24 THE ANALYST. INORGANIC ANALYSIS. The Determination of Platinum and Iridium in Platinum Ores. Leidie and Quennessen. (Journ. Pharm. Chim. , 1901, xiv., 351-355.)-The following modi- fication of LeidiB's method (ANALYST, xxvi., 108) is recommended when only the amount of platinum and available iridium in an ore are required : Five grammes of the ore are repeatedly extracted with hot aqua regia (nitric acid, specific gravity 1.32, 1 part; hydrochloric acid, specific gravity 1.18, 3 parts) until nothing more dissolves. The extract is evaporated to dryness at 1 0 5 O to 110' C., and the residue dissolved in a small quantity of water and filtered. The filtrate is then brought to about 70" C., and treated with sodium nitrite until the liquid becomes neutral to turmeric, when sodium carbonate is added little by little until no more precipitate is formed. The liquid is then filtered, after being heated to boiling, in order to remove the foreign metals.Sodium hydroxide is now added to the filtrate until the reaction is alkaline, and the osmium and ruthenium expelled as volatiIe peroxides by means of a current of chlorine, which is passed through the liquid maintained at 70" to 80' C.THE ANALYST. 25 The liquid is next neutralized with hydrochloric acid, and any chlorides that may have been formed reconverted into nitrites by the addition of more sodium nitrite. The solution will now only contain alkali salts and the ,double nitrites of platinum, palladium, iridium, and rhodium. Estimation of the Platinum.-At this stage in the original method the liquid was saturated with ammonium chloride, but, instead of this, from 30 to 35 per cent.of potassium chloride are introduced. The double nitrites of rhodium and iridium with potassium, which are precipitated, are insoluble in solutions of alkali chlorides. The nitrites of platinum and palladium in the filtrate are converted into chlorides with hydrochloric acid, and the mass dissolved in boiling water. A slight excess of formaldehyde is now added, and the metals precipitated on boiling the liquid are ignited, reduced in a current of hydrogen, and dissolved in aqua regia. The residue left on evaporation of this solution is dissolved in water and treated with a reducing agent, to convert palIadic chloride into palladous chloride, after which the platinum can be precipitated with ammonium chloride, and estimated in the usual way.Estimation of the Iridium.-In a portion of liquid from which the osmium and ruthenium have been expelled, the double nitrites of rhodium and iridium are con- verted into chlorides by means of aqua regia containing but little nitric acid. The iridium is then precipitated by means of ammonium chloride, and the precipitate washed with ammonium chloride solution, ignited, reduced in hydrogen to the metallic form, and weighed. The iridium could also be precipitated by means of potassium chloride as the double chloride 2KCLIrC1,. The potassium chloride is removed by washing with water, and the residue reduced, dried at 115" C., and weighed. The authors point out that this method only gives the available iridium, since any which is present in the free state or in combination with the osmium is insoluble in q u a regin, and is left in the residue in the preliminary extraction.C. A. M. The Estimation of Tin in White-metal Alloys. Frederick Ibbotson and Hmry Brearley. (Chem. News, 84, 167.)-Tin may be readily estimated by reducing the hot solution of the chloride, cooling in an atmosphere of GO,, and titrating with iodine and starch. The reduction can be conveniently effected by means of iron, but any excess added must be dissolved completely. If antimony is present, it will be precipitated as metal, and cannot be dissolved again, but as cold acid solutions of stannic chloride are not reduced by antimony, which readily reduces them on heating, the solution may be directly titrated with iodine as usual, very good results being obtained.The antimony may be filtered off after the titration and estimated, but the results obtained are rather low. An improvement on the above method consists in reducing the stannic chloride with finely-powdered metallic antimony. The reduction of 0-15 gramme of tin is complete after one minute's boiling, and the excess of antimony which remains undissolved acts as a safeguard during the cooling, since it reduces any tin which may have become oxidized whilst the solution is still hot. Cold solutions of stannous26 THE ANALYST. chloride take up oxygen less readily. The test analyses given show that the reduction is complete. The influence of various substances likely to be present in an ordinary analysis on the above method was examined, and it was found that the presence of iron, chromium, nickel, zinc, manganese, aluminium, bismuth, phosphorus, and sulphur is without effect on the results.The quantity of hydrochloric acid present should always be about one-fifth of the total volume. If copper is present, it will be reduced to the cuprous state, but accurate results may nevertheless be obtained if the iodine is added drop by drop to the vigorously agitated solution, so as to prevent the formation of a local excess of iodine. It is also advisable to have rather more hydrochloric acid present, up to about one-third of the total volume. Cobalt apparently gives very slightly high values. Lead is without influence if sufficient hydrochloric acid is present to prevent the formation of lead iodide.The presence of arsenic completely vitiates the results, whether the tin is reduced with iron or with antimony. Mercury is reduced to the metallic state, but is not oxidized in cold solutions. If molybdenum or tungsten is present, a coloured lower oxide is formed, but this is not appreciably reoxidized by the iodine, and the starch blue can be readily distinguished. A. G. L. The Determination of Sulphur in Iron and Steel. William A. Noyes and L. Leslie Helmer. (Journ. Amer. Chem. Xoc., xxiii., 675.)-The authors recom- mend the following two methods as being more rapid and accurate than those now in use : I n the first, 5 grammes of the steel are added in portions at a time to 200 C.C. of water and 8 C.C.of bromine contained in a flask, cooling after each addition. The solution is finally boiled for a moment to expel excess of bromine, filtered, the filtrate added to 130 C.C. of 10 per cent. ammonia contained in a 500 C.C. flask, the liquid made up to the mark, and well mixed. Three hundred C.C. are then filtered off, evaporated in a wide beaker to 100 c.c., and precipitated with 10 C.C. of barium chloride solution, after the addition of 1 drop of hydrochloric acid of specific gravity 1.12. After the precipitate has settled, it is filtered off, ignited, and weighed as usual. The filter-paper containing the original insoluble residue is burnt with some sodium carbonate, the oxidation being completed by adding a little potassium nitrate. The melt is dissolved in water, the solution filtered, acidified with hydro- chloric acid, and precipitated with 5 C.C.of barium chloride solution. Three-fifths of the weight of barium sulphttte found here is added to the main quantity of barium sulphate found, and the total represents the sulphur in 3 grammes of iron. I n the second method, 5 grammes of the iron, mixed with 7 grammes potassium chlorste, are added in portions at a time to 120 C.C. nitric acid (specific gravity 1-20) and 1 gramme potassium bromide, cooling if necessary. When solution is complete, the liquid is evaporated to dryness in a dish, the residue dissolved in hydrochloric acid and filtered after diluting, both filtrate and residue being treated as in the first method. I n the bromine method, the amount of sulphur left in the insoluble residue is larger than in the nitric acid method, but no evaporation to dryness is required, and practically no fumes are given off during solution.Test experiments, in which known amounts of ammonium ferrous sulphate wereTHE ANALYST. 27 added to sulphur-free iron, and to steel containing known amounts of sulphur, gave satisfactory results. The methods also gave concordant and higher results than Blair’s aqua regia method, but the values obtained by using the latter are rather irregular. Since the quantities of sulphur left in the residues will probably prove nearly constant for any given class of steel, the authors believe that it will be possible to add an empirical correction to the quantity found in the solution, thus shortening the process very considerably.A. G. L. Ammonium Persulphate as a Substitute for Lead Peroxide in the Colori- metric Estimation of Manganese. Harry E. Walters. (Proc. Engiizeers’ XOC. , Western Pennsylvania, xvii., 257.)--In the following method the manganese is oxidized by means of ammonium persulphate in the presence of a silver salt : Samples of 0.2 gramme each of the test-steel and of a steel of known percentage of manganese are placed in test-tubes, and 10 C.C. of nitric acid (specific gravity 1.20) added to each. The tubes are heated in a water-bath till solution is complete and nitrous fumes are driven off. Fifteen C.C. of a solution of silver nitrate containing 0.02 gramme of AgNO, are added, and then immediately 1 gramme of solid ammonium persulphate, and the tubes again warmed, when oxidation will commence.After mother half-minute’s heating the tubes are placed in a cold-water bath. After oxidation is over and the tubes are cool, they are compared as usual. Pig-irons and slags are treated in the same way as steels, except that slags should be moistened before adding the acid, and the solution filtered from suspended matter. In this case, the 15 C.C. of silver nitrate solution are hsed to wash the filter- paper. In dissolving shot or chilled iron samples, a little persulphate may be added to aid the solution of the sample. The advantages claimed for the method are that it does away with the decanta- tion of the solution from the excess of lead dioxide, that less acid is used, and that the oxidation is effected more rapidly at a lower temperature.The results quoted show that the two methods furnish practically identical results. I n experimenting with the method, it was found that solutions of ammonium persulphate could not be used instead of the solid, and also that if the salt was perfectly dry the reaction was incomplete. The author makes it a practice to add 10 C.C. water to each 1 lb. bottle of the dry salt a day or two before use. A. G. L. Detection of Small Quantities of Nickel in the Presence of Cobalt. H. Ditz. (2eits.f. angew. Chem., 1901, 894.)-Advantage is taken of the property of tartaric acid to prevent the precipitation of cobalt ; but, instead of precipitating with hydrogen sulphide, as in Villiers’ process (ANALYST, 1900, 305), the nickel is precipitated as chromate. The neutral solution is placed in a capacious flask, and potassium chromate is added in slight excess.It is heated nearly to boiling, and 5 to 10 grammes of Rochelle salt are added. After this is all dissolved, the liquid is boiled vigorously for several minutes; then it is allowed to cool and diluted with water, if it be28 THE ANALYST. coloured very intensely. Any cobalt chromate precipitated is thus brought into solution again; but if nickel be present, a brown precipitate settles to the bottom of the flask, where it can be readily detected in the green liquid, even though the amount of nickel be very small. The precipitate contains a small proportion of cobalt, A. M. The Quantitative Separation and Determination of Uranium. Edward I?.Kern. (Journ. Amer. Chem. SOC., xxiii., 685.)-As the result of his investigation, the author finds that the separation of uranium from iron, nickel, cobalt, zinc, and manganese is best e€fected by means of either a, saturated solution of sodium carbonate, in which the uranium is completely soluble on prolonged boiling, or else by extracting the iron with ether and separating uranium from the other metals by m a n s of a saturated solution of ammonium carbonate. In this method, which is to be preferred to the first, the solution of the mixed chlorides in hydrochloric acid (of specific gravity 1-10> is extracted three times with ether free from alcohol, which has previously been shaken with hydrochloric acid of the same density. From the alkalies and alkaline earths, uranium may be separated by electrolysis in the presence of sodium acetate and acetic acid ; by repeated precipitation with ammonis, in hot solutions containing ammonium chloride ; or by precipitation with ammonium phosphate in the presence of ammonium acetate, the precipitate in this case being boiled for at least fifteen minutes before filtering.The pyrophosphate should be heated to low redness in a porcelain crucible; should reduction occur, it may be remedied by moistening with strong nitric acid and reigniting. The author also finds that the reduction of U,O, to UO, is unreliable as a, check, and that the most rapid method of determining uranium consists in reducing the sulphate solution to U(SO,), with zinc, aluminium, or magnesium, and titrating with potassium permanganate.In hydrochloric acid solution, or with stannous chloride as a reducing agent, the reduction is carried beyond UCl,, sometimes reaching UCI,, and the results are consequently utterly unreliable. The preceding methods were tried practically in estimating uranium in pitchblende, those based on the ether separation giving the best results. A. G. L. The Precipitation and Separation of Silver Electrolytically. W. H. Fulweiler and Edgar F. Smith. (Joarn. Amer Chem Soc., xxiii., 582.)-A series of experiments was made in which silver was deposited electrolytically from cyanide solutions (2 to 5 grammes KCN per 125 c.c.) containing, in addition to the silver, copper, copper and cadmium, copper, cadmium and zinc, and copper, cadmium, zinc, and nickel.The time required to heposit 0.1 gramme of silver at a temperature of 55" to 80" C . was three to six hours, using a current of about 0.02 amperes for a surface of 100 square centimetres, the voltage being 1.2. The results obtained were satisfactory ; but if both cadmium and zinc are present, it appears to be essential to heat the liquid to 75" or 80" C. before passing the current to prevent the precipitation of some of the cadmium, and the presence of nickel slightly retards the deposition of the silver. A. 0. L.THE ANALYST. 29 The Electrolytic Separation of Mercury from Copper. C. Boscoe Spare and Edgar F. Smith. (Jourrc. Amer. Chem. SOC., xxiii., 579.)-In reply to some cnticisrns on this method the authors give results of a series of experiments in which mercury was separated from copper electrolytically in potassium cyanide solution (2 to 6 grammes KCN per 125 c.c.), the quantity of copper present being from one-third to eleven times the weight of the mercury.The current used was 0.1 to 0.6 amperes for a surface of 125 square centimetres, the voltage being 1.1 to 1.7, the temperature of the liquid 60" to 65" C., and the time required to deposit 0.12 gramme of mercury, two to five hours. All the results obtained were satisfactory, and the deposit was always perfectly free from copper. Some experiments in which mercury was deposited from solutions containing copper and cadmium, and copper, cadmium and zinc, also gave good results. A. G. L. The Electrolytic Method applied to Uranium. L. Gavit Kollock and Edgar F.Smith. (Joum. Anzer. Chem. SOC., xxiii., 607.)--The authors obtained good results by electrolyzing solutions of uranium acetate containing a little free acetic acid, as well as solutions of uranium sulphate and nitrate. The uranium is deposited as hydrated protosesquioxide, ignited and weighed as U,O,. The dilution used was 125 C.C. for 0.1 to 0.2 gramme U,O,, the temperature 65" to 75", and the duration of each experiment four to seven hours. The current employed for 107 square centi- metres of cathode surface was 0.05 to 0.55 amptires and 4 to 16 volts in the case of the acetate solutions, and 0.02 to 0.04 amperes and 2 to 4 volts for the sulphate and nitrate solutions. Good results were also obtained in separating uranium from barium, calcium, magnesium, and zinc in solutions of the acetates, provided that evaporation of the liquid was prevented as far as possible, as otherwise the uranium hydrate tends to enclose the separated salts, which cannot then be washed out. This method, however, cannot be used to separate uranium from iron, nickel, or cobalt.A. G. L. The Electrolytic Determination of Molybdenum. L. Gavit Kollock and Edgar I?. Smith. (Journ. Anzer. Chem. SOC., xxiii., 669.)-The authors found that no deposition took place on passing a, current through neutral solutions of sodium molybdate heated to 75" C. On adding 2 drops of concentrated sulphuric acid, however, the liquid turned dark blue, and then, as this colour gradually disappeared, the cathode became coated with a black deposit of the hydrated sesquioxide.Concordant results could not be obtained by merely washing and drying this deposit, but very satisfactory results were found by dissolving the deposit in dilute nitric acid, evaporating the solution to dryness, and heating the residue of molybdic acid. The electrolyte contained 0.1 to 0.25 gramme MOO, in 125 C.C. ; the current used was 0.02 to 0.04 ampere and 2 volt8 for a cathode surface of 107 square centi- metres ; and the time required for the deposition was two to four hours. It was also found that, instead of acidulating thle solutions with sulphuric acid, acetic acid (1 C.C. of 29 per cent.) could be used; the time required in this case is rather longer, and it is best to add the acetic acid gradually. Since the presence30 THE ANALYST. of sodium acetate was found not to interfere with the electrolysis, the method was applied to the analysis of molybdenite as follows: The mineral was fused with sodium carbonate and nitrate, the melt dissolved in water, the solution filtered from insoluble oxides, and the filtrate electrolyzed after acidulating with acetic acid and boiling off the carbonic acid.In the liquid left after the electrolysis, the sulphur was estimated as barium sulphate. The results obtained were fairly satisfactory. A. G. L. On the Solubility of certain Metallic Oxides in Solutions of Sodium or Ammonium Salicylate. (Zeit. anal. Chem., 1901, XI., 459-462. j-Con- centrated solutions of sodium or ammonium salicylates possess the power of dis- solving freshly-precipitated metallic oxides, the amount dissolved depending on the nature of the oxide.Copper hydroxide is exceedingly soluble even in the cold. Iron hydroxide and aluminium hydroxide, however, do not dissolve until the temperature reaches 80" C., and are both much less soluble than copper hydroxide. On adding a slight excess of sodium hydroxide to the solutions in sodium salicylate, the red iron solution becomes lighter in colour, but there is no change in the copper or aluininium solutions. In the presence of a large excess of sodium hydroxide the iron solution gives a precipitate of ferric hydroxide, the copper solution changes from emerald green to bright blue. A slight excess of ammonium hydroxide does not change the appearance of m y of the ammonium salicylate solutions, whilst a large excess of ammonia produces a blue coloration in the copper solution, but does not give any precipitate with either the iron or aluminium solutions.J. Wolff. Hydrogen sulphide completely precipitated the iron as sulphide. The solubility of copper hydroxide in concentrated sodium salicylate solution depends upon the formation of copper-sodium salicylate, which the author has obtained in a crystalline condition. A solution ol this salt is reducible by glucose. OH C,H, - COONa The author explains the formation of this compound by the following equation : 2C H +cu< = ' 0 4 \ ~ ~ OH ! 3- 2H,O. c u I /O C,H, - COONa. C. A. M. Estimation of Chlorates in Electrolytic Bleach and Chlorate Liquors. H. Ditz. (Clzem. Zeit., 1901, xxv., 727.)-The process which the author has already described (ANALYST, 1900, sxv., SO) for the iodometric analysis of mixtures of chlortltes and hypochlorites has been reinvestigated and modified slightly to suit the purposes indicated in the title.The apparatus remains 8s before, but the main bottle should hold 1.5 litres, and the washing flask, which is now a cylinder, rtboutiTHE ANALYST. 31 20 C.C. I t has been found that the one hour’s standing is not necessary, the reaction being finished well within five minutes. The volume of strong hydrochloric acid must not be reduced below 50 c.c., and 500 or 600 C.C. of diluting water should be employed in order to hinder any action between that acid and the potassium iodide. The method now runs as follows : If a chlorate liquor is to be analysed, a volume containing about 0.1 gramme of KClO, should be taken; if a determination of chlorate in mixtures of chlorates and hypochlorites, the volume should be such that between 40 and 50 c.c, of decinormal thiosulphate shall be required in the final titration. If, in the latter case, 25 C.C.or less of the liquor are sufficient, 50 C.C. of strong hydrochloric acid are employed; if the liquor is weaker and more must be taken (e.g., 50 c.c.), 100 C.C. of acid and a correspondingly larger quantity of diluting water are used. The liquor is brought into the flask with 10 C.C. of 10 per cent, potassium bromide solution, and after the washing tube has been two-thirds filled with 5 per cent. potassium iodide, the 50 C.C. of strong HC1 are run in. The stopper of the washing tube is immediately inserted, and the apparatus allowed to rest for five minutes.Then 500 or 600 C.C. of water are introduced and 20 C.C. of 5 per cent. potassium iodide. After thorough shaking, the contents of the side tube are blown and rinsed back into the flask, and the liquid is titrated with thiosulphate. Hypo- chlorites must be determined separately, and their proportion deducted. Commercial chlorate liquors can be run hot into the flask as soon as the free chlorine has been boiled off, and the whole process can be finished in half an hour. It is quicker than, and quite as accurate as, the ferrous sulphate method, which latter cannot be used if much hypochlorite is present, and in which the reagent needs standardizing every day. F. H. L. Estimation of Phosphoric Acid in Soil by the Gotz Process. E.Gully. (Chew. Zeit., 1901, xxv., 419.)-By making certain modifications in this process, the author has rendered it simpler in operation, capable of dealing with larger quantities of phosphoric acid, and more suitable for the analysis of soils. The vessel in which the mechanical agitation is conducted is about 17 centimetres long, holding 65 to 70 C.C. The narrow part holds altogether 0.4 c.c., is 40 to 45 millimetres long, and is graduated into eighty equal divisions, two of which contain 5 cubic millimetres, and hold ammonium phospho-molybdate equivalent to about 0.23 milligramme of P205. The whole graduated tube thus holds a precipitate equal to 18.45 milli- grammes of P,O, ; but as Von Juptner has already shown in the case of iron analysis, the yellow precipitate does not consolidate itself evenly in the tube, which must be calibrated with mercury, and checked on a large number of blank experiments with pure disodium phosphate.(Gully appends to his paper a table of the readings of a tube in which 80 divisions are equal to 18-43 milligrammes of P,O,, whereas 10, 20, and 40 are equal to 2.22, 4.42, and 9.10 milligrammes respectively.) In carrying out an analysis of soil, a quantity of the sample (from 2 to 10 grammes), which contains less P,O, than the maximum capacity of the tube, is brought into acid solut.ion, and the silica removed. Hydrochloric acid is driven off by repeated evaporations with strong nitric acid, the dry residue is taken up in 1-19 nitric acid, and the solution is diluted to 50 C.C.with the same liquid. The graduated portion32 THE ANALYST. of the agitation tube is filled with 25 per cent. ammonium nitrate solution, carefully avoiding air bubbles, 25 C.C. of the soil solution are pipetted in, another 10 C.C. of ammonium nitrate are introduced, and the whole is placed for at least ten minutes in a water-bath at 65" C. It is next taken out, 25 C.C. of Finkener's* molybdate solution are added, and thoroughly agitated (by hand) for one minute, watching that the liquid in the graduated portion of the tube mixes with the rest. The tube is then filled up with the ammonium nitrate solution, and put back in the water-bath, where it should remain immersed up to its neck for twenty minutes. I t is next taken out, adjusted inzthe centrifugal machine, and revolved for four minutes at a speed of 1,100 to 1,200 revolutions per minute.The volume occupied by the pre- cipitate is finally read off and calculated into P,O,. If it is preferred to use less than 25 C.C. of the soil solution for precipitation, that quantity of fluid must be made up by adding the requisite amount of nitric acid. It is necessary that the nitric acid should have the specific gravity specified, and that the prescribed volume of it should be present; in all other respects, notably the temperature of the water-bath, adherence to the details given above is essential. To show the accuracy of the process, Gully gives a table of sixty-five moorland soils containing from 0.03 to 0.33 per cent. of P,05 when analysed by the ordinary gravimetric method, the measurement figures for which show maximum differences of +0.009 and -0.01 per cent.; and he points out that, when the discrepancy is larger, the fault is probably due to the gravimetric process, which contains greater possibilities of error than the modified Gotz method.F. H. L. Determination of Phosphoric Acid as Phospho-Molybdic Anhydride. A. Seyda. (Chew. Zeit., 1901, xxv., 759-768.)-The investigations described with great minuteness in this lengthy article fall into three divisions, including attempts to simplify the Meineke-Woy process (ANALYST, 1897, xxii., 250 and 333) for determining phosphoric acid with molybdenum only, a study of the effect of adding citric acid to the molybdate reagent in order to obtain a precipitate free from uncombined molybdic acid, and the elaboration of a method of converting the Wagner-Stutzer combined molybdenum-magnesium process into one in which the magnesian precipitation is avoided.The results of the various experiments prove : (1) In the determination of phosphoric acid as ammonium phospho-molybdate, the only source of error is a contamination of the precipitate with free molybdic acid-Woy's idea, that '' alkali " may be present being quite incorrect. (2) Using excess of molybdate solution con- tamination of the first precipitate with molybdic acid cannot be avoided with absolute certainty-whence it follows that no mere agitation process can be considered universally applicable. (3) If the phosphoric acid is only precipitated once, such contamination is most likely to be prevented by mechanically agitating the mixed liquids for fifteen minutes at the atmospheric temperature, while an addition of 20 C.C.of 10 per cent. citric acid solution to the whole is advisable. In this case, if iron is present, a temperature of 20° C. must not be exceeded, and the liquid must be filtered .It One hundred and sixty grammes of ammonium molybdate dissolved in 795 C.C. of cold water, and 320 C.C. of 0.925 ammonia, poured into a mixture of 1,710 C.C. of 1.2 nitric acid with 1,205 C.C. of water, preventing any rise of temperature by constant cooling.TRE ANALYST. 33 fifteen minutes after the end of the agitation. In the absence of iron the temperature may be allowed to rise to 30" C. (4) The precipitate is best purified from molybdic acid by dissolving it in ammonia, adding 50 to 100 C.C.of " dilute molybdate " (see below) and reprecipitating hot with nitric acid. In certain circumstances the process of purification must be repeated several times; if so, the strengths and volumes of the reagents employed must be kept as specified. ( 5 ) The precipitate can be filtered hot, and washed with The filtrate and washings should be preserved for twenty-four hours to note any imperfect precipitation of phosphoric acid. (6) The precipitate should be rinsed down from the sides of the Gooch crucible as it creeps up with 95 per cent. alcohol. (7) The yellow precipitate is only to be considered as completely converted into phospho-molybdic anhydride when it exhibits a uniform black and crystalline appearance all through.The reagents required for the '( Wagner-Meineke-Woy " process are as follows : ( l a ) Wagner-Stutxer Molybdate : 150 grammes of ammonium molybdate are dis- solved in about 600 c . ~ . of warm water in a 2-litre flask; when cool, 1 litre of 1.19 nitric acid is introduced, and to the clear mixture 400 grammes of solid ammonium nitrate are added in small portions at a time, shaking repeatedly. The liquid is diluted to the mark, kept for several days and then filtered. So prepared, separation of molybdic acid will not occur. ( l b j Citric Acid Molybdate: 10 grammes of pow- dered citric acid are dissolved in 1 litre of the above and filtered after twenty-four hours. (2) Nitric Acid: 25 per cent., specific gravity 1.15; stored in a bottle with a rubber cork bearing a 20 C.C.pipette. (3) Ammonia: 8 per cent., specific gravity 0-967; stored in a bottle with a 10 C.C. pipette. (4) Dilute Molybdate: a solution containing 0.1 per cent. of ammonium molybdate and 10 per cent. of ammonium nitrate, in a wash-bottle. ( 5 ) Washing Liquid: 5 per cent. ammonium nitrate and 1 per cent. nitric acid. (6) AZcohoZ: 95 per, cent. in a wash-bottle. The original solution in which phosphoric acid has to be determined is pre- cipitated with the molybdate according to Wagner's directions, placing the beaker for twenty or thirty minutes on the water-bath and then letting it settle. [It would appear that 60" C. is the critical temperature" for the deposition of molybdic acid, and the liquid should be kept below that point.The rule applies even when citric acid is present, as that reagent loses its inhibitory power at the same temperature.] When clear, the liquid can be either filtered at once, or safely left for hours, even overnight. The precipitate is collected on a paper, well washed with the washing liquid, and the bulk of it is rinsed into a wide flat beaker with the dilute molybdate solution, which is ready in a wash-bottle. The residue on the paper is then dissolved by allowing 10 C.C. of the ammonia to run over its surface, but the paper is preferably not perforated. Finally the filter is washed three times with cold dilute molybdate. The ammoniacal solution should be quite clear, and always is so, as if the molybdate reagent is properly prepared iron is never thrown down with the yellow precipitate. The beaker is covered with a glass, placed on gauze, heated till the liquid begins to boil, removed from the flame, and 20 C.C.of the nitric acid are added boiling hot. After two or three gentle agitations the vessel is stood on the water-bath till the solution is clear, and then set aside. I t is ready for filtration in five or ten minutes. washing liquid" at 60" or 80" C. This reagent is much more permanent than the foregoing.34 THE ANALYST. Filtration is done in a Gooch crucible, the precipitate being with hot washing liquid. Ignition is performed in a modified described by Woy. collected and washed form of the apparatus F. H. L. The Examination of Enamel and the Fusibility of Silicates. E. Kochs and F. Seyfert.(Zeits. f. angew. Chem., 1901, 719.)-The authors give a rule for calculating from their composition a fusibility coefficient for the silicates used for enamelling iron. The ratio of the number of molecules of alumina to the number of molecules of flux is divided by the ratio of the number of molecules of silica to the number of molecules of alumina,. Or, expressing the rule algebraically : where q = function of fusibility. T = per cent. Al,O,. K = ,, ,, SiO,. F=sum of the figures obtained by dividing the percentage of the The higher the function q is, the higher is the melting-point of the enamel. With the aid of this formula it is possible to calculate the addition or additions which must be made to raise or lower the function to any required extent. Examples are given of a number of actual determinations and experiments.A. M. other (flux) constituents each by its molecular weight. The Behaviour of Borax when Distilled with Methyl Alcohol. E. Polenske. (Arb. Kaiserl. Gesundheit, 1900, xvii. , 564-568; through Zeit. fur UTatersuchung. der Nahr. und Genussmittel, 1901, iv., 801.)-When borax is distilled with methyl alcohol, 57 to 59 per cent. of the boric acid contained in the borax passes over in the distillate; about 50 per cent. comes over readily, the remainder slowly. From the methyl alcohol remaining in the distillation-flask, sodium metaborate, NaBO, + 5CH30H, crystallizes out. It was found that on fusion 1 molecule of borax would drive out the carbon dioxide from 3 molecules of sodium carbonate. The melt obtained by fusing together 1 molecule of borax and 1 molecule of sodium! carbonate, and which had the composition of sodium metaborate, when distilled with methyl alcohol, gave a considerable quantity of boric acid in the distillate.Distillates from the melts containing more sodium carbonate were free from boric acid. When the different melts were dissolved in methyl alcohol, the substance NaBO, + 5CH,OH always separated out. Sodium metaborate, on distillation with methyl alcohol, gives off boric acid until the residue has the composition Na,,B,0,7( = 5Na,0 + 4B20,). This residue must be considered as a mixture of sodium metaborate and soda. The latter prevents the formation of boric esters and the possibility of distillation. w. P. s. Determination of Boric Acid. H.Luhrig. (Phclrm. Centralh., 1901, xlii., 50-56 ; through Zeit. fiir Untemuch. Nahr. und Geizussmittel, 1901, iv., 801, 802.)-THE ANALYST. 35 The author confirms the accuracy of Jorgensen’s method for the estimation of boric acid (see also the ANALYST, 1900, xxv., 101, 102). By combining Gladding’s method with that of Jorgensen’s, less accurate results were obtained than with the latter done. w. P. s. The Quantitative Determination of Ozone. A. Ladenburg and R. Quasig. (Ber. deutsch. chern. Ges., xxxiv., 7, 1184.)-It was recently pointed out by Laden- burg (Bericlzte, xxxiv., 631) that the present methods for the estimation of ozone suffer from the want of a check, and are consequently unreliable. It was also shown (Zoc. cit.) that by direct weighing of the ozone, even when mixed with oxygen, such a check could readily be obtained.Evidently this method ihelf might be used for the estimation of ozone, but it is not generally applicable, as it demands certain condi- tions which cannoi always be fulfilled in practice. By its help, however, it is now possible to determine the reliability of the various titrimetric methods in common use, and the examination of the potassium iodide method forms the subject of the present paper. The authors reserve to themselves the right to examine other methods in the same way later on. I n the potassium iodide method it is usual to lead the gas through the sqaeous solution of the salt acidified with an equivalent amount of sulphuric or acetic’acid, and titrate the liberated iodine with sodium thiosulphate solution.Proceeding in this way, the authors always obtained results about 50 per cent. too high, and consequently adopted the method in which the ozone is led through a neutral solution, which is only acidified just before the titration. This method gives very accurate results and has the advantage that there is no return of the blue colour, which always occurs in the first method after the lapse of several hours, and necessitates the addition of a further considerable amount of sodium thiosulphate. This result is quite contrary to that found by Brunck (Berichte, xxxiii., 1835) and to the current opinion. The high results obtained by the first method may be due to a catalytic action of the ozone on the oxygen and the solution of potassium iodide, or else to the formation of hydrogen peroxide during the reaction according to the equation- which represents the quantities found almost quantitatively. The actual determinations were carried out as follows : Dry oxygen was passed through a glass bulb of a capacity of nearly iJ litre, provided with two stopcocks, until the weight of the closed bulb remained constant.Ozone was then passed through for about twenty minutes, and the bulb again weighed. The difference between the two weighings, multiplied by 3, gives the weight of ozone present. The bulb was then connected to two wash-bottles containing potassium iodide solution, and the ozone slowly sucked through them by means of a pump. When a neutral solution of potassium iodide was employed, only a trace of iodine was found in the second wash-bottle.An equivalent quantity of sulphuric acid was then added, and the liberated iodine titrated in the usual way. 40, + lOHI + H,O = 51, + H,O, + 5H,o + 30,, A. G. L.36 THE ANALYST. On the Determination of Hydrogen in Gas Mixturas. Francis C. Phillips. (Journ. Amr. Chem. SOC., xxiii., 355.)-In burning a mixture of hydrogen and oxygen by means of palladium asbestos, it is usual to heat the tub? containing the asbestos in order to prevent con- densation of the water formed. If a naked flame is used for this purpose there is danger of burn- ing hydrocarbons with the hydrogen. In the form of appsratus shown, the heating is conveni- ently effected by means of boiling water contained in the cup-shaped vessel b, b, 18 centimetres deep and 12 centimetres wide, which rests on three supports riveted to the inside of the brass cylinder a, a, open at both ends, 30 centimetres long and 18 centimetres wide, the rims being strengthened by wires.One of the supports is shown at c. The cup is silvered on the outside and the cylinder on both sides. The polished silver surfaces serve to prevent radiation of the heat to the neighbouring gas-measuring apparatus. The cup is heated by a small flame placed underneath. The tube containing the asbestos has a total length of 70 centimetres and an internal diameter of 3 millimetres. A. G. L. Method for Preparing Strictly Tenth-Normal, Fifth-Normal, eta, Hydro- chloric or Nitric Acid. (Journ. Amer. Chem. Soc., xxiii., 343.) -The author has applied his method of making standard sulphuric acid (Journ.Amer. Chem. Xoc., xxiii., 12) to the preparation of standard hydrochloric and nitric acid. To prepare decinormal hydrochloric acid, 12,487 grammes pure crystallized copper sulphate are dissolved in 500 C.C. of water, the copper is precipitated electro- lytically, and the solution transferred to a 1-litre graduated flask. A solution containing exactly 12-215 grammes crystallized barium chloride is then added, and the solution made up to the mark ; after adding 2.6 C.C. of water to make up for the volume occupied by the precipitate, the whole is mixed, allowed to settle, and the clear liquid siphoned off and filtered through a dry filter. To prepare decinormal nitric acid, 13.076 grammes barium nitrate are sub- stituted for the barium chloride. Solutions prepared in this way were found to be of exact strength and free from barium, but contained small quantities of sulphuric acid (up to 0.0320 gramme per litre).For many purposes it might be better to add a small excess of barium salt to avoid this. RichardR. Meade. (Cf. ANALYST, xxvi., 51.) A. G. L. Methods of Standardizing Acid Solutions. Cyril G. Hopkins. (Joum. Amer. Chem. Soc., xxiii., 727.)-The object of the investigation was to find the most exact method of standardizing solutions of hydrochloric or sulphuric acid. In the silver chloride method used, the hydrochloric acid was precipitated by a very slight excess only of silver nitrate, the precipitate filtered on a Gooch crucible and dried at 130" to 150" C. In the ammonium sulphate method excess of ammonia was addedTEE ANALYST. 37 to the sulphuric acid, the solution evaporated, and the residue dried at 120" C. Both these methods gave excellent results. For the standardization by means of sodium, the latter was weighed out under petroleum in a weighing-bottle, dissolved in alcohol, and, after diluting the solution with water, titrated with the acid solutions, using phenolphthalein as indicator. The results agreed well amongst themselves, but were higher than those obtained by the first two methods, owing, probably, to impurities i n the sodium. Incidentally they provided a means of comparing these values wit4 each other, and showed them to be in excellent agreement. The borax method, in which air-dried crystallized borax is titrated with the hydrochloric acid, using dimethyl orange as indicator, gave values which showed that the borax lost water on continued exposure to the air. This was verified by exposing the salt to air and weighing from time to time. Consequently the method is not reliable. Values obtained by electrolyzing solutions containing known weights of copper sulphate, titrating the acid liberated with sodium hydrate solution, and using this to standardize the hydrochloric acid, agreed well amongst themselves, but still were not nearly so good as those obtained by the first two methods, the maximum varia- tion of the results found in the copper sulphate method 'being 0.048 milligrammes HC1 in 1 C.C. solution, in the silver chloride method 0°002 milligramme HCI, and in the ammonium sulphate method 0.004 milligramme H,SO, in 1 C.C. solution respectively. Consequently the author concludes that these two methods are by far the most suitable for standardizing acid solutiom'. A. G. I;.