THE ANALYST. 369 INORGANIC ANALYSIS. The Iodometric Estimation of Arsenic and Antimony in the Presence of Copper. F. H. Heath. (Zeits. Anorg. Chenz., 1908, 59, 87-93.)-A solution con- taining copper, antimony and arsenic in their higher states of oxidation is acidified with 1 to 2 grams of citric acid. For quantities of copper up to 0.3 gram, dissolved in 50 to 100 C.C. of liquid, 3 to 5 grams of potassium iodide are added, and the liberated iodine is at once titrated with sodium thiosulphate solution. The cuprous iodide is filtered off on asbestos, and the filtrate heated to boiling after adding 1 C.C. of bromine to destroy tetrathionates. If the liquid does not become clear after some time, it is cooled, 0.5 C.C. of bromine added, and again boiled for some time until its volume is reduced to 60 C.C.About 40 C.C. of water and 2 grams of potassium iodide are then added, and the liquid is boiled down to 50 C.C. to reduce the arsenic and antimony. The cooled liquid is exactly decolourised by means of a solution of sulphur dioxide, after adding starch solution. Sodium bicarbonate is then added in excess, and arsenic and antimony are titrated together with iodine solution. The results obtained are fairly satisfactory, both for copper and for the sum of arsenic and antimony present. A. G. L. Estimation of Sulphur and Arsmic. W. C. Ebaugh and C. B. Sprague. (Eng. and Min. Joum., 1908, 85, 1048; Chem. Zeit. Rep., 1908, 32, 349.)-The370 THE ANALYST. following modification of the methods of Eschka and of Pearce is recommended for the estimation of arsenic and sulphur in ores, rocks, foods, etc.: 0.5 gram of the substance is mixed with a sufficient quantity of a mixture of soda and zinc oxide (1 : 4) in a porcelain crucible, which is then heated for fifteen to twenty minutes in a muffle.The mass is taken up with water, filtered, the filtrate acidified, and the sulphate precipitated with barium chloride. The filtrate is acidified with acetic acid, treated with silver nitrate, and boiled, the precipitated silver arsenate filtered off, washed, and dissolved in nitric acid, and the silver titrated with ammonium t hiocyanate. C. A. M. Note on the Electrolytic Estimation of Bismuth. J. Peset. (Zeit.9. Anal. Chew%., 1908, 47, 401.)-About 0.04 to 0.08 gram of a bismuth salt is dissolved by warming with 3 C.C. of concentrated sulphuric acid and 4 to 5 C.C.of water. The liquid is then diluted to 140 C.C. with water, any cloudiness being neglected, and electrolysed for eighteen to twenty-four hours at a temperature of 50' C., using a current of N.D,,, = 0.002 to 0.1 ampihe at 2 volts. The anode is rotated to prevent the formation of bismuth peroxide. As soon as a test with ammonium sulphide shows the whole of the bismuth to be deposited, a known weight (0.08 to 0.15 gram dissolved in 10 C.C. of water) of cadmium sulphate is added, and the electrolysis con- tinued for another eight hours with a current of 0.2 to 0.3 amperes at 2.5 to 3.5 volts. The cadmium is deposited over the bismuth, and protects it from oxidation during the subsequent washing and drying.A. G. L. Estimation of Cerium in Presence of Other Rare Earths by Means of Potassium Ferricyanide. P. E. Browning and H. E. Palmer. (Zeits. Anorg. Chem., 1908, 59, 71-73.)-An excess of a standard solution of potassium ferricyanide is added to the solution of a ceric salt. Potassium hydroxide is then added, the precipitated ceric dioxide is filtered off, and the potassium ferrocyanide formed is titrated in the filtrate, after acidifying with dilute sulphuric acid. A blank deter- mination should always be carried out on the ferricyanide solution, which keeps uncbanged for a week or even longer. Results obtained by this method on solutions of cerium containing other rare earths show errors of up to about +I per cent. A. G. L. Detection and Quantitative Estimation of Chlorites and Hypochlorites in Commercial Chlorate.B. Carlson and J. Gelhaar. (Chem Zeit., 1908, 32, 604-605, 633-634.)-To detect hypochlorites and chlorites in commercial chlorate, 5 grams of the sample are dissolved in 100 C.C. of cold water. HFpochlorite is present if a blue colour is produced on adding a few drops of a solution of iodide of potassium and starch, as little as 0.000002 per cent. of hypochlorite (on the solution) producing a faint blue colour after a few minutes; if a few drops of a weak solution of sodium thiosulphate are added, the instantaneous disappearance of the colour is especially marked. If hypochlorite is absent, 2 C.C. of sulphuric acid are added. The appearance of a blue colour now indicates the presence of chlorite, 0.0001 per cent.of chlorite producing a blue colour after a few seconde, whilst, with the quantity ofTHE ANALYST. 371 acid used, chlorate only produces a colour after sixty minutes. Samples of chlorate which are odourless are also free from chlorite or hypochlorite. Hypochlorite is estimated quantitatively by adding an excess of a standard solution of arsenious acid, and titrating back with iodine solution, as usual. To estimate chlorite, the calculated quantity of arsenious acid necessary to reduce the hypochlorite is added to another portion of the sample. The solution is then diluted to 250 c.c., heated to 95" C., and acidulated with sulphuric acid. The chlorite is then titrated with a solution of sodium indigo-sulphonate (containing 6 grams of the salt per litre) which has been standardised on a pure solution of chlorite, the strength of the latter being determined iodometrically. The authors state that commercial samples of chlorate are almost invariably free from both chlorite and hypochlorite, but may contain chloride, chromate, iroq and calcium.A. G. L. Estimation of Chromium as Silver Chromate. F. A. Gooch and L. H. Weed. (Zeits. Anorq. Chem., 1908, 59, 94-96.)-A solution containing chromium as potassium dichromate is heated to boiling, and a considerable excess of silver nitrate solution is added. The whole is again boiled, rendered slightly alkaline with am- monia, and then slightly acid with acetic acid, and allowed to stand for thirty minutes. The silver chromate is then filtered off on asbestos, washed first with dilute silver- nitrate solution, and finally with 20 to 30 C.C.of water, and dried at 135' C., or gently ignited. The presence of sodium or animonium The results obtained are good. nitrate is without influence. A. G. rJ. Separation and Estimation of Cobalt and Nickel. E. Pozzi-Escot. (Bull. SOC. Chim., 1908 [4.], 3, 776-777.)-The double molybdate of nickel and ammonium, the formation of which affords a means of detecting nickel in the presence of cobalt (ANALYST, 1908, 33), is completely insoluble in a saturated solution of ammonium molybdate, whereas the corresponding double salt of cobalt and ammonium is soluble. On this fact is based the following rapid method of estimating cobalt and nickel in the presence of each other : The solution containing the two metals is divided into two parts, in one of which the cobalt and nickel are estimated together by one of the known electrolytic methods. The other portion is concentrated, nearly neutralised (though still left slightly acid), treated with a very large excess of a saturated solution of ammonium molybdate, and with about 10 grams of ammonium chloride, and heated to about 60" to 70" C.with continual stirring. The liquid is then cooled, and the precipitated nickel filtered off, washed with a cold saturated solution of ammonium chloride containing Bome ammonium molybdate, and dissolved in hot water. From this solution the nickel is precipitated by means of potassium hydroxide and bromine, the nickel oxide washed and dissolved in an acid, and the nickel estimated electrolytically.The difference between the result and that obtained in the previous electrolytic estimation gives the amount of cobalt. C. A. M.372 THE ANALYST. Estimation of Graphite. F. Browne. (Chem. N~ZOS, 1908, 98, 51.)-The author has observed that when ferric oxide is heated in an open dish with graphite at a bright red heat, all the carbon is burnt off, and the oxide of iron is partially converted into magnetic iron oxide. The error due to this loss of oxygen can be avoided by regulating the temperature as roughly measured by the eye, and the author has based on these observations the following method for the determination of graphite : Five grams of magnetic iron oxide, prepared by heating ferric oxide in a covered crucible at a pale red heat for two to three hours, are heated in an open dish, with occasional stirring, at a temperature not exceeding a pale redness for one hour, and then weighed. Half a gram of the (i graphite " is then stirred in, the mixture heated as before for one to two hours, stirring once or twice meanwhile, and then cooled and weighed.The oxide remains unaltered, and the loss of weight after the addition of the graphite represents the carbon. Another determination can then be made on the same used sample of oxide. The mineral matter in coal may be similarly estimated, though the results are somewhat higher than the '( ash '' of coal determined by simple ignition of the sample. A. R. T. Estimation of Iron and Vanadium in Presence of each Other. G. Edgar. (zeits. Anorg. Chem, 1908,69,74-78.)-The solution of vanadic acid and ferric sulphate is rendered slightly acid with sulphuric acid and reduced as usual with sulphur dioxide, the vanadium being reduced to V,O, and the iron to FeO.The liquid is then titrated with potassium permanganate, a t first in the cold and, finally, at 70" to 80" C. The same solution is then run over a long column of amalgamated zinc in a Jones reductor, the reduced solution being caught in a flask containing ferric sulphate solution. After adding phosphoric acid, to lessen the colour of the iron salt, the contents of the flask are again titrated with potassium permanganate solution. From the two results obtained, tho quantities of iron andvanadium present are calculated. The test results quoted aru good for both elements.The zinc reduces vanadium to V,O, and the iron to FeO. A. G. L. Volumetric Estimation of Iron and Chromium by means of Titanous Chloride. S. B. Jatar. (Journ. SOC. Chem. Ind., 1908, 27, 673-674.)-The titration of chromium alone (in the form of bichromate) by means of titanous chloride can be best accomplished by the use of an indicator prepared by adding a few drops of thio- cyanate solution to B dilute ferrous sulphate solution. If a drop of this indicator be mixed on a white plate with a drop of the titration mixture, a red coloration is produced as long as any bichromate is present. For the estimation of iron and chromium in ferro-chrome and chrome iron ores, the following process may be used. The principles involved are : (1) Hydrogen peroxide oxidises chromic salts in alkaline solutions to chromates, these becoming bichromates on acidifying the solutions ; (2) Hydrogen peroxide in acid solutions reduces bichromates to chromates, ferric salts being unchanged; (3) When titrating a mixture of bichromate and a ferric salt, for all practical purposes the bichromate ie reduced first, the colour change marking the reduction, the ferric salt being left in solution. The first of these considerations is involved in bringing the iron and chromium in an ore into solution, whilst theTHE ANALYST.373 process depends on the second and third. About 0.5 gram of the finely-ground ore is fused with sodium hydroxide, with subsequent addition of sodium peroxide ; the fused mass is extracted with water, the solution is boiled to remove all traces of hydrogen peroxide, carefully acidified with dilute sulphuric acid, again boiled, and diluted to a volume of 500 C.C.Fifty C.C. of this solution are transferred to a small flask and titrated with standardised titanous chloride solution until the solution assumes a clear violet colour, showing that all the bichromate is reduced; a, few more drops of titanous chloride are added to make sure, then a few drops of thiocyanate solution, and the solution is further titrated until the coloration disappears. The volume of titanous chloride used represents- the bichromate and iron combined. A second 50 C.C. of the original solution are treated with hydrogen peroxide ; after a short time the mixture is boiled for about fifteen minutes to expel the excess of peroxide, and the solution, now violet in colour, is cooled.A few drops of thiocyanate solution are added, and the solution is titrated with titanous chloride solution until it becomes quite clear and violet in colour. The volume of titanous chloride used represents the amount of iron only; this volume, subtracted from that required for the first titration, gives a measure of the bichromate. w. P. s. Estimation of Nickel and Chromium in Steel. E. D. Campbell and W. Arthur. (Joum. dmer. Chem SOC., 1908, 30, 1116-1120.)-0ne gram of the steel is dissolved in 10 to 15 C.C. of nitric acid (specific gravity 1*2), with the addition of a, little hydrochlorio acid if necessary. The solution is evaporated to the point of fuming with 6 to 8 C.C. of sulphuric acid (1 : 1).The residue is heated with 30 t o 40 C.C. of water until dissolved, and filtered if necessary; 13 grams of sodium pyrophosphate, dissolved in 60 to 70 C.C. of water heated to 60" to 70' C., are then added, the whole is cooled to room temperature, and dilute ammonia, added gradually with vigorous stirring, until the greater part of the ferric pyrophcsphate is dissolved and the solution has assumed a greenish tinge. At this stage the liquid should be alkaline to litmus, but should not smell of ammonia. The solution is then warmed and stirred until the pyrophosphate is completely dissolved, any brownish colour being destroyed by carefully adding dilute sulphuric acid. The solution is then cooled, and the nickel titrated as usual with silver nitrate and potassium cyanide.If copper is present, it will be estimated together with the nickel; but steels oontaining copper can generally be dissolved in dilute sulphuric acid, leaving the copper in the residue, which is filtered off. The iron in the filtrate is oxidised with hydrogen peroxide before proceeding as above. If chromium is present as well as nickel, the sulphuric acid solution of the steel obtained as above is boiled and treated with potassium permanganate solution until a, permanent precipitate of manganese dioxide is obtained. The liquid is cooled and filtered, and chromium estimated by adding a known weight of ferrous ammonium sulphate in excess of that required to reduce the chromium, and titrating back with potassium permanganate solution. The chromium is then again oxidised to chromic acid by boiling with potassium permanganate, and the filtered liquid is titrated for nickel as before, after adding sodium pyrophosphate and ammonia.374 THE ANALYST.The advantage of using sodium pyrophosphate instead of citric acid to keep the iron in solution is stated to lie in the fact that the end-point is more easily seen. A. G. L. The Estimation of Tungsten and Chromium in Steel. F. W. Hinrichsen and L. Wolter. (Zeits. Anorg. Chem., 1908, 59, 183-197.)-The authors show that, in the absence of chromium, Von Knorre’s benzidine method (ANALYST, 1(307,32, 131) gives good results for tungsten in steel. If chromium also is present, however, the results obtained for tungsten are low if the chromium, which invariably contaminates the precipitate, is estimated and deducted.The presence of tungsten also interferes with the estimation of chromium by precipitation with ammonia, high results being obtained. For such steels a modification of the Berzelius method gives good results, the procedure being as follows : Two grams of the sample are oxidised by repeated evaporation with nitric acid. The residue is ignited and fused with 16 grams of sodium peroxide. The melt is dissolved in water, the solution filtered through asbestos, and the filtrate evaporated to about 150 C.C. The separated silica is filtered off on aRbestos, and the filtrate made up to 250 C.C. ; 50 C.C. of the filtrate are then acidified with nitric acid, made alkaline with ammonia, heated to boiling, and precipitated with mercurous nitrate.The washed precipitate is ignited, treated with hydrofluoric and sulphuric acids to remove silica, and weighed as WO, + Cr,O,,. The mixed oxides are then fused with sodium carbonate and a little potassiuni carbonate, the aqueous solution of the melt is evaporated with sulphuric acid to remove nitrous acid, and in the solution obtained the chromium is estimated iodometrically, tungstic acid being kept in solution by the addition of sodiuni phosphate, so as not to obscure the end-point. A. G. L. Separation of Tungsten from Chromium and Estimation of Tungsten in Steel containing Chromium. G. v. Knorre. (Zeits. anal. Chem., 1908, 47, 337-366.)-The author agrees with Hinrichsen’s conclusion (Stnhl. u. Eisen, 1907, No. 40) that his own benzidine method for estimating tungsten in steel (ANALYST, 1907, 32, 131) gives unsatisfactory results in the presence of chromium.He also shows that Rose’s method, in which chromic hydroxide is separated from tungstic acid by means of ammonia, is useless, and that the modification oi Berzelius’s mercurous nitrate method proposed by Hinrichsen is too tedious. He has now worked out the following modification of the benzidine method for the case of steels containing chromium and tungsten, the results obtained being very exact : From 1.5 to G grams of the steel are heated in a flask with dilute hydrochloric acid until the action ceases. The liquid is nearly neutralised with a solution of sodium hydroxide, about 10 C.C. of -& sulphuric acid and 30 to GO C.C. of benzidine solution (containing 20 grams of benzidine and 25 C.C.of fuming hydrochloric acid per litre) are added, and the liquid is well cooled. After standing for twenty minutes, the mixture of undissolved metal, benzidine tungstate, and benzidine sulphate (which last renders filtration easy) is filtered off, washed with dilute benzidine solutiori, and ignited. The oxides obtained are fused with sodium carbonate, the melt extracted with water, and the insoluble iron oxide washed with dilute sodiumTHE ANALYST. 375 carbonate solution. The alkaline filtrate ie then rendered just acid with hydro- chloric acid, using methyl-orange as indicator. If much chromium is present, the liquid is boiled for some time, to convert the tungstic acid into metatungstic acid, after which it is cooled, treated with sulphur dioxide, to reduce the chromic acid, and precipitated with an excess of benzidine solution.The tungstic oxide obtained after filtration and ignition of the benzidine metatungstate is quite free from chromium. If only a little chromium is present, the neutralised solution of the melt may be treated with 10 C.C. of & sulphuric acid and about 0.4 gram of hydroxylamine hydrochloride ; the whole is mixed and at once precipitated with an excess of benzidine solution, the benzidine tungstate being filtered off, washed, and ignited as above. The presence of hydroxylamine prevents the oxidising action of the chromate on the bemidine, but the latter must be added to the liquid before the chromate has been reduced by the hydroxylamine. The second procedure is quicker than the first, but the precipitate may be contaminated with traces of chromium.9. G. L. The Influence of Fine Grinding on the Water and Ferrous-Iron Content of Minerals and Rocks. W. F. Hillebrand. (Jozm. Anzer. Chem SOC., 1908, 30, 1120-1131.)-Confirrning the work of Nauzelius (Soeriges Geol. Undcrs. Arsbok, 1907, 1, No. 3) and others, the author shows that many minerals take up hygroscopic water and undergo extensive oxidation when subjected to prolonged fine grinding before analysis, The oxidation appears to be due chiefly to local heating of the particles in presence of air, and is greatest when mechanical grinders are employed. The author recommends that the portions of minerals used for fusions should simply be crushed, without grinding, so as to pass a sieve with 80 meshes to the inch, and that, if a finer powder is used for separate estimations, a moisture estimation should be carried out on it, and a correction applied.For alkalis, the portion may generally be ground after weighing. For ferrous iron, the portion may be ground under alcohol before weighing, or, preferably, the coarse powder may be decomposed as far as possible by a short boiling, in an atmosphere of carbon dioxide, with hydro- fluoric and sulphuric acids, and any undecomposed residue left after the titration ground under water and again treated. Pratt’s method appears to give more exact results than Cooke’s (United States Geological Swuey, Bull. 305, p. 138). A. G. L. Colorimetric Estimation of Phosphorus in Steel. G.Misson. (Chew,. ZeLt., 1908, 32, 633.)-Results accurate to about 0-005 per cent. of phosphorus are obtained in ten minutes by the following method : One gram of the steel is dissolved in 20 C.C. of nitric acid (specific gravity 1.2) in a 100 C.C. flask, provided with a mark at 80 C.C. The solution is boiled, 10 C.C. of a 0.8 per cent. solution of potassium permanganate are added, and the solution is again boiled for a few seconds. Ten C.C. of a hydrogen peroxide solution, prepared by gradually adding 40 grams of sodium peroxide to a well-cooled mixture of 100 C.C. of nitric acid and 900 C.C. of water, are then added, followed by 10 C.C. of a solution of ammoniuin vanadate, containing 2.345 grams of the salt and 20 C.C. of nitric acid per litre, and the liquid376 THE ANALYST.again boiled until the hydrogen peroxide is completely decomposed. The liquid is diluted to 60 or 65 c.c., cooled, 10 C.C. of a 10 per cent. ammonium molybdate solution are added, and the solution is diluted to the 80 C.C. mark. The colour is then compared with that of a standard steel, preferably in Eggertz tubes. The colour obtained is due to a compound of the probable formula : (NH,),PO,,(NH,)VO,, 16Mo0,. The standards keep well for about fourteen days. A. G. L. Volumetric Estimation of Phosphoric Acid in Superphosphate. S. Kohn. (ChenL. Zeit., 1908, 32, 475-476.)-Twenty grams of the superphosphate are extracted with 1,000 C.C. of water; 50 C.C. of the solution are filtered, diluted with water to 350 c.c., and titrated with standard sodium hydroxide solution free from carbonate, methyl orange being used as indicator.The quantity of alkali used corresponds to the free phosphoric acid. A considerable excess of a neutral solution of calcium chloride and a little phenolphthalein are then added, and the titration continued until a distinct red colour is obtained. The quantity of alkali so used corresponds to the combined phosphoric acid, according to the equation : CaH,P,O,+ 2CaC,I, + 4NaOH = Ca,P,Os + 4NaCl+ 4H,O. The results obtained by the new method agree within about 0.3 per cent. with those given by the citrate method. The method cannot be applied to mixed fertilisers containipg ammonium salts. A. G. L. Examination and Valuation of Red Lead. P. Beck. (Zeit. anal. Chem., 1908, 47, 465-492.)-Many reducing-agents may be used to render red lead com- pletely soluble in, e.y., dilute nitric acid, such as sugar, oxalic acid, lactic acid (cf.ANALYST, 1907, 32, 395), methyl alcohol, formaldehyde, glycerim, phenylhydrazine, hydroxylamine salts, etc., and especially hydrogen peroxide. Thus, for 100 grams of red lead, from 80 to 100 C.C. of nitric acid (specific gravity P4j with 15 to 20 C.C. of hydrogen peroxide (90 per cent.) are required, whilst the dissolved lead may be separated by the addition of 35 C.C. of strong sulphuric acid previously diluted to 100 C.C. When the decom- position of the red lead is complete, the liquid is diluted to 500 c.c., and kept boiling for an hour, to destroy the excess of hydrogeD peroxide and facilitate the separation of iinsoluble constituents.After standing for some hours in a moderately warm place, the insoluble matter may be filtered off, and the filtrate tested for impurities after separation of the lead as sulphate. In estimating the barium sulphate it is sufficient for technical purposes to dry and weigh the insoluble matter. The results are only approximate, however, since the37 7 THE ANALYST. barium sulphate is never pure, and also dissolves to a considerable extent in the nitric acid. For the estimation of the lead peroxide Topf's method (Zeit. a d . Chenz., 1887, 26, 296) gives the most reliable results, since, owing to the large quantities of the sample that may be used, the chance of error is reduced. When only a small quantity is available, however, distillation with hydrochloric acid in a current of carbon dioxide is preferable, especially for works' laboratories.From 0.3 to 0-4 gram of red lead is placed in a distillation flask, c, d, provided with a tubulure, a, closed by means of a rubber stopper, b, through which passes a glass tube, c, reaching nearly to the bottom, and having at cl a disc-like projection to protect the lower part of the stopper. The open end of the flask is connected by means of rubber tubing with two absorption tubes (not shown), each of which contains 35 C.C. of a 5 per cent. solution of potassium iodide (free from iodate). The flask, e , contains sufficient (100 c.c.) hydrochloric acid of specific gravity 1.14 to 1-15 to decompose the whole of the red lead, and is connected by indiarubber tubing at f with the distillation flask.The hydrochloric acid is forced into the latter by the pressure of the purified carbon dioxide entering the upper flask, the distillation flask being then cautiously heated, a slow current of the gas passed through the apparatus during the distillation, and a more rapid current for ten minutes after all the chlorine has been driven over. The contents of the absorption tubes are then washed into a beaker, solid sodium bicarbonate added to neutralise the free acid, and the liquid diluted to about 750 C.C. and titrated with i>G sodium arsenite solution, with starch solution (or preferably zinc-iodide starch solution) as indicator. The advantages of this method over the original Bunsen-Yresenius method are that there is no loss of chlorine, that the latter is completely expelled by the carbon dioxide, and that there is no frothing.The results are closely concordant, and the chlorine does not act upon the rubber to any appreciable extent in the presence of the carbon dioxide. C. A. 11. Estimation of Lead in Alloys. W. Elborne and C. M. Warren. (CheViL. Nezus, 1908, 98, 1.)-The following simple method is applicable to all lead alloys soluble in the solvent used, and is based on the insolubility of lead chloride in absolute alcohol, the chlorides of antimony, tin, arsenic, bismuth, iron, nickel, cobalt, manganese, zinc, copper, cadmium, aluminium, and magnesium being soluble. One gram of the alloy in fine shavings is treated with 50 C.C. of concentrated hydrochloric acid, and a slow current of chlorine gas, obtained by the interaction of manganese peroxide and strong hydrochloric acid in the cold, passed through the cold liquid for twenty-four hours, the liquid being occasionally agitated.This treatment readily dissolves type- and stereo-metals. More refractory alloys will dissolve on boiling the solution and repeating the treatment with chlorine. When solution is complete, the liquid is evaporated to dryness on a water-bath, and the cold residue extracted with absolute alcohol by stirring the mixture and immediately pouring off the supernatant liquid through a dry weighed filter-paper. This extraction is repeated three or four times, allowing five minutes for the final extraction with alcohol. The residue and filter-paper containing the lead chloride are then dried at 100" C.for three hours and weighed. A. 1%. T.378 THE ANALYST. New Method of Separating Lithium Chloride from the Chlorides of the other Alkalis, and from the Chloride of Barium. L. Kahlenberg and F. C. Krauskopf. (Journ. Amw. Chem. SOC., 1908, 30, 1104-1115.)-The authors show that the chlorides of sodium, potassium, rubidium, caesium, and barium are practically insoluble in pyridine containing less than 5 per cent. of water, while lithium chloride is quite soluble even in pure pyridine. The chlorides of strontium, calcium, and magnesium occupy an intermediate position. In order to separate lithium chloride from the first-named chlorides, the aqueous solution is evaporated just to dryness, and the residue, which should not weigh more than 2 grams, is treated with 25 C.C.of boiling pyridine for a few minutes, any large crystals being broken up with a glass rod. The liquid is then decanted through a small filter, the residue washed twice with a little hot pyridine, dissolved in hot water, the solution again evaporated, and the extraction repeated a second, and, if necessary, a third time. From the filtrates the pyridine is removed by distillation, the lithium converted into anhydrous sulphate, and weighed. The method gives good results. A. G. L. The Use of Organic Acids and Anhydrides as Standards in Alkalimetry, and a Comparison between Swecinic Acid, Arsenious Oxide, and Silver Chloride as Standards in Iodometry and Alkalimetry. J. K. Phelps and L. H. Weed. (Zeits. Anoiy. Chem., 1908, 59, 114-119, 120-12G.)-Continuing their previous work on the use of succinic acid as a standard for volumetric work (ANALYST, 1907, 32, 230), the authors now show that, in titrating sodium hydroxide solutions with phenolphthalein as indicator, succinic acid and anhydride, malonic acid, benzoic acid, and phthalic acid and anhydride all give results as accurate a s those obtained by the use of hydrochloric acid standardised by means of silver nitrate, or arsenious acid (by the potassium iodide and iodate method).On account of their easy solubility in water, succinic acid and malonic acid are the most convenient to use of the organic standards. They are readily obtained by hydrolysis of their esters with water acidulated with nitric acid, and crystallisation of the solution obtained.Crystals of succinic acid can be kept for a year over sulphuric acid or calcium chloride without any succinic anhydride being formed. A. G. L. The Application of the Cobaltic Nitrite Method to the Estimation of Potassium in Soils. W. A. Drushel. (Zcits. Anory. Chem., 1908, 59,97-101.)-The dried soil is treated on the water-bath with an excess of hydrochloric acid. The solution is filtered and evaporated to dryness. Bases are then removed wit;h sodium carbonate, or arnmonia, and ammonium oxalate. Tlie filtrate is again evaporated to dryness, and the residue gently ignited. The alkalis left are dissolved in a little water, acidified with a few drops of acetic acid, and evaporated to pastiness with excess of a solution of sodium cobaltic nitrite. The mass is stirred with cold water, and filtered through asbestos ; the precipitate is washed with sodium chloride solution, and then treated with an excess of hot dilute standard potassium perrnanganate solution.The excess of the latter is destroyed by standard oxalic acid solution, the excess of which is exactly titrated with potassium permanganate solution. (Clf. ANALYST, 1908, 35,) A. G. L.THE ANALYST. 379 A Short Volumetric Method for the Estimation of Sulphuric Acid. T. Cooksey. (Proc. Boy. SOC., New South Wales, 1908.)-The method proposed depends on the volumetric estimation of the excess of barium ohloride left in the solution after precipitation of the sulphuric acid as barium sulphate. The solution containing the sulphuric acid to be estimated should be free from bases or metals other than the fixed alkalis, and must not contain carbonates, phosphates, silicates, or borates. Any free acidity is neutralised by the addition of sodium hydroxide solution, using phenolphthalein as indicator, and the solution is then treated with an excess of neutral stsndardised barium chloride solution, a quantity of alcohol equal to one-half the volume of the final total solution being also added. After the addition of a little more phenolphthalein, the excess of the barium chloride is titrated back with standardised sodium carbonate solution ; the solution is stirred constantly during the titration, and the end-point is taken when the pink coloration obtained is permanent.The addition of the alcohol causes the precipitated barium carbonate to settle rapidly, and enables the end-point of the titration to be observed sharply and distinctly. The difference between the quantity of barium chloride added and that found by the titration is a measnre of the sulphuric acid present. I n the estimation of sulphuric acid (sulphates) in a solution containing magnesium or other metals, it is advisable to precipitate these by the addition of a mixture of potassium hydroxide and potassium carbonate, filter off the precipitate, and carry out the estimation of the sulphate on the filtrate. Calcium, and possibly strontium, can also be estimated by titration with sodium carbonate solution, if alcohol be added to the solution. The solution containing the calcium salt, and free from carbonates, borates, phosphates, and silicates, is rendered neutral to phenolphthalein, and is then titrated with standardised sodium carbonate solution, small quantities of alcohol being added from time to time with constant stirring, until the proportion of spirit in the mixture at the end of the titration is about one-half of the whole. If sulphates are not present, the whole of the alcohol may be added at the beginning of the estimation ; otherwise, it must be added gradually so as not to precipitate calcium sulphate. w. P. s.