INORGANIC CHEMISTRY. 39 I n o r g a n i c Chemistry. True Atomic Weight of Hydrogen. By G. HINRICHS (COW@. rend., 117, 663-666).-The author has applied his method of limits (Abstr., 1893, ii, 316) to the results of Reiser, Cooke and Richards, Ditbmar and Henderson, and Morley, making use of some hitherto unpublished data supplied by the last. The application is somewhat; difficult, because in each series of experiments the weight of materials employed has varied but little. It is clear, however, that the di- vergencies from the value H = 1 are functions of the weight of hydrogen employed, and if equal weight is attached to the work of Dittmar, Keiser, and Morley, the divergencies from this value fall on a curve which is represented by the formula 7 = - 5000 (5' - h'), where h is the weight of hydrogen and 7 the divergence from unity.h 5-240 A J ~ S ~ ' ~ ~ A C T S OF OHEMICAL PAPERS. For the limit h = 0, T,I = 0, and hence E = 1 exactly when 0 = 16 (compare Zoc. cit.) . Relation between the Precipitation of Chlorides by Hydro- chloric acid and the Reduction of the Freezing Point. Ry R. EFGEL (Compt. rend., 117, 485--488).-The law that 1 mol. of hydrochloric acid will precipitate from its saturated solution a t 0" 1 atom of chlorine in the form of the chloride of a univalent or bivalent metal holds good for temperatures as high as 75'. The law also holds good for the double chloride CuC1,,2NH4Cl, 1 mol. of which is precipitated by 4 mols. of hydrochloric acid, Van't Hoff has shown that the molecular osmotic pressure of hydro- chloric acid, 1.98, is practically the same as that of sodium chloride, 1.89, and therefore 1 mol.of the acid should precipitate 1.05 mol. of the salt. I n the case of the chlorides of bivalent metals, the mole- cular reduction of the freezing point is not double that, of the chlorides of univalent metals. The author finds, however, that the molecular reduction produced by chlorides of univalent metals remains prac- tically the same for various concentrations, and for different chlorides varies only between 35 and 40. With chlorides of bivalent metals, on the contrary, the molecular reduction increases with the concen- tration and becomes practically double that of the univalent chlor- ides. The molecular reduction of the double salt CuCI2,2NH4Cl also increases with the concentration and tends to become four times as great as that of univalent chlorides. The following table shows the relation between the nature of the chloride, the molecular reduction of the freezing point, and the number of molecules of hydrochloric acid required to precipitate 1 mol.of the salt. C. H. B. Atoms of Ratio of chlorine the in the molecular Mols. molecule. reductions. HC1. Chlorides of univalent metals . . 1 1 1 bivalent .. 2 2 2 Doubii chloride, CuC12,2NH~C1. . 4 4 4 In this connection, it is important to observe, however, that sodium hydroxide, which has practically the same molecular osmotic pressure as hydrochloric acid, precipitates only 4 mol. of sodium chloride, bromide, or iodide from a saturated solution. When the number of molecules of water in combination with a molecule of a chloride at the freezing point of its saturated solution is calculated, i t appears that the product of this number into the reduc- tion of temperature is a constant which has double the value in the case of chlorides of bivalent metals that it has i n the case of chlor- ides of univalent metals.Molecules Freezing of water. point. Product. Ammonium chloride.. . . . . 12.4 15.5 19.2 Potassium chloride. . . . . . . 16.6 11.4 18.9 Potassium iodide . . . . . . . . 8.5 22.0 18.7IKORGANIO CHEMlSTRY. 41 Molecules Freezing of water. point. Product. Sodium bromide. . . . . . . . 8.1 24.0 19.4 Ammonium iodide . . . . . . . 6.4 27.5 17.6 Strontium chloride. . . . . . . 22.9 17.0 38.9 Calcium chloi-ide . . . . . .. . 10.7 37-0 39.6 C. R. B. Ozone. By A. WOLKOWICZ (Zeit. an0rr.g. Chem., 5, 264-265).- From a consideration of the position of oxygen in the periodic system, it is concluded that ozone has the constitution O:O:O, just as sulphurous anhydride is 0:S:O. Ozone would then be the anhydride of an acid H204, corresponding with HzSOa. This acid, like sulph- urous acid, is unknown, but potassium tetroxide, Kz04, is its potassium salt, and resembles the sulphite in its property of reducing perman- ganate. C. F. B. DecomFosition of Nitrous acid in Nitric acid Solution. By B. LILJENSZTERN and L. MARCHLEWSK~ (Zeit. anorg. Chem., 5, 288-292). -By passing a stream of carbonic anhydride through 60 per cent. nitric acid to which 4 per cent. of nitric peroxide had been added, and determining the ratio of nitrous to total nitrogen in the gases carried over, it was shown that the nitric acid contained nitrous acid in addition to nitric peroxide.Montemartini’s statement ( R e d Accad. Linc., 1, 63), that the reaction HN02 + HNO, = 2N0, + H,O is the only one that takes piace in nitric acid of over 30 per cent. strength, is, therefore, erroneous. Conversion of Yellow Phosphorus into Red Phosphorus. By J. W. RETGERS (Zeit. anorg. Chem., 5, 211-230 ; compare Pedler, Trans., 1890, 599; Retgers, Abstr., 1893, ii, 457; Muthmann, ibid., 458) .-It is generally admitted that yellow phosphorus crystnllises from the usual organic solvents in dodecahedra ; the crystals are best obtained by cooling a hot solution. The author cannot substantiate the text-book statement that yellow phosphorus crystallises from essential oils in octahedra.Thus, whilst it is admitted that yellow phosphorus is crystalline, this substance has, on several occasions, been described as amorphous. This is not surprising in view of the fact, that a thin layer of yellow phosphorus between two thin glasses has all the appearance of an amorphous substance. When phosphorus is heated, it passes through three stages :-(1) it becomes yellow to brown, but remains transparent and regular ; (2) a granular, undoubtedly crystalline, segregation occurs (formation of red, opaque phosphorus) ; (3) a graphitic, chocolate-coloured phos- phorus is produced. The second stage corresponds with the first production of a true modification. The same series of changes is produced by light-not merely superficially, as stated in text-books-a fact which the author calls in evidence of his contention that red phosphoims is not amorphous.The author has not been able to determine the character of the white crust which forms on phosphorus which has been submerged C. F. B.42 ABSTRACTS OF CHEMICAL PAPERS. in water, but he inclines towards the opinion that i t is a hydrate rather. than a modification. I t is pointed out that the properties of an amorphous modification are never intermediate between those of two crystalline modifications. I f red phosphorus were amorphous, this rule would be violated ; for the specific gravity of red phosphorus is 2,148, whilst that of regular, yellow phosphorus is 1.826, and that of chocolate-coloured, hexagonal phosphorus is 2.34.Some remarks in reply to Muthmann (Abstr., 1893, ii, 458) are included in the paper. A. G. B. Volatility of Pyrophosphoric acid. By G. WATSON (Chem. News, 68, 199-200).-The author has made a series of experiments, heating weighed quantities of orthophosphoric acid for various periods, at fixed temperatures, and examining the products; from which he concludes that orthophosphoric acid requires a temperature above 230-235" for its complete dehydratiou into pyrophosphoric acid ; that at 255-260" it is completely converted into pyrophosphoric acid, which is, moreover, volatile at that temperature ; that at 290-300" metaphosphoric acid is beginning to form. Action of the Electric Arc on the Diamond, Amorphous Boron, and Crystallised Silicon.By H. MOISSAN (Conzpt. rend., U7, 423-425).-1n the electric arc, a t a somew hat high temperature, the diamond becomes incandescent, swells up without melting, and hecomes covered with black masses, consisting en tirely of hexagonal lamella of graphite, which is easily converted into graphitic oxide. If the diamond is placed in a small carbon crucible in the electric furnace previously described, and is subjected to the action of an arc produced by a current of 70 volts and 400 amphres, the crystal first breaks up into small fragments along the planes of cleavage, and then at a higher temperature swells up and is completely converted into graphite, which yields yellow graphitic oxide. It follows that at the temperature of even a moderately intense electric arc, the stable form of carbon is graphite.When heated in a carbon envelope at the temperature of the oxy-hydrogen blow-pipe, the diamond is sometimes covered with an adherent black mass, which slowly dissolves in a mixture of potassium chlorate and nitric acid, but which is not graphite. Amorphous boron, prepared by means of magnesium, volatilises without fusion in the electric arc, the extremities of the electrodes being converted into partially-crystallised boron carbide. Crystallised silicon, when heated in the arc, first melts and then boils, the extremities of the electrodes at the end of the experiment being covered with pale green crystals of carbon silicide. The phenomena in the arc were observed by projecting on a screen by means of an intense arc an image of the arc of lower intensity in which the substances were heated.C. H. B. D. A. L. Silicon Carbide. By 0. M~HLHAEUSEK (Zeit. anorg. Chem., 5, 105-125) .-A mixture of finely-powdered coke, sand, and salt wasTNORQANIC CHEMISTRY. 43 packed around a carbon core in an oblong fire-brick box. The core was connected with the terminals of a transformer, whereby it could be heated until the mixture surrounding it became white hot. A transverse section of the mass after the reaction showed that the core was surrounded by ( a ) a zone of adhering graphite, ( b ) a zone of crystalline silicon carbide, ( c ) a zone of amorphous silicon carbide, ( d ) a zone containing pockets of fibrous material, (e) a zone of the original mixture only slightly altered, and ( f ) a hard skin consisting almost entirely of salt.The graphite has all the properties of the natural mineral, and the same crystalline form as that of the silicon carbide, a fact which indi- cates that the latter compound is a t first formed in this zone and sub- sequently loses its silicon by volatilisation. The outer portion OF this zone, when washed, left 33.71 per cent. of variously-coloured crystals, which contain 30.49 per cent. of silicon and 68.26 per cent. of carbon. These crystals differed considerably in appearance from those in the zone of silicon carbide ; they were frequently parti-coloured, green, violet, and red being the prevailing tints, The crystalline silicon carbide (carborundicm) constituted the chief product of the reaction. The mass was easily broken up in a mortar, when the separate crystals were found to be bluish or yellowish- green; their size varied from a diameter of several millimetres to merely microscopic dimensions.The largest crystals occurred i n fissures i n the mass, where: apparently, there was room for them to form, and i t remains an open question whether they had Reparated from a, fused mass or were a product of sublimation. To purify silicon carbide it may be heated to dull redness in oxygen, then boiled with potash solution, washed, digested with hydrochloric acid, again washed, and finally treated with hydrofluoric and sulphuric acids. After this treatment, its composition corresponds with the formula Sic. Silicon carbide is insoluble in all acids, but is attacked by molten alkalis.It burlis very slowly in oxygen when very finely divided, and if strongly heated in a platinum crucible it becomes a greenish- golden mass of great beauty, a portion of it being burnt during the ignition, Its specific gravity at 25" is 3.22, but when finely divided it remains suspended in water for months. The amorphous silicon carbide found in the third zone has all the properties of the crystalline specimens. The fibrous material found in fissures and pockets in the fourth zone consists of silicon, aluminium, and carbon with a little lime and magnesia; it would t h u s appear to be a polycarbide oE silicon and aluminium. Much salt was volatilised during the reaction; the gas evolved consisted largely of carbon monoxide. Crystallised Carbon Silicide.Bg H. MOISSAN (Compt. rend., 117, 425428) .-When carbon is dissolved in fused silicon in a wind Furnace, crystals of carbon silicide several mm. in length can be abfained, and it follows that the two elements combine readily in a It is attacked by hot ferric oxide. A. G. B.44 ABSTRAOTS OF CHEMICAL PAPERS. fused medium at 1200-1400°. The silicide can, however, be pre- pared much more easily by heating in an electric furnace a mixture of 12 parts of carbon and 28 parts of silicon, the product being first heated with a mixture of nitric and hydrofluoric acids, and after- wards with a mixture of nitric acid and potassium chlorate. As a rule, the crystals are yellow, but by operating in a closed crucible with silicon as free as possible from iron, transparent, sapphire-blue crystals are obtained.If iron silicide is heated with excess of silicon in the electric furnace, and the product is treated successively with aqua regia, nitric and hydrofluoric acids, and nitric acid and potassium chlorate, crystals of carbon silicide are obtained, and the result is similar with a mixture of iron, silicon, and carbon, or of iron, silica, and carbon. The silicide is also produced by heating carbon and silica in the electric furnace, or by allowing the vapour of carbon to come into contact with vapour of silicon, when it is obtained in almost colour- less, very hard and brittle, prismatic needles. Carbon silicide prepared by any of these methods has the com- position CSi, and is colourless when free from iron. It is very dis- tinctly crystalline, and sometimes forms regular, hexagonal lamells, which occasionally, though very rarely, show triangular impressions and parallel strim They act strongly on polarised light.The crystals are very hard, ar_d scratch chrome steel and rubies. Sp. gr. = 9-12. Carbon silicide is not affected by oxygen at lOOO", nor when heated in air by a Schloesing's blowpipe. Sulphur vapour at 1000" is without action, and chlorine at 600" attacks the compound very slowly, although its action is complete a t 1200". Fused potassinm nitrate or chlorate, boiling sulphuric, nitric, and hydrochloric acids, aqua regia: and mixtures of nitric and hydrofluoric acids do not attack it. Fused lead chromate oxidises the silicide, but repeated treatment is necessary in order to obtain complete combustion of the carbon.Fused potassium hydroxide gradually converts it into potassium carbonate and silicate. C. H. B. Volatility of Ammonium Chloride. By I(. KRAUT (Zeit. anorg. Chem., 5, 258--279).-When ammoninm chloride is heated in a platinum basin in a water bath, an appreciable amount is lost by volatilisation; 50 per cent. of the whole if the heating is continued for 270 hours. C. F. B. Hemihydrate of Calcium Sulphate. By A. L. POTILITZIN (J. Rum. Chena. Xoc., 25, 'LO7--21O).-The hydrate 2CaS04,H,0 may be formed by dehydrating gjpsum at 98-99O, the water being lost very slowly ; or it may be formed by exposing anhydrous calcium sulphate to the air. The absorption of moisture is at first rapid, but the rate soon diminishes. The hemihydrate is capable of taking up small quantities of moisture from the atmosphere, and large quanti- ties when allowed to remain over water. This excess of water is only partially lost when the hydrate is placed over sulphuric acid.J. W.INORGIANIC CHEMISTRY. 45 Double Halojids of Caesium with Zinc and Magnesium. By H. L. WELLS and G. F. CAMPBELL ( Z e i t . anorg. CIwm., 5, 273-277).- The following salts were prepared, and crystallise in colonrless prisms or plates. 3CsC1,ZnC12 ; SCsBr,ZnBr2 ; 3CsT,ZnIz ; 2CsC1,ZnC12 ; 2CsBr,ZnBr2 ; 2CsI,Zn12 ; CsCl,MgCl, + 6H20 ; CsBr,MgBr, + 6H20, Zinc probably also forms salts of the 1 : 1 type, but the solutions are so syrupy that pure crystals cannot be obtained. No double chloride of caesium and beryllium could be prepared.Double Haloids of Caesium and Cadmium. By H. L. WELLS and P. T. WALDEN ( Z e i t . anorg. Chem., 5, 266-272).-The following salts were prepared, and are all colourless. 3CsBr,CdBr2 ; 3CsT,CdIz ; 2CsC1,CdCl2 ; 2CsBr,CdBra ; 2CsI,Cd12 ; CsCl,CdC12 ; CsBr,CdBrz ; CsI,Cd12 + H20. Those of the 1 : 1 type, and 2Cs12,Cd12, can be recrystallised from water ; the others, when so treated, yield generally salts of the 1 : 1 type. Occlusion of Gases by Metallic Oxides. By T. W. RICHARDS and E. F. ROGERS (Amer. Chem. J., 15, 567-578).-As in the case of cupric oxide, the oxides of zinc and nickel, and especially of mag- nesium, when Frepared by ignition of the nitrates, are found to contain occluded gas ; but the oxides of cadmium, mercury, lead, and bismuth retain no gas when prepared in this way.No gas is in any case retained when the oxide is prepared by ignition of the carbonate. The occluded gas was liberated by dissolving the oxide in hydro- chloric acid, and was measured and avalysed ; it consisted mainly of oxygen and nitrogen, together with a very small quantity of a gas that dissolved in caustic potash, and was assumed to be carbonic anhydride [no reason is, however, given why this gas should not have been an oxide of nitrogen]. Oxides which still contained a trace of nitrate were found to be devoid of occluded gas ; hence this gas must, i n the other cases, have been derived from the decomposition of the last trace of nitric acid. The amount retained varied with the physical nature of the oxides, the more compact oxides retaining more gas, whilst those which form very fine powders retained no gas at all.This occlusion of gases affects the atomic weights of a number of elements of which the oxides, as prepared by ignition of the nitrates, have been used in determining the atomic weights, C. 3'. B. C. F. B. C. F. B. Compounds of Hydroxylamine with Metallic Carbonates. By H. GOLDSCHMIDT and K. L. SYNGROS ( Z e i t . anorg. Chem., 5, 129- 146).-Dihydroxylawtine zinc carbonate, Zn(NH30)2C03 is prepared by dissolving hydroxylamine hydrochloride and zinc chloride in water, adding an equivalent quantity of sodium carbonate, and passing air through the solution, when the new compound is precipitated. It is a white, microcrystalline powder, OE sp. gr. 2.50 at 18", insoluble in water. The function of the air in the above method of preparation is to remove carbonic anhydride from the solution.Cryoscopic ex- periments indicate that the zinc exists in the solution partly as the ion 2;n(NHd0)*, and partly as dissociated zinc hydrogen carbonate ;46 ABSTRACTS OF CHEMICAL PAPERS. the removal of a portion of the carbonic anhydride determines the molecular rearrangement. When sodium carbonate solution is added to a solution containing ferrous chloride and hydroxylamine hydrochloride, a dark red coloration is produced, and the passage of hydrogen through the liquid determines the formation of a nearly black precipitate which contains ferrous oxide, hydroxylamine, and carbonic anhydride, but cannot be dried without decomposition. A cryoscopic examination of the solution indicates the presence of the ion Fe(NH,0)2.When manganous chloride is mixed with hydroxylamine hydro- chloride and sodium carbonate, a slight precipitate is first. formed ; this was filtered off, and a current of air was passed through the solu- tion. A nearly white precipitate was thrown down, which dried to a grey powder, having the composition 4MnC0,3NH30,2H2O. A cryo- scopic investigation showed that in this case no complex ion of metal and hydroxylamine exists in the solution. The precipitate obtained by similar treatment of nickelous chloride varies in camposition with the time during which air is passed through the liquid, and in no case is it of a very definite character. Cadmium chloride gives a slight precipitate when mixed with hydroxylamine hydrochloride and sodium carbonate in solution, and when this has been filtered, the filtrate spontameously deposits the dihydroxylamiiie cadmium chloride, Cd(NH30)oC1,, previously pre- pared by L.de Bruyn and Crismer (Abstr., 1890, 558). Thisis some- what soluble in cold water, and crystallises from hot water in white prisms ; its sp. gr. is 2.72 at 18'. A. G. B. Preparation of Anhydrous Crystalline Metallic Silicates. By H. TRAUBE (Bey., 26, 2735-2736).-When amorphous zinc silicate, prepared by precipitating a sol'ution of zinc sulphate with sodium silicate, is heated with boric acid for 10 days at a high tem- peratare, it is converted into a white, crystalline powder, which has the coniposition ZnSi03, and is insoluble in acids. The optical pro- perties of the crystals show that they belong to the rhombic systlem, and the product is therefore a zinc-pyroxene isomorphous with the mineral enstatite.E. C.-R. Lead .Oxide as a Mordant. By A. BONNET (Con@. rend., 117, 518--519).-When cotton is mordanted with an alkali plumbate and then washed with a large quantity of water, dissociation takes place, and the fibre becomes charged with lead peroxide, which partially oxidises and destroys it. A similar. change takes place with plumbites, except that the fibre is not oxidised, lead monoxide being deposited. I f the fibre thus treaied is immersed in solutions of camyeachy wood. it is dyed black; with shumac, it becomes green; with old fustic, a bright yellow. Tannin and catechu are also strongly attracted. By immersing the tissue thus mordaiitecl with lead oxide in hot neutral solutions of other salts, the tissue can be impregnated in a similar manner with oxides of gold, silver, mercury, vanadium,INORGANIC CHEMISTRY.4 i manganese, chromium, iron, cobalt, nickel, and zinc, double decom- posigon taking place, and a-lead salt going into solution. C. H. B. The True Atomic Weight of Copper. By G. HINRICHS (Zeit. aworg. Chenz., 5, 293--298).-A protest against alleged errors in the atomic weight determinations of Stas and his followers, a recent de- termination by Richards of the atomic weight of copper (Abstr., 1893, ii, 12) being selected ns a specimen of such faulty niethods. [In hhe. opinion of the abstractor, the author does not substantiate all his objections, and the alternative " method of limits " which he proposcs seems open to damaging criticism.] Colour, &c., of Cupric Chloride Solutions.By N. N. TZUCHANOFF (J. Rzcss. Chem. Xoc., 25, 151--152).--The author has investigated the connection between the colour and the electric con- ductivity of aqueous solutions o€ cupric chloride of diff ererit strengths at constant temperature. In weak solutions of a blue colour, the conductivity increases rapidly with the concentration, but the rate of increase falls off as the solution becomes gyeen. A maximum con- dncti vity is finally reached, the subsequent decrease with iiicreasing concentration being accompanied by a change o€ tint to yellowish- brown. J. W. C. F. B. Cssium Cuprichlorides. By H. L. WELLS and L. C. DUPEE (Zeit.anorg. Chem., 5, 300-303) .-Four salts were obtained. 2CkCl,CuC12 ; yellow, rhombic prisms. 2CsCI,CuC12 + 2H20 ; bluish- green crystals, turning yellow in the air. 3CsC1,2CuC12 ; brown, triclinic crystals. CsCI,CuC12 ; hexagonal prisms, red by transmitted, black by reflected, light. By H. Ti. WELLS (Zeit. anorg. Chem., 5, 306--308).-These were prepared by boiling a solution of cmsium and cupric chlorides in hydrochloric acid with copper wire, and then allowing the solution to crystallise. CsC1,2CuC1 and 3CsC1,2CuCl; colourless crystals, turning yellowish when dried. SCsC1,CuCl + HJ3 ; prismatic crystals, veiy pale yellow in colour. Caesium Cupribromides. By H. L. WELLS and P. T. WALDEN (Zeit. anorg. Chem., 5, 304-305) .-Two salts were obtained. 2CsBr,CuBr2 ; black, rhoinbic crystals with a greenish shade.CsBr,CuBr2 ; black, hexagonal crystals with a shade of bronze ; c r p - tallisation from water converts these into the first-mentioned salt. Oxides contained in Cerite, Samarskite, Gadolinite, and Fergusonite. By W. GIBBS (Amer. Chem. J., 15, 546-566).-A description is given of various attempts that were made in order to separate the oxides contained in certain specimens of gadolinite, samarskite, and fergusonite. The methods employed were those of fractional precipitation or crys tallisation, the salts used being the oxalates, oxychlorides, the double salts with sodium sulphate, an.t with the various cobaltamine sulphates, the lactates, and the double C. F. B. Caesiurn Cuprochlorides. C. F. B. C. I?. B.48 ABSTRACTS OF CHEMIUAL PAPERS.salts with mercurous nitrate pEus mercuric oxide, with acid molybdates, and with phosphotungstates and phosphomolybdates. The methods were often only imperfectly worked out, and the whole paper is somewhat disconnected, and cannot be satisfactorily ab- stracted. As a means of determining the " mean atomic mass " of a fraction, it is recommended to precipitate with oxalic acid, carefully mix t h e dried oxalates in a mortar, and convert a weighed portion by igni- tion into oxides. C. F. B. Chemical Behavionr of Glass. By F. FOERSTER (Ber., 26, 2915--2922).-1n continuation of his previous work on the action of water and of aqueous alkali and salt solutions on glass (Abstr., 1892, 1401), the author has now investigated the behaviour of acids towards glass.Round flasks of various kinds were employed, the glass, unless otherwise stated, being a calcium-alkali glass, such as is usually applied to chemical purposes ; the acid was allowod to act for six hours at 100". With the same kind of glass, the action, which is always less than that of pure water, is independent both of the nature of the axid and also of its concentration between the limits N/1000 and 10K. With concentrated acids at 160-190", the nature of the acid is without effect, whilst the amount of change decreases with increasing con- centration. The same result was obtained with hydrochloric acid at 260-270'. Emmerling's results, which led him to the contrary con- cIusion, me probably incorrect. Acids appear, therefore, to be in- different towards glass ; the action which actually takes place is due simply to the water which is always present, and which dissolves alkali; this rapidly attacks the glass, but the stronger the acid the more quickly will the alkali be neutralised.This view receives ad- ditional support from the fact that, in comparison with water, acids dissolve larger quantities of alkali and less silica from the glass. The difference in behaviour towards concentrated acids between glass and other calcium-alkali silicates is noteworthy ; the compound NhO,SiO, is more readily decomposed by concentrated than by dilute acids, whilst sodium silicate with the composition N%0,3Si0 behaves like glass in this respect. The reaction is conditioned both by the nature of the bases, and by the relative proportion of silica present ; t h i s is shown from the fact that, towards acids, lead crystal glass behaves like calcium-alkali glass, but their action on flint glass, which is poorer in silica, increases with the conceiitration and differs according to the nature of the acid.Jena thermometer glass I P , which contains zinc, calcium and sodium, is more rapidly acted on by concentrated than by dilute hydrochloric acid at 190", whilst a calcium-sodium glass of equivalent composition behaves in the usual manner. Pure sulphuric acid attacks caloium-alkali glass less rapidly than pure water at 100"; the action slowly increases with rising temperature ; the vapour acts comparatively vi.gorously, and the glass becomes covered with a network of alkali snlphate crystals. Carbonic anhydride resembles acids in its action; the " weathering '' of glam is chiefly caused by the action of moisture.Glass of every kind combines chemically with more or less water, a graduated seriesINOROANlC CHEMISTRY. 49 of compounds being obtained, which form connecting links between fresh glass and the substances that are found in solution. The paper concludes with tables showing (1) the composition of the six better kinds of glass employed for chemical purposes ; (2) their rela- tive stability towards water, alkalis, and alkaline carbonates. The best glass mentioned has the following composition :-K20 = 6.2 ; NhO = 6.4; CaO = 10.0; MnO = 0.2; A1203 + Fe,O, = 0.4; SiOz = 76.8 per cent. ; the ratio RzO : RO : Si02 = 0.95 : 1 : 7.16 ; it contains 10.4 mols.of alkali in 100 mols., and is, therefore,almost identi- cal with the glass used by Stas in his atomic weight determinations. J. B. T. Manganese. By 0. PRELINGER (Monatsh., 14, 353-370).- Manganese amalgam is prepared by passing an electric current (11 volts ; 22- 23 C.C. electrolytic gas per minute) from a cathode of pure mercury (20 c.c.) through a saturated aqueous solution of pure manganous fchloride (75 c.c.) to an anode of carbon or platinum- iridium contained in a porous vessel. The temperature rises to 70°, and after 5-6 hours the mercury assumes a pasty consistence. The paste is quickly washed in rmning water without undergoing ap- preciable decomposition, the excess of mercury squeezed out through linen, and the residue dried over calcium chloride in an atmosphere of hydrogen.The solid amalgam thus obtained is broken into small pieces and subjected to great pressure (2000 kilos. per sq. cm.) for several hours. The peripheral portion of the pressed cake is broken off, and the central portion broken up and again compressed. The operation is repeated until samples punched from the centres of successive discs yield constant analytical results. Analyses of three distinct prepara- tions yielded practically identical results, pointing to the composition Mn,Hg5. Manganese-mercuny, Mn2Hg5, has a slate-grey colour, and assumes a metallic lustre when rubbed or cut. At ordinary temperatures, it oxidises very slowly in the air, metallic mercury being elimi- nated. At 100-llO", it decomposes slowly into its elements, It decomposes water and acids at the ordinary temperatures.The sp. gr. is 12.828, a number greater than that, 12.533, calculated from the sp. gr. of its constituents, so that contraction takes place in the formation of the compound. Manganese-mercury is electro-positive t o manganese, so that heat is probably absorbed in its formation. A solution of the compound in mercury is not attacked by dry air, but is quickly oxidised by moist air to manganic oxide, Mn203, which forms a fine dust on the surface of the liquid. When manganese-mercury is gently ignited in a, stream of pure dry hydrogen (free from oxygen), pure manganese remains behind, uncontaminated with mercury or hydrogen. The manganese forms a grey, porous mass, which may be ground into a slate-coloured powder, perfectly stable in air.If the temperature does not reach dull redness in the ignition, however, the metal is set free in a pulver- d e n t spontaneously-inflammable form. The metal cannot be ma,& t o exhibit lustre. The sp. gr. of the powdered metal is 7.4212 and $the atomic volume (at. wt. = 548) 7.385. Manganese powder is It corresponds with the cadmium compound Cd2Hg,.50 ABSTRACTS OF CHEMIUAL PAPERS. neither magnetic nor magnetisable at ordinary temperatures. It attacks water, slowly in the cold, rapidly when heated. It has little action on concentrated sulphuric acid in the cold, bmt attacks it rapidly on heating gently, sulphurous anhydride being formed. It liberates hydrogen from dilute sulphuric acid. The action on con- centrated nitric acid is explosively violent. It displaces hydrogen from hydrochloric and acetic acids and sodium hydroxide. From aqueous ammonium chloride it liberates ammonia and hydrogen.I t reduces metals with great rapidity from warm solutions of their salts (preferably sulphates) ; the excess of manganese may be dis- solved by ammonium chloride. Arsenic, antimony, copper, lead ~ bismuth, tin, iron, nickel, cobalt, chromium, cadmium, and zinc are thus reduced. The reduction of the magnetic metals may be con- veniently demonstrated by means of a compass needle, the latter being deflected after, although not before, the addition of the man- gan ese . JN. W. Iron Nitride. By G. J. FOWLER (Chem. Neics, 68, 152-153).- By exposing iron, reduced from the hydroxide by hydrogen, t o a fairly rapid current of ammonia, at a temperature slightly above the melting point of lead, until its weight becomes constant, iron nitride, Fe,N, is obtained as a feebly magnetic, grey powder. It dissolves in hydrochloric acid, yielding ferrous chloride, ammonium chloride, and hydrogen; it burns in chlorine to ferric chloride and nitrogen; ammonia or ammonium salts are produced when it is heated in hydrogen at the temperature of its formation; in steam a t loo", or in hydrogen sulphide at 200", or by treatment with hydrogen per- oxide and sulphuric acid.By simply heating, it is resolved into iron and nitrogen, but remains unchanged in nitrogen at &O", or when heated in carbonic oxide or with phenol a t 200" ; with ethylic iodide, however, in a, sealed tube at 200°, ferrous iodide, ammonium iodide, ethylene, and bydrogen are formed.Its heat of formation is about 3 cal. These results agree with many of Stahlschmidt's observa- tions (Ann. Phys. Chem., 125, 37). Constitution of Cobalt, Chromium, and Rhodium Bases. By S. M. JORGENSEN (Zeit. anorg. Chem., 5, 147-196; compare Abstr., 1892, 782, 783).-The first part of this paper deals with the views put. forward by A. Werner concerning the structure of ammonio- metallic salts (Abstr., 1893, ii, 379); the author does not regard them as being a t variance with the views promulgated by himself and Blomstrand. The author next describes the best method of preparing the croceocobalt salts from xanthocobalt chloride, and a series of salts, which he terms flavocobalt salts, isomeric with the croceocobalt salts.Elavocobalt nitrate, (N02),Co ( NH3),N03, is prepared by dissolving carbonatotetraminecobalt nitrate (Zeit. anorg. Chem., 2, 282) in cold dilute nitric acid, and adding crystalliped sodium nitrite. After being warmed for a few minutes on the water bath, the liquid is cooled, and dilute nitric acid is added. A mixture of the acid and normal salts crystallises, and may be converted entirely into D. A. L.INORGANIC CHEMISTRY. 51 the normal salt by washing with 9-5 per cent. alcohol. The flavo- cobalt nitrate dissolves in 33 parts of cold water, whilst the croceo- cobalt nitrate requires 400 parts of cold water for its solution. A detailed comparison of the reactions of croceo- and flavo-cobalt nitrates follows, and the dinitrate, (NO,),CO(NH,)~NO~,HNO~ ; the sulphate, [ (N02)2Co(NEI,)4]aS04 ; the chromate, the dichromate, [ (N0,),Co(NH,),],Cr20i ; the platinosochloride, [ (NOz) $0 (NH,),] zPt C14 ; the pZatinochloride, and the aurochloride, (N02),Co(NH3)4AuCl~, are described.To pre- pare the xanthocobalt salts, chloropurpureocobalt chloride is warmed with water and ammonia ; the liquid is filtered from some precipitated cobalt oxide, cooled, and neutralised with dilute hydrochloric acid ; sodium nitrite and hydrochloric acid are next added, whereon a red precipitate is thrown down. This is an isomeride of xanthocobalt chloride, and is provisionally named nitritocobalt chZoride ; it is readily converted into the xantho-salt by dissolving it in water con- taining a few drops of ammonia, and adding strong hydrochloric acid, which throws down the santhochloride.The reactions of the xantho-salts are detailed, and the subhates, 4[ (NO,) CO(NH,)~,SO~),~H~SO~ and NO,-Co(NH,),SO~, are described. Xanthocobalt chloride dissolves in 50 parts of cold water : nitrito- cobalt chloride dissolves in 200 parts. Differences between the r a c - tions of the two salts are described, and it is shown that the nitrito- salt is not a roseo-salt. The author has prepared a luteorhodium rhodium chloride, Eh(NH,)6C13,RhC13, and a chloropu~~ureorhodiunz rhodizinz chloride, 3RhC1 (NH,) sC 12,2RhC13. The following salts are described :-LuteocobaZt cobaltinitrite, Co( NH,) ,CO(NO~)~ ; xant hocobalt cobaltinitrite, [NO,*Co(NH,),],,[Co(NOL),],; croceocobalt cobaltinitrite, [ (N 0,) ,Co (NH,) 4] SCo (N 0,) 6 ; $avocoba Zt cobaltnitrite, [ ( N0,)2Co(NH3)~]3C~(K02)6 ; luteocobalt diaminecobnlti- nitrite, Co (NH,) J (NO2),( N H3)2Co (NO,),], ; xanthocobalt diamine- cobalt nitrite, N02Co(NH3)5[ (N0,),(NH~)2Co(N02)2]2 ; croceocobalt dianzinecobazt nitrite, ( N02)2Co ( NH,) 4 ( N02)2(NH3) ?C o (NO,) 2 ; $a uo- coba Zt diaminecobalt nitrite, ( NO,),Co (NH,),( NO2) , (NH,) ~ C O (NO,), ; triaminecobalt nitrate, CO(NH~),(NO~)~,~H~O ; dichrocobalt chloride, C o (N H3) , (OH) ,GI ; nitrotrium inecoba It nitrite, N 02*Co ( NH,) , (NO,) ; and chloronitrotetraminecobalt chZoride, (NO,) C1Co(NH3)4C1. [N02)2Co(NHJ),],Cr04 ; [ (NO,),CO(NH,),]~P~CI, ; A.G. B. Atomic Weight of Molybdenum. By E. F. SMITH and P. MAAS ( Z e i t . anorg. Chem., 5, 280-282) .-Pure sodium molybdate was heated in a current of hydrochloric acid: the reaction Na2MoOJ + 4HCI = 2NaC1 + MoO,,ZHCl + H20 took place, and the residue, which consisted of pure sodium chloride, was weighed.The atomic52 ABSTRACTS OF OHEMICAL PAPERS. weight was found to be 96.087 (0 = 16) as a mean of 10 determina- tions, which varied from 96.130 to 96.031. C. P. B. Tin and Stannic Oxide. By F. EM~CH (Honatsh., 14,345-352). W h e n pure block tin is heated in a stream of air or in an open porcelain crucible, drops of molten metal ooze after some hours from beneath the superficial crust of stannic oxide, and, becoming super- ficially oxidised, gradually assume curious forms resembling bulbs and worms. After 12 hours' heating, the oxide is crystalline, and consists of microscopic, needle-like, rhombic plates of sp.gr. 7-01. The crystah seem to be formed directly from the metal, as the amorphous oxide prepared from " mstastannic acid " may be heated in contact with tin in an indifferent atmosphere without undergoing change of form. They are not due to the oxidation of metallic vapour, since the metal is not volatile at the temperatures employed. When tin containing traces of iron (0.05 per cent.) is heated in the manner described above, most of the iron is eliminated with the first portions of stannic oxide, a mere trace (0.001 per cent.) being left in the tin. The stannic oxide assumes a colour varying from brown to yellow with the amount of iron, thus serving as a good qualitative test for that impurity.The oxide formed after the iron has been eliminated is snow-white. JN. W. Thorium Compounds. By P. JANBBSCH, J. LOCKE, and J. LESINSKY (Zeit. anorg. Chem., 5. 283-287).-This is a preliminary paper. I t contains a detailed account of the means adopted to obtain pure thorium oxalate from thorite and orangite. From the hydroxide the bromide (? ThBrl + 10HzO) and iodide were prepared; they form very deliquescent crysbals. C. F. B. Double Haloids of Antimony and Rubidium. By H. L. WHEELER (Zeit. anorg. Chewt., 5,253-263 ; and Amer. J. Sci., [3], 46, 269-279).-The double haloids enumerated below have been pre- pared by concentrating solutions in 10 per cent. halogen acids of the two haloids in varying proportions ; the proportion by molecules of riibidium t o antimony 1ialoYd in the solution is given by the numbers in square brackets.The 23 : 10 salts might equally well have a 16 : 7, 9 : 4, or 7 : 3 formula; the 3 : 2 salts are isomeric with the corresponding arsenic compounds. RbC1,2SbC13 + HaL0 [l : 10, 6, or 81 ; colourless, monoclinic plates melting at 77", a : b : c = 1.699 : 1 : 0.820 ; /3 = 89" 2d#. RbCl,SbCl, (1 : 4 or 31 ; colourless, monoclinic crystals, a : b : c = 1.732 : 1 : 1.085; p = 65" 34'. 3RbC1,2SbC13 [l : 1+] ; yellow rhombohedra, exhibiting rhombo- 23RbC1,1OSbCl, [l, 4, or 6 : 13 ; colourless, 3RbBr,2SbBr3 C2.3 or 5 : 11 ; lustrous, yellow, hexagonal plates. 3RbI : 2Sb13, red rhombohedra. hedral tetartohedry. hexagonal plates. 23RbBr,10SbBr3 [6, 8, or 13 : I] ; lustrous, yellow needles. C. F. B.NINERALOQlCAL CHI!NISTRY. 53 Recovery of Osmium from Residues. By W. GULEWITSCH ( Z e i t . aizorg. Chenz., 5, 126 -128).-The osmium residues ohtnined in histological laboratories are the solution of osmium tetroxide drained or filtered from the treated objects, and the de'6ris of the objects themselves. The solutions are treated with zinc and hydro- chloric acid, the osmium is collected, washed successively with hydrochloric acid, water, alcohol, and ether, and then allowed to dry at a low temperature The osmium is then transferred to the hind portion of a combustion tube which has a constriction containing an asbestos plug in front of the osmium; the forward portion of the tube is bent into a U, and the end is drawn out to a fine tube which is bent over and placed in a flask ; the U -tube and the flask are sur- rounded by a freezing mixture. A current of dry oxygen is passed through the apparatus, the osmium is gently heated, and the re- sulting tetroxide condensed in the (J bend; if the process be properly conducted, none of the tetroxide will be carried over into the flask. The osmium cannot be directly sublimed as tetroxide from the objects because of the organic matter which they contain. These residues are warmed iri a retort with 10 times their weight of aqua regia, and the resulting solution is distilled until two-thirds have passed over, the receiver being well cooled. The distillate is again distilled uutil two-thirds have passed over, the second distillate is reduced by zinc, and the precipitated osmium is treated as described above. A large excess of ziiic must be used, and the liquid must be warmed, in order that the osmium may be precipitated in a condition in which it will be retained by a filter. A. G. B.