THE ANALYST. INORGANIC ANALYSIS. The Iodometric Determination of Gold in Dilute Solutions. Ralph N. Max- son. (Zeits. Anorg. Chem., xxxvii., 80.)--The author has made a number of experi- ments, comparing the relative accuracy of Gooch and Morley’s method (Am. JO~~YPZ. Sci. SiZZ., viii., 1899, 261), which depends on the reaction 3KI + AuCl, = 3KC1+ AuI + I,, with that proposed by Rupp (Berichte, xxxv., Z O l l ) , in which an excess of arsenious oxide is allowed to act on the gold chloride solution according to the equation- 3As20, + 4AuC1, + 6H,O = 3As205 + 12HC1+ ~ A u , the excess of As,O, being then titrated with iodine. He finds that in the last method the gold is only precipitated if sufficient bicarbonate is present to neutralize all the acid, and shows that the first method is by far the more accurate, the error being only about one-tenth of that occurring with the second method when similar quantities of gold (from 9 to 0.05 milligrammes) are being determined.A. G. L. The Detection of Traces of Lead and Manganese. A. Trillat. (A?z?z. de Chim. anal., 1903, viii. , 408-410.)-The tetramethyl base of diphenylmethane, CH,[C,H,N(CH,),],, gives in acetic acid solution a bright blue colouration with certain metallic peroxides (lead, manganese and copper), due to the formation of the corresponding hydrol, CH.OH[C,H,N (CH3),I2. The reagent is prepared by heat- ing for an hour on the water-bath a mixture of 30 grammes of dimethylaniline, 10 grtlmmes of formaldehyde, and 200 C.C. of water rendered acid with 10 grammes of sulphuric acid. After cooling, the liquid is treated with a large excess of sodium hydroxide, and the excess of dimethylaniline expelled by means of a rapid current of steam. The crystals that form on now cooling the liquid are purified by a single recrystallization from alcohol, 15 to 20 grammes being eventually obtained.For the preparation of the reagent 5 grammes of the crystals are dissolved in 100 C.C. of water containing 10 grammes of pure acetic acid, and this solution is kept in a stoppered bottle in the dark. I n testing for lead the substance is converted into sulphate azd treated with a hot saturated solution of sodium hypochlorite (2 drops to 0.01 gramme of dry residue). The chlorine is expelled by ignition, the reagent added directly to the residue in the dish, and the colour produced compared with that given by water containing a known quantity of lead peroxide in suspension.The test is capable of detecting 1 part of lead peroxide in 3,000,000, and a clear reaction is obtained with rain-water that has been left for forty-eight hours in contact with granulated lead. To insure the absence of manganese in certain cases, all that is necessary is to wash the ash after formation of the sulphates. In applying the test to manganese, the ash of the substance should be treated with pure sodium hydroxide to form the peroxide, and the test continued in the same way as described above. I t is necessary to prove that copper is absent. By this means the author has obtained the blue coloration with 0.01 gramme of the ash of potato, beetroot, dahlia root, various leguminous plants,THE ANALYST.13 extract of oak, chestnut, etc. The ash of laccase and of an extract of RzLssula delica gave an intense reaction, and in the case of the latter, at least, the author infers that the manganese is present in the form of a peroxide compound. C. A. M. On the Addition of Litharge in the Assay of Lead. A. Copalle. (An?z. de Chim, anal., 1903, viii., 412-415.)-1n determining the amount of lead in poor ores by the dry method, it is usual to add to the fusion mixture a definite weight of litharge containing a known amount of lead, which is subsequently deducted from the amount found. As the amount of lead in the litharge is determined by fusion with alkalies alone, the conditions are not identical with those of the actual assay, and the author has therefore made experiments to determine the influence of different gangues, such as occur in lead ores, upon the litharge assay.In each case 25 grammes of litharge were fused under identical conditions in the presence of 25 grammes of different gangues, and the following amounts of lead found as the average of three or more closely concordant results : Litharge alone, 89.48 ; with quartz, 89.12 ; with pyrites, 87.60 ; with blende, 88-68 ; with natural barium sulphate, 88.32 ; with calcite, 89.12 ; and with clay, 88-96-per cent. Hence the author concludes that, in order to obtain the true amount of lead in the assay of a given lead ore, it is necessary in standardizing the litharge used, to add to it gangues of the same kind and quantity as are present in the lead ore, and that this is especially the case when the latter contains sulphur.C. A. M. On the Separation of Antimony and Tin by Means of Oxalic Acid. F. Henz. (Xeits. Anorg. Chem,, xxxvii., 1.)-According to the author, Rossing’s modification of Clarke’s oxalic acid method (Zeitschr. analyt. Chem., xli., 1) does not effect a perfect separation, a small quantity of tin sulphide being always precipitated with the antimony sulphide, even when the operation is repeated. He found it possible, however, to separate the two metals completely by first oxidizing the solution of their sulpho-salts with hydrogen peroxide. The method is carried out by placing the sulphide solution, which must not contain more than 0.3 gramme of the alloy, in a 500 C.C.beaker, adding 6 grammes (one-third of the sum of the weight of the tartaric and oxalic acids) pure potash and 3 grammes tartaric acid (ten times the weight of the alloy), and then twice as much 30 per cent. hydrogen peroxide as is necessary to completely decolorize the liquid. After heating to boiling for a few minutes, the liquid is allowed to cool somewhat, and a hot solution of 15 grammes of oxalic acid added. The liquid, the volume of which should now be 80 to 100 c.c., is then boiled vigorously for ten minutes, after which hydroqen sulphide is led into the solution, which is still kept. boiling. Fifteen minutes after the antimony has commenced to come down the liquid is diluted to 250 C.C. with boiling water ; after another fifteen minutes the heating is interrupted, and ten minutes later the current of hydrogen sulphide is also %topped.The precipitate is filtered on to a, tared Gooch crucible, and is washed twice with a 1 per cent. solution of oxalic acid, and then with very dilute acetic acid; both washing liquids should be at the boiling-point and saturated with hydrogen sulphide.14 THE ANALYST. The Gooch crucible is placed in t\l wide tube sealed to a narrow tube at one end and fitted with a perforated rubber stopper at the other. A current of dry carbon dioxide is led through the tube, which is at the same time heated, first to 100" to 130" C. until the precipitate is dry, and then to 300" C. until it has been completely converted into Sb,S,. The heating may be done with a naked flame; for exact work it is best, however, to heat the part of the tube containing the crucible in a small drying- oven designed for the purpose.A Rose's crucible-lid may be placed against the stopper inside the tube to protect it against radiated heat. Ground-glass stoppers may also be used, but it is difficult to make them absolutely tight at first. The author prefers this method of determining antimony to all others, although he blso obtained good results by converting the sulphide to tetroxide, and by depositing the antimony electrolytically from a sulphide solution containing sodium sulphite or, preferably, potassium cyanide. The results obtained by this last method were, however, always slightly high. The tin in the oxalic acid filtrate is determined most readily by evaporating to a small bulk, and electrolysing the solution at 60" to 80" C.with a current of 0.2 to 0.3 amphe and 2 to 3 volts, 5 C.C. sulphuric acid (1 : 1) being added after six hours, after which the electrolysis is continued for another eighteen hours. The deposition of the tin is complete under these conditions, and the deposit is easily handled. The author found that this method gives the best results for tin, electrolysis from a sulphide solution giving deposits of an unsatisfactory nature. If the tin has been originally obtained as sulphide during the course of an analysis, it may be dissolved in acid ammonium oxalate, a few C.C. of sulphuric acid being added after the electro- lysis has been in progress for some hours. As regards the method of separation, practically theoretical results are obtained for the antimony, but the values found for the tin are generally somewhat low (maximum error 0-0036 gramme on 0.2 gramme tin).A. G. L. Qualitative and Quantitative Estimation of Traces of Antimony in the Presence of Large Quantities of Arsenic. (Chem. News, Ixxxviii., 261.)-The author describes the following methods in order of increasing sensitive- ness : The platinum-tin couple method consists in immersing a piece of tin for half an hour in a 25 per cent. hydrochloric acid solution of the antimony contained in a small platinum crucible, or in a dent made in a platinum sheet. The piece of tin should be pointed, the pointed end dipping into the liquid, and should be bent over so as to again touch the platinurn at another point.Even with only 0.04 milli- gramme antimony per C.C. a black stain will be produced; and as $T C.C. of liquid suffices for the test, 0-002 milligramme antimony may be recognised in this way, whilst 0.001 milligramme will give a dark stain in one hour. If arsenic is present, it must be thoroughly oxidized by evaporation with nitric acid, the residue being taken up in hydrochloric acid ; its quantity must be less than 5 milligrammes per C.C. If zinc is substituted for tin, the test is rendered far less sensitive. By using silver instead of platinum the method is made more sensitive, OoOO1 milligramme antimony giving a dark stain, even in presence of 1 milligramme of arsenic. In this case the G. Denigha.T E E ANALYST. 15 duration of the test should not exceed five minutes.An ordinary silver coin may be used instead of pure silver, In the czesium salts method, a reagent is used consisting of 1 gramme cesium iodide and 3 grammes cesium chloride dissolved in 10 C.C. water, with the addition of a drop of 10 per cent. ammonia. I n examining stains obtained on platinum by means of tin, or deposited on porcelain from the flame of a Marsh apparatus, the stain should be treated with nitric acid, and the solution evaporated in a small porcelain crucible on the water-bath. On the dry residue a very small drop (0.001 c.c.) of the caesium reagent is placed and spread out by means of a fine glass rod; another small drop of 10 per cent. sulphuric acid is then added and mixed with the reagent. A red colour is developed after a short time, which should not disappear on adding a drop of sulphurous acid, and the intensity of which is a measure of the amount of antimony present. If, at least, 0.01 milligramme is present in the original dry residue, it may be dissolved in 0.1 C.C.of 25 per cent. hydrochloric acid; on placing a drop of the cesium reagent in the centre of this drop and waiting for two minutes, well-defined hexagonal crystals of the double iodide of czsium and antimony appear, which may be observed under the microscope. The quantity of arsenic present should not exceed 5 milligrammes per 01 c.c., as otherwise iodine may be liberated by the reduction of the arsenic acid. A. G. L. A Rapid Method for the Determination of Arsenic in Arsenopyrite. J. L. Danziger and W.H. Buokhout. (Sch. of 1CliYzes Quart., xxiv., 400.)-Half a gramme of the finely-ground ore is mixed with 10 to 15 grammes sodium peroxide in an iron crucible, the whole covered with a layer of peroxide, and heated very gently until the mass has sintered together. The heat is then raised to bright redness for five minutes with occasional agitation of the melt, which is then treated with 100 C.C. of boiling water, the whole being transferred to a 200 C.C. flask and made up to the mark; 100 C.C. are decanted through a dry filter-paper, made strongly acid with concen- trated hydrochloric acid, and boiled down to 75 C.C. Another 75 C.C. of hydrochloric acid (specific gravity 1.2) are then added, the solution is allowed to cool to room temperature, 3 grammes of potassium iodide in saturated aqueous solution are added, and the liberated iodine is titrated with standard sodium thiosulphate solution, drops of starch solution on a white tile being used as indicator.The thiosulphate is best standardized against arsenic acid; the value thus found is about 1 per cent. higher than that calculated from its iodine or copper value. The starch indicator should not be added to the liquid itself, as it would be hydrolysed by the large amount of acid present. Should antimony be present in the ore, it is removed from the acidified extract of the melt by Bunsen’s method, the filtrate containing the arsenic being then oxidized with potassium chlorate and hydrochloric before applying the above method. The authors state that even when considerable quantities of arsenic are present in the solution the precipitate of antimony pentasulphide obtained is free from arsenic.The method gives good concordant results, slightly higher than those obtained with other processes. A. G. L.16 THE BNALYST. A Study of the Quantitative Determination of Antimony.-Part 11.: An Investigation of the Herroun-Weller Volumetric Method for the Determination of Antimony. Lewis A. Youte. (Sch. of Mines Quart., xxiv., 407.)-From the author's work it appears that antimony is quantitatively oxidized to the antimonk state by nitric acid or potassium chlorate in hydrochloric acid solution; that the solution obtained may be boiled to expel chlorine without volatilizing any antimony ; and that complete reduction to the antimonious state is effected by boiling with sulphur dioxide in an open or closed vessel, by sulphur dioxide and potassium bromide, and by potassium iodide in cold hydrochloric acid solution.Using the Herroun-Weller method for the determination of antimony, in which the oxidized solution is reduced with potassium iodide and hydrochloric acid in the cold, the liberated iodine then being titrated with sodium thiosulphate solution, the author obtained results which are concordant, but show a consistent loss of 1 per cent. of the antimony. The cause of this loss is as yet unexplained. A. G. L. The Determination of Manganese in the Presence of Iron. G. v. Knorre. (Zeit. f. angew. Chem., 1903, xxxviii., 905.)-The author has already proposed a method for the estimation of manganese by oxidizing with persulphuric acid, filtering off the precipitated manganese dioxide and titrating (see ANALYST, 1902, 68).He now finds that this may be simplified by omission of the filtration, a fact which is of special importance in the case of irons and steels containing less than 0.5 per cent. of manganese, because theae give precipitates which are liable to pass through the filter. I t is, however, essential that before the titration the persulphuric acid be entirely destroyed by prolonged boiling, otherwise the results would be too high, especially if there be nickel or copper in ths sample that is being examined. Hydrochloric acid must not be present in the solution, but nitric acid has no ill effect. If there be a large proportion of phosphoric acid present, the results will be too low, but the amounts that occur in the analysis of ordinary iron and iron ores have no appreciable effect.Great excess of acid must be avoided, as it prevents the complete precipitation of the manganese. The sample is dissolved in sulphuric acid or nitric and sulphuric acid, methyl orange is added, and the mid is neutralized with ammonia or potash. Then ammonium persulphate is added in excess, 3 or 4 grammes to each gramme of iron, the solution is diluted to about 300 c.c., and 20 C.C. of dilute sulphuric acid (specific gravity 1.17) are added. The liquid is now boiled for fifteen to twenty minutes, with occasional shaking to prevent bumping; it is cooled and a slight excess of a standard solution of hydrogen peroxide is added.When the manganese dioxide is completely redis- solved, the excess of hydrogen peroxide is estimated by titration with permanganate. A. M. Electrolytic Separation of Iron and Xanganese. J. Koster. (Ber. , 1903, xxxvi., 2716-2719.)-1ron can be readily separated from considerable quantities of manganese by the following modification of Classen's process : The solution con- taining the metals in the form of double oxalates with ammonium (prepared by theTEE ANALYST. 17 addition of about 10 grammes of ammonium oxalate) is electrolysed without the application of heat, and as soon as a deposit of manganese peroxide appears on the anode, a few C.C. of a 10 per cent. solution of phosphorous acid are added to reduce it. This addition is repeated &B required, but should not be made more frequently than is necessary, since otherwise the reduction of the iron is retarded ; and in any case no more phosphorous acid should ba introduced after the manganese has been kept in solution for two hours, since the bulk of the iron will then have been deposited.As a rule 5 C.C. of the 10 per cent. solution sufficiently retard the deposi- tion of manganese. After complete separation of the iron the platinum basin is washed with water without interrupting the current, and finally with pure alcohol, any slight particles of manganese deposit being removed by means of a soft brush. Results agreeing excellently with the theoretical amounts were thus obtained under the following conditions : Voltage, 3 to 4 ; amperage, NDIo,= 1.5 to 2 ; ordinary temperature; time, 5 to 8 hours, according to the amount of phosphorous acid used.C. A. M. The Electrolytic Separation of Manganese and Iron, of Aluminium and Iron or Zinc, and of Zinc and Iron. Hollard and Bertiaux. (Annd. Chim AnaZyt., viii., 324.)-Manganese and iron may be separated by electrolysing the solution of their sulphates, to which a little ammonium sulphate and 20 to 25 C.C. of a saturated solution of sulphur dioxide in water has been added before making it slightly ammoniacal. The determination is carried out at 40" C., using a current of 1 ampere. The cathode should have a large surface, whilst the anode surface should be as small as possible. Under these conditions the iron deposited is free from manganese, whilst the sulphur dioxide present prevents the formation of manganese dioxide, which would carry iron oxides down with it, The iron on the cathode, how- ever, contains a little platinum dissolved off the anode, and for this reason cannot be weighed directly.It is dissolved in dilute sulphuric acid, the solution filtered from platinum and titrated in an atmosphere of carbon dioxide with potassium per- manganate. To determine the manganese in solution the current is again passed through, but in the reverse direction, so as to have a large anode surface. The solution is kept at 90" to 95" C. At the end of several hours the current is inter- rupted by withdrawing the cathode, the anode being left in it for another thirty minutes to reduce any higher oxide of manganese to the state of dioxide.The peroxide is then washed with hot water into a beaker, filtered on to asbestos, and washed with hot water. Filter and precipitate are then treated in a beaker with hydrochloric acid and potassium iodide, and the liberated iodine is titrated with thiosulphate. The results obtained in this way are fairly good, the largest error being 5 milligrammes on 0.2 gramme iron. Iron may be separated from aluminium in the same way as from manganese, except that ammonium citrate is also added to the solution, good results being obtained. The method may also be used to separate nickel and aluminium, the nickel originally deposited being dissolved in sulphuric acid, the liquid filtered, and electrolysed after being made alkaline with ammonia, but the results obtained are not very satisfactory.18 THE ANALYST.To separate iron from zinc, sulphurous acid is added to the solution of the sulphates, which is then exactly neutralized with sodium hydrate ; potassium cyanide solution is then added, and the solution electrolysed in the cold, using 8 current of 0.3 ampere and a cathode covered with copper. The zinc deposited always contains a little iron, which can be determined with potassium permanganate as above. A. G. L. The Electrolytic Precipitation of Nickel from Phosphate Solutions. Walter T. Taggart. (Journ. Anter. Chem. SOC., xxv., 1039.)-The author shows that nickel can be deposited completely and free from phosphorus from phosphate soh- tions, obtained by precipitating nickel sulphate solutions with a solution of dieodium hydrogen phosphate of specific gravity 1.038, dissolving the precipitate in phosphoric acid of specific gravity 1.347, and diluting with water.The conditions of the electrolysis were : N.D7,,, 3 amperes ; tension, 6 volts ; dilution, 175 C.C. ; tempera- ture, 50" to 70' C. ; time, nine hours; sodium phosphate, 135 C.C. ; phosphoric acid, 6.75 C.C. ; nickel, 0.3249 gramme. By raising the temperature to 88" C., the time may be shortened to five hours, working with a current density of 2.25 amphres. A. G. I;. On a Method for the Separation of Cobalt from Nickel,.and the Volumetric Determination of Cobalt. a. L. Taylor. (Chem. News, lxxxviii., 184.)-The author again calls attention to his modification of Rose's method for the separation of nickel and cobalt (ANALYST, xxvii., 285), which has been found to work well in the assay of cobalt 01188.Either calcium or barium carbonate may be used, provided only that the solution is neutral. Zinc ehould be absent, as it prevents the pre- cipitation of the cobalt peroxide, the composition of which approximates to the formulae Co901* and Co,Oll, and is sufficiently constant to enable the chlorine evolved on treating it with hydrochloric acid to be used as a measure of its quantity. A. G. L. The Analysis of Ferrosilicon. H. Lidhplm. (Zed. f. angew. Chem., 1903, xliv., 1060.)-The author overcomes the great difficulty in dissolving these alloys by fusing them with sodium peroxide. The reaction is so violent that it is better to add some cmbonate to moderate it. The ferrosilicon (0.2 to 0.3 gramme) is mixed with five times the quantity of sodium and potassium carbonate in a nickel crucible, which is then gently heated to drive off moisture.To the dry,mass twice as much sodium peroxide is added, and the mixture is gradually fused, the ignition being continued until no solid lumps can be seen. When cool the crucible is placed in a covered beaker of warm water to dissolve out the mass. The solution is widified with hydrochloric acid, and iron and silicon are determined in the usual way. A. M. The Determination of Uranium and Uranyl Phosphate by means of the Zinc Reductor. 0. €3. Pulman, Jun. (Zeds. Anorg. Chem., xxxvii., 113.)-The author has made a number of determinations, using Kern's method of reduction (Journ.THE ANALYST. 19 Amer. Chem. Soc., xxiii., 716), and finds that, contrary to Kern’s statement, the reduction, if carried out in an atmosphere of carbon dioxide, gives oxides lower than the dioxide.In contact with air, however, the solution is immediately oxidized to the dioxide stage. He therefore prefers to work in an atmosphere of air. The reductor used is constructed from a 50 C.C. burette filled to R height of 1 inch with broken glass; above this is placed a pad of glass wool, and then an 18-inch layer of 20-mesh amalgamated zinc. The uranium must be in the form of sulphate; the solution, which should measure about 100 to 150 c.c., should contain one-sixth of itsvolume of concentrated sulphuric acid. I t is heated to 100’ C. before being poured through the reductor, which has previously been warmed by pouring hot dilute sulphuric acid (1 : 6) through it.The reductor is washed first with dilute sulphuric acid and then with 200 C.C. hot water, and the reduced liquid at once titrated with & potassium permanganate solution. For 0.2 gramme UO,, the passage of the liquid through the reductor should take about eight to ten minutes, and for 0-3 gramme UO, about thirty minutes. As this modified method gave very good results, it was applied to the determina- tion of uranium in uranyl phosphate precipitated from a boiling solution containing ammonium acetate and acetic acid. This precipitate is so finely divided that it passes through a double filter-paper or through a layer of ordinary asbestos ; it may be retained by using only the fine asbestos fibres which remain in suspension when the asbestos is shaken up with water and allowed to stand for one minute, but the filtration and washing are tedious, and the results obtained by passing the sulphuric acid solution of the precipitate through the reductor are only moderately satisfactory.Suction must be applied to the receiving flask to effect this. A. G. L. The Estimation of Titanium. J. Watson Bain. (Joum. Amer. Chem. Soc., xxv., 1073.)-The author has examined various methods for the determination of titanium, and concludes that Gooch’s method is the best. Baskerville’s method (Joum. Soc. Chem. Ind., xix., 419), in which sulphur dioxide is passed into the neutralized solution and the titanium then precipitated by a few minutes’ boiling, is superior to the methods of Arnold, founded on the insolubility of the phosphotitanate of iron in dilute hydrochloric &id, and of Hilger and Haas, in which the mixed oxides of titanium and other metals are reduced in a current of hydrogen and then treated with dilute acid.Excellent results are obtained by Walker’s method, in which hydrogen peroxide is added to the titanium sulphate solution, which is then slowly poured into ammoniacal hydrogen peroxide; but as this operation must be repeated three times, the method requires too much time for ordinary work. A. G. L. The Determination of Small Amounts of Potassium in Aqueous Solutions. F. K. Cameron and G. H. Failyer. (Journ. Anzer. Chem. SOC., xxv., 1063.)--The method described depends on the pink colour produced by adding potassium iodide to a solution of platinum salt, and gives especially good results with solutions con- taining one to ten parts of potassium per million, although its usefulness is not20 THE ANALYST.confined to these limits. Solutions in which potassium is to be determined must be free from ammonia and organic matter ; their removal is effected, if necessary, by evaporation with sulphuric acid and ignition of the residue. This residue is then treated with a little hydrochloric acid and an excess of platinic chloride, and the whole evaporated to a stiff paste. The excess of platinum is then removed by six or eight washings with 95 per cent. alcohol, the precipitate being filtered on to asbestos, using gentle suction to remove the alcohol. I t is then dissolved in a small quantity of hot water, a drop of concentrated hydrochloric acid added after cooling, and then an excess of potassium iodide solution.The red colour developed is compared with a standard after, at least, four hoilrs' standing. The colour should be allowed to develop in as strong a solution as possible, but if the liquid is diluted afterwards, the diluted liquid should be allowed to stand a t least one hour before a reading is taken. The standard solution used for comparison contains 2 parts of potassium per million. It is contained in a cylinder fitted with side-arm near the bottom, by means of which the height of the liquid in the cylinder may easily be varied. The authors have also found that, by adding a little alcohol to the liquid in which the pink colour has been developed, and heating, a strong, clear, yellow colour is produced.By matching this yellow colour against that of a standard solution, a check on the reading obtained with the pink colour is given. Both readings agree well with each other and with the theoretical values. The method is applicable even when as much as 50 parts sodium, 50 parts magnesium, and 50 parts calcium are present together with 50 parts potassium per million, the values found ranging from 48.0 to 53.5 parts per million. A. G. L. On the Occurrence of Iron in Sulphur. R. v. Hasslinger. (Momtshefte f. Chew,., xxiv., 729.)-The author shows that all commercial foyms of sulphur, from the native to the very purest product, leave, on distillation, a black residue consisting only of iron and carbon, probably in the proportions indicated by the formula FeC,.This black residue may also be obtained by simply boiling the sulphur for some time. I t appears to be a decomposition product of a volatile compound of iron, carbon, and sulphur, and is always formed when a bituminous material (e.g., asphalt) is distilled with an iron compound and sulphur. It is only-very slowly decomposed at the temperature of boiling sulphur. The author was only able to obtain a sulphur which did not give this product by cautiously oxidizing pure hydrogen sulphide. A. G. L. On a New Method for the Preparation of Pure Iodine. Launcelot W. Andrews. (Amer. Chenz. Journ., xxx., 428.)-Potassium iodide is powdered together with one and four-tenths times its weight of potassium bichromate, both salts having previously been separately fused to insure dryness.The mixture is introduced into a, wide tube closed at one end, and is heated for some time to 200° C. in a current of dry air to expel the last traces of moisture. The interior of the tube is then carefully dried with cotton, a plug of glass wool is placed above the mixture, and a second short tube is slipped over the open end of the first one, so as to fit it as closely as possible.THE ANALYST. 21 The mixture is then gradually heated to fusion, the tube being kept in an inclined position. 5K,Cr,O, + 6KI = 8K,CrO, + Cr,O, + 61. Iodine is liberated according to the equation- At the end of the operation the upper end of the tube carrying the sublimate is cut off, Even if the iodine used contains bromide or chloride, the iodine obtained is pure, since these Salt8 are not decomposed when fused with potassium bichromate.A. G. L. The Use of Potassium Tetroxalate in Volumetric Andysisis. 0. Kiihling. (Zeit. f. mzgezu. Chem., 1903, xxxvii,, 1031.)-Objections have been raised by Duprh and others to the employment of this substauce on the ground that it is not of constant composition, but is liable to contain varying proportions of water of crystallization. The bad results obtained by some workers are ascribed by the author to their having dried the compound over sulphuric acid, whereas it should have been used in the air-dry condition. The following method yields a product possessing a satisfactory degree of purity : Commercial oxalic acid is crystallized twice from hot hydrochloric acid having a specific gravity of about 1.07, and then three times from boiling distilled water.The product thus obtained, which must give no precipitate with silver nitrate and leave no residue on ignition, is used to make a cold saturated solution by allowing it to stand for a day with a quantity of water insufficient entirely to dissolve it. Of this solution one-fourth is taken and titrated with a concentrated solution of pure caustic potash, using phenolphthalein as indicator; the other three-quarters are then mixed with it again and stirred well, whereupon a crystalline precipitate forms. This should be redissolved by heating the liquid, and be allowed to crystallize out again on cooling. The crystals are filtered off on hardened filter-paper and again recrystallized from a small quantity of hot water, filtered off, pressed between hardened filter-paper, and allowed to dry in the air until they no longer adhere to the sides of a glass vessel. This usually requires about two days. The composition of the resultant product is constant, and agrees very closely with that calculated from the formula C204HK,C20,H,,2H,0. I n a series of determinations, in which the tetroxalate was compared by means of permanganate solution with a thiosulphate solution standardized against specially purified iodine, the agreement was very satisfactory. A. M. Hydrogen Peroxide in Crystals. R. Willstatter. (Berichte, xxxvi., 1828 ; through Pharnz. Jown., 1903, lxxi., 777.)-Various salts are described in which hydrogen peroxide appears to replace the water of crystallization. For instance, beautiful prisms, having the formula (NHJ2S04.H,0,, separate out if ammonium sulphste be dissolved in a 30 per cent. solution of hydrogen peroxide and placed over sulphuric acid in a desiccator. The crystals smell like ozone, effloresce in air, but keep well in closed vessels. Sodium sulphate forms a similar compound, Na,S0,H,0.~H,02, as do also alum, aluminium sulphate, borax, and sodium acetate.22 THE ANALYST, These crystals give up their hydrogen peroxide to ether and other solvents, and may therefore be of use in carrying out reactions with peroxides in indifferent solvents. They may, perhaps, replace the persulphates and percarbonstes for technical purposes. w. P. s.