26 ABSTRACTS OF CHEMICAL PAPERS. An a 1 y t i c a1 C h e mi s t ry. New Gas-volumeter of General Applicability. By F. GANTTER (Zed. anal. Chem., 32, 553-564).-The principle of the method consists in the measurement of a liquid expelled by the evolved gas. The cylinder D and tubcf are filled to the point off' with a, liquid in which the gas to be liberated in not soluble : g is then clased. For carbonic anhydride, a, solution of calcium chloride of sp. gr. 1.4 servesANALY TIOAL OHEMISTRY. 27 well. The substance is weighed into A, and the reagent placed in B, and A and B are then plunged into the water-jacket E. When equilibrium of temperature is attained, pressure is equalised by opening e for a moment. The tube f' is now placed in an empty flask of known capacity, and 9 is opened.The gas is next evolved by allowing the reagent to flow into A, and the contents of A are boiled, the internal pressure being all the time maintained below that of the atmosphere by keeping the point of f' at a lower level than the surface of the liqnid in D, but as soon as cooling commences f' must be raised so as to produce a small excess of internal pressure, and its point must he kept plunged i n the liquid in the flask. The reaction vessels are again plunged into the water in E, and when cooling is complete f ' is lowered until the manometer C shows that the pressure is the same as at the outset ; g is then closed, f' removed from the flask, and the nnoccupied portion of the flask measured by filling up to the mark from a burette. It is necessary that the temperature of I3 should be exactly the same at the end as at the beginning, and if i t is desired to avoid reduction of the gas volume to normal temperature and pressure, an experiment with a pure substance will give in half an hour a factor for direct calculation from the observed volume under the existing conditions.I n estimating ammonia by decomposition with hypobromite, i t is best to put the hypobromite in A, and the ammoniacal solution in B, and, finally, to rinse R thrice with the hypobromite by inclining the apparatus so that the liquid flows from A through a and b to B. Nitric acid can also be estimated by decomposition with copper and sulphuric acid, but for this purpose a special laboratory vessel is required, from which the air can be expelled by hydrogen.Borax as a Basis for Acidimetry. By T. SALZER (Zeit. anal. Chem., 32, 529-537; see Rimbach, Abstr., 1893, ii, 232).-The author suggested the use of borax as a standard alkali in 1857, but the impossibility of obtaining solutions of normal strength operated against its adoption at the time. It is, however, far preferable to work with weak solutions, since the tendency of boric acid to redden litmus disappears on dilution. Methyl-orange may be used as indi- cator in the titration of mineral acids, but in dilute solutions is far less sensitive than litmus : for organic acids it is inadmissible. The borax used must be free from octahedral crystals ; it should lose 4r.1 per cent. of water when ignited. A decinormal solutioir (19.0872 grams per litre) is of convenient strength, and acids must be diluted to a strength not exceeding N/10 before tit,ration.The acid is to be coloured a pale yellowish-red with sensitive litmus tincture, and the borax solution run into i t until the colour just changes to a bluish- red, and no longer returns to yellowish-red on stirring. Mineral acids, oxalic, acetic, and tartaric acids give satisfactory results ; phosphoric acid behaves like a monobasic acid. These acids give practically the same results when the acid is run into the borax as when the pro- cess is reversed. Citric acid can equally well be titrated by %he borax solution, but when the citric acid is run into the borax more acid is consumed, seemingly from formation of borocitric acid. M. J. S. 31. J. S.28 AESTRACTS OF OHEMICAL PAPERS. Phenolphthalei’n as an Indicator.By R. T. THOMSON (J. SOC- Chem. Ind., 12, 432--433).-The author has found that boric acid may be accurately titrated with standard solution of sodium hydroxide, using phenolphthaleyn a8 indicator, but it is necessary to add to the liquid about one-third-of its bulk of glycerol. The amount of boric acid contained in borax may be estimated in the same manner, after first adding sufficient sulphuric acid to combine with the soda, methyl- orange serving as indicator. The excellent results obtained by the author are probably due to the greater viscosity of the liquid, caused by introducing the glycerol, L. DE K. Extraction of the Gases Dissolved in Water. By J. ROBSON (J. Xoc. Chem. Ind., 11, 504).-The apparatus used is a modification of that of Kreusler for the estimation of nitric acid.The gases are boiled out from the water in a flask from which the air has been com- pletely expelled by steam, and are collected in a vertical cylinder of 150 C.C. capacity, closed at its lower end by a cork, and at its upper end connected by a rubber tube to a Schifi’s nitrogen measurer. Through the cork of this cylinder pass two tubes, the one connected to the long narrow neck of the extraction flask, and the other with a reservoir of boiling distilled water at a higher level. A figure shows the arrangement of the apparatus and the mode of using it. M. J. S. Estimation of Dissolved Oxygen. By G. ROMIJN (Rec. Tmv. Chirn., 12, 241--247).-The sample of water is taken from any required depth beneath the surface by means of a pipette of known volume.The pipette is furnished with three-way stopcocks above and below, the upper stopcock having above it a small bulb with a fiducial mark on its free neck, and of known volume, about one- twentieth the capacity of the pipette. The pipette is filled by the aid of a small pump, the liberation of dissolved gases from the water in the pipette, owing to reduction of pressure due to the action of the pump, is avoided by having the aperture of the lower stopcock o€ double the diameter of that next the bulb. The upper three-way tap allows of the water sampled being drawn through the pipette in quantity su5cient to ensure the purity of the sample. The pipette being full and the upper bulb empty, the lower tube is connected with a reservoir containing sodium hyposulphite (Na2S02) solution, and a little of the solution is r u n through the stopcock to wash out the tube and tap ; the bulb is then placed in communication with the pipette, and, the lower cock being turned, hyposulphite soliition is run in until the displaced water reaches the mark on the neck of the bulb; the taps are now turned off at both ends of the pipette, and the whole set aside for about ten minutes.The same volume of iodine solution is then introduced in the same manner, tbe lower tap and tube washed with water, the remaining contents of the pipette transferred to a flask, and the residual iodine determined by thiosulphate in the usual way. The hyposulphite solution is kept in a bottle connected with a small hjdrogen generator, arranged so as to keep an atmosphere ofANALYTICAL OEEMISTRT.29 hydrogen under pressure over the solution, the pressure being sufficient to force the solution into the pipette. I€ V be the volume o€ the pipette, *u the volume of the bulb between the mark and the upper tap, then the volumt of water ulbimately used for the titration is (V - vj2jV. Let 0 be the titre of oxygen, in cubic centimetres per lifre of water, then 0 = (A - B)a. I n this formula, A represents the number of cubic centimetres of thiosulphate used, B the number of cubic centi- metres of thiosulphate which would be used for water free from oxygen, and a is a constant depending on the dimensions of the pipette and the titre of the thiosulphate solution.B can best be determined by an experiment with water saturated with air a t a known temperature, of which the oxygen titre is known from Roscoe and Lunt’s or Winkler’s determinations. The constant 1000 V is the factor for In this formula, -- (V- u)2 8000 V 1.43 (v - v)2 r. m = - --’ calculating’from the volume of water used, the figure for a litre of water ; 8/1-43 is the factor reducing the weight-equivalent of oxygen to the corresponding volume equivalent ; and r is the factor used in reducing the actual volume of thiosulphate used to the equivalent volume of normal solution. A correction is required if the water contains any substance such as hydrogen sulphide, which itselE absorbs iodine, or when only about 2 C.C. of oxygen per litre is found and the deficiency is caused by a rapid decomposition of organic matter ; this is found by a preliminary experiment to determine the iodine so absorbed.Detection of Iodic acid in Nitric acid. By E. PIESZCZECK (Chem. Centr., 1893, ii, 337; from Apoth. Zeil., 8, 322).-The iodic acid is reduced with hydrogen sulphide, tin filings, or sulphurous anhydride. It is most convenient to use tin, in which case 10 O.C. of the acid is gently warmed with a few scraps of tin, and, after a few minutes, shaken with a small quantity of chloroform, which remains colourless in the absence of iodine or iodic acid. E. C. It. W. T. Influence of Ammonia on the Estimation of Hydrochloric acid in the Stomach Contents. By H. STHAUSS (Chem. Centr., 1893, ii, 379-S80 ; from Berlin KZin. Woch., 30, 398--402).-See this vol., ii, 21.Estimation of Phosphorus in Steel and Iron containing Silicon. By J. SPELLER and S. KALMAN (Zeit. anal. Chem., 32, 538-550) .-In the precipitation of phosphoric acid by molybdate from solutions containing silica, a, certain quantity of a silicomolybdate is apt to separate. This precipitation is much favoured by the presence of ammonium salts, especially the nitrate ; potassium nitrate is without influence. The lower the tempxature at which the preci- pitation takes place, the smaller the amount of silicomolybdate formed, so that by proceeding in the following manner a correct determination can be made even in presence of silica. Of steel, 3.3 grams is dis- solved in 50-55 c . ~ . of nitric acid (1.2 sp- gr.) in a *.litre conical30 ABSTRACTS OF UEIEMICAL PAPERS.flask, at first in the cold, but finally with vigorous boiling. When all nitrous vapours are expelled, the solution is treated with 8 C.C. of a 3 per cent. solution of potassium permanganate, and boiled until the red colour disappears. The precipitated manganic oxide is re- duced by adding 4 C.C. of a 10 per cent. solution of potassium nitrite, and, after cooling to 55-60', the phosphoric acid is precipitated by adding 80 C.C. of molybdate, or, equally well, it may be cooled corn- pletely before adding the molybdate, and then warmed to 40". The mixture is now kept for two hours at 36-40", with an occasional shake ; the precipitate which forms is collected, and since i t has a tendency to pass through the filter, this should be prepared by pouring upon it a thin pulp of paper fibres.The molybdate solution is made by dis- solving 150 grams of ammonium molybdate to a litre, and pouring it into a litre of nitric acid (1.2 SP. gr.). The precipitates, after being washed, first with a mixture of 100 vols. of molybdate solution, 20 vols. of nitric acid, and 80 vols. of water, and then with a 10 per cent. solu- tion of ammonium nitrate, are dissolved from the filter with a little dilute ammonia, and the solution evaporated and gently ignited in small porcelain basins until the ammonium salts are expelled. Cast iron requires 60-80 C.C. of nitric acid to dissolve 3.3 grams, and should be very finely powdered. I t usually contains so much phosphorus that only one-half to one-fiPth of the solution need be pre- cipitated, and if so much silica is present that part of it separates as a gelatinous precipitate, the solution should be made up to a known volume, and run through a filter before measuring out an aliquot part.. M. J. S. Estimation of Arsenic and Phosphorus in Iron Ores. By J. PA~TINSON and H. S. PATTINSON (J. Soc. Chem. Ind., 12, 119-121). --The authors recommend treating 3 grams of the ore with hydro- chloric acid. After rendering the silica insoluble by evaporation, the residue is taken up with a very little acid, and then diluted with hot water to 50 C.C. A 50 per cent. solution of sodium thiosulphate is now run in until the iron is completely reduced to the ferrous state. The sulphurous acid is expelled by boiling, and, after cooling, 5 C.C. of st,rong hydrochloric acid is added.A little powdered zinc sulphide is now introduced, which causes the precipitation of any arsenic as trisulphide. The precipitate is first washed with hot 5 per cent. hydrochloric acid to remove any lead, and, after washing with water, it is digested with ammonium sulphide. The solution is evaporated to dryness on the water bath, the residue oxidised with bromine and a little nitric acid, and the arsenic finally precipitated, and weighed, as magnesium ammonium arsenate. The filtrate from the arsenic trisulphide is boiled to expel hydrogen sulphide, and diluted t o 250 C.C. ; a little ferric chloride is added, and the liquid neutralised with calcium carbonate. This will cause a precipitate of ferric phosphate, which must be washed on a filter with lukewarm water. To estimate the phosphorus, it must be dissolved in dilute nitric acid, and heated w i t h solution of ammonium molybdate ; this causes the precipitation of ammonium phosphomolybdate, which is t'her, collected and weighed.The test analyses are certainly extremely satisfactory,AN ALP TICAL CHEMISTRY. 31 but, as no silicates were introduced, no evaporation to dryness in presence of hydrochloric acid was required, and therefore there could be no loss of arsenic. L. DE K. Estimation of Boric acid in Boronatrocalcite. By G. A. LE ROT (Chew,. Cent?*., 1893, ii, 291 ; from Bull SOC. ind. Rouen, 21, 62) .-2.5-5 grams of the finely-powdered substance is decomposed in a reflux apparatus with a few C.C. of sulphuric acid diluted with au equal volume of water.The presence of a small quantity of hydrochloric acid facilitates the decomposition. The cooled mixture is filtered, the filter washed with acidified water, the filtrate neutralised with soda, heated to boiling, and again filtered. The filtrate is acidified with hydrochloric acid, boiled to expel carbonic anhydride, and then made up to a known bulk with water free from carbonic acid. The solution contains free boric acid and alkali sulphate and chloride. 20 C.C. of the solution is placed by the side of the same bulk of water, and an equal number of drops of a 10 per cent. solution of Poirrier’s Orange I11 added to each. The acid solution is then titrated with a solution of sodium hydroxide until it acquires the same colour as the aqueous solution of the indicator.This titration gives the quantity of free hydrochloric acid. A second 20 C.C. of the acid solution is titrated with sodium hydroxide in the presence of Orange I1 until it acquires a dark-red colour. The difference in the two titrations gives the quantity of boric acid. The sodium hydroxide must be free from carbonates, silicates, and aluminates. Analysis of Silicon Carbide. By 0. M~HLHAEUSER (Zeit anal. Chem., 32, 564--567).-The analysis of this substance, which is - being produced in America as a substitute for emery and bort under the name “ Carborundum,” presents peculiar difficulties, in conse- quence of its extreme hardness. After trituration in an agate mortar, i t is necessary to submit the powder to elutriation, and accurate results in the carbon determination can only be obtaiiied with that which remains in suspension for at least five minutes.The carbD,n is best estimated by combustion with 20 parts of lead chromate (the addition of potassium dichromate causing the oxidation to proceed with explosive rapidity) ; the silicon by fusion with potassium sodinm carbonate for about six hours, during which the heat should be raised very -gradually. A very pure specimen gave the following numbers :- C 30.2, Si 69.1 (Al, Fe)20, 0.46, CaO 0.15, MgO 0.09 per cent. The formula S i c requires 30 per cent. of carbon. E. C. R. M. J. S. By R. WEG- SCHEIDER (Monatsh., 14, 315-322).-The author has found that j, order to obtain accurate results in estimating copper as cuprous sulphide by heating the precipitated cnpric sulphide in a current of hydrogen, the temperature must not be allowed to rise too high.A temperature of about 650” (dull red heat) is the highest which can be safely employed, as at higher temperatures than this a portion of the sulphide is reduced to metallic copper, hydrogen sulphide Estimation of Copper as CUprous Sulphide.32 ABSTRACTS OF CHEMICAL PAPERS. being evolved. The substitntion of coal-gas for hydrogen does not give good results. When hydrogen sulphide is employed instead of hydrogen, a portion of the cupric sulphide remains unreduced. Separation of Metals in Alkaline Solution by means of Hydrogen Peroxide. By P. JANNASCH (Ber., 26, 2329-2331 ; compare Abstr., 1893, ii, 492).-Copper peroxide is precipitated on adding hydrogen peroxide to a solution of cupric hydroxide in dilute ammonia, but the reaction does not occur in presence of even slight excess of ammonia or ammonium salts.The precipitation is, in any case, incomplete, 1-5-2 per cent. of copper remaining dissolved. Copper peroxide forms voluminous, dark olive-green flocks, and appears to attack porcelain, as it was always found to contain silica. It is possible, by the help of hydrogen peroxide, to show in one solution the various oxidation stages of copper. On mixing 10 per cent. copper sulphate solution (3 c.c.) with 20 per cent. soda (4 c.c.), cupric hydroxide is precipitated ; this is dissolved in 10 per cent. tartaric acid (3 c.c.) and 2 per cent. hydrogen peroxide solution (15 c.c.) added; copper peroxide is precipitated, but dissolves on heating, and on cooling the solution, cuprous oxide is deposited.Separation of Metals in Alkaline Solution by means of Hydrogen Peroxide. By P. JANXASCH and J. LESINSKY (Bw., 26, 2331-2334, 2334-2336 ; compare preceding abstract) .-Separation of Lead from Copper.-Lead nitrate (0.5 gram) and copper (0.3 gram) are dissolved in water (50 c.c.) and concentrated nitric acid (10 c.c.), and the solution treated at ordinary temperatures with a mixture (80-125 c.c.) of 3 parts of hydrogen peroxide (2 per cent.) and 1 part of strong ammonia. Saturated solution of ammonium carh- onate ( 5 c.o.) is added, and the precipitate is washed 4 or 5 times with a mixture of hydrogen peroxide (1 part), concentrated ammonia (1 part), and water (6-8 parts) ; when all the copper is separated, the washing is continued with dilute ammonia (1 : 8) at 60-80", and finally with warm water, in which the lead hydroxide is completely insoluble.The lead precipitate is dried and treated in the manner previously described (Abstr., 1893, ii, 493). The copper in the filtrate is determined in the usual manner, with due regard to the presence in the solution of nitric acid and ammonium salts. A large excess of hydrogen peroxide causes the lead to be precipitated in dense, crystalline plates. Mixtures of lead nitrate and copper sulphate may be separated in a similar manner if the precipitation of lead sulphate is prevented by the addition of acetic acid and ammonia. The analytical results agree closely with the theoretical. Separation of Lead from Zinc.-Lead nitrate (0.5 gram) and zinc oxide (0.2 gram) are dissolved in concentrated nitric acid (2 c.c.) and water (50 c.c.) ; the solution is treated with R mixture of 2-3 per cent.hydrogen peroxide (40 c.c.) and concentrated ammonia (15 c.c.) ; saturated ammonium carbonate solution (5 c.c.) is then added, and the liquid well stirred, the lead oxide is collected, washed first with dilute ammonia and finally with cold water, and treated in the manner already described (Zoc. cit.). The concentrated filtrate A. H. J. B. T.AXALYTICAL CHEM ISTRI’. 38 is treated with pure sodium hydroxide (5 grarns) and boiled until free from ammonia; i t is then acidified w i t h hydrochloric acid, and the ziiic precipitated with sodium carbonate and determined in the usual manner, care being taken to free i t from silica or nlumina.The lead oxide may also be precipitated in the crystalline form (see above. Separation of Lead from Nickel.-Lead nitrate (0.5 gram) and potas- sium nickel sulphate (0.4 gram) are dissolved i n concentrated nitric acid ( 5 c.c.) and water (50 c.c.), glacial acetic acid ( 5 c.c.) and strong ammonia (15 c.c.) are then added, and the lead precipitated with a mixture of hydrogen peroxide (75 c c.) and concentrated ammonia (20 c.c.) ; the addition of ammonium carbonate is unnecessary. The lead oxide is precipitated in a flocculent condition, and is treated in the manner described i n the separation of lead from copper (see above). The solution containing the nickel is evaporated t o dryness, treated with concentrated hydrochloric acid (1 5 c.c.), evaporated to dryness again, the residiie dissolved in dilute hFdro- chloric acid, the silica separated; the filtrate (250-300 c.c.) is then boiled, treated with aqueous hydroxylamine hydrochloride (4 grams), and precipitated with 15 per cent.soda (60-70 c.c.) in the usual manner. hydroxylamine, in contradistinction to ammonia, causes the nickel t o be precipitated in a form which admits of rapid filtration. The numerical results show a tolerably close agreement with the theoretical. J. €3. T. Estimation of Manganese by means of Potassium Perman- ganate. By A. GORGEU (Bull. Xoc. Chim., [3], 9, 490--496).-When manganese is estimated, as in Guyard’s method, by adding a solution of potassium permanganate to an almost neutral solution of man- ganons chloride, heated at 80”, the results are too low, partly in consequence of the acidity of the liquid.Donath’s modification, in which the solution of the manganous salt is added to a solutioii of the permanganate mixed with sodium carbonate, also gives in- exact results, unless the solutions are mixed very slowly, especially towards the end of the reaction. If Guyard’s process is modified by adding precipitated calcium carbonate before the permanganate, the results are less exact than by Donath’s method. According to Guyard, t bree manganese permanganates, MnjOI0, Mn6011, and Mn7012, can be obtained by mixing potassium perman- ganate and manganous chloride i n different proportions. E e seems, however, to have overlooked the fact that hydrogen chloride is liberated at the same time.The author finds that the oxide Mn,O,, is never formed, even when the liquid is kept neutral, and the two oxides Mn6011 and Mn,Olo are only formed when the acid that is set tree is continually neutralised with calcium carbonate. The author criticises Guyard’s views as to the constitution of these oxides, and contends that there is no evidence that they are manganese perman- ganates, and also that Guyard has advanced no evidence of the exist- ence of manganese mauganates. C. H. B. VOL. LXVI. ii. a34 ABSTRACTS OF OHEMICAL PAPERS. Estimation of Oxide of Iron and Alumina in Mineral Phos- phatss. By A. SMETHAM (J. Xoc. Chem. Ind., 12, 112--116).--The author uses Glaser's process, and checks the results by his ammonium acetate method.The analyses should agree within 0.1 per cent. The ammonium acetate process is carried out as follows :-2 grams of the sample is evaporated with strong hydrochloric acid, and the residue is taken up with 10 C.C. of dilute acid, and filtered from any siliceous residues. After boiling with a few drops of bromine, the liquid is allowed to cool, and mixed with dilute ammonia until a permanent precipitate is produced, which is then again dissolved by cautious addition of hydrochloric acid. Large excess of ammonium acetate is now added, and the precipitate is collected on a filter, well washed, burnt, and weighed. It theoretically consists of ferric and aluminium phosphates, but in practice it will be found to contain variable quan- tities of calcium phosphate.It must, therefore, be quantitatively tested for phosphoric anhydride, calcium oxide, and ferric oxide ; the alumina. then being found by difference. The author conducts the analysis as follows :-The precipitate is dissolved in hydrochloric acid, diluted with water, and, after adding some citric acid, boiled with ammonium oxalate. Dilute ammonia is then added until the liquid is neutral to litmus paper, and subsequently acetic acid to slightly acid reaction. The calcium oxalate is collected after Pome time, and ignited t o carbonate as usual. The filtrate from the lime is mixed with magnesia mixture, the precipitate being afterwards purified by redissolving and reprecipitating, and from the filtrate from the magnesium ammonium phosphate, the iron is precipitated with ammonium snlphide and treated as usual.The calcium carbonate is calculated to oxide ; the magnesium pyrophosphate to phosphoric anhydride ; to their sum is added the weight of the ferric oxide, and Electrolytic Estimations and Separations. By G . VORTMANN (Monatsh., 14, 536-552) .-The author describes the behaviour on electrolysis of salts of zinc, iron, cobalt, and nickel, to which solutions of an alkaline tartrate and an alkaline hydroxide have been previously added. Of these metals, the three first named may be quantitatively deposited from the alkaline solution ; whilst, under the same conditions, the last named remains in solution, thus affording a method for the separation of nickel from zinc, iron, and cobalt. Iron may be sepa- rated froni zinc in alkaline tartrate ,solution by employing a cathode of platinum, whereby the iron is deposited with only a trace of zinc.By dissolviiig and redepositing two or three times, the iron is obtained entirely free from zinc. To estimate eincin pi'esence of iron, potassium cyanide is first added to the solution, whereby the iron is converted into potassium ferrocyanide ; sodium hydroxide is then added and the current passed, when the zinc is deposited, the iron remaining iii solution. The authoi- further describes a method of estimating cobalt, nickel, and coppci. when present with much iron. A solution of the metal , i n which the iron is present as a ferric salt, is placed in a platinum dish and treated with excefis of ammonia, and the current passed, where- the deficiency represents the alumina.L. DE K.ANALYTICAL OHEMISTRY. 35 upon, wit,houf the necegsity of filtering off the ferric hydroxide, cobalt. nickel, and coppsr are thrown down on the cathode in a well-adhering mass. I n the separation of copper and iron, the latter metal is best oxidised wit,h nitric acid. G. T. M. Separation of Nickel from Cobalt. By H. HERRENSCHNTDT and E. CAPELLE (Zed. anal. Ohem., 32, 607-610; from Le Cobalt et le Nickel, Rouen, 1888).-The potassium nitrite process is not capable of giving a complete separation. Trace9 of nicks1 can be detected in presence of much cobalt by microscopic examination, in consequence of the different colonrs of cobaltic and nickelic hydroxides, that of cobalt being a pale brown, whilst nickelic hydroxide is black.On examining, with a magnifying power of 180 to 200 diameters, the precipitate thrown down by excess of sodium hypochlorite from the nearly neutralised solution, a completely black field will be seen if the nickel amounts to 1 per ccnt. of tlie cobalt, but when only traces are present, tbey are exhibited as black spots on a brown field. Traces of cobalt can be separated from much nickel by suspending the hydr- oxides in water or an alkali, and passing chlorine, when the nickel dissolves completely, the cobalt remaining undissolved. The practical method of estimating the two metals in ores, &c., is, therefore, as follows :-After removal of the metals precipitable by hydrogen sulphide and ammonium carbonate, an excess of ammonium sulphide is added, then a small excess of acetic acid, and the liquid is boiled.The precipitate is rapidly filtered off, and washed with hot ammonium acetate. It is then dissolved i n nitric acid, the solution evaporated to dryness, taken up with water, filtered from sulphur, and divided into two parts. One part is evaporated with sulpburic acid, ignited very gently, and the residue of meta,llic sulphates weighed. The other half is treated wihh potassium cyanide and bromine, and the well-washed precipitate treated wit>h chlorine. The solution containing the nickel is feebly acidified with hydrochloi-ic acid, b d e d to expel chlorine, and poured into boiling soda. The precipitate is washed, dissolved in hydrochloric acid, and converted into sulphate, as above, for weighing.When zinc is present, the mixed sulphates of nickel and zinc are dissolved, and the metals precipitate1 by hydrogen sulphide, after adding ammonium acetate and acetic acid. The pre- cipitate is treated with dilute hydrochloric acid, potassium cyaiiide added to the neutralised solution, and the zinc thrown down by potassium sulphide and weighed. The corresponding quzntity of sulphate is d-educted from the weight of the mixed snlphates. M. J. S. Estimation of Chromium in Ferrochromium and Steel. By J. CLARK (J. SOC. Chem. Ind., 11, 501-504; 12, 340--341).-FFinely pulverised ferrochromium, if rich i n chrornium, is very readily oxidised when heated with a mixture of magnesia and sodium hydr- oxide (2 parts of the former to 3 of the latter). The crucible con- taining the intimate mixture is heated for half an hour with the tip of a small bunsen burner flame, and subsequently for another half hour to dull redness.The contents are then boiled with water, hydrogen36 ABSTRAOTS OF CHEMICAL PAPERS. peroxide is added to reduce any manpanate, and, after boiling for sonlo minutes, the soliition is filtered. The insoluble matter may still contain chromium. It should be ignited, ground in an agate mortar, and again fused with the magnesia-sods mixture, and this fusion map even need to be repeated a third time. The solutions are acidified with sulphuric acid, and the chromic acid titrated by Penny’s process. For poor alloys, ignition in n porcelain crucible over the blowpipe or in a muffle with 3 parts of calcium hydroxide should precede the ignition with magnesia-soda mixture.I n this case about 3 parts of sodium hydrogen carbonate should be added to the solution before filtration, for the purpose of removing the calcium. Second and third fusions are required in this method also. The powdered ferro- chromium masy also be prepared for the fusion with magnesia-soda by a preliminary heating in sulphur vapour, or, better, vapour of carbon bisulphide, in a boat in a porcelain tube. The sulphides produced are very bulky, and, in the case of sulphur, a slight loss results from the violeuce of the action. The sulphides may be a t once mixed with 8 parts of magnesia-soda, or be first roasted in a porcelain crucible, and then fused with 5 parts of the mixture. Chromium steel is best attacked by the carbon bisulphide method. The residue from the first fusion with magnesia-soda will usually yield a very small amount of chromate on a second fusion. In cases where the alloy can be dissolved in dilute hydrochloric acid, the solution IS treated with ammonia until a permanent precipitate is produced, and then with a small excess of sodium phosphate and about double that amount of s d i u m thiosulphate or, preferably, sulphite, and boiled for 10 minutes. The precipitate, which contains all the chromium, is washed, gently ignited, ground, and fused with 8 parts of magnesia-soda. Equally good results are obtained when the hydrochloric acid solution is neutralised with sodium carbonate, and boiled with sodium sulphite as long as sulphurous anhydride comes off. The precipitate, which contains all the chromium as basic sulphite, is ignited and fused as above. Estimation of Cyanides in Gas Refuse. By W. LEYBOLD (Zeit. antcl. Clzem., 32, 571-57d).-The powdered substance is treated with soda to convert the cyanide into ferrocyanide, and a portion of the solution is evaporated to dryness with excess of sulphuric acid, and heated until the excess of acid is expelled. The ferric sulphatt. which remains is dissolved in sul1)huric acid, reduced with zinc, and titrated with permanganate. A possible objection that), in the presence of organic matter, iron might be dissolved by the alkali in other forms than as ferrocyanide, is met by the consideration t h a t there is always formed at, the same time an alkaline sulphide, by which such iron would be precipitated. M. J. S. M. J S.