INORGANIC CHEMISTRY. 111 I n o r g a n i c C h e m i s t r y . Sulphur Tetroxide. By D. CARNEGIE (Chem. News, 64, 158- 159).-A criticism of Traube’s work on the electrolysis of 40 per cent. aqueous sulphuric acid (Abstr., 1891, 978). The author, whilst admit- ting that the substance formed cannot be a heptoxide of sulphur, comments on the absence of direct evidence in favour of the existence of a tetroxide. The ratio 1 : 5 of active oxygen to sulphuric anhydride, which Traube considers t o prove the existence of the tetroxide SO, in the electrolysed solution, would be equally well112 ABSTRAOTS OF OEEMIOAL PAPERS. explained on the hypothesis of the existence of a substance having the composition S207,H20.,,xH20. The existence of such a substance would not only accord with Berthelot’s results (Abstr., 1878, 469), but would harmonise with the known existence of tungsten and molybdenum heptoxides, and with the tendency of peroxides of the type M20, to form stable compounds with hydrogen peroxide. Moissmn’s perchromic acid (Abstr., 1884, 20 ; comparo Berthelot, Abstr., 1889, 350) might then be regarded as a compound, Cr207,H202,H20, of the same type. With regard to the low amount of active oxygen shown by the iodometric method, the author points out that this cannot be accountJed for by the presence of sodium hydrogen carbonate, since the latter has practically no action on iodine, but that it can be accounted for by the presence of free alkali in the potassium iodide, and that this supposition accords with the higher results which were obtained on working with acid instead of neutral solutions.JN. W. Azoimide. By T. CURTIUS (Bey., 24, 3341-3349; compare Abstr., 1891, 56).-Ethereal salts of benzoic acid react with hydrazinc hydrate accordiiig to the equation PheCOOR + N2H4,H20 = COPh*NH.NH2 + ROH + H20. The benzoylhydrazine, when treated with nitrous acid, yields benzoylazoimide, and on digesting this with sodium ethoxide it IS decomposed quantitatively into sodium nitride, NaN3, and ethyl benzoate. The salts of azoimide may also be prepared from hippurylhydrazine, by the action of nitrous acid; in this case, a compound is formed which was previously termed nitrosohippurylhydrazine ; it is, however, n diazo-compound with the formula COPh*NH*CH,*CO*NH*N:N*OH (see below) ; it cannot be converted into hippurylnzoimide by elimin- ation of water, but on treatment with ammonia in alcoholic solution, it is decomposed into hippuramide and ammonium nitride ; the hip- puramide combines with hydrazine hydrate to form hippuryl- hy drazine. Argentic nitride, AgN,, has been previously described; it is soluble in ammonia, from which it crystallises in long, colourless needles ; i t is exceedingly explosive.Mercurous nitride, HgN,, is precipitated in microscopic needles which are insoluble in water ; it is more stable than either the silver or lead salts, becomes yellow on exposure to light, and yields a black, insoluble compound with aqueous ammonia. Plztmbic nitride, PbN6, is prepared by adding plumbic acetate to a solution of the sodium or ammonium salts ; it is soluble in excess of the precipitant, but insoluble in water in the cold, and more sparingly soluble in boiling water than plumbic chloride, which it closely resembles. It crystallises from water in long, colourless, lustrous needles, which explode violently on gently warming and decompose gradually when heated with water o r acetic acid.Sodium nitride is most readily prepared in the manner described above, but may also be obtained by adding soda to solution of the free acid or of the ammonium salt; it is readily soluble in water, in- soluble in alcohol or ether, has a slight alkaline reaction and aINOROANIO CHEMISTRY. 113 saline taste, The compound is neit,her ~olatile nor hygroscopic ; its solution may be evaporated to dryness without undergoing any change, and i t only explodes when heated to a comparatively high temp- erature.Ammonium nitride, N4Hi, obtained as above, is rea,dily purified by adding ether to the alcoholic solution, or it may be crystallised from alcohol, irom which it is deposited in plates closely resembling am- monium chloride, but not belonging to the regular system. It is excessively volatile, and explodes violently when heated in a combus- tion tube with cupric oxide in a current of air; it may, however, be sublimed by cautiously heating at a little aboye loo", although violent explosions occur if it is rapidly heated, Hydrazine nitride, NsH5, is prepared by adding hydrazine hydrate to ammonium nitride or to the free acid; it crystallises i n long, lustrous prisms, or in plates, and is sparingly soluble in alcohol. By detonation, or on rapidly heating, the compound explodes violently, but it will burn quietly with a smoky, slightly yellow flame ; if the combustion takes place on a metallic surface, every trace of oxide on it will be reduced arid the metal will appear bright and polished.The formation of diazohippuramide, COPh*NH*CH,-CO*NH*N:OH, has already been described ; it combines with ammonia. aniline, hydr- azine, and similar compounds with elimination of axoiniide, whilst the action of water, alcohol, aldehydes, or acidyl hydrazines causes nitrogen to be evolved ; for example, the action of aniline on diazo- bippurlylamide gives rise t o hippurylanilide, aniline nitride, and water ; the reaction with alcohol is represented by t.he equation, NHBz~CH2*CO*NH*N2*OH + EtOH = N2 + NHBz*CH2*CO*NH*OEt + HzO.On treating diazohippurylamide with benzoylhydrazine, nitrogen and water are eliminated, and a compound is formed which has the formula NHBz*CH,*CO*NH*NH*NH.Bz ; this is exceedingly stable towards acids and alkalis, and attempt8 to hydrolyse it have hitherto been unsuccessful; its properties and mode of formation prove it to be a derivative of trinmide, NH2*NH*NHz, which has not as yet been obtained in the free state. J. B. T. The Colour of Nitric Acid. By L. M~CHLEWSKI (Ber., 24, 3271--327G).-It is well known that as water is added to red fuming nitric acid, the colour changes through green to blue and finally dis- appears. I n explanation of this, it has always been assumed that the red acid i s a solution of nitrogen peroxide, N204, in nitric acid, and that the water added decomposes the peroxide with formation of nitric and nitrous acids.The solution of nitrous acid in nitric acid is blue, and this, with the red solution of still undecomposed peroxide, gives a green colonr. As more water is added, this excess of peroxide is decomposed, and nothing is then left but a blue solution of nitrous acid. The author has investigated the matter experimentally in the follow- ing manner :-The gases contained in the coloured acid were expelled by means of carbonic anhydride and collected i n concentrated sulph- uric acid, with which nitrous acid forms nitrosyl hydrogen sulphate,114 ABSTRACTS OF CHEMICAL PAPERS. but with nitrogen peroxide forms a mixture of nitrosyl hydrogen gulphate with nitric: acid in moiecular proportion.The total nitrogen was determined with the nitrometer, and the reducing power by titra- tion with permanganate. Both results were calculated to trioxide, and from the ratio of the second to the first a conclusion could be drawn as to the constitutlion of the mixture of gases absorbed. I f the gas were pure trioxide, the ratio would of course be 1 ; if it were pure peroxide, 0.5, since the reducing power of the peroxide is only half that of the trioxide. The results obtained were astonishing, for although they showed that the blue acid contained pure trioxide, yet they also showed that the green acid did not contain more than a t most mere traces nf the peroxide. The author was consequently led to suspect that his results were vitiated by the presence of nitric oxide, NO, in the coloured acids examined; and in fact when the experiments were repeated, the gases that escaped absorption in the strong sulphuric acid being passed through a strongly acid per- manganate solution, the permanganate was perceptibly reduced.The reduction was, of course, due to nitric oxide, and a great deal of this gas must have been present originally, for had only a small quantit>y been there, i t would have formed nitrosyl hydrogen sulphate wit.h the peroxide also present and the concentrated snlph- uric acid, and would thus have been absorbed. The solutions investigated were made by passing the gaseous oxide into pure nitric acid. Solutions obtained by mixing pure liquid peroxide with nitric acid of different strengths are now being in- vestigated.C. F. B. Boron Phosphoiodides. By H. MOISSAN (Conz-pt. rend., 113, 624--627).-Melted phosphorus acts with great energy on boron triiodide, and if red phosphorus is heated in the vapour of the iodide, decomposition takes place with incandescence. If, however, a solution of the iodide in carbon bisulphide is mixed with a similar solution of phosphorus, great care being taken t o avoid the presence of moisture, the reaction takes place more slowly. The mixture is sealed up in a flask and kept a t the ordinary temperature of the laboratory ; it is at first clear, but has a red colour. I n a few minutes a brown precipi.- tate begins to separate, and the reaction is complete in about three hours.The product is filtered through glass wool, washed with carbon bisulphide, and dried in a vacuum, the apparatus being filled with carbonic anhydride until the latter is removed by the pump, The product is boron phosphodiiodide, BPI,, an amorphous, homo- geneous, deep-red powder. When heated in a vacuum, it melts at 190-200", and will remain in superfusion a t the ordinary temperature for a long time ; in a vacuum, it begins to volatilise at 170-200", and condenses on the cold part of the tube in distinct red crystals. It is only very slightly soluble in carbon bisulphide, and seems t o he com- pletely insoluble in benzene, phosphorus trichloride, and carbon tetrachloride. It is extremely hygroscopic, and decomposes very rapidly in moist air. In presence of a large excess of water, it be- comes yellow, without, apparent development of heat, and hydriodic, phosphorous, and boric acids are formed, a small quantity of phosphineINORGANIC CHEMISTRY.115 being evolved, and a small quantity of a yellow substance with an odour of phosphorus being deposited. With a very small quantity of water, the yellow precipitateis produced i n larger quantity, and a distinct quantity of phosphonium iodide is formed. Boron phosphodiiodide, wheir heated in hydrogen sulphide, yields boron sulphide, phosphorus sdphide, and hydrogen iodide without any free iodine. Dilute nitric acid yields phosphoric acid and boric acid, whilst strong nitric acid produces the same result, but with in- candescence. Siilphuric acid (even Nordhausen) has 110 action in the cold, b u t , on heating, free iodine, hydrogen iodide, and sulphurous anhydride are evolved.Phospliorous trichloride and carbon tetra- chloride have no action even in sealed tubes at 100". Chlorine pro- duces incandescence, with formation of boron chloride, iodine chloride, and phosphorus pentschloride. When heated in oxygen, the compound burns and yields iodine, boric anhydride, and phosphoric anhydride. Sodium has no action in the cold, but decomposition takes place at t>he melting point of the metal. Powdered magnesium reacts with incandescence. When thrown into mercury vapour, the phosphodiiodicle takes fire a t once. I n presence of carbon bisulphide the behaviour of metals is different; magnesium or sodium a t the ordinary temperature produces a red compound, PBT, whilst silver or mercury in the cold, or more rapidly a t loo", yields a maroon- coloured compound with the properties of boron phosphide, BP.Boron phosphoiodide, BPI, is obtained by heating the preceding compound in hydrogen, and is an amorphous, red powder, somewhast less hygroscopic than the diiodide. It volatilises in a vacuum a t 210-250" without previous fusion, and condenses in orange-yellow crystals. Strong nitric acid decomposes i t with development of heat and without incandescence, iodine being liberated. Concentrated sulphuric acid has no action in the cold, but, OIL heating, iodine, sulphurous anhydride, and boric acid are formed. When heated out of contact with air, it decomposes a t a temperature below dull redness with evolution of Fapours of iodine and boron phosphide.Mercury in excess, in presence of dry carbon bisulphide, yields mercuric iodide and boron phosphide a t the ordinary temperature. Boron phoThide, BP, resembles the phosphoiodide B P I in its general properties. It can be obtained by heating the phosphoiodide in hydrogen, and if the heating is continued, a residue of the composi- tion B,P, is left. C. H. B. Reducing Action of Graphitoidal Silicon. By H. N. WARREN (Chenz. News, 64, 75).-When the oxides of easily reducible metals, such as lead, copper, and silver, are heated to dull redness with powdered graphitoydal silicon, they are reduced to the metal, and if the silicon is in excess, a metallic silicide is formed. The oxides of chromium, tungsten, and molybdenum may also be reduced in thia way.In some cases, the reduction takes place with explosive violence ; when, for instance, a small quantity of a mixture of equal paxts of finely-divided silicon, aluminium, and litharge was melted before the blowpipe, the explosion was so violent as to indent the supporting brick. JN. W.116 ABSTRACTS OF CBEMICAL PAPERS. Solubility of Sodium Carbonate and Sodium Hydrogen Carbonate in Solutions of Sodium Chloride. By K. REICR (Monatsh., 12, 464--473).-The solubility of sodium carbonate a t 15" in a solution of sodium chloride of gradually increasing concentration at first diminishes and then increases. The solubility y as a function of the quantity x of sodium chloride in 100 parts of water is expressed by the formula y = 61.406 - 2.091077~ + 0.0554932 - 0*00029'7357~~. Accordingly, the minimum lies near cc = 23.15 where y = 39.05. On passing carbonic anhydride through the saturated solution, the amount of bicarbonate precipitated increases with the quantity of sodium chloride in solution ; but a maximam cannot be recognised.G. T. M. Allotropic Silver. By M. C. LEA (Phil. Mag. [ 5 ] , 32, 337-3421. -The blue form of allotropic silver is capable of change into an intermediate yellow form which seems t o be identical with that, into which the gold-coloured form passes under the influence of various causes (Abstr., 1891, 803). The change takes place at about B O O , both with lumps of the blue silver and with films. By the action of sulphuric acid, however, blue silver can be converted into yellow silver at the ordinary temperature, and consequently with retention of all its active properties.40 grams of sodium hydroxide and 40 grams of yellow or brown dextrin are dissolved in 2000 C.C. of water, and 28 grams of silver nitrate is added in successive very small quantities, with frequent -agitation. The solution is slightly turbid. and is deep-green by re- flected' light, red by transmitted light. The precipitate tbat forms spontaneously or is produced by the addition of acetic acid, dilute nitric acid, and many neutral substances, consists of blue silver, but if sulphuric acid is added, the precipitate, when dried in films, is blue, green, yellowish-green, or yellow, according to the proportion of acid used. When the silver solution is mixed with an equal volume of a mixture of i s 5 C.C. of sulphuric acid and 92.5 C.C.of water, the precipitate consists wholly of yellow silver, bnt with higher propor- tions of acid the product dries with a coppery shade. The lustre of )the product, diminishes as the proportion of acid used for precipitation is increased. Conversely, it solution which would yield yellow silver under or- dinary circumstances can be made to yield blue silver by the addition of an alkali, and it is immaterial whether the alkali is added to the ferrous tartrate or the silver mixtnre or to a mixture of the two. There is, therefore, a tendency for acids to produce the yellow pro- .duct, and alkalis the blue product., but it is a tendency only, and both forms can be obtained from neutral solutions ; for instance, ferrous tartrate and silver tartrate yield gold-coloured silver, but ferrous citrate and silver citrate produce the blue variety.When sodium hypophosphite is added to silver nitrate, there is no recluction, but if phosphoric acid is added so that hypophosphorous .acid is liberated, it transient red colour appears, and red and blueINORGANIC CHEMISTRY. 117 stains are formed on the side of the vessel. Phosphorous acid gives similar though somewhat less distinct results. The blue silver obtained by adding the soda-dextrin silver solution to an equal volume of water containing 4 per cent. of sulphuric acid is not only constant in character but is one of the forms most sensi- tive to light. When this form is exposed to light, it first becomes more distinctly blue, then yellowish-brown, and finally is converted into the golden-yellow intermediate or crystalline form, with great brilliancy and lustre.It is noteworthy that the action of light on this blue varietyat first. increases its sensitiveness to reagents such as potassium ferricpanide, and afterwards reduces it. This is well shown if one part of a film is covered with an opaque substance, another part with a translucent substance, whilst the third is left uncovered, and the three are ex- posed simultaneously to bright sunlight for about five hours and afterwards treated with dilute ferricyanide solution. The author regards this phenomenon as analogous to the reversing action ob-- served with gelatinobromide plates. The production of reduced silver is direct when an ordinary silver.compound is converted into metal without formation of a sub-salt, and indirect when the silver compoufid is first reduced to a, sub-salt, and the latter is afterwards reduced to the metal. It would seem that only under the latter conditions is there any formation of allo- tropic silver. If, in any of the three principal methods of producing allotropic silver, the action is interrupted by the addition of hydro- chloric acid, a dark, chestnut-brown or purple-brown mixture of the subchloride with the photochloride is obtained, and from it beautiful rose-red photochloride can be obtained by treatment with cold dilute nitric acid after complete removal of the hydrochloric acid. This result is only obtained by interrupting the reaction before it is c v - .plete, and if the hydrochloric acid is added after complete reduction, only grey, normal silver is precipitated. Ifi every case examined, silver subchloride is obtained as one of the products when a reaction resulting in the formation of allotropic silver is interrupted by the addition of hydrochloric acid before reduction is complete. The rich and varied colour of silver sub-salts would seem to indi- cate that in these compounds the metal existls in an allotropic form, but, on the other hand, the greater activity of allotropic silver and its lower sp. gr. would tend to show that the allotropic form has a simpler molecular structure than the normal metal. Colloidal Silver. By E. A. SCHNEIDER (Ber., 24, 3370-3373).- Collo'idal silver prepared according to Carey Lea's method, by the reduction of silver nitrate with ferrous citrate, cannot be purified by dialysis alone ; the better plan is to separate the solid colloid from the mother liquor as completely as possible by filtration, then to dissolve the silver in a little water and allow this solution to dialyse.On adding hydrochloric acid to aqueous collojdal silver solutions, silder and argentic chloride are precipitated, the proportion of the latter being greater with increasing quantities of hydrochloric acid ; the mixed precipitate is extracted with ammonia, the argentic chloride * C. H. B.118 ABSTRAOTS OF OEIEMIOAL PAPERS. precipitated by acidification with nitric acid, and hydrochloric acid added t o the filtrate ; if the quantity of hydrochloric acid originally employed was small, a further precipitate is produced, showing tho presence of a silver snbchloride.The mixed precipitate of silver and argentic chloride was always rose-red. Nitric and sulphuric acids react with colloydal silver solutioiis in a similar manner. No evolution of hydrogen could be detected even when sutlicient hydrochloric acid was added to the silver solution to produce a considerable precipitate of argentic - chloride ; this may indicate the presence of nrgentous oxide : neither was oxygen evolved when, in consequence of the small quantity of hydrochloric acid em- ployed, t.he precipitat,e consisted of almost pure silver. J. B. T. Direct Combination of Chlorine and Bromine with Metals, By H. GAUTIER and G. CHARPY (Compt.rend., 113, 597-600).-Well cleaned wires of various metals, 2 mm. in diameter, were kept in the dark in contact with dry bromine for a definite length of time at 15" and 100". The percentage loss of weight in each case is given in the following table :- At 15' in At 15' in At 100' in 8 days. 4 months. 8 days. Magnesium ...... 0.0 0.0 0.19 Copper .......... 0.371 1- 740 6.62 Zinc ............ 0.289 0.48 7 0.63 Iron ............ 0.210 0.440 23.27 Silver. .......... 0.003 0.540 - Aluminium under similar conditions combines energetically with bromine and becomes incandescent, a burning fragment running about on the surface of the bromine like potassium on water. With liquid chlorine in sealed tubes at the ordinary temperature, the results are similar, the percentage losses being : magnesium, 0.0 ; zinc, 0.0 ; iron, 0.740 ; copper, 3.241 ; silver, 0.673.Potassium, sodium, and aluminium seem to be unaffected by liquid chlorine at its boiling point, but a t -20" aluminium combines with the halogen with incandescence. Magnesium and aluminium, when placed in bromine-water, produce a regular evolution of hydrogen, and, after some time, an oxybromide separates, the reactions being analogous to the decomposition of water by iodine in presence of aluminium. With zinc, iron, and copper, there is no evolution of gas, and a wire 2 mm. in diameter and 50 mm. in length disappears in seven t o eight hours in excess of bromine-water. It seems probable that in these cases the decomposition of water by the halogen is accelerated by the presence of' the metal, the latter being converted into an oxide which is attacked by the hydracid formed. C.H. B. I n presence of water, the results are very different. Lithium Copper Chloride. By A. CHASSEVANT (Compt. rend., 113, 646-648) .-When a concentrated solution of lithium chloride isINORGANIO OHEMISTRY. 119 added to 8 cancentrated solution of an equivalent quantity of cupric chloride, a magma of crystals of the latter salt is formed, but if the midurc is evaporated in a vacuum over phosphoric acid or on a water-bath at loo", the ci-ystals dissolve, the liquid acqnires a, brownish-red colour and deposits transparent, gzrnet-red crystals of the double chloride 2CuClZ,2LiC1,5H,O. When exposed to the air, they decompose, and become opaque, crystals of cupric chloride separating, whilst the lithium chloride deliquesces.If heated rapidly to 130", they melt in their water of crystallisation arid form a deep-brown, almost black, solution. A t tt higher temperature, the salt, like lithium chloride under similar conditions, decomposes and evolves chlorine. When heated slowly i n an oven, or in a current of dry ail-, the crystals become anhydrous a t 100-120", but some hydrochloric acid is likewise given off, and, on treatment with water, a residue of cupric oxychloride is left. The double salt can be obtained as an anhydrous, chamois-coloured powder by heating it a t 120" in a current of dry air mixed with dry hydrogen chloride. It is decomposed by water, but can be recrystallised from a concentrated solution of lithium chloride.C. H. B. Formation of Saline Hydrates at High Temperatures. By G. ROUSSEAU (Conzpt. Tend., 113, G43--648).---When the hydrated sodium ferrite obtained a t 800", and previously described, is allowed t o remain in contact with glycerol for several days, and is washed first with this liquid and afterwards with absolute alcohol, the dried residue contains only 9-68 per cent. of water, instead of the 14.5 per cent, that it contains when water is used for washing. The ferrite, 14Fe,0,,13H20,Naz0, when treated with glycerol in a similar manner, loses 2.79 per cent. of water. If either of these compounds is heated at 100" with glycerol, its colour rapidly becomes darker, and if diges- tion is prolonged, the whole of the water and alkali is removed, and anhydrous ferric oxide remains.The sodium manganites, such as 12Mn02,4H,0,Na,0, are not affected i n a similar way by glycerol. The author considers that these results establish. his previous con- clusions and support the view that the alkaline oxide replaces part of the water. C. H, B. Crystallised Ferric Oxychlorides. By G. ROCSSEAU (Compt. rend., 113, 542-5444 .-Very concentrated solutions of ferric chloride, containing ahout 80 per cent. of Fe2Cls, maintained at a temperature of lW--220" for some time, yield a crystallised feryic oxychloride, 2Fe,03,Fe2Cl, + 3H,O. By prolonged boiling with water, this compound is gradually converted into a ferric hydroxide, (Pe,O, + H20)3, which retains the same crystalline form as the oxgchloride. Solutions containing 85 to 911 per cent.of the anhydrous ferric chloride have been heated in sealed tnbes, together with a fragment of marble or giobertite. Between 225" and 280", red-brown lamella: of the oxychloride 2Fe,03,1?e2C16 are obtained. Between 300" and 340°, large plates of a brownish-black oxychloride 3Fe2O3,Fe2Cl6 are formed.120 ABSTRAOTS OF CHEMICAL PAPERS. The author has been unable to study the reaction at higher tem- peratures, but believes that a, series of oxychlorides of the type (Fe,O3),,Fe,CI6 would be formed, in which the proportion of EeZO, would increase with the temperature. The anhydrous oxychlorides are very sparingly soluble in dilute mineral acids. When boiled with water i n presence of marble for 150-200 hours, they lose all their chlorine as hydrochloric acid, ferric oxide of a fine brownish-red colour remaining. The optical properties of these oxychlorides have been determined by Fouqu6. They occur in prisms giving longitudinal extinction.The plans of their optic axes is transverse, and the bisectrix is positive. It would be interesting to observe whether the rhombic oxychlor- ides retain their form during the concentration of ferric oxide in the molecule, or whether they assume, at a temperature near a red heat, the hexagonal form characteristic of ferric oxide. On the latter hypothesis, the hydrolysis of the hexagonal oxychloride, by boiling with water, should give a new method for the synthesis of hzmatite, allowing the determination of the degree of polymerisation of this mineral oxide. W. T. Action of Water on Glass.By E. PFEWFER (Ann. Phys. Chein. [ Z ] , 44, 239-264) .--The author has taken advantage of electrolytic conductivity for the purpose of determining the amount of substaiice dissolved from glass by water at low temperatures. Ordinary chemical methods give very uncertain numerical results, on account of the difficixltg in determining the exact magnitude of the large surface which must be exposed to the action of water, and even these results are obtained ander conditions diverging considerably from those of laboratory practice. Water first of all dissolves practically pure alkali (potassium or sodium hydroxide) out of the glass, and this afterwards exerts its own influence by dissolving silica. The author estimates the amount of alkali dissolved by determining the electrical conductivity of the solution, to which it is proportional. As the molecular conductivities of potassium hydroxide and sodium hydroxide lie close together (220 x 10-7 and 200 x according to Kohlrauschj, no great error is comrnitted in estimating the total amount dissolved on the assumption that each alkali dissolves proportionately to the extent to which i t is contained in the glass.When silicates are formed in the solution, the conductivity falls. The experiments were made by exposing cylinders of good Thuringian glass with known surface to the action of water contained in glazed porcelain vessels; the temperature in the three series of observations made being lo", 30", and 30". For one sort of glass at a given temperature, it was found that A0 = A w / o is a constant, A being the increase in conductivity per hour, w the volume of water, and o tbe surface of the cylinder.When the glass has been exposed for some time to the action of water at a temperature of 60", the value of A,, for 2G" falls considerably. Aha("") for the first day is muchINORGANIC CEEMISTRY. 122 greater than A,-,(2o) afterwards. With the specimen of glass examined, it was found that a t 20" one to two millionths of a milligram were dissolved out by 1 C.C. of water per square centimeter in an hour. No silica is dissolved a t 10" or 20". Ati 30°, however, a considerable falling off of A. with the time is observed, due in all probability to this cause. The values of AL, (reduced to go*) for the various tern- peratures are as follows :- 10".20". 30". A. .... .. .. 25 100 673 Prolonged treakment of the glass at a low temperature does n o t appreciably affect its solubility a t a higher temperature. J. W. Stannibromides. By LETEUR (Compt. rend., 113, 540-542).- The stannibromides of the alkali metals and magnesium are yellow, well-crystallised substances. Concentrated solutions have the same colour, but a t a certain state of dilution, the colour disappears. The anhydrous stnnnibromides of potassium and ammonium only suffer change in very moist a i r ; others are very deliquescent. Concen- trated solutions may be heated without decomposition ; on dilution, decomposition occurs with the formation of hydrobromic acid and the deposition of hydrated tin dioxide. Alcohol decomposes these corn- pounds, slowly in the cold, more rapidly on heating ; benzene has no action on them.The general method for the preparation of the stannibromides consists in mixing concentrated solutions of the two bromides and evaporating the mixture in R vacuuni or in dry air. The ammonium salt, (NH4)2SnBr6, forms sulphur-yellow ocka- hedra belonging to the cubic system; it decrepitates when heated, and volatilises with partial decomposition. The sodium salt, Na,SnBr, + 6H20, for.ms yellow prisms of the monoclinic system, having a position of extinction in polarised light at lFjo to the longer axis. It is very deliquescent, but effloresces rapidly over sulphuric acid OY in a vacuum. It decomposes, when heated, with evolntion of water and stannic bromide. The lithium salt is probably Li,SnBr, + 6H20, but the water cannot be accurately determined owing to the extreme deliquescence of the compound.It forms small, yellow, prismatic needles whicli act on polarised light, and appear to belong to the monoclinic system. Over sulphnric acid, these crystals effloresce. giving a citron-yellow, crystalline powder tending towards the composition Li2SnRr, + 5H,O. The magnesium salt, MgSnBr, + 10H20, gives small, sulphnr- yellow, monoclinic crystals. The ordinary form is a prism showing the faces q1 and h, with modifications on the angle a. Macles are frequent. The study of the alkaline earthy stannibromides is being nolv carried on. W. T. VOL. LX1I. 72 Extinction occurs a t an angle of 60".122 ABSTRACTS OF CIHEMICAL PAPERS. Dissolution of Bismuth Chloride in a Saturated Solution of Sodium Chloride : Basic Bismuth Salicylate.By H. CAUSSE (Compt. rend., 113, 547-549) .-Sodium chloride, like ammoniiiin chloride, may be employed instead of free acid to prevent the dis- sociation of bismuth salts by water. Hence, in presence of sodium chloride, hydrochloric acid may be completely neutralised by bismuth carbonate or oxide. 100 C.C. of hydrochloric acid solution containing 3.0775 grams HC1 is left in contact with 3 grams of bismuth oxide until no further solution takes place ; the remaining oxide is collected and weighed. 1.50 grams of the oxide are dissolved, requiring 0.4775 gram of the acid t o form BiCI,; the remaining 2.60 grams of the hydrochloric acid are required to maintain the equilibrium in the solution.With 100 C.C. of acid containing 6.155 grams HCl, 6.00 grams of oxide are dissolved, and 3.117 grams of free acid remain. With 100 C.C. of acid containing 9.2325 grams HCI, 10 grams of oxide are dissolved, and 4.557 grams of acid remain. Each of these solutions is saturated with common salt, and then treated with bismuth oxide. The quantities of bismuth oxide dissolved as compared with the quantities required to neutralise the hydrochloric acid present with production of bismuth trichloridc are respectively 6.80 : 6.584, 13.25 : 13.160, and 20.25 : 19.70. The numbers given above do not show that any definite relation exists between the free acid and the amount of bismuth chloride formed. To ascertain whether such a relation exists at the experi- mental limit, the author treats 50 grams of oxide with 50 C.C.of saturated hydrochloric acid containing 22.80 grams of acid. 47.50 grams of oxide are dissolved, and 5.18 grams of free acid remain. 5.40 grams of acid would be uncombined if the reaction were as follows: Bi,O, + 8HC1 = S(BiCl,,HCl) -l- 3H?O ; helice tRe author concludes that under these circumstances a definite salt is formed. To obtain basic bismuth salicylate, 40 C.C. of conccritrated hydrochloric acid is saturated with bismuth oxide in presence of 500 C.C. of saturated sodium chloride solution ; to another SO0 C.C. of brine are added 9 grams of soda and 22 grams of sodium salicylate, the two solutions are mixed, and the precipitate formed is washed with water containing a few drops of nitric acid.The basic sali- cylate, C,H,BiO4,IIZO, obtained, forms microscopic prisms, and has properties similar to those of the normal salicylate previously described. It is decomposed by heat with loss of the whole of its salicylic acid, which may also be completely eliminated by boiling concentrated alcohol. The constitution of t h i s salt may be represented by the formula OH*CsH~*COO*Bi(OH)z, which accounts for its ready hydrolysis. W. T.INORGANIC CHEMISTRY. 111I n o r g a n i c C h e m i s t r y .Sulphur Tetroxide. By D. CARNEGIE (Chem. News, 64, 158-159).-A criticism of Traube’s work on the electrolysis of 40 per cent.aqueous sulphuric acid (Abstr., 1891, 978). The author, whilst admit-ting that the substance formed cannot be a heptoxide of sulphur,comments on the absence of direct evidence in favour of theexistence of a tetroxide.The ratio 1 : 5 of active oxygen to sulphuricanhydride, which Traube considers t o prove the existence of thetetroxide SO, in the electrolysed solution, would be equally wel112 ABSTRAOTS OF OEEMIOAL PAPERS.explained on the hypothesis of the existence of a substance having thecomposition S207,H20.,,xH20. The existence of such a substancewould not only accord with Berthelot’s results (Abstr., 1878, 469),but would harmonise with the known existence of tungsten andmolybdenum heptoxides, and with the tendency of peroxides of the typeM20, to form stable compounds with hydrogen peroxide. Moissmn’sperchromic acid (Abstr., 1884, 20 ; comparo Berthelot, Abstr., 1889,350) might then be regarded as a compound, Cr207,H202,H20, of thesame type.With regard to the low amount of active oxygen shown by theiodometric method, the author points out that this cannot beaccountJed for by the presence of sodium hydrogen carbonate, sincethe latter has practically no action on iodine, but that it can beaccounted for by the presence of free alkali in the potassium iodide,and that this supposition accords with the higher results which wereobtained on working with acid instead of neutral solutions.JN. W.Azoimide.By T. CURTIUS (Bey., 24, 3341-3349; compareAbstr., 1891, 56).-Ethereal salts of benzoic acid react with hydrazinchydrate accordiiig to the equation PheCOOR + N2H4,H20 =COPh*NH.NH2 + ROH + H20. The benzoylhydrazine, when treatedwith nitrous acid, yields benzoylazoimide, and on digesting this withsodium ethoxide it IS decomposed quantitatively into sodium nitride,NaN3, and ethyl benzoate.The salts of azoimide may also be prepared from hippurylhydrazine,by the action of nitrous acid; in this case, a compound is formedwhich was previously termed nitrosohippurylhydrazine ; it is, however,n diazo-compound with the formula COPh*NH*CH,*CO*NH*N:N*OH(see below) ; it cannot be converted into hippurylnzoimide by elimin-ation of water, but on treatment with ammonia in alcoholic solution,it is decomposed into hippuramide and ammonium nitride ; the hip-puramide combines with hydrazine hydrate to form hippuryl-hy drazine.Argentic nitride, AgN,, has been previously described; it is solublein ammonia, from which it crystallises in long, colourless needles ; i tis exceedingly explosive.Mercurous nitride, HgN,, is precipitated in microscopic needleswhich are insoluble in water ; it is more stable than either the silveror lead salts, becomes yellow on exposure to light, and yields a black,insoluble compound with aqueous ammonia.Plztmbic nitride, PbN6, is prepared by adding plumbic acetate to asolution of the sodium or ammonium salts ; it is soluble in excess ofthe precipitant, but insoluble in water in the cold, and moresparingly soluble in boiling water than plumbic chloride, which itclosely resembles.It crystallises from water in long, colourless,lustrous needles, which explode violently on gently warming anddecompose gradually when heated with water o r acetic acid.Sodium nitride is most readily prepared in the manner describedabove, but may also be obtained by adding soda to solution of thefree acid or of the ammonium salt; it is readily soluble in water, in-soluble in alcohol or ether, has a slight alkaline reaction and INOROANIO CHEMISTRY.113saline taste, The compound is neit,her ~olatile nor hygroscopic ; itssolution may be evaporated to dryness without undergoing any change,and i t only explodes when heated to a comparatively high temp-erature.Ammonium nitride, N4Hi, obtained as above, is rea,dily purified byadding ether to the alcoholic solution, or it may be crystallised fromalcohol, irom which it is deposited in plates closely resembling am-monium chloride, but not belonging to the regular system.It isexcessively volatile, and explodes violently when heated in a combus-tion tube with cupric oxide in a current of air; it may, however, besublimed by cautiously heating at a little aboye loo", although violentexplosions occur if it is rapidly heated,Hydrazine nitride, NsH5, is prepared by adding hydrazine hydrateto ammonium nitride or to the free acid; it crystallises i n long,lustrous prisms, or in plates, and is sparingly soluble in alcohol. Bydetonation, or on rapidly heating, the compound explodes violently,but it will burn quietly with a smoky, slightly yellow flame ; if thecombustion takes place on a metallic surface, every trace of oxide onit will be reduced arid the metal will appear bright and polished.The formation of diazohippuramide, COPh*NH*CH,-CO*NH*N:OH,has already been described ; it combines with ammonia. aniline, hydr-azine, and similar compounds with elimination of axoiniide, whilst theaction of water, alcohol, aldehydes, or acidyl hydrazines causesnitrogen to be evolved ; for example, the action of aniline on diazo-bippurlylamide gives rise t o hippurylanilide, aniline nitride, and water ;the reaction with alcohol is represented by t.he equation,NHBz~CH2*CO*NH*N2*OH + EtOH = N2 + NHBz*CH2*CO*NH*OEt + HzO.On treating diazohippurylamide with benzoylhydrazine, nitrogen andwater are eliminated, and a compound is formed which has theformula NHBz*CH,*CO*NH*NH*NH.Bz ; this is exceedingly stabletowards acids and alkalis, and attempt8 to hydrolyse it have hithertobeen unsuccessful; its properties and mode of formation prove it tobe a derivative of trinmide, NH2*NH*NHz, which has not as yet beenobtained in the free state.J. B. T.The Colour of Nitric Acid. By L. M~CHLEWSKI (Ber., 24,3271--327G).-It is well known that as water is added to red fumingnitric acid, the colour changes through green to blue and finally dis-appears. I n explanation of this, it has always been assumed that thered acid i s a solution of nitrogen peroxide, N204, in nitric acid, andthat the water added decomposes the peroxide with formation ofnitric and nitrous acids. The solution of nitrous acid in nitric acidis blue, and this, with the red solution of still undecomposed peroxide,gives a green colonr.As more water is added, this excess of peroxide isdecomposed, and nothing is then left but a blue solution of nitrous acid.The author has investigated the matter experimentally in the follow-ing manner :-The gases contained in the coloured acid were expelledby means of carbonic anhydride and collected i n concentrated sulph-uric acid, with which nitrous acid forms nitrosyl hydrogen sulphate114 ABSTRACTS OF CHEMICAL PAPERS.but with nitrogen peroxide forms a mixture of nitrosyl hydrogengulphate with nitric: acid in moiecular proportion. The total nitrogenwas determined with the nitrometer, and the reducing power by titra-tion with permanganate. Both results were calculated to trioxide, andfrom the ratio of the second to the first a conclusion could be drawn asto the constitutlion of the mixture of gases absorbed.I f the gas werepure trioxide, the ratio would of course be 1 ; if it were pure peroxide,0.5, since the reducing power of the peroxide is only half that of thetrioxide. The results obtained were astonishing, for although theyshowed that the blue acid contained pure trioxide, yet they alsoshowed that the green acid did not contain more than a t mostmere traces nf the peroxide. The author was consequently ledto suspect that his results were vitiated by the presence of nitricoxide, NO, in the coloured acids examined; and in fact when theexperiments were repeated, the gases that escaped absorption inthe strong sulphuric acid being passed through a strongly acid per-manganate solution, the permanganate was perceptibly reduced.The reduction was, of course, due to nitric oxide, and a great deal ofthis gas must have been present originally, for had only a smallquantit>y been there, i t would have formed nitrosyl hydrogensulphate wit.h the peroxide also present and the concentrated snlph-uric acid, and would thus have been absorbed.The solutions investigated were made by passing the gaseous oxideinto pure nitric acid.Solutions obtained by mixing pure liquidperoxide with nitric acid of different strengths are now being in-vestigated. C. F. B.Boron Phosphoiodides. By H. MOISSAN (Conz-pt. rend., 113,624--627).-Melted phosphorus acts with great energy on borontriiodide, and if red phosphorus is heated in the vapour of the iodide,decomposition takes place with incandescence.If, however, a solutionof the iodide in carbon bisulphide is mixed with a similar solution ofphosphorus, great care being taken t o avoid the presence of moisture,the reaction takes place more slowly. The mixture is sealed up in aflask and kept a t the ordinary temperature of the laboratory ; it is atfirst clear, but has a red colour. I n a few minutes a brown precipi.-tate begins to separate, and the reaction is complete in about threehours. The product is filtered through glass wool, washed withcarbon bisulphide, and dried in a vacuum, the apparatus being filledwith carbonic anhydride until the latter is removed by the pump,The product is boron phosphodiiodide, BPI,, an amorphous, homo-geneous, deep-red powder.When heated in a vacuum, it melts at190-200", and will remain in superfusion a t the ordinary temperaturefor a long time ; in a vacuum, it begins to volatilise at 170-200", andcondenses on the cold part of the tube in distinct red crystals. It isonly very slightly soluble in carbon bisulphide, and seems t o he com-pletely insoluble in benzene, phosphorus trichloride, and carbontetrachloride. It is extremely hygroscopic, and decomposes veryrapidly in moist air. In presence of a large excess of water, it be-comes yellow, without, apparent development of heat, and hydriodic,phosphorous, and boric acids are formed, a small quantity of phosphinINORGANIC CHEMISTRY.115being evolved, and a small quantity of a yellow substance with anodour of phosphorus being deposited. With a very small quantityof water, the yellow precipitateis produced i n larger quantity, and adistinct quantity of phosphonium iodide is formed.Boron phosphodiiodide, wheir heated in hydrogen sulphide, yieldsboron sulphide, phosphorus sdphide, and hydrogen iodide withoutany free iodine. Dilute nitric acid yields phosphoric acid and boricacid, whilst strong nitric acid produces the same result, but with in-candescence. Siilphuric acid (even Nordhausen) has 110 action in thecold, b u t , on heating, free iodine, hydrogen iodide, and sulphurousanhydride are evolved. Phospliorous trichloride and carbon tetra-chloride have no action even in sealed tubes at 100".Chlorine pro-duces incandescence, with formation of boron chloride, iodinechloride, and phosphorus pentschloride. When heated in oxygen, thecompound burns and yields iodine, boric anhydride, and phosphoricanhydride. Sodium has no action in the cold, but decompositiontakes place at t>he melting point of the metal. Powdered magnesiumreacts with incandescence. When thrown into mercury vapour, thephosphodiiodicle takes fire a t once. I n presence of carbon bisulphidethe behaviour of metals is different; magnesium or sodium a t theordinary temperature produces a red compound, PBT, whilst silveror mercury in the cold, or more rapidly a t loo", yields a maroon-coloured compound with the properties of boron phosphide, BP.Boron phosphoiodide, BPI, is obtained by heating the precedingcompound in hydrogen, and is an amorphous, red powder, somewhastless hygroscopic than the diiodide.It volatilises in a vacuum a t210-250" without previous fusion, and condenses in orange-yellowcrystals. Strong nitric acid decomposes i t with development of heatand without incandescence, iodine being liberated. Concentratedsulphuric acid has no action in the cold, but, OIL heating, iodine,sulphurous anhydride, and boric acid are formed. When heated outof contact with air, it decomposes a t a temperature below dull rednesswith evolution of Fapours of iodine and boron phosphide. Mercuryin excess, in presence of dry carbon bisulphide, yields mercuric iodideand boron phosphide a t the ordinary temperature.Boron phoThide, BP, resembles the phosphoiodide B P I in its generalproperties.It can be obtained by heating the phosphoiodide inhydrogen, and if the heating is continued, a residue of the composi-tion B,P, is left. C. H. B.Reducing Action of Graphitoidal Silicon. By H. N. WARREN(Chenz. News, 64, 75).-When the oxides of easily reducible metals,such as lead, copper, and silver, are heated to dull redness withpowdered graphitoydal silicon, they are reduced to the metal, and ifthe silicon is in excess, a metallic silicide is formed. The oxides ofchromium, tungsten, and molybdenum may also be reduced in thiaway. In some cases, the reduction takes place with explosiveviolence ; when, for instance, a small quantity of a mixture of equalpaxts of finely-divided silicon, aluminium, and litharge was meltedbefore the blowpipe, the explosion was so violent as to indent thesupporting brick.JN. W116 ABSTRACTS OF CBEMICAL PAPERS.Solubility of Sodium Carbonate and Sodium HydrogenCarbonate in Solutions of Sodium Chloride. By K. REICR(Monatsh., 12, 464--473).-The solubility of sodium carbonate a t 15"in a solution of sodium chloride of gradually increasing concentrationat first diminishes and then increases. The solubility y as a functionof the quantity x of sodium chloride in 100 parts of water is expressedby the formulay = 61.406 - 2.091077~ + 0.0554932 - 0*00029'7357~~.Accordingly, the minimum lies near cc = 23.15 where y = 39.05.On passing carbonic anhydride through the saturated solution, theamount of bicarbonate precipitated increases with the quantity ofsodium chloride in solution ; but a maximam cannot be recognised.G. T.M.Allotropic Silver. By M. C. LEA (Phil. Mag. [ 5 ] , 32, 337-3421.-The blue form of allotropic silver is capable of change into anintermediate yellow form which seems t o be identical with that, intowhich the gold-coloured form passes under the influence of variouscauses (Abstr., 1891, 803). The change takes place at about B O O ,both with lumps of the blue silver and with films. By the action ofsulphuric acid, however, blue silver can be converted into yellow silverat the ordinary temperature, and consequently with retention of allits active properties.40 grams of sodium hydroxide and 40 grams of yellow or browndextrin are dissolved in 2000 C.C.of water, and 28 grams of silvernitrate is added in successive very small quantities, with frequent-agitation. The solution is slightly turbid. and is deep-green by re-flected' light, red by transmitted light. The precipitate tbat formsspontaneously or is produced by the addition of acetic acid, dilutenitric acid, and many neutral substances, consists of blue silver, but ifsulphuric acid is added, the precipitate, when dried in films, is blue,green, yellowish-green, or yellow, according to the proportion ofacid used. When the silver solution is mixed with an equal volumeof a mixture of i s 5 C.C. of sulphuric acid and 92.5 C.C. of water, theprecipitate consists wholly of yellow silver, bnt with higher propor-tions of acid the product dries with a coppery shade. The lustre of)the product, diminishes as the proportion of acid used for precipitationis increased.Conversely, it solution which would yield yellow silver under or-dinary circumstances can be made to yield blue silver by the additionof an alkali, and it is immaterial whether the alkali is added to theferrous tartrate or the silver mixtnre or to a mixture of the two.There is, therefore, a tendency for acids to produce the yellow pro-.duct, and alkalis the blue product., but it is a tendency only, and bothforms can be obtained from neutral solutions ; for instance, ferroustartrate and silver tartrate yield gold-coloured silver, but ferrouscitrate and silver citrate produce the blue variety.When sodium hypophosphite is added to silver nitrate, there is norecluction, but if phosphoric acid is added so that hypophosphorous.acid is liberated, it transient red colour appears, and red and bluINORGANIC CHEMISTRY.117stains are formed on the side of the vessel. Phosphorous acid givessimilar though somewhat less distinct results.The blue silver obtained by adding the soda-dextrin silver solutionto an equal volume of water containing 4 per cent. of sulphuric acidis not only constant in character but is one of the forms most sensi-tive to light. When this form is exposed to light, it first becomesmore distinctly blue, then yellowish-brown, and finally is convertedinto the golden-yellow intermediate or crystalline form, with greatbrilliancy and lustre.It is noteworthy that the action of light on this blue varietyat first.increases its sensitiveness to reagents such as potassium ferricpanide,and afterwards reduces it.This is well shown if one part of a film iscovered with an opaque substance, another part with a translucentsubstance, whilst the third is left uncovered, and the three are ex-posed simultaneously to bright sunlight for about five hours andafterwards treated with dilute ferricyanide solution. The authorregards this phenomenon as analogous to the reversing action ob--served with gelatinobromide plates.The production of reduced silver is direct when an ordinary silver.compound is converted into metal without formation of a sub-salt,and indirect when the silver compoufid is first reduced to a, sub-salt,and the latter is afterwards reduced to the metal.It would seemthat only under the latter conditions is there any formation of allo-tropic silver. If, in any of the three principal methods of producingallotropic silver, the action is interrupted by the addition of hydro-chloric acid, a dark, chestnut-brown or purple-brown mixture of thesubchloride with the photochloride is obtained, and from it beautifulrose-red photochloride can be obtained by treatment with cold dilutenitric acid after complete removal of the hydrochloric acid. Thisresult is only obtained by interrupting the reaction before it is c v - .plete, and if the hydrochloric acid is added after complete reduction,only grey, normal silver is precipitated.Ifi every case examined,silver subchloride is obtained as one of the products when a reactionresulting in the formation of allotropic silver is interrupted by theaddition of hydrochloric acid before reduction is complete.The rich and varied colour of silver sub-salts would seem to indi-cate that in these compounds the metal existls in an allotropic form,but, on the other hand, the greater activity of allotropic silver and itslower sp. gr. would tend to show that the allotropic form has asimpler molecular structure than the normal metal.Colloidal Silver. By E. A. SCHNEIDER (Ber., 24, 3370-3373).-Collo'idal silver prepared according to Carey Lea's method, by thereduction of silver nitrate with ferrous citrate, cannot be purified bydialysis alone ; the better plan is to separate the solid colloid fromthe mother liquor as completely as possible by filtration, then todissolve the silver in a little water and allow this solution to dialyse.On adding hydrochloric acid to aqueous collojdal silver solutions,silder and argentic chloride are precipitated, the proportion of thelatter being greater with increasing quantities of hydrochloric acid ;the mixed precipitate is extracted with ammonia, the argentic chloride*C.H. B118 ABSTRAOTS OF OEIEMIOAL PAPERS.precipitated by acidification with nitric acid, and hydrochloric acidadded t o the filtrate ; if the quantity of hydrochloric acid originallyemployed was small, a further precipitate is produced, showing thopresence of a silver snbchloride.The mixed precipitate of silver andargentic chloride was always rose-red.Nitric and sulphuric acids react with colloydal silver solutioiis in asimilar manner. No evolution of hydrogen could be detected evenwhen sutlicient hydrochloric acid was added to the silver solution toproduce a considerable precipitate of argentic - chloride ; this mayindicate the presence of nrgentous oxide : neither was oxygen evolvedwhen, in consequence of the small quantity of hydrochloric acid em-ployed, t.he precipitat,e consisted of almost pure silver.J. B. T.Direct Combination of Chlorine and Bromine with Metals,By H. GAUTIER and G. CHARPY (Compt. rend., 113, 597-600).-Wellcleaned wires of various metals, 2 mm.in diameter, were kept in thedark in contact with dry bromine for a definite length of time at 15"and 100". The percentage loss of weight in each case is given in thefollowing table :-At 15' in At 15' in At 100' in8 days. 4 months. 8 days.Magnesium ...... 0.0 0.0 0.19Copper .......... 0.371 1- 740 6.62Zinc ............ 0.289 0.48 7 0.63Iron ............ 0.210 0.440 23.27Silver. .......... 0.003 0.540 -Aluminium under similar conditions combines energetically withbromine and becomes incandescent, a burning fragment runningabout on the surface of the bromine like potassium on water. Withliquid chlorine in sealed tubes at the ordinary temperature, the resultsare similar, the percentage losses being : magnesium, 0.0 ; zinc, 0.0 ;iron, 0.740 ; copper, 3.241 ; silver, 0.673.Potassium, sodium, andaluminium seem to be unaffected by liquid chlorine at its boilingpoint, but a t -20" aluminium combines with the halogen withincandescence.Magnesiumand aluminium, when placed in bromine-water, produce a regularevolution of hydrogen, and, after some time, an oxybromide separates,the reactions being analogous to the decomposition of water by iodinein presence of aluminium. With zinc, iron, and copper, there is noevolution of gas, and a wire 2 mm. in diameter and 50 mm. in lengthdisappears in seven t o eight hours in excess of bromine-water. Itseems probable that in these cases the decomposition of water by thehalogen is accelerated by the presence of' the metal, the latter beingconverted into an oxide which is attacked by the hydracid formed.C.H. B.I n presence of water, the results are very different.Lithium Copper Chloride. By A. CHASSEVANT (Compt. rend.,113, 646-648) .-When a concentrated solution of lithium chloride iINORGANIO OHEMISTRY. 119added to 8 cancentrated solution of an equivalent quantity of cupricchloride, a magma of crystals of the latter salt is formed, but if themidurc is evaporated in a vacuum over phosphoric acid or on awater-bath at loo", the ci-ystals dissolve, the liquid acqnires a,brownish-red colour and deposits transparent, gzrnet-red crystals ofthe double chloride 2CuClZ,2LiC1,5H,O. When exposed to the air, theydecompose, and become opaque, crystals of cupric chloride separating,whilst the lithium chloride deliquesces.If heated rapidly to 130",they melt in their water of crystallisation arid form a deep-brown,almost black, solution. A t tt higher temperature, the salt, likelithium chloride under similar conditions, decomposes and evolveschlorine. When heated slowly i n an oven, or in a current of dry ail-,the crystals become anhydrous a t 100-120", but some hydrochloricacid is likewise given off, and, on treatment with water, a residue ofcupric oxychloride is left. The double salt can be obtained as ananhydrous, chamois-coloured powder by heating it a t 120" in a currentof dry air mixed with dry hydrogen chloride. It is decomposed bywater, but can be recrystallised from a concentrated solution of lithiumchloride. C.H. B.Formation of Saline Hydrates at High Temperatures. ByG. ROUSSEAU (Conzpt. Tend., 113, G43--648).---When the hydratedsodium ferrite obtained a t 800", and previously described, is allowedt o remain in contact with glycerol for several days, and is washedfirst with this liquid and afterwards with absolute alcohol, the driedresidue contains only 9-68 per cent. of water, instead of the 14.5 percent, that it contains when water is used for washing. The ferrite,14Fe,0,,13H20,Naz0, when treated with glycerol in a similar manner,loses 2.79 per cent. of water. If either of these compounds is heatedat 100" with glycerol, its colour rapidly becomes darker, and if diges-tion is prolonged, the whole of the water and alkali is removed, andanhydrous ferric oxide remains.The sodium manganites, such as 12Mn02,4H,0,Na,0, are notaffected i n a similar way by glycerol.The author considers that these results establish.his previous con-clusions and support the view that the alkaline oxide replaces part ofthe water. C. H, B.Crystallised Ferric Oxychlorides. By G. ROCSSEAU (Compt.rend., 113, 542-5444 .-Very concentrated solutions of ferric chloride,containing ahout 80 per cent. of Fe2Cls, maintained at a temperatureof lW--220" for some time, yield a crystallised feryic oxychloride,2Fe,03,Fe2Cl, + 3H,O. By prolonged boiling with water, thiscompound is gradually converted into a ferric hydroxide, (Pe,O, +H20)3, which retains the same crystalline form as the oxgchloride.Solutions containing 85 to 911 per cent.of the anhydrous ferricchloride have been heated in sealed tnbes, together with a fragmentof marble or giobertite. Between 225" and 280", red-brown lamella:of the oxychloride 2Fe,03,1?e2C16 are obtained. Between 300" and340°, large plates of a brownish-black oxychloride 3Fe2O3,Fe2Cl6 areformed120 ABSTRAOTS OF CHEMICAL PAPERS.The author has been unable to study the reaction at higher tem-peratures, but believes that a, series of oxychlorides of the type(Fe,O3),,Fe,CI6 would be formed, in which the proportion of EeZO,would increase with the temperature.The anhydrous oxychlorides are very sparingly soluble in dilutemineral acids. When boiled with water i n presence of marble for150-200 hours, they lose all their chlorine as hydrochloric acid,ferric oxide of a fine brownish-red colour remaining.The optical properties of these oxychlorides have been determinedby Fouqu6.They occur in prisms giving longitudinal extinction.The plans of their optic axes is transverse, and the bisectrix ispositive.It would be interesting to observe whether the rhombic oxychlor-ides retain their form during the concentration of ferric oxide in themolecule, or whether they assume, at a temperature near a red heat,the hexagonal form characteristic of ferric oxide. On the latterhypothesis, the hydrolysis of the hexagonal oxychloride, by boilingwith water, should give a new method for the synthesis of hzmatite,allowing the determination of the degree of polymerisation of thismineral oxide.W. T.Action of Water on Glass. By E. PFEWFER (Ann. Phys. Chein.[ Z ] , 44, 239-264) .--The author has taken advantage of electrolyticconductivity for the purpose of determining the amount of substaiicedissolved from glass by water at low temperatures. Ordinarychemical methods give very uncertain numerical results, on accountof the difficixltg in determining the exact magnitude of the largesurface which must be exposed to the action of water, and even theseresults are obtained ander conditions diverging considerably fromthose of laboratory practice.Water first of all dissolves practically pure alkali (potassium orsodium hydroxide) out of the glass, and this afterwards exerts itsown influence by dissolving silica.The author estimates the amountof alkali dissolved by determining the electrical conductivity of thesolution, to which it is proportional. As the molecular conductivitiesof potassium hydroxide and sodium hydroxide lie close together(220 x 10-7 and 200 x according to Kohlrauschj, no greaterror is comrnitted in estimating the total amount dissolved on theassumption that each alkali dissolves proportionately to the extent towhich i t is contained in the glass. When silicates are formed in thesolution, the conductivity falls.The experiments were made by exposing cylinders of goodThuringian glass with known surface to the action of water containedin glazed porcelain vessels; the temperature in the three series ofobservations made being lo", 30", and 30".For one sort of glass ata given temperature, it was found that A0 = A w / o is a constant, Abeing the increase in conductivity per hour, w the volume of water,and o tbe surface of the cylinder. When the glass has been exposedfor some time to the action of water at a temperature of 60", thevalue of A,, for 2G" falls considerably. Aha("") for the first day is mucINORGANIC CEEMISTRY. 122greater than A,-,(2o) afterwards. With the specimen of glass examined,it was found that a t 20" one to two millionths of a milligram weredissolved out by 1 C.C. of water per square centimeter in an hour.No silica is dissolved a t 10" or 20". Ati 30°, however, a considerablefalling off of A. with the time is observed, due in all probability tothis cause.The values of AL, (reduced to go*) for the various tern-peratures are as follows :-10". 20". 30".A. .... .. .. 25 100 673Prolonged treakment of the glass at a low temperature does n o tappreciably affect its solubility a t a higher temperature. J. W.Stannibromides. By LETEUR (Compt. rend., 113, 540-542).-The stannibromides of the alkali metals and magnesium are yellow,well-crystallised substances. Concentrated solutions have the samecolour, but a t a certain state of dilution, the colour disappears. Theanhydrous stnnnibromides of potassium and ammonium only sufferchange in very moist a i r ; others are very deliquescent. Concen-trated solutions may be heated without decomposition ; on dilution,decomposition occurs with the formation of hydrobromic acid and thedeposition of hydrated tin dioxide.Alcohol decomposes these corn-pounds, slowly in the cold, more rapidly on heating ; benzene has noaction on them.The general method for the preparation of the stannibromidesconsists in mixing concentrated solutions of the two bromides andevaporating the mixture in R vacuuni or in dry air.The ammonium salt, (NH4)2SnBr6, forms sulphur-yellow ocka-hedra belonging to the cubic system; it decrepitates when heated,and volatilises with partial decomposition.The sodium salt, Na,SnBr, + 6H20, for.ms yellow prisms of themonoclinic system, having a position of extinction in polarised lightat lFjo to the longer axis. It is very deliquescent, but efflorescesrapidly over sulphuric acid OY in a vacuum.It decomposes, whenheated, with evolntion of water and stannic bromide.The lithium salt is probably Li,SnBr, + 6H20, but the watercannot be accurately determined owing to the extreme deliquescenceof the compound. It forms small, yellow, prismatic needles whicliact on polarised light, and appear to belong to the monoclinic system.Over sulphnric acid, these crystals effloresce. giving a citron-yellow,crystalline powder tending towards the composition Li2SnRr, +5H,O.The magnesium salt, MgSnBr, + 10H20, gives small, sulphnr-yellow, monoclinic crystals. The ordinary form is a prism showingthe faces q1 and h, with modifications on the angle a. Macles arefrequent.The study of the alkaline earthy stannibromides is being nolvcarried on. W. T.VOL. LX1I. 72Extinction occurs a t an angle of 60"122 ABSTRACTS OF CIHEMICAL PAPERS.Dissolution of Bismuth Chloride in a Saturated Solution ofSodium Chloride : Basic Bismuth Salicylate. By H. CAUSSE(Compt. rend., 113, 547-549) .-Sodium chloride, like ammoniiiinchloride, may be employed instead of free acid to prevent the dis-sociation of bismuth salts by water. Hence, in presence of sodiumchloride, hydrochloric acid may be completely neutralised by bismuthcarbonate or oxide. 100 C.C. of hydrochloric acid solution containing3.0775 grams HC1 is left in contact with 3 grams of bismuth oxideuntil no further solution takes place ; the remaining oxide is collectedand weighed. 1.50 grams of the oxide are dissolved, requiring 0.4775gram of the acid t o form BiCI,; the remaining 2.60 grams of thehydrochloric acid are required to maintain the equilibrium in thesolution.With 100 C.C. of acid containing 6.155 grams HCl, 6.00 grams ofoxide are dissolved, and 3.117 grams of free acid remain. With100 C.C. of acid containing 9.2325 grams HCI, 10 grams of oxide aredissolved, and 4.557 grams of acid remain. Each of these solutionsis saturated with common salt, and then treated with bismuth oxide.The quantities of bismuth oxide dissolved as compared with thequantities required to neutralise the hydrochloric acid present withproduction of bismuth trichloridc are respectively 6.80 : 6.584,13.25 : 13.160, and 20.25 : 19.70.The numbers given above do not show that any definite relationexists between the free acid and the amount of bismuth chlorideformed. To ascertain whether such a relation exists at the experi-mental limit, the author treats 50 grams of oxide with 50 C.C.of saturated hydrochloric acid containing 22.80 grams of acid.47.50 grams of oxide are dissolved, and 5.18 grams of free acidremain. 5.40 grams of acid would be uncombined if the reactionwere as follows: Bi,O, + 8HC1 = S(BiCl,,HCl) -l- 3H?O ; helice tReauthor concludes that under these circumstances a definite salt isformed.To obtain basic bismuth salicylate, 40 C.C. of conccritratedhydrochloric acid is saturated with bismuth oxide in presence of500 C.C. of saturated sodium chloride solution ; to another SO0 C.C. ofbrine are added 9 grams of soda and 22 grams of sodium salicylate,the two solutions are mixed, and the precipitate formed is washedwith water containing a few drops of nitric acid. The basic sali-cylate, C,H,BiO4,IIZO, obtained, forms microscopic prisms, and hasproperties similar to those of the normal salicylate previouslydescribed. It is decomposed by heat with loss of the whole of itssalicylic acid, which may also be completely eliminated by boilingconcentrated alcohol. The constitution of t h i s salt may be representedby the formula OH*CsH~*COO*Bi(OH)z, which accounts for its readyhydrolysis. W. T