318 ABSTRACTS OF CHEMICAL PAPERS INORGANIC ANALYSIS. Estimation of Air, Water-Vapour, and Nitrous Oxide in Mixtures of these Three Constituents. 6. A. Burrell and G. W. Jones. ( J . I n d . und Eng. Chem,, 1916, 8, 735.)-An application of a method already described by Burrell and Robertson (ANALYST, 1915, 40, 162) to this particular purpose. The gas is admitted to a vacuous bulb connected to a manometer and provided with a stop- cock.The mixture is liquified by immersing the bulb in liquid air, and the air is withdrawn by means of a Topler pump and measured. The residual gas is sub- jected to a temperature of - 78" C. (C0,-alcohol slush) and the nitrous oxide with- drawn and measured. Finally the partial pressure of the water-vapour is measured. The ratio of this pressure to the barometric pressure multiplied by 100 grms.gives the percentage of water-vapour in the original sample. Three samples of American '' laughing gas " analysed in this way showed from 95.6 to 96.2 per cent, of nitrous oxide, air 2.0 to 2.1 per cent., and water-vapour 2 per cent. G. C. J. Estimation of Aluminium. C. F. Sidener and E. Pettijohn. (J. Ind. and Eng. Chem., 1916, 8, 714-716.)-As the result of experiments described in the paper, it is shown that boiling for one minute precipitates aluminium quantitatively,IKORGANIC ANALYSIS 319 whilst longer boiling may lead to partial re-solution.Aluminium hydroxide need not be washed free from ammonium chloride before ignition. The excess of am- monia present when precipitation is made should be as small as possible, and never more than 1 or 2 C.C.in 250 C.C. of solution. Aluminium hydroxide, freshly pre- cipitated, is soluble to a slight extent in water and to about the same extent in dilute ammonium nitrate solution. Precipitates of alumina, when large, require to be heated with the blowpipe for as long as forty minutes, and the ignited alumina is strongly hygroscopic.G. C. J. Iodimetrie Estimation of Gold. L. Vanino and F. Hartwagner. (Zeitsch. anal. Chem., 1916, 55, 377-388.)-The method described by Peterson (ANALYST, 1899, a, 108) is modified by the author in order to render it capable of yielding more accurate results. The gold chloride (AuCl,.HCl) is treated with a mixture of potassium iodide and potassium iodate, and the liberated iodine is titrated wit)h thiosulphate solution.The reactions taking place are shown by the following equa- tions : (1) 6AuCl3-HC1 + 18KI=6AuI + 121 + lSKCl + 6HCl; (2) 6HC1+ 5KI + KIO,= 61 + 6KCl+ 3H,O ; (3) 181 -I- 18Na2S,03=18NaI + 9Na2S,0,. Three molecules of thiosulphate correspond to 1 atom of gold. Using quantities of gold chloride equivalent to 2-10 mgrms. of gold, the error in the results obtained varies from - 0.9 to + 4.5 per cent, The method may be applied to the analysis of the salt AuCl,.2H20; the quantity of iodine liberated by the salt from potassium iodide (Peterson's method) is a measure of the quantity of gold present.The difference between this quantity of iodine and that liberated when the salt is treated wit,h a mixture of iodate and iodide gives Free acid must not be present in the salt.the amount of hydrochloric acid present. w. P. s. Precipitation of Magnesium Salts by Ammonium Carbonate. F. Fiehter and R. Osterwalder. (Zeitsch. anal. Chem., 1916, 55, 389-392.)-The authors have investigated the nature of the precipitate obtained when a soluble magnesium salt is treated with ammonium carbonate. Ordinary ammonium carbonate when dissolved yields a solution containing a mixture of ammonium carbonate and ammonium hydrogen carbonate.If this solution is added to magnesium sulphate solution a precipitate of hydrated magnesium carbonate, MgC03.3H,0, forms gradually, the precipitation being accelerated by warming the mixture. Crystals of hydrated magnesium carbonate appear wit,hin about thirty minutes when mag- nesium sulphate solution is mixed with ammonium carbonate solution to which ammonia has been added previously in such quantity that the proportion of ammonia to carbon dioxide in the solution corresponds with the formula (NH,),CO,; if this mixture is heated to 40" C., an amorphous precipitate of basic magnesium carbonate (magnesia a h ) forms, which changes into the hydrated carbonate when the mixture is cooled.Magnesium sulphate solution treated with ammonium carbonate solu- tion containing an excess of ammonia yields a precipitate of the basic carbonate, and this, after the lapse of a few hours, changes into the hydrated carbonate; if, however, the precipitation is made from a hot solution, the basic carbonate remains as such when the mixture is cooled.A double carbonate, having the composition320 ABSTRACTS OF CHEMICAL PAPERS (NH,),CO,~MgCO3*4H2O, is obtained when one equivalent of magnesium sulphste (in solution) is treated with ten equivalents of ammonium carbonate, (NH,),CO, (also in solution). w. P. S. Volumetric Estimation of Thiosulphate in the Presence of Sulphite. A . Sander. (Zeitsch. a w l . Chem., 1916, 55, 340-342.)-Bodnar has described a method (ANALYST, 1914, 39, 101) for the estimation of thiosulphate in the presence of sul- phite which depends on the reaction of the thiosulphate with silver nitrate.This method can be rendered much more simple and rapid by substituting mercuric chloride for the silver nitrate. The sulphite forms a soluble complex with the mercuric chloride, Na,SO, + HgCI, = HgCl-NaSO, + NaCl, whilst the thiosulphate reacts according to the equations : Na,S,O, + HgCl,=HgS,O, + 2NaC1; HgS,O, + H,O=HgS + H,SO,; and 2HgS+ HgCl,=Hg,S,Cl,.The solution containing the sulphite and thiosulphate is simply treated with an excess of mercuric chloride solution, and the mixture titrated with FG sodium hydroxide solution, using methyl- orange as indicator. The acidity of the mjxt,ure is due to the liberated sulphuric acid, which in turn is a measure of the thiosulphate present.A quantity of am- monium chloride should be added to the mixture before the titration is commenced in order to prevent the precipitation of mercuric oxide by the added alkali. (See also ANALYST, 1916, 84.) w. P. s. Detection of Tungsten.PI. L. Hartman. (Chern. News, 1916, 114, 45-46.) -The reduction teat and especially its limit of sensitiveness have been investigated. I n solutions containing only tungsten salts the test is very sensitive, and as little as 1 mgrm. of tungstic acid may readily be detected. The most satisfactory re- ducing agent is metallic tin, and at least 10 per cent. of concentrated hydrochloric acid should be added to the liquid to be tested.For testing ores, t)he following procedure is recommended: About 0.2 grm. or more of the finely powdered ore is boiled with concentrated hydrochloric acid until half the acid has evapora'ted. The solution is diluted with an equal volume of water, a piece of mossy tin added, and the mixture warmed unless a blue colour develops in the cold. If this test gives negative results, about 0.5 grm. of the ore is fused with 4 grms. sodium carbonate, the melt dissolved by boiling water in the crucible, the solution acidified with a n equal volume of concentrated hydrochloric acid, a piece of tin added, and the mixture warmed if necessa'ry. This test will detect 2 per cent. of tungsten with certainty. Vanadium gives a blue colour on reduction, but tartaric acid will effect, the reduction in the case of vanadium, whereas it will not reduce tungstic acid. G. C. J.