IN0 RGXNIC CHE M ISTR T. It I n o r g a n i c C h e mi s t r y. Revised Hydrochloric Acid Tables. By G. LUSGE and L. MARCHLEWSKI ( Z e i t . any. Chem., 1891, 133-1 35).--The author9 have cortstructed a very useful table, giving the specitic gravities (reduced by Kohlrausch's formuln to vacuum and 4" C.), the corresponding degrees of Baume's and l'waddell's hydrometers, the percentage of hydrogen chloride by w-eight, the percentage of weight of 18", 19"' 20°, 21", and 22" acid, and the number of kilograms of the said acids contained in 1 litre. The specific gravities were taken with the utmost care in a specific gravity bottle, provided with art accurate thermometer, and are guaranteed to be accurate witliiu 0.0001. The acid was analgsed by titvation with N/5 solution of sodium hydroxide, which had been staiidardised with N/5 hydrochloric acid. This acid was standardised with pure sodiu tit carbonate, and the results were verified by a gi*nvimetrie estimation as silver chloride.The titrations were done with the greatest care, sud the experimentd error never exceedod 0.02 per cent. with the weaker acids, arid 0.05 per cent. with the stronger ones. The weaker acids were weighed iu a Winkler's stop-cock pipette: the fiiruing ones in sealed glass bulbs. The Place of Fluorine in the Classification of the Elements. By H. h~OlSYhN (Bull. Soc. C h h . [ 3 ] , 5, 8~0--885).-Whilst fluorine in many respects behaves as the most energetic member of the chlorine family, in some ways it is markedly distinguished from chlorine, and seems almost to bear a remote analogy to oxygen ; as instances in point are cited, the ready combustibility of charcoal in flnorine, the dissimilarity of calcium flwoi-ide from the chloride, and its L.D E K.r2 ABSTRACTS OF CHEMICAL PAPERS. resemblance to the oxide, the solubility of silver fluoride in water, and the stability of aluminium fluoride towards water. It is pointed out also that,, whilst the cornpounds of fluorine with the non-metallic elements, including its organic compounds, are uuiformly inore volatile than the corresponding compounds of chlorine, its metallic salts ~equire, a s a rille, a higher temperature for fusion than tlie corrc- spoilcling chlorides. Comparative numerical data are given. Js. W. Action of Fluorine on Phosi7horuS Trifluoride. By H.MOISSAN (Bull. Soc. C h k . [3], 5, 880).-In order to stucly the action or fluorine on other gaseous substances, the two gases were admitted by separate tubes to the central poi.tioii of a platinum tube 15 cm. long, the ends of which were closed by transparent plates of fluor- spar, and the resulting gaseous product was conducted to tlie water or mercury troagh by a third taube near the end. When fluorine comes in contact with phosphorus trifluoride, a yellow tlame of low temperature is seen, and the product of the reaction is mainly phosphorus pentafluoride, the residue consisting of unchanged tri- fluoride. JN. W. Persulphates. By BJEHTIIELOT (Conzpt. rend., 112, I481 -1483 ; see also Berthelot, Gompt. rend., 86, ‘LO, 71, and 277; Marshall, Trans., 1891, 771; MendelAeE, HdZ.Sot; Chirrt., 38, 168; Traube, Her., 22, 1518, and 24, 1764).-The mixture of peraulphuric and sulpliuric acids obtained by electr.ilgsis, when neutralised by baryta- water or potash solution, yields a quantity of the normal persulphate, which is greater when the operation is carried on at a very low temperature. The neutralised solution rapidly becomes acid again, 11m-ticularly on heating; a t the same time losing oxygen. l’he quantity of active oxygen is proportional to the quaiititmy of sulphuric acid formed in the decomposition. Barium persulphate is soluble, and as neutral as the thiosulphate or permanganat’e ; its decomposition proceeds accordiug t o the equation BaS,08 + H,O = BaS04 + II,SO, + 0. In t,iti.ating the active oxygen by the usual reagents,it is necessary to take into account the hydrogen peroxide often coexistent with pei*sulphuric acid, or capable of being formed during the dilutions and other operations. The proportion of hjdiuogen peroxide present has been found to vary from 0 to 2 mols.for each mol. of persuiph- uric acid. Viewing the whole of the active oxygen present in the extreme case as united to the sulphur, the complex acid would have the formula S,O,,nH,O. There is no proof of the existence of a lieu tral and anhjdrous compound, such as the holoxide, SO4, of Trauba. The salts of persulphuric acid are distinct and definite, and com- parable with the permanganates, perchlorates, permolybdates, a i d Thiosulphates. By A. FOCK and I(. KrXss (Ber., 24,3016-3017). --Potassiwn calcium tliiosidphnte, 3K,S20,,CaS20, + 5H20, is obtained in colourless, monosymmetric crystals, by evaporating on the water- bat8h a mixture of potassium and calcium thiosulpbate solutions. pertungstates.NT* 1’.INORGANIC CHEMISTRY. 13 The axial relation of the crptals is a : 5 : c = 1.7610 : 1 : 0.89.31 ; /j = 80" 2'. Potassium sfmitiurn thiosulpnhate, K2S2O3,Si S?Os + 5H20, is pre- pared in a similar manner to the calcium salt ; it is readily soluble in water, and crystallises in slender, lustrous, silky needles, which are not, well adapted for crystallogi*aphic measurements. J. B. T. Ammonium Dithionate Hydrochloride. By A. FOCK and K. Kr,iiss (Ber., 24, 3017--.3018).-0n mixing solutions of ammonium dithiouate and ammonium chloride in molccular proportions, and evaporating, a compound of the formula (NH,)2S,06,H C1, is deposited in rhombic crystals.The axial relation is a : b : c = 0.09827 : 1 : 0.9612. The same substance is also formed in presence of two or more mole- cular proportions of ammonium chloride. J. B. T. Revised Nitric Acid Tables. By G. LUNGE and H. REY (Zeit. any. Cham., 1891, 165-1 70).-The tables are coiistructed on similar lines to the rerised hydrochloric acid tables (compare this POI., p. l l ) , m d similar processes were employed in their construction, but instead of weighing the acid in glass bulbs, or in a Winkler's stop- Q cwk pipette, a new form of weighing pipette, shown in the figure, RW used.14 ABSTRACTS OF CHEMICAL PAPERS. Above the stop-cock a is fixed a bulb b, of about 2 cm.diameter, connected with a second stop-cock c. The lower part of the pipet'te fit,s into a glass tube d, which has n ground neck. When the pipette is wanted for use, the stop-cock a is closed ; suction with the mouth is applied to the top, and the stop-cock c immediately closed, by which nieans R diminished pressure will be obtained inside the bulb / I . The lviint of the pipette is now dipped into the acid, the stop-cock a opened, and the acid allowed t o ascend a s far as the stop-cock, which is then iirimediately turned off. After wiping the pipette, i t is fixed into the tube d and weighed. The pipette is now tnrned round, PO as to open communication between the vhannels e and f . The stop-cock c is opened, water is squirted into the bulb b, and then allowed to flow rhrough cr, into the pipette.The diluted acid runs into the tube d , the air of which escapes through e and f. The acid is finaily emptied into a beaker and titrated. By using this apparatus, any loss of iiitric fumes is avoided. L. r)E K. Compounds of Sulphur and Phosphorus. By J. MAI (Annalcn, 265, 19.L-208).-Th object of these experiments was to ascertain wIiet.her the boiling point of tripbosphorus hexasnlphide, P3S6, was sufficient.ly below that. of stannous chloride (SOSO), and above that of pliosphorus peiitasulpliide (51 So), to allow of its advantageous tdmploymeut as a heating vapour in pyrochemic,zl investigations. When a n intimate rriixtnre of sulphur (2 atoms) and amorphous ~)hcsphorus (1 atom) is cmef rilly heated in an atmosphere of carbonic ;inhydride, and the product theii distilled, a yellow, crystalline sub- stmce (b.p. abont 410°), which seems to contain the cornpound P,&, passes over first; the later fractions consist of a mixture OE tt.iphosphorus hexasulphide and phosphorus pentasulphide, which cannot be separated into its constituents, as they boil at the same temperature (about 508'). Phosphorus pentasulphide boils at 332--340" under a pressure of 10-11 mm. ; triphosphorus hexasulphide, P3Ss, boils a t 335-340" under the sa9e conditions. When the compound of the composition P4S3, prepared from its clements by the usual met,hod, is distilled under a pressure of 11 mm., H yellow liquid passes over between 230" and 240°, and amorphous phosphorus remains i t 1 the flask ; the distillate is almost completely soluble in carbon bisLIphide, from which it is g~adually deposited in ciystals of the composition P4S3.The twiling point of t h i s compound uilder the ordinary pressure would be about 408-418" ; it begins to soften a t 130-135", h u t does not melt completely until the tempers- ture has risen above 160". When phosphorus (2 atoms) is melted with sulphur (3 atoms), and the pyoduct submitted to distillation uiider a pressure of 11 mm., alniost the whole passes over between 285" and 335" ; the distillate slowly solidifies to a yellowish, resinous, pasty mass, which, when digested with carbon bisiilphide under pressure, yields a crystalline compound of the composition PAS, ; the mother liquors from this com- pound contain a crystalline substance, which seems to be a mixture of phosphorus trisclpl~ide with PAS,.I?. s. K.IN0 dGIAS1C CHEMIS rRY. 15 New Method of Preparing Carbon Oxysulphide. 337 J. NuHicsm (Ber., 24, 2967--2974).--Carbon oxysulphide is obtained when cnrbonyl chloride is passed through concentrated sulphuric azid to dry it, and then through a tube 50 cm. long. filled with ignited asbestos well mixed with finely pulverised cadmium eulphide, the tube being placed in a combustion furnace and heated. The author 6nds that even when no external heat; is applied a sin;tll quantity of carbonyl sulphide is formed, but, the most, favouiable teu- yerature for its formation anpears to be 260-28O". The gas thus prodnced was found on analysis to contain COS, 94.87 per cent.; CO, 3.98 per cent.: air, 1.15 per cent.A quantity of crystals, which were identified as cadmiurn chloride, were obsez-ved in a tube previously charged with a layer of cadmium sulp5ide and heated in tt flame during the passage of a current of csarbonrl chloride; the reaction, therefore, appears to be a double decomposition. A. R. L. Allotropic Silver. 111. Blue Silver. By M. C. T ~ E A ( P l d . Uag. [ 5 ] , 31, 497-504 ; compare Abstr., 1891, 803).-When silver nitrate is added to a solution of dextrin made alkaline with sodium 01- potassium hxdroxide, the alkali being kept in moderate excess, the silver is first precipitated as the ordinary brown oxide, but the colour gradually changes to rcddish-chocolate, and the silver begins to di+ solve, and in a few minutes has completely dissolved to form an almost black liquid, a few drops of which.when added to a large quantity of water, give a splendid red liquid, that, spectroscopic clxnmination shows to be a perfect solution. Different specimens of d ~ x t r i n behave differentJy, and the common brown form seems to give the best results. Convenient proportions are 20 grams of sodium Irjdroxidc and 20 grams of dextrin in 1000 C.C. cf water, 14 grams of silver nitrate previously dissolved in a small quantity of water being gradually added. It is interesting to observe that allotropic silver can be formed and can exist in solution in neutrill, acid, and alkaline liquids. The alka- line solution spontaneonsly deposits allotropic silver after some time; and precipitation is immediate on the addition of dilute nitric or sulpll- uric acid.Even with a laige excess of acetic acid, however, precipita- tion is incomplete. When the precipitate has once formed, ~t is almost insolublc, but on washing dissolves to a very small extent, forming a rose-red liquid. A small quantity of sodium phosphate, when added to the alk~.line solution, precipitates the whole of the allotropic silver with a rub?- copper coloui-, which after p~olonged washing changes to deep Nile- #IY~II, and becomes slightly soluble, forming a wine-red solution. The various f o m s of allotropic silver show different body and surface colours, which tend to be complementary. 'The precipitate by sodium phosphate, €or example, when spread thickly on paper, dries with a bright-green, metallic colour, but after it has heen changed to deep green by washing, films dry with a dark gold or cwpper surface-colour. Although the dry substance has a11 the appearance of a metal, it16 ABSTRACTS OF CHEXICAL PAPERS.mRy contain 8 to 10 per cent. OF organic matter that cannot be rcmovetl, even by prolonged washing with hot water under pressure. Four specimens of carefully purified material contained respectively 93.77, 94.27, 92.86, and 96.64 per cent. of metallic silver. When the allo- tropic forms are heated, a vapour is evolved which condenses in small, brownish drops, with an empyreumatic odour. The residue consists of bright white metallic silver, and when dissolved in nitric acid, leaves a residue of black flakes of carbon. When the allotropic silver is dissolved in dilute nitric acid, and the silver precipitated by hydrochloric acid, the filtrate, on evaporation, leaves a small residue of a, yellowish, gummy substance.Tannin reduces silver nitrate to allotropic silver more readily than does clextrin. and gives better results in presence of sodium or potassium carbonate than with the caustic alkalis. 24 grams of anhydrous sodium carbonate are dissolvod in 1200 C.C. of water, and mixed with 72 C.C. of a filtered 4 per cent. solution of tannin. 24 grams of silver nitrate dissolved in a small quantity of water is added gradually, and solution of the silver takes place imme- diately. After standing for a day or two, the intensely dark-coloured liquid may be poured off from a small quantity of black precipitate. On adding a small quantity of dilute acid, the allotropic silver is precipitated, and in films dries with a bluish, steel-grey, surface- colour.Blue allotropic silver (inclnding the green and steel-grey varieties) shows endless variations, and cannot be reduced to one type. Slight dif'ferences in the conditions of formation result in very different products. Of ten products obtained by the action of tannin and sodium carbonate in various proportions, several were easily and completely soluble in ammonia, whilst some dissolved slightly, and others not at all. Some of the specimens, insoluble in water, when treated with phosphoric acid, did not dissolve, but on washing away the acid, were found to have become soluble in water ; other insoluble spacimens did not behave in this way.Some of the solutions were scarcely affected by acetic acid, whilst others were partially, and others almost completely, precipitated. The films OIL paper var.y greatly in their sensitiveness to light, an3 in the ease with which they are converted into the yellow, inter- mediate form. The least sensitive modification is that precipitated by nitric acid ; it dries with a steel-grey colonr. The modifications precipitated by acetic acid tend to have a greenish metallic surface, and are more sensitive. Permanency varies greatly in different speci- mens, and is increased by thorough washing. The blue, grey, and green forms are related to the black, or dark- grey normal silver, and tend to pass into it, whilst the gold-coloured modification tends to change into bright white metallic silver on t h e surface, with dark, or even black, silver underneath.Tannin likewise reduces silver nitrate in prescnce of lithium, ammonium, magnesium, barium, calcium, and strontium carbonates. The product with strontium carbonate is dark-red whilst moist, but dries with a rich, bluish-green, metallic surface-colour in thick films, but is red and transparent in thin films.INORGANIC CHEJILSTRT. 17 The intermediate golden-yellow form produced during the passage of the pold-coloured allotropic form to white normal silver has none of the properties of the original form, except its colour and Instre. It is hard and tough, is not converted into normal silver by friction and high pessure, and offers as much resistance as normal silver to the action of oxidising and chlorinating agents.The difference between the original form and the int,ermediate form lies in the fact that the latter. has a crystallinc structure. which becomes evident on treating the film with ferric chloride solution. When the blue form is gently heated, it also first becomes yellow, and then changes to white normal silver. A film on glass begins&o change from blue to yellow a t about 180", and the same change IS produced by the action of light, some specimens reqniriiig a few hours' exposure to sunlight, whilht others require several days. The author considers that in these allotropic forms the silver exists in a state of atomic divihion, and that this is true also of the pyro- phoric forms of various metals. He points out that there is a differ- ence between cheniical and mechanical division, and that a metal may form N compact mass, and yet be in a state of atomic division, whilst, on the other hand i t may bein a fine state of inechatlical division, and yet possess a high degree of molecular complexity Characteristics of the Alkaline Earths.By G. BHCGELMANP.- (Zsit. anal. Chem., 30, 579--580).-1n his attempts to prepare baryta by igniting the hydroxide in graphite crucibles, the author has been unable to obtain a pure product, the crucible being always attacked. The supposition oE the dimorphism of barium oxide, and the assumed catalytic action of platinum (Abstr., 1890, SSO), must theretorc be Rbandoned, since the substance obtained as a felted mass of needles showing chromatic polarisation c m no longer be regarded as baryta.Since a similar action occurs with lime and strontia, all statements respecting the specitic gra\ ity, &c., of the alkaline earths prepared in other than platinum vessels, must be withdrawn. 'Ylie statements of Fresenius and others, that strontium carbonate fuses while decomposing, cannot be confirmed. Much contraction occurs, but the oxide produced retains the form of the original mass OF car )on- ate, and is, a t most, slightly sintered. Composition of a Boiler Incrustntion. By A. CaR1s.r (Zeit. any. Chem., 1891, 77).-The author communicates an analysis of a curious boiler deposit, which owed its origin partly t o the use of an animal o r a vegetable lubricating oil. CaO, 11-09; MgO? 9.79; Fe,O?,, 5 60 ; A1203, 1.10 ; PbO, 0.98 ; CuO, trace ; SO2, 16.00 ; SO3, 1.71 ; fatty acids, 22.62 ; neutral fats, 25.84 ; moisture, 2.69 ; combined C.H. B. M. J. $, water and organic matter, 5-22 per cent. L. DE K. Action of Hydrogen Peroxide and of Water saturated with Carbonic Anhydride on Magnesium. By G. G~ORGIS (Gazzetta, 21, 510-514) .-According t o Weltzien (dnnalen, 138, 1:32), hydrogen peroxide acts slowly on magnesium, forming an alkaline liquid, which contains the normal hydroxide, Mg(OH),, and on evaporation to dry- VOL. LXlt. C18 ABSTRACTS OF CHEMICAL PAPERS. ness leaves a white, strongly alkaline mass, completely sola\de in waiLr. The author has also observed the slow action of IiTdrogeu peroxide 0 1 1 maqnesiiim and the alkalinity of t.he solution, but fiiidv that, after.remaining for a few days, acicular crystals arc! deposited, which efiervesce on treatment with hydrochloric acid. On treating metallic magnesium in a vacuum w i t h hjdrogen peroxide free from carbonic aiihyclride, the liquid after a time becomes slightly alkaline, and on evaporating to dr-yness in a vacuum, a flocculent precipitate mparates, which is only very sparingly soluble in water free from carbonic anhydride. On the other hand, when magnesium is treated with distilled water saturated with pure carbonic anhydride, the liquid soon becomes strongly alkaline, and the m t k l is energetically attacked with evolution of hydrogen ; the reaction slackens very gradually, and gas ceases to come off after about 10 or 12 hours. The solutioii, after :L time, deposits acicular crystals havirig the composition AJgCO, -t 3H,O.The evolution of hydrogen by the actioii of an aqueous solution of cmbonic anhydride on magnesium is noteworthy ; the reaction also Iwovides nu easy method for the preparation of the normal carboilate of magnesium. Cuprammonium Oxide. By PHuD'HuJi \I (Chem. Centr., 1891, ii, 339-340 ; from Mnn. Sci. [4], 5, ~jt31).--C'upramtrionium oxide is i t more powerful oxidking agent than hydrogen peroxide; it acts i w r e rapidly on cellulose, aud converts it into oxycellulose ; it also decoloriscs indigo-blue mow rapidly than hydrogen peroxide, With " mercerised " cotton wool, i t acis like Ilydrogeo peroxide. I n preparing it, by agitating copper turiiiiip with ammonia and air, both cupric oxide atid nitrous an hycliide are forlt,ed : the Ditrous anhydride is, however, only formed a t a later st,iige of the reaction.Ammonium nitrite oxidises copper, atid nitric peroxide is formed in the tirst instalice; ammonia is also liberated by the reaction, and when this occurs, free nitrogen is formed instead of nitric peroxide, The liquid becomes intensely blue ; an excess of water precipitates cupric hydroxide ; dilute acetic acid ptecipitates cupric oxide as soon S. B. A. A. as the ammonia is exactly neutrdisecl. J. W. 1,. Action of Ferric Chloride on Metallic Sulphides. By CAMMEREK (Chern. Centr., 1891, ii, 370; from Berg. Hutlen Zeit., 50, 2cil-- 664, 282-28-4.).-1n a sealed tube, ferric chloride reacts with both cupric and cuprous sulphides, cupric chloride, ferrous chloride, and mlphur being formed.With pi-ecipii ated ferrous sulphide OF iron pyrites, ferrous chloride and sulphur are the products ; with copper i)yrites, cupric chlo~ide, ferrous chloride, and sulphur are formed. With arsenious sulphide, ferrous chloride, sulphur, and nraeriiouu chloride are tirst formed ; the latter, howevei., is furtlier oxiclised to arsenic anhydride by the continued action of t-xcess of ferric chloride. With eitliey precipitated or native antimony trisulphide, antinronious iAloride, ferrous cliloride, and sulphur are formed. With stannous -;ulphide, staiiiiic chloride, ferrous cliloI*ide, aid sulphur are formed. J. IV, 1 1 .IXL?ROXNlC CHEMISTRY. 19 Manganese Tetraahloride. By H. M. VERNON ( P l i i l . LMCLg. [ 5 ] , 31, 469-484).--Pisher (Trans., 1878, 409) endeavoured to show thab t,he dark-brown liquid obtained on dissolving manganese dioxide in hydrochloric acid contains manganese tetrachloride.Picker-ing, who repeated Fisher's experiments, aiid made, in addition, a number of his own (Tram., 1879, 6.54), thought that the evidence was in fayour of the existenre of Mn,Cl, in the solution, and not of MnCII. The author now attempts to decide between these views by experi- ments on the rate a t which the " available " cliloi*iiie is evolved at different, temperatures. If Mn2C16 is really formed in the solution, half the available chlorine should be much rno1.e easily removed by means of a current of air passed through the solution than the other half, whereas. if MnCI, is present, the rate of removal should be more ~egular.The curvt s obtained from the experimental results at various temperaturt-s sliow no break a t the point where half the available chlorine has been driven out, so the author concludes that when MrisOr, Mn,03, OY MnO, is dissolved in hydrochloric acid, the onlF higher chloride formed is MnC14. The brown solution is much more stable at -26" than at ordiiiary temperatures. Pickeriug found that the presenceoF MnC1, in the solutioii from the beginning rendered it more stable. This the author accounts for on the supposition that the MnC14 dissociates directly into illnC1, and CI,, so that the presence in excess of MnCl,, one of the products of dissociation, would diminish the actual amount of disso iation. Calorimetric Researches on the Condition of Silicon and of Aluminium in Cast Iron.By F. OSMOND (Cornpt. ~d., 113, 474 -476).-The heat developed on dissolving 1 gram of cast ii.or.i con- taining different amounts of :;ilicon in a saturated solution of the double chloride of copper and ammoniuni was measured. From the results, i t appears that when the silicon is present in sufficient quan- tity,. it conrbines with the iron with the development of heat and formation of a cotupound which is dissociated by an excess of iron, and therefore only exists wlieii the amount of silicon i n the alloy is sufficiently great. Hence the difference between the quantities of heat found and those calcolated in the above experiments changes sign when the amount of silicon present reaches a percentage some- where betwecn 4.1 and 7.3.Samples of c,ast iron containing aluminium were also examined in the above manner. The results show that alumiiiiuni dissolves ill cast iron with the absorptioii of heat. Attempts to Prepare Metallic Chromium from Chromic Fluoride. By W. P. EVASS (Zeit. any. C h e w , 1891, 18-2Oj.-Tht. mthor tried the effect ot' nwtallic sodium, metallic zinc, aiid mixed carbon and silica on chromic fluoride a t a very high temperature, in the hope of getting metallic chromium. 1. Action of Sodium.-The metal was heated a t 400' in a porce- lain tube, and the fluoride passed over it. The latter was prepttrzd by distilling a mixture of lead chwnictte, calcium fluoride, and snlph- uric acid. A very &ong reactiun-tcok place and the bube becarns i .2 J. W. H. C.20 ABiTRACTS OF CHEJllCAL PAPERS. red-hot, but was soon stopped up. After cooling, the excess of sodium was dissolved out by water, and the residue examined for metallic chromium. But very little was found, and this was chiefly in combination with silicon. The experiment was repeated in an iron apparatus, in which tile va1)our of about 70 grams of metallic sodium was brought in contact, at R temperature of 900-1000°, with the fluoride. The chief products obtained were amorphous, green chromic oxide, bright-green particles of sodium-chromic fluoride, and a grey- i S l l , Lritrtle, spongy mass, which was chiefly located near the fluoride clelivery tube. On analysis, i t proved to be an alloy of sodium witli %bout 25 per cent. of chromium intermixed with a little chromic oxide and a trace of iron.The experiment was repeated, and, t h i s time, to avoid contamination with iron, in a Hessian crucible. The crucible and the delivery tubes were well lined inside with a mixture of 6 parts of alumina and 1 part of potassium chloride. After being. kept red-hot for balf an hour, the experiment came to ;i prc- iilttture end through the tube gettiag stopped up. After cooling, there was found adhering to the tube a greyish, stalactitic mixss, which foy the greater part dissolved in hot water. The insoluble por- tion consisted of fairly pure, crystalline, metallic chromiurn. 2. Action of Zinc.-About 2010 griims of zinc was heated to boiling under a layer of salt, and the fluoride M'HS passed in through a porcelain delivery tube.'l'he greater part of the fluoride was taken up by the salt with formation of sodiiini chromate. The zinc: regulus dissolved slowly in nitric acid, which gradually acquired a greenish tinge. No chromium could be detected in the insoluble residue. There was, however, good reason to believe that. an alloy of zinc with about 0*67-1*60 per cent. of chromium had been formed. In the porcelain tube, however, a lustrous, steel-like substance wls found, which on analysis proved to be practically pure metalli chromiii m. 3. Action of Carbon and Silica -The fluoride was passed through a strongly heated porcelain tube containing small granules of a mix- ture of 9 parts ot' amorphous silica and 4 parts of powdered char- coal. Abundant vapours of silicon fluoride made their escape, and on openiiig the tube, sharp, shining, hexagonal crystals of chromic oxide were foiind; besides this, an almoht black, hrittle mass, wllich on analysis yielded Cr, 30.13 ; SiO:, 48.53 ; C, 11-39; 0, 9-95, COU- taining, therefore, 8.39 per cent.of metallic chromium. From these experiments, it seems that although there is not the slightest doubt atbout the reduction of the chromic fluoridn by the piocesses described, it is not possible to prepare the metal 011 the large scale in t h i s way. Pure Bismuth. By A. CLASSEN (J. p r . Chem. [2], 44, 411- 514 ; compare Abstr., 1891, 271, 525,1324).--This is another chapter in the controversy between Schneider and Classen as to pure bismuth. Against Schneider's aualyses of commercially pure bismuth (Abatr., 1891,1524), the author puts in two analyses, the one of '' chemically llnre bismuth for scientific investigations," which gave sonle 10 grarll,s lead chloride from 500 grams of the metal, the other of '' bismuth L.DE K.,\.XISERALOGICAL CHENISTRT. 21 pnrissimnm,” from the same factory, which contained, besides lead (ufidetermined), 1-56 per cent. of copper and 0.45 per cent. o.f iron. The nnthor’s electroljtic bismuth melts at 264” ; “bismuth puriss.” from Tromsdorff melted at 272.8’. and another sample at 273”; ‘‘ absolutely pure bismuth ” from Tromsdorff melted a t 265-266” ; ‘‘ bismuth pnriss.” from Schering melted a t 269-270”. This is further evidence of the impurity of commercial “ pure b.ismuth.” The rest of the paper deals with Marignac’s method and material, and is purely polemical.A. G. B.IN0 RGXNIC CHE M ISTR T. ItI n o r g a n i c C h e mi s t r y.Revised Hydrochloric Acid Tables. By G. LUSGE and L.MARCHLEWSKI ( Z e i t . any. Chem., 1891, 133-1 35).--The author9 havecortstructed a very useful table, giving the specitic gravities (reducedby Kohlrausch's formuln to vacuum and 4" C.), the correspondingdegrees of Baume's and l'waddell's hydrometers, the percentage ofhydrogen chloride by w-eight, the percentage of weight of 18", 19"'20°, 21", and 22" acid, and the number of kilograms of the said acidscontained in 1 litre. The specific gravities were taken with theutmost care in a specific gravity bottle, provided with art accuratethermometer, and are guaranteed to be accurate witliiu 0.0001.Theacid was analgsed by titvation with N/5 solution of sodium hydroxide,which had been staiidardised with N/5 hydrochloric acid. This acidwas standardised with pure sodiu tit carbonate, and the results wereverified by a gi*nvimetrie estimation as silver chloride. The titrationswere done with the greatest care, sud the experimentd error neverexceedod 0.02 per cent. with the weaker acids, arid 0.05 per cent. withthe stronger ones. The weaker acids were weighed iu a Winkler'sstop-cock pipette: the fiiruing ones in sealed glass bulbs.The Place of Fluorine in the Classification of the Elements.By H. h~OlSYhN (Bull. Soc. C h h . [ 3 ] , 5, 8~0--885).-Whilst fluorinein many respects behaves as the most energetic member of thechlorine family, in some ways it is markedly distinguished fromchlorine, and seems almost to bear a remote analogy to oxygen ; asinstances in point are cited, the ready combustibility of charcoal inflnorine, the dissimilarity of calcium flwoi-ide from the chloride, and itsL.D E Kr2 ABSTRACTS OF CHEMICAL PAPERS.resemblance to the oxide, the solubility of silver fluoride in water, andthe stability of aluminium fluoride towards water. It is pointedout also that,, whilst the cornpounds of fluorine with the non-metallicelements, including its organic compounds, are uuiformly inore volatilethan the corresponding compounds of chlorine, its metallic salts~equire, a s a rille, a higher temperature for fusion than tlie corrc-spoilcling chlorides.Comparative numerical data are given.Js. W.Action of Fluorine on Phosi7horuS Trifluoride. By H.MOISSAN (Bull. Soc. C h k . [3], 5, 880).-In order to stucly the actionor fluorine on other gaseous substances, the two gases were admittedby separate tubes to the central poi.tioii of a platinum tube 15 cm.long, the ends of which were closed by transparent plates of fluor-spar, and the resulting gaseous product was conducted to tlie wateror mercury troagh by a third taube near the end. When fluorinecomes in contact with phosphorus trifluoride, a yellow tlame of lowtemperature is seen, and the product of the reaction is mainlyphosphorus pentafluoride, the residue consisting of unchanged tri-fluoride. JN. W.Persulphates. By BJEHTIIELOT (Conzpt.rend., 112, I481 -1483 ;see also Berthelot, Gompt. rend., 86, ‘LO, 71, and 277; Marshall,Trans., 1891, 771; MendelAeE, HdZ. Sot; Chirrt., 38, 168; Traube,Her., 22, 1518, and 24, 1764).-The mixture of peraulphuric andsulpliuric acids obtained by electr.ilgsis, when neutralised by baryta-water or potash solution, yields a quantity of the normal persulphate,which is greater when the operation is carried on at a very lowtemperature. The neutralised solution rapidly becomes acid again,11m-ticularly on heating; a t the same time losing oxygen. l’hequantity of active oxygen is proportional to the quaiititmy of sulphuricacid formed in the decomposition.Barium persulphate is soluble, and as neutral as the thiosulphate orpermanganat’e ; its decomposition proceeds accordiug t o the equationBaS,08 + H,O = BaS04 + II,SO, + 0.In t,iti.ating the active oxygen by the usual reagents,it is necessaryto take into account the hydrogen peroxide often coexistent withpei*sulphuric acid, or capable of being formed during the dilutionsand other operations.The proportion of hjdiuogen peroxide presenthas been found to vary from 0 to 2 mols. for each mol. of persuiph-uric acid. Viewing the whole of the active oxygen present in theextreme case as united to the sulphur, the complex acid would havethe formula S,O,,nH,O. There is no proof of the existence of a lieu traland anhjdrous compound, such as the holoxide, SO4, of Trauba.The salts of persulphuric acid are distinct and definite, and com-parable with the permanganates, perchlorates, permolybdates, a i dThiosulphates. By A.FOCK and I(. KrXss (Ber., 24,3016-3017).--Potassiwn calcium tliiosidphnte, 3K,S20,,CaS20, + 5H20, is obtainedin colourless, monosymmetric crystals, by evaporating on the water-bat8h a mixture of potassium and calcium thiosulpbate solutions.pertungstates. NT* 1’INORGANIC CHEMISTRY. 13The axial relation of the crptals is a : 5 : c = 1.7610 : 1 : 0.89.31 ;/j = 80" 2'.Potassium sfmitiurn thiosulpnhate, K2S2O3,Si S?Os + 5H20, is pre-pared in a similar manner to the calcium salt ; it is readily soluble inwater, and crystallises in slender, lustrous, silky needles, which arenot, well adapted for crystallogi*aphic measurements. J. B. T.Ammonium Dithionate Hydrochloride. By A.FOCK and K.Kr,iiss (Ber., 24, 3017--.3018).-0n mixing solutions of ammoniumdithiouate and ammonium chloride in molccular proportions, andevaporating, a compound of the formula (NH,)2S,06,H C1, is depositedin rhombic crystals. The axial relation is a : b : c = 0.09827 : 1 : 0.9612.The same substance is also formed in presence of two or more mole-cular proportions of ammonium chloride. J. B. T.Revised Nitric Acid Tables. By G. LUNGE and H. REY (Zeit.any. Cham., 1891, 165-1 70).-The tables are coiistructed on similarlines to the rerised hydrochloric acid tables (compare this POI.,p. l l ) , m d similar processes were employed in their construction,but instead of weighing the acid in glass bulbs, or in a Winkler's stop-Qcwk pipette, a new form of weighing pipette, shown in the figure,RW used14 ABSTRACTS OF CHEMICAL PAPERS.Above the stop-cock a is fixed a bulb b, of about 2 cm.diameter,connected with a second stop-cock c. The lower part of the pipet'tefit,s into a glass tube d, which has n ground neck. When the pipetteis wanted for use, the stop-cock a is closed ; suction with the mouthis applied to the top, and the stop-cock c immediately closed, by whichnieans R diminished pressure will be obtained inside the bulb / I . Thelviint of the pipette is now dipped into the acid, the stop-cock a opened,and the acid allowed t o ascend a s far as the stop-cock, which is theniirimediately turned off. After wiping the pipette, i t is fixed intothe tube d and weighed. The pipette is now tnrned round, PO as toopen communication between the vhannels e and f . The stop-cock cis opened, water is squirted into the bulb b, and then allowed to flowrhrough cr, into the pipette.The diluted acid runs into the tube d ,the air of which escapes through e and f. The acid is finaily emptiedinto a beaker and titrated. By using this apparatus, any loss ofiiitric fumes is avoided. L. r)E K.Compounds of Sulphur and Phosphorus. By J. MAI (Annalcn,265, 19.L-208).-Th object of these experiments was to ascertainwIiet.her the boiling point of tripbosphorus hexasnlphide, P3S6, wassufficient.ly below that. of stannous chloride (SOSO), and above thatof pliosphorus peiitasulpliide (51 So), to allow of its advantageoustdmploymeut as a heating vapour in pyrochemic,zl investigations.When a n intimate rriixtnre of sulphur (2 atoms) and amorphous~)hcsphorus (1 atom) is cmef rilly heated in an atmosphere of carbonic;inhydride, and the product theii distilled, a yellow, crystalline sub-stmce (b.p. abont 410°), which seems to contain the cornpoundP,&, passes over first; the later fractions consist of a mixture OEtt.iphosphorus hexasulphide and phosphorus pentasulphide, whichcannot be separated into its constituents, as they boil at the sametemperature (about 508').Phosphorus pentasulphide boils at 332--340" under a pressure of10-11 mm. ; triphosphorus hexasulphide, P3Ss, boils a t 335-340"under the sa9e conditions.When the compound of the composition P4S3, prepared from itsclements by the usual met,hod, is distilled under a pressure of 11 mm.,H yellow liquid passes over between 230" and 240°, and amorphousphosphorus remains i t 1 the flask ; the distillate is almost completelysoluble in carbon bisLIphide, from which it is g~adually deposited inciystals of the composition P4S3.The twiling point of t h i s compounduilder the ordinary pressure would be about 408-418" ; it begins tosoften a t 130-135", h u t does not melt completely until the tempers-ture has risen above 160".When phosphorus (2 atoms) is melted with sulphur (3 atoms), andthe pyoduct submitted to distillation uiider a pressure of 11 mm.,alniost the whole passes over between 285" and 335" ; the distillateslowly solidifies to a yellowish, resinous, pasty mass, which, whendigested with carbon bisiilphide under pressure, yields a crystallinecompound of the composition PAS, ; the mother liquors from this com-pound contain a crystalline substance, which seems to be a mixture ofphosphorus trisclpl~ide with PAS,.I?. s. KIN0 dGIAS1C CHEMIS rRY. 15New Method of Preparing Carbon Oxysulphide. 337 J.NuHicsm (Ber., 24, 2967--2974).--Carbon oxysulphide is obtainedwhen cnrbonyl chloride is passed through concentrated sulphuricazid to dry it, and then through a tube 50 cm. long. filled with ignitedasbestos well mixed with finely pulverised cadmium eulphide, thetube being placed in a combustion furnace and heated. Theauthor 6nds that even when no external heat; is applied a sin;tllquantity of carbonyl sulphide is formed, but, the most, favouiable teu-yerature for its formation anpears to be 260-28O".The gas thusprodnced was found on analysis to contain COS, 94.87 per cent.;CO, 3.98 per cent.: air, 1.15 per cent. A quantity of crystals,which were identified as cadmiurn chloride, were obsez-ved in a tubepreviously charged with a layer of cadmium sulp5ide and heated intt flame during the passage of a current of csarbonrl chloride; thereaction, therefore, appears to be a double decomposition.A. R. L.Allotropic Silver. 111. Blue Silver. By M. C. T ~ E A ( P l d .Uag. [ 5 ] , 31, 497-504 ; compare Abstr., 1891, 803).-When silvernitrate is added to a solution of dextrin made alkaline with sodium01- potassium hxdroxide, the alkali being kept in moderate excess, thesilver is first precipitated as the ordinary brown oxide, but the colourgradually changes to rcddish-chocolate, and the silver begins to di+solve, and in a few minutes has completely dissolved to form analmost black liquid, a few drops of which.when added to a largequantity of water, give a splendid red liquid, that, spectroscopicclxnmination shows to be a perfect solution. Different specimens ofd ~ x t r i n behave differentJy, and the common brown form seems togive the best results. Convenient proportions are 20 grams of sodiumIrjdroxidc and 20 grams of dextrin in 1000 C.C. cf water, 14 grams ofsilver nitrate previously dissolved in a small quantity of water beinggradually added.It is interesting to observe that allotropic silver can be formed andcan exist in solution in neutrill, acid, and alkaline liquids.The alka-line solution spontaneonsly deposits allotropic silver after some time;and precipitation is immediate on the addition of dilute nitric or sulpll-uric acid. Even with a laige excess of acetic acid, however, precipita-tion is incomplete. When the precipitate has once formed, ~t is almostinsolublc, but on washing dissolves to a very small extent, forming arose-red liquid.A small quantity of sodium phosphate, when added to the alk~.linesolution, precipitates the whole of the allotropic silver with a rub?-copper coloui-, which after p~olonged washing changes to deep Nile-#IY~II, and becomes slightly soluble, forming a wine-red solution.The various f o m s of allotropic silver show different body andsurface colours, which tend to be complementary.'The precipitateby sodium phosphate, €or example, when spread thickly on paper,dries with a bright-green, metallic colour, but after it has heenchanged to deep green by washing, films dry with a dark gold orcwpper surface-colour.Although the dry substance has a11 the appearance of a metal, i16 ABSTRACTS OF CHEXICAL PAPERS.mRy contain 8 to 10 per cent. OF organic matter that cannot be rcmovetl,even by prolonged washing with hot water under pressure. Fourspecimens of carefully purified material contained respectively 93.77,94.27, 92.86, and 96.64 per cent. of metallic silver. When the allo-tropic forms are heated, a vapour is evolved which condenses in small,brownish drops, with an empyreumatic odour.The residue consistsof bright white metallic silver, and when dissolved in nitric acid,leaves a residue of black flakes of carbon. When the allotropicsilver is dissolved in dilute nitric acid, and the silver precipitated byhydrochloric acid, the filtrate, on evaporation, leaves a small residueof a, yellowish, gummy substance.Tannin reduces silver nitrate to allotropic silver more readily thandoes clextrin. and gives better results in presence of sodium orpotassium carbonate than with the caustic alkalis. 24 grams ofanhydrous sodium carbonate are dissolvod in 1200 C.C. of water, andmixed with 72 C.C. of a filtered 4 per cent. solution of tannin.24 grams of silver nitrate dissolved in a small quantity of water isadded gradually, and solution of the silver takes place imme-diately.After standing for a day or two, the intensely dark-colouredliquid may be poured off from a small quantity of black precipitate.On adding a small quantity of dilute acid, the allotropic silver isprecipitated, and in films dries with a bluish, steel-grey, surface-colour.Blue allotropic silver (inclnding the green and steel-grey varieties)shows endless variations, and cannot be reduced to one type. Slightdif'ferences in the conditions of formation result in very differentproducts. Of ten products obtained by the action of tannin andsodium carbonate in various proportions, several were easily andcompletely soluble in ammonia, whilst some dissolved slightly, andothers not at all.Some of the specimens, insoluble in water, whentreated with phosphoric acid, did not dissolve, but on washing awaythe acid, were found to have become soluble in water ; other insolublespacimens did not behave in this way. Some of the solutions werescarcely affected by acetic acid, whilst others were partially, andothers almost completely, precipitated.The films OIL paper var.y greatly in their sensitiveness to light, an3in the ease with which they are converted into the yellow, inter-mediate form. The least sensitive modification is that precipitatedby nitric acid ; it dries with a steel-grey colonr. The modificationsprecipitated by acetic acid tend to have a greenish metallic surface,and are more sensitive.Permanency varies greatly in different speci-mens, and is increased by thorough washing.The blue, grey, and green forms are related to the black, or dark-grey normal silver, and tend to pass into it, whilst the gold-colouredmodification tends to change into bright white metallic silver on t h esurface, with dark, or even black, silver underneath.Tannin likewise reduces silver nitrate in prescnce of lithium,ammonium, magnesium, barium, calcium, and strontium carbonates.The product with strontium carbonate is dark-red whilst moist, butdries with a rich, bluish-green, metallic surface-colour in thick films,but is red and transparent in thin filmsINORGANIC CHEJILSTRT. 17The intermediate golden-yellow form produced during the passageof the pold-coloured allotropic form to white normal silver has noneof the properties of the original form, except its colour and Instre.Itis hard and tough, is not converted into normal silver by friction andhigh pessure, and offers as much resistance as normal silver to theaction of oxidising and chlorinating agents. The difference betweenthe original form and the int,ermediate form lies in the fact that thelatter. has a crystallinc structure. which becomes evident on treatingthe film with ferric chloride solution.When the blue form is gently heated, it also first becomes yellow,and then changes to white normal silver. A film on glass begins&ochange from blue to yellow a t about 180", and the same change ISproduced by the action of light, some specimens reqniriiig a fewhours' exposure to sunlight, whilht others require several days.The author considers that in these allotropic forms the silver existsin a state of atomic divihion, and that this is true also of the pyro-phoric forms of various metals.He points out that there is a differ-ence between cheniical and mechanical division, and that a metal mayform N compact mass, and yet be in a state of atomic division, whilst,on the other hand i t may bein a fine state of inechatlical division, andyet possess a high degree of molecular complexityCharacteristics of the Alkaline Earths. By G. BHCGELMANP.-(Zsit. anal. Chem., 30, 579--580).-1n his attempts to preparebaryta by igniting the hydroxide in graphite crucibles, the author hasbeen unable to obtain a pure product, the crucible being alwaysattacked.The supposition oE the dimorphism of barium oxide, andthe assumed catalytic action of platinum (Abstr., 1890, SSO), musttheretorc be Rbandoned, since the substance obtained as a felted massof needles showing chromatic polarisation c m no longer be regardedas baryta. Since a similar action occurs with lime and strontia, allstatements respecting the specitic gra\ ity, &c., of the alkaline earthsprepared in other than platinum vessels, must be withdrawn. 'Yliestatements of Fresenius and others, that strontium carbonate fuseswhile decomposing, cannot be confirmed. Much contraction occurs,but the oxide produced retains the form of the original mass OF car )on-ate, and is, a t most, slightly sintered.Composition of a Boiler Incrustntion.By A. CaR1s.r (Zeit.any. Chem., 1891, 77).-The author communicates an analysis of acurious boiler deposit, which owed its origin partly t o the use of ananimal o r a vegetable lubricating oil. CaO, 11-09; MgO? 9.79;Fe,O?,, 5 60 ; A1203, 1.10 ; PbO, 0.98 ; CuO, trace ; SO2, 16.00 ; SO3,1.71 ; fatty acids, 22.62 ; neutral fats, 25.84 ; moisture, 2.69 ; combinedC. H. B.M. J. $,water and organic matter, 5-22 per cent. L. DE K.Action of Hydrogen Peroxide and of Water saturated withCarbonic Anhydride on Magnesium. By G. G~ORGIS (Gazzetta, 21,510-514) .-According t o Weltzien (dnnalen, 138, 1:32), hydrogenperoxide acts slowly on magnesium, forming an alkaline liquid, whichcontains the normal hydroxide, Mg(OH),, and on evaporation to dry-VOL.LXlt. 18 ABSTRACTS OF CHEMICAL PAPERS.ness leaves a white, strongly alkaline mass, completely sola\de inwaiLr. The author has also observed the slow action of IiTdrogeuperoxide 0 1 1 maqnesiiim and the alkalinity of t.he solution, but fiiidvthat, after. remaining for a few days, acicular crystals arc! deposited,which efiervesce on treatment with hydrochloric acid. On treatingmetallic magnesium in a vacuum w i t h hjdrogen peroxide free fromcarbonic aiihyclride, the liquid after a time becomes slightly alkaline,and on evaporating to dr-yness in a vacuum, a flocculent precipitatemparates, which is only very sparingly soluble in water free fromcarbonic anhydride.On the other hand, when magnesium is treatedwith distilled water saturated with pure carbonic anhydride, the liquidsoon becomes strongly alkaline, and the m t k l is energetically attackedwith evolution of hydrogen ; the reaction slackens very gradually, andgas ceases to come off after about 10 or 12 hours. The solutioii, after:L time, deposits acicular crystals havirig the composition AJgCO, -t3H,O.The evolution of hydrogen by the actioii of an aqueous solution ofcmbonic anhydride on magnesium is noteworthy ; the reaction alsoIwovides nu easy method for the preparation of the normal carboilateof magnesium.Cuprammonium Oxide. By PHuD'HuJi \I (Chem. Centr., 1891,ii, 339-340 ; from Mnn. Sci. [4], 5, ~jt31).--C'upramtrionium oxide isi t more powerful oxidking agent than hydrogen peroxide; it actsi w r e rapidly on cellulose, aud converts it into oxycellulose ; it alsodecoloriscs indigo-blue mow rapidly than hydrogen peroxide, With" mercerised " cotton wool, i t acis like Ilydrogeo peroxide.I n preparing it, by agitating copper turiiiiip with ammonia andair, both cupric oxide atid nitrous an hycliide are forlt,ed : the Ditrousanhydride is, however, only formed a t a later st,iige of the reaction.Ammonium nitrite oxidises copper, atid nitric peroxide is formed inthe tirst instalice; ammonia is also liberated by the reaction, andwhen this occurs, free nitrogen is formed instead of nitric peroxide,The liquid becomes intensely blue ; an excess of water precipitatescupric hydroxide ; dilute acetic acid ptecipitates cupric oxide as soonS. B.A. A.as the ammonia is exactly neutrdisecl. J. W. 1,.Action of Ferric Chloride on Metallic Sulphides. By CAMMEREK(Chern. Centr., 1891, ii, 370; from Berg. Hutlen Zeit., 50, 2cil--664, 282-28-4.).-1n a sealed tube, ferric chloride reacts with bothcupric and cuprous sulphides, cupric chloride, ferrous chloride, andmlphur being formed. With pi-ecipii ated ferrous sulphide OF ironpyrites, ferrous chloride and sulphur are the products ; with copperi)yrites, cupric chlo~ide, ferrous chloride, and sulphur are formed.With arsenious sulphide, ferrous chloride, sulphur, and nraeriiouuchloride are tirst formed ; the latter, howevei., is furtlier oxiclised toarsenic anhydride by the continued action of t-xcess of ferric chloride.With eitliey precipitated or native antimony trisulphide, antinroniousiAloride, ferrous cliloride, and sulphur are formed.With stannous-;ulphide, staiiiiic chloride, ferrous cliloI*ide, aid sulphur are formed.J. IV, 1 1 IXL?ROXNlC CHEMISTRY. 19Manganese Tetraahloride. By H. M. VERNON ( P l i i l . LMCLg. [ 5 ] ,31, 469-484).--Pisher (Trans., 1878, 409) endeavoured to show thabt,he dark-brown liquid obtained on dissolving manganese dioxide inhydrochloric acid contains manganese tetrachloride. Picker-ing, whorepeated Fisher's experiments, aiid made, in addition, a number ofhis own (Tram., 1879, 6.54), thought that the evidence was in fayourof the existenre of Mn,Cl, in the solution, and not of MnCII.The author now attempts to decide between these views by experi-ments on the rate a t which the " available " cliloi*iiie is evolved atdifferent, temperatures.If Mn2C16 is really formed in the solution,half the available chlorine should be much rno1.e easily removed bymeans of a current of air passed through the solution than the otherhalf, whereas. if MnCI, is present, the rate of removal should be more~egular. The curvt s obtained from the experimental results at varioustemperaturt-s sliow no break a t the point where half the availablechlorine has been driven out, so the author concludes that whenMrisOr, Mn,03, OY MnO, is dissolved in hydrochloric acid, the onlFhigher chloride formed is MnC14. The brown solution is much morestable at -26" than at ordiiiary temperatures.Pickeriug found that the presenceoF MnC1, in the solutioii from thebeginning rendered it more stable.This the author accounts for onthe supposition that the MnC14 dissociates directly into illnC1, andCI,, so that the presence in excess of MnCl,, one of the products ofdissociation, would diminish the actual amount of disso iation.Calorimetric Researches on the Condition of Silicon and ofAluminium in Cast Iron. By F. OSMOND (Cornpt. ~d., 113, 474-476).-The heat developed on dissolving 1 gram of cast ii.or.i con-taining different amounts of :;ilicon in a saturated solution of thedouble chloride of copper and ammoniuni was measured. From theresults, i t appears that when the silicon is present in sufficient quan-tity,. it conrbines with the iron with the development of heat andformation of a cotupound which is dissociated by an excess of iron,and therefore only exists wlieii the amount of silicon i n the alloyis sufficiently great.Hence the difference between the quantities ofheat found and those calcolated in the above experiments changessign when the amount of silicon present reaches a percentage some-where betwecn 4.1 and 7.3.Samples of c,ast iron containing aluminium were also examined inthe above manner. The results show that alumiiiiuni dissolves illcast iron with the absorptioii of heat.Attempts to Prepare Metallic Chromium from ChromicFluoride. By W. P. EVASS (Zeit. any. C h e w , 1891, 18-2Oj.-Tht.mthor tried the effect ot' nwtallic sodium, metallic zinc, aiid mixedcarbon and silica on chromic fluoride a t a very high temperature, inthe hope of getting metallic chromium.1.Action of Sodium.-The metal was heated a t 400' in a porce-lain tube, and the fluoride passed over it. The latter was prepttrzdby distilling a mixture of lead chwnictte, calcium fluoride, and snlph-uric acid. A very &ong reactiun-tcok place and the bube becarnsi . 2J. W.H. C20 ABiTRACTS OF CHEJllCAL PAPERS.red-hot, but was soon stopped up. After cooling, the excess ofsodium was dissolved out by water, and the residue examined formetallic chromium. But very little was found, and this was chieflyin combination with silicon. The experiment was repeated in an ironapparatus, in which tile va1)our of about 70 grams of metallic sodiumwas brought in contact, at R temperature of 900-1000°, with thefluoride.The chief products obtained were amorphous, green chromicoxide, bright-green particles of sodium-chromic fluoride, and a grey-i S l l , Lritrtle, spongy mass, which was chiefly located near the fluorideclelivery tube. On analysis, i t proved to be an alloy of sodium witli%bout 25 per cent. of chromium intermixed with a little chromic oxideand a trace of iron. The experiment was repeated, and, t h i s time,to avoid contamination with iron, in a Hessian crucible. Thecrucible and the delivery tubes were well lined inside with a mixtureof 6 parts of alumina and 1 part of potassium chloride. Afterbeing. kept red-hot for balf an hour, the experiment came to ;i prc-iilttture end through the tube gettiag stopped up. After cooling,there was found adhering to the tube a greyish, stalactitic mixss,which foy the greater part dissolved in hot water. The insoluble por-tion consisted of fairly pure, crystalline, metallic chromiurn.2. Action of Zinc.-About 2010 griims of zinc was heated to boilingunder a layer of salt, and the fluoride M'HS passed in through aporcelain delivery tube. 'l'he greater part of the fluoride was takenup by the salt with formation of sodiiini chromate. The zinc: regulusdissolved slowly in nitric acid, which gradually acquired a greenishtinge. No chromium could be detected in the insoluble residue.There was, however, good reason to believe that. an alloy of zincwith about 0*67-1*60 per cent. of chromium had been formed.In the porcelain tube, however, a lustrous, steel-like substance wlsfound, which on analysis proved to be practically pure metallichromiii m.3. Action of Carbon and Silica -The fluoride was passed througha strongly heated porcelain tube containing small granules of a mix-ture of 9 parts ot' amorphous silica and 4 parts of powdered char-coal. Abundant vapours of silicon fluoride made their escape, and onopeniiig the tube, sharp, shining, hexagonal crystals of chromicoxide were foiind; besides this, an almoht black, hrittle mass, wllichon analysis yielded Cr, 30.13 ; SiO:, 48.53 ; C, 11-39; 0, 9-95, COU-taining, therefore, 8.39 per cent. of metallic chromium.From these experiments, it seems that although there is not theslightest doubt atbout the reduction of the chromic fluoridn by thepiocesses described, it is not possible to prepare the metal 011 thelarge scale in t h i s way.Pure Bismuth. By A. CLASSEN (J. p r . Chem. [2], 44, 411-514 ; compare Abstr., 1891, 271, 525,1324).--This is another chapterin the controversy between Schneider and Classen as to pure bismuth.Against Schneider's aualyses of commercially pure bismuth (Abatr.,1891,1524), the author puts in two analyses, the one of '' chemicallyllnre bismuth for scientific investigations," which gave sonle 10 grarll,slead chloride from 500 grams of the metal, the other of '' bismuthL. DE K,\.XISERALOGICAL CHENISTRT. 21pnrissimnm,” from the same factory, which contained, besides lead(ufidetermined), 1-56 per cent. of copper and 0.45 per cent. o.f iron.The nnthor’s electroljtic bismuth melts at 264” ; “bismuth puriss.”from Tromsdorff melted at 272.8’. and another sample at 273”;‘‘ absolutely pure bismuth ” from Tromsdorff melted a t 265-266” ;‘‘ bismuth pnriss.” from Schering melted a t 269-270”. This isfurther evidence of the impurity of commercial “ pure b.ismuth.”The rest of the paper deals with Marignac’s method and material,and is purely polemical. A. G. B