhydrous when heated a t 1'70'; sp. gr. = 2.532. Praseodymium selenate, Pr,(SeO,), + 8H20, ig obtained by allowing a solution of the oxide in selenic acid to crystallise over sulphuric acid ; sp. gr. = 3.094. It crystallises, with 5H20, in prisms when the solution is concentrated on the water-bath; the anhydrous salt is' obtained by heating the preceding salt a t 200' ; sp.gr. = 4.305. The double potassium salt, Pr2(SeO4),,3K2SeO, + 4H20, is similar to the corresponding sulphate. A double salt with ammonium selenate could not be obtained. Pruseodyinium dithionnte, Pr2(Sa06)s.+ 12H,O, obtained from praseodymium sulphate and barium dithionate, crystallises from the syrupy solution mhsn it is exposed over sulphuric acid, rapidly deli-100 ABSTRACTS OF CHEMICAL PAPERS. quesces on exposure t o the air, and evolves sulphurous anhydride when warmed. The acid selenite, Pr,(SeO,),,H,SeO, + 3H20, obtained by pre- cipitating the sulphate with sodium selenite and then adding a solution of selenic acid, crystallises in slender needles. The carbonate, Pr2(C0,), + 8H,O, obtained by treating the hydroxide, suspended in water, with carbonic anhydride, or by precipitating the chloride with ammonium carbonate or potassium hydrogen carbonate, crystallises in small, lustrous scales.The oxalate, Pr2(C204), + lOH,O, is a crystalline precipitate, soluble in concentrated acids, and does not form double salts with alkaline oxalates. The acetate, with 2H,O, crystallises in slender needles at the ordinary temperature, and with 1+H20 in small, slender needles at the temperature of the water-bath. The propionate, when evapor- ated over sulphuric acid, crystallises, with 3H20, in large prisms, and on the water-bath, with lH,O, in large, thin, lustrous leaflets. Behaviour of Thallium in Acid Solution towards Hydrogen Bulphide in Presence of Arsenic, Antimony, and Tin.By JOSEF LOCZKA (Chem. Centr., 1898, i, 657 j from Magyar Chrnicci Folybirat, 3).-Thallium is not precipitated by hydrogen sulphide in acid solu- tions, but in presence of dissolved arsenic, antimony, or tin, red precipitates are formed which contain thallium. The larger the amount of acid present, the less the amount of thallium precipitated, and, with a large excess of acid, no thallium is thrown down. The arsenic precipitate, TlAsS,, occurs naturally in a crystalline form in the realgar from Alchar in Macedonia. By ROBERT MELDRUM (Chern. News, 1898,78, 209--210).-In these experiments, 7 feet of bright copper wire, &-inch in diameter, was exposed, in a test tube, to 100 C.C. of the water ; lake water, waters with and with- out free ammonia, and distilled water were all found to dissolve copper to a certain extent, some hundredths per 100,000 even in 24 hours.I n the lead experiments, pieces of the same lead piping, 2 inches long and $-inch bore, were closed at one end and each frequently filled and agitated with a different water during various periods. One water with a total hardness of 3 * 5 O and a permanent hardness of 3 9 O , dissolved considerably more lead than another with a total hardness of 18.6O and permanent hardness of 5 ’ ; a relationship that was not altered by charging both with carbonic anhydride. But the activity of the first was destroyed by thoroughly agitating with precipitated calcium carbonate and filtering. Cause of the Retention and Release of Gases Occluded by Metallic Oxides.By THEODORE W. RICHARDS (Amer. Chern. J., 1898, 20, 701--732).--Scott has recently stated (Trans., 1897, 559) that copper oxide prepared by exposing the nitrate to a “full red heat” in a muffle-furnace, contains only one-tenth the volume of occluded gas previously found by the author to be present in ignited copper oxide (Abstr., 1891, 805); the statement is, however, not a E. C. R. E. W. W. Action of Water on Metallic Copper and Lead. D. A. L.INORGANIC CHEMISTRY. 101 contradiction of the latter's results, since these were obtained at much lower temperatures, in most cases not exceeding 700". Experiments are now described which confirm the conclusions formerly arrived at : that copper oxide prepared by the ignition of the nitrate by Hampe's method, contains between 4-5 times its volume of occluded gas, and that the latter is partially expelled at a red heat.Nearly the whole of the gas is retained until the temperature rises above 850', when nine-tenths of it is rapidly evolved ; a t higher temperatures, the oxide is partially decomposed, with loss of oxygen. The amount of gas retained by the oxide below 850' does not depend on the time of ignition, although it is dependent on it above this temperature. When zinc oxide, prepared by heating the pure nitrate during a long period a t 280°, is ignited a t temperatures between 660' and 880', the amount of gas retained diminishes as the temperature rises, other con- ditions being the same ; continued heating a t one temperature causes a slow evolution of the gas. The total amount of gas occluded by zinc oxide varies greatly according to the method used for preparing the latter, and according to its physical condition. Whilst the actual amount of nitrogen occluded by the same specimen of zinc oxide is independent of the temperature a t which it is subsequently ignited, the quantity of oxygen retained decreases as the temperature is increased. Richards and Rogers' results (Abstr., 1894, ii, 45) are thus confirmed, whilst Morse and Arbuckle's statement to the contrary (Abstr., 1897, ii, 334) is shown to be incorrect.In the case of copper oxide, the more rapid evolution of occluded oxygen a t high tempera- tures is still more marked, but it is not so evident in the case of magnesium oxide. From his results, the author concludes that the occluded gas is due t o a residue of basicnitrate imprisoned in the oxides in question; that the oxygen escapes more rapidly than the nitrogen when this residue is decomposed, is explained by assuming that the loss takes place, not by diffusion, but by chemical transference.According to this hypothesis, the particles of copper oxide in immediate contact with the imprisoned gases dissociate into copper and oxygen, and then remove a portion of the occluded oxygen by combining with it ; the oxygen is subsequently transferred from molecule to molecule by changes similar to those assumed to take place by Clausius's theory of electro- lysis, until it finally escapes. This view is supported by the fact that the occluded oxygen is lost much more rapidly by copper oxide than by the oxides of zinc and magnesium, which are far less easily dissociated. That metallic copper is liberated when copper oxide is ignited a t a red heat is shown by a great increase in the latter's conductivity ; under the same conditions, the conductivity of zinc oxide was much less changed, whilst that of magnesium oxide re- mained unaffected.It also appears th& when copper oxide is heated in a vacuum a t 790", oxygen is evolved as long as i t is rapidly removed, cuprous oxide being formed. The apparent volatility of cadmium oxide a t high temperatures is probably due to its initially dissociating into oxygen and cadmium, which sublimes and i s then reoxidised. W. A. D.102 ABSTRACTS OF CHEMICAL PAPERS. Action of Hydrogen Phosphide on Cupric Sulphate. By E.RUB~NOVITCH (Compt. rend., 1898,127,270-273).-When hydrogen phosphide is allowed to act on cupric sulphate solution without any special precautions, the composition of the product is complex and variable, but if the pure gas, prepared from phosphonium chloride, is allowed to act on the copper salt in complete absence of oxygen, a black product is obtained which consists solely of the phosphide P2Cu5 + H,O. One-third of the phosphorus in the hydrogen phosphide is converted into phosphoric acid, and the acid of the cupric salt is found in the free state in the liquid 10CuS04+5Ph,+6H20= 2(P2Cu5 + H,O) + H3PO4 + 10H2S04. The changes in the colour of the solution indicate that part of the cuppic salt is first reduced to cuprous salt. The black phosphide begins to lose water at 80' and is completely dehydrated at 150°, the colour changing to red-brown. It oxidises much more readily when anhydrous than when hydrated ; in contact with the liquid in which it was formed, it is readily oxidised by air with liberation of metallic copper and formation of more phosphoric acid.Nitric acid and bromine attack the phosphide readily ; potassium per- manganate oxidises it, and concentrated sulphuric acid dissolves i t with liberation of hydrogen phosphide. C. H. B. Aluminium as a Reducing Agent. By LBON FRANCK ( C h m Zeit., 1898, 22, 236--245).-When aluminium in a finely divided state is heated to redness in an atmosphere of hydrogen and phosphorus vapour, a vigorous reaction takes place, and a dark-grey, infusible phosphide is produced having the composition AI3P7.An intimate mix- ture of aluminium and phosphorus, heated to white heat in a current of hydrogen, yields another phosphide, A15P3, which forms a yellowish- grey, infusible mass. Dark blue, shining needles are obtained in small quantity by heating aluminium and phosphorus in sealed tubes, but their exact composition has not been determined. Two other phos- phides, A1,P and A1,P2, which have a metallic appearance and a crys- talline fracture, are obtained by the use of the electric furnace. All the above phosphides give off phosphine when treated with water or acids. A mixture of aluminium and sodium metaphosphate, when heated t o redness in a current of hydrogen, yields half the phosphorus in tho free state, whilst half remains combined with the metal, forming AI,P3 ; if silica be added to the mixture, the whole of the phosphorus is liberated.Calcium phosphate behaves in a similar manner, but calcium metaphosphate containing calcium sulphate cannot be employed in this experiment, as the presence of the latter salt causes explosions when the mixture is heated. Aluminium, like magnesium, decomposes carbonic anhydride and carbonic oxide, forming the oxide of the metal and free carbon. A mixture of the metal and lampblack, when heated to a white heat, gives rise to a certain amount of aluminium carbide, which, however, cannot be separated from the excess of carbon. The carbonates OF lithium, sodium, potassium, calcium, strontium,INOI.1CIANIC CHEMISTRY. 103 and barium are reduced on heating with aluminium, the products being alumina, carbon, and the free elements ; the latter alloy with the excess of aluminium, and in all cases a small amount of aluminium carbide is formed, Aluminium reduces the metallic oxides when heated with them t o sufficiently high temperatures; in the case of the oxides of silver, copper, and lead, the reaction proceeds with explosive violence.The oxides of iron, manganese, cobalt, nickel, chromium, and molybdenum are all partially reduced ; boron and silicon are easily obtained from their oxides. A mixture of powdered aluminium and sodium peroxide, when moistened with water, ignites spontaneously. Formation of Alums by Electrolysis. By J. LEWIS HOWE and E. A. O'NEAL (J. Amer. Chem. Xoc., 1898,20, 759--765).-The preparation of manganese alums, the primary object of the research, was not effected ; various alums were, however, prepared by passing currents of 0.02-0.1 9 amperes through solutions of sulphates of the metals together with excess of sulphuric acid.The anode consisted of a platinum dish or crucible, and the cathode, which was immersed in a porous cup, of a platinum wire. Iron ammonium alum was prepared by passing the current through a solution of the sulphates of ammonium and iron acidified with sulphuric acid; the alum is formed a t the anode, I n a similar manner, iron potassium alum, iron rubidium and iron casium alums, cobalt rubidium and cobalt cssium alums were prepared. Chromium ammonium alum was formed when the cathode dipped into a solution of ammonium chromate acidified with sulphuric acid.With manganese aulphate and the sulphates of ammonium, rubidium, and cssium, no satisfactory results could be obtained, although Piccini has just stated that he obtained caesium manganese alum by electro- lytic oxidation (Abstr., 1898, ii, 521). From solutions containing ruthenium, no alum was obtained, but it was observed that ruthenium tetroxide was evolved at the positive pole. Actionof Water and Saline Solutions on Metallic Iron. By ROBERT MELDRUM (Chem. News, 1898, '78, 202-203).-Numerous ex- periments have led the author to the conclusion that the oxidation of iron in water takes place in the absence of ammonia, carbonic an- hydride, or of bacteria and other forms of life, but it is uncertain whether the action is due to the water itself or to dissolved oxygen.The experiments were made with piano wire, in glass bottles free from lead, in the light ; generally, a white cloud formed immediately round the metal, which extended and in 15 minutes the whole liquid became cloudy and then yellow ; a yellow precipitate forming in 3 or 4 hours. Sometimes the internal surface of the glass was coated with an iridescent film, Erom which glistening plates became detached, and were found in the precipitate of hydrated oxides; this occurred when the surface of the iron exposed was great in proportion to the volume of water, and when there mas free contact with the air. I n bottles half full of solutions of pure salts, activity was observed in the following cases, the solutions employed being of 1 per cent, G.T. M. G. W. F. H.104 ABSTRACTS OF CHEMICAL PAPERS, strength-ammonia, sodium carbonate, sodium hydrogen carbonate, sodium hydrogen phosphate, potassium hydrogen phosphate, sodium metaphosphate, sodium pyrophosphate, and potassium borate, whilst sodium biborate and potassium biborate gave white precipitates. The following do not show any action : calcium hydroxide (saturated solu- tion), sodium peroxide, sodium hydroxide, potassium hydroxide, barium hydroxide, strontium hydroxide (5 per cent. solution). The experi- ments extended, in some cases, to 4 weeks, It seems that alkaline potassium salts are more active than the corresponding sodium salts. D. A. L. A Double Iron Tungsten Carbide. By P. WILLIAMS (Corn@. rend., 1898, In, 410--412).-0n heating a mixture of tungstic an- hydride, 150 grams ; iron, 250 grams, and petroleum coke, 80 grams, in a carbon crucible in an electric furnace with an arc from a current of 900 amperes and 45 volts, an iron tungsten carbide, most probably 2Fe3C,3W,C, is obtained.The product is treated two or three times with hot concentrabed hydrochloric acid, and the double carbide, which is highly magnetic, is separated by means of a magnet. It is then treated with aqua regia, which dissolves about half the crystals and leaves a residue of definite composition. The double carbide thus obtained forms very brilliant, microscopic prisms with the colour of pyrites ; it scratches glass easily, but not rock crystal ; sp. gr. = 13.4 at 18'. At a red heat, it is not affected by hydrogen or by hydrogen sulphide, but is attacked by chlorine, bromine, iodine, and oxygen ; it is not affected by water vapour a t the softening point of glass, but oxidises slowly in moist air.Gaseous hydracids have no action, but their solutions decompose the carbide in sealed tubes at about 250' : nitric and sulphuric acids dissolve it rapidly, and it is also decomposed by alkalis.ind by oxidising agents. C. H. B. Tungsten Silicide. By EMILE VIGOUROUX (Compt. rend., 1898, 12'7, 393-395).-Tungsten silicide, W,Si,, forms beautiful, steel-gray plates with a metallic lustre; sp. gr. = 10.9. It is attacked by chlor- ine, with incandescence between 200' and 300' ; by bromine, without incandescence, below a red heat, and by iodine, without incandescence, above a red heat.I n oxygen, it burns very brilliantly a t about 500°, and in air it oxidises below a red heat, but does not burn ; nitrogen is without action at any temperature. The ordinary acids, including hydrofluoric, have no action even on heating. Aqua regia is practi- cally without action, but a mixture of nitric and hydrofluoric acids attacks the silicide violently, even at the ordinary temperature. Alkalis in solution have little action, but fused alkali hydroxides and carbonates decompose the silicide very readily, whilst fused potassium nitrate acts with somewhat less energy. Tungsten silicide is prepared by heating in the electric furnace a mixture of 100 grams of silicon wibh 230 grams of tungsten oxide obtained by heating ammonium tungetate. The heavy, brittle, crys- talline product is suspended in dilute hydrochloric acid (1 in lo), and connected with the positive pole of a battery of two or three cells, a carbon plate placed in the liquid being connected with the negativeINORQANTC CHEMISTRY. 105 pole.The metal dissolves, and the silicide, which collects a t the bottom of the vessel, is treated successively with aqua regia, ammonia, and hydrofluoric acid, and any carbon silicide that may be present is separated by means of methylenic iodide. By GRBGOIRE N. WYROUBOFF and AUGIUSTE VERNEUIL (Compt. rend., 1898, 127, 412--414).-The mineral is dissolved in the usual way, and the solu- tion, which must contain sufficient acid to prevent precipitation of the phosphates, is Precipitated with half the quantity of oxalic acid necessary for complete precipitation, The oxalates are washed until free from phosphoric acid, converted into carbonates by means of a hot solution of sodium carbonate (1 : lo), and some sodium hydroxide added to insure complete precipitation of the thorium.The carbonates are washed until free from oxalic acid, dissolved in just the necessary quantity of hydrochloric acid, and mixed with successive small quan- tities of barium peroxide suspended in water until the liquid gives no precipitate with hydrogen peroxide. The precipitated peroxide con- tains all the thorium, together with 20 to 30 per cent. of impurities; it is washed and dissolved in cold concentrated hydrochloric acid, barium eliminated by means of sulphuric acid, enough water added to yield a solution containing 15 per cent.of acid, and the bases precipi- tated with oxalic acid. The oxalates are washed, and treated with a highly concentrated solution of ammonium carbonate mixed with sufficient ammonia to form the normal salt. By two or three succes- sive treatments, all the thorium is dissolved, and the solution is precipitated by means of sodium hydroxide, the precipitate well washed and dissolved in not more than the requisite quantity of nitric acid, and the liquid poured into sufficient water to yield a solution containing not more than 2 per cent. of thorium, Excess of hydrogen peroxide is then added, and the precipitate is well washed. The precipitate is dissolved in nitric acid and reprecipitated with hydrogen peroxide in order to eliminate all the cerium.It is next dissolved in hydrochloric acid, precipitated with oxalic acid, and the oxalate decomposed by pure sodium hydroxide. After careful washing, the precipitate is again dissolved in hydrochloric acid and precipitated with ammonia. This final precipitate is well washed, dissolved in nitric acid, and the nitrate crystallised. I n treating 5 tonnes of monazite, the authors never obtained, either in the hydrogen peroxide precipitate or in the ammonium carbonate solution of the oxalates, any element corresponding with the Russium of Chrustchoff. C. H. B. By EMIL WOHLWILL (Zeit. Elektro- chern., 1898,4,379,402, and 421).-Solutionsin potassium cyanidecannot be used, becauoe silver and copper are deposited along with the gold. When neutral solutions of auric chloride or of hydrogen aurichloride, HAuCl,, are used, chlorine is evolved and the gold anode is not attacked.With hydrochloric acid alone, even so dilute as 0.4 gram per litre, or with solutions of hydrogen aurichloride acidified with hydrochloric acid, no chlorine is evolved, and the gold is dissolved ; C. H. B. Extraction of Thorium on EL large Scale. From this point, all the reagents must be pure. Electrolytic Gold Refining.106 ABSTRACTS OF CHgMTCAL PAPERS. the chlorides of the alkalis or of ammonium have the same effect as hydrochloric acid. I n order that the gold anode shall dissolve, it is, therefore, necessary that the conditions permit of the formation of AuCI, ions. The behaviour of gold chloride solutions towards silver nitrate confirms Hittorf's conclusion as to the existence of these ions ; reddish-yellow precipitates are invariably formed on adding the silver nitrate ; they are, however, much less stable than the analogous silver platinochloride.The dissolution of the anode is also promoted by raising the temperature to 60' or TO', and in n hot solution containing about 3 per cent. of hydrochloric acid, an anodic current density of 3000 ampfires per square metre may be employed without separation of free chlorine ; this is more than sufficient for practical purposes. At the cathode, a fairly high current density may be employed without injury to the gold deposit, and without necessitating the use of too concentrated a gold solution ; the deposit is usually coarsely crystal- line, and sufficiently adherent to permit of thorough washing without loss.The greater the concentration of the gold solution, the better the deposit; 30 grams of gold per litre is, however, sufficient, even with 3000 ampires per square metre, No trouble is experienced from the formation of dendritic deposits, so that the electrodes may be placed close together, and a considerable output be obtained from a fimall plantl. Of the impurities in the gold, platinum and palladium pass into solution, but are not deposited a t the cathode ; after they have accumu- lated sufficiently, the solutions are worked up to recover them. Silver remains, mainly as a mud of silver chloride; the small p r t which dissolves in the hot acid solution does not pass into the deposited gold.The refined gold is very seldom less than 999.8 fine, not infrequently it is 1000. The E.M.F. required with a current density of 1000 amperes per square metre is less than 1 volt. On the assumption that only tervalent gold exists in the solution, 2-45 grams should be deposited per ampkre hour. The author's experi- ments, as well as the results obtained in large scale working, show that more nearly 3 grams per ampere hour are obtained. A considera- tion of the loss of gold a t the anode gives the explanation of this. Part of the loss is due to the separation of fine particles of gold which are found, for the most part, in the silver chloride mud. These par- ticles are equal to about one-tenth of the amount of gold deposited on the cathode, and are not merely mechanically detached from the anode, but are due to the formation of aurous chloride at its surface, which subsequently decomposes into metallic goId and auric chloride.In support of this view, the author shows that the particles are much purer than the anode itself; that aurous chloride may be detected in the solution, and that, under certain circumstances, small crystals of gold separate out throughout the mass of the solution. The presence of aurous gold with three times the dectro-chemical equivalent of the auric gold readily accounts for the large deposit observed. The part of the gold dissolved from the anode in the aurous state, which decom- poses into metallic gold and &uric chloride, never reaches the cathode, and, therefore, the loss at the anode is always greater than the deposit at the cathode. A series of experiments with varying anodic currentNINERALOGICAL CHEMISTRY. 107 densities shows that the formation of univalent gold ions diminishes as the curreil t density increases. With 15 00 amperes per square metre, 2-48 grams of gold per ampere hour were deposited (instead of 2-45), and the loss at the anode and gain at the cathode, in two experiments, were (1) 105.2 : 104.5 ; (2) 107.7 : 105. Here, therefore, tervalent gold was present almost exclusively, Such results are only obtained when the liquid is well stirred; otherwise, owing to the accumulation and decomposition of aurous chloride at the anode, the loss there is much in excess of the cathode gain. I n experiments with very small current densities, the non-electrolytic solution of gold in the hot, acid auric chloride solution must be allowed for. This appears to be reversible, a sheet of gold losing weight when the temperature is raised, and gaining it when it is lowered again. After allowing for this chemical attack, the author obtained with 1 ampere per square metre a deposit of 4.33 grams per ampere hour, and an anodic loss of 6.01 grams, that is, 72.5 per cent. of the gold had dissolved in the aurous state. T. E. Claims of Davyum to Recognition as an Element. By JOHN W. MALLET (Amsr. Chrn. J., 1898, 20, 776--783).--Kern (this Journal, 1877, ii, 278 and 712) has stated t h a t platinum ores contain a small quantity of a new metal, “davyum,” characterised by its chloride combining with that of sodium to form a sparingly soluble double salt. I n the hope of separating this metal, the author has examined the sparingly soluble residue remaining after clearing a solution of platinum ore in aqua regia from osmiridium by decanta- tion, removing the platinum by adding ammonium chloride, adding a n excess of common salt, evaporating to dryness, and extracting the residue with a minimum of water. It was found, however, to consist only of quartz, zircon, iridium, rhodium, and osmium. From a large quantity of rose-coloured crystals of sodium chloride recovered during the manufacture of platinum, a very small amount of a metallic substance was isolated, which, judging from an analysis of the double chloride, appeared to have an atomic weight of about 151.5, corresponding with the value 154 attributed by Kern to davyum. Its solution also in aqua regia gave a red coloration with potassium thiocyanate, a reaction characteristic of davyum. Since, however, it was found to consist merely of rhodium and iridium with a trace of iron, the existence of davyum is considered as very doubtful. W. A. D.