318 THE ANALYST. INORGANIC ANALYSIS. The Separation of Mercury from Molybdenum and Tungsten by Hydra- zine. (D. Chem. Ges. Ber., 1904, xxxvii., 2210; through Chem. Zeit. Rep., 1904, xviii., 205.)-Mercury can be separated from molyb- denum by the addition ol hydrazine sulphate to a hot ammoniacal solution of the mixture. The precipitation of molybdenum is prevented by the addition of tartaric or citric acid. In the separation of mercury from tungsten the addition of citric acid is unnecessary, and tungsten is finally precipitated by heating with fuming nitric acid to destroy the hydrazine sulphate. P. Jannasch and W. Bettges. H. A. T. The Determination of Palladium by Hydrazine. P. Jannasch and W. Bettges. (D. Chem. Ges. Ber., 1904, xxxvii., 2210; through Chem. Zeit. Rep., 1904, xviii., 205.)-To a hot, slightly acid solution of palladium aqueous solution of hydrazine sulphate is added.Reduction takes place very rapidly, with evolution of nitrogen, and palladium is precipitated. The washed and dried precipitate is ignited in a current of hydrogen. Palladium has thus been separated from K, Na, Mg, Zn, Fe. H. A. T. Determination of Palladium, H. Erdmann and 0. Makowka. (Berichte, xxxvii., No. 11; through Chem. News, xc., S9.)-By means of acetylene, either in aqueous solution or as a gas, palladium may be readily separated in cold, strongly acid solutions from platinum, iridium, rhodium, and gold. The yellowish-brown precipitate obtained is non-explosive; it is readily soluble in ammonia, and on ignition with a little ammonium nitrate in a current of hydrogen yields the pure metal.The presence of copper does not hinder the reaction, as is the case when hydraeine is used as the precipitating agent, A. G. L. The Microchemical Detection of Gold b y Means of the Colloidal Coloration of Silk Fibre. Julius Donau. (Momtshefte f. Chem., xxv., 545.)-The test depends on the red colour developed in a silk thread previously treated with a, reducing agentTHE ANALYST. 319 when dipped into a solution of auric chloride and exposed to the air. Cotton threads similarly treated may also be used; the reaction in this case, however, is not so sensitive, and the colour produced may be blue instead of red. The silk threads used are best prepared from raw silk by soaking it in 10 per cent. sodium or potassium hydroxide for three or four hours, well washing, and then soaking for several hours in either a solution of tannin or of a mixture of stannous chloride and pyrogallol. To carry out the test a pellet of wax is fastened to one end of the prepared thread, which is then cut off so as to leave a length of 1 centimetre.It is then dipped in a vertical position into a drop of the gold solution and examined under a microscope, using a magnification of 200. Using a silk thread treated with stannous chloride and pyrogallol, 0.000002 milligramme of gold may be recognised ; tannin threads are somewhat less sensitive. The presence of other metals, except iron, arsenic, and antimony, does not prevent the reaction ; silver gives a yellow colour, which is hidden, however, by the red of the gold, and may in a6y case be removed by treating the thread with nitric acid.Silicic acid and strong mineral acids prevent the reaction. To isolate the gold for the purpose of the test, the following procedure is recommended : The suspected solution is evaporated almost to dryness with lead filings; the residue is washed with water by decantation, and treated with aqua regia; a little sulphuric acid is then added to the decanted aqua regia solution, which is next evaporated to fuming, diluted, and filtered to remove lead ; the filtrate is evaporated nearly to dryness, and a drop tested as above. Gold could be detected in this way on 5 or 10 grammes of a mixture of iron and copper oxides with sand to which one ten-millionth part its weight of gold had been added.A. G. L. On the Examination of Tinned Wares. Utz. (Oesterr. Chem. Zeit., vii., 271.) -The tinning used is examined for lead by heating 0-5 gramrne of the scrapings with 7 to 8 C.C. of. strong sulphuric acid in a small flask for a short time. If the sample dissolves completely, no lead is present ; the solution obtained is occasionally dark in colour from the presence of organic matter. If lead is present, a white precipitate of lead sulphate is formed, which is filtered off and determined after diluting the sulphuric acid with 20 C.C. of a 5 per cent. ammonium oxalate solution, a little water, and an equal volume of alcohol. A. G. L. Determination of Minute Quantities of Arsenic in Copper Ores and Metallurgical Products. (Journ. SOC. Chem. Ind., xxiii., 524.)- The author uses a modification of the Marsh apparatus described by Chittenden and Donaldson (Amer. Chem.Journ. ; ChenzicaZ Pews, January 14, 1881). The low results always obtained by the presence of iron in the solution introduced into the apparatus are obviated by the use of sufficient copper to form a ‘‘ couple ” with the zinc. Forty grammes of zinc and 0-05 to 0.08 gramme copper are employed in each experiment. Two grammes of the finely ground sample are treated with 20 to 30 drops of T. C. Cloud. The arsenic deposited in the tube is finally weighed.320 THE ANALYST. sulphuric acid, then a sufficient quantity of nitric acid, and the mixture digested on the water-bath until decomposed. The excess of sulphuric acid is largely evaporated, the cooled residue.dissolved in water, boiled and filtered.The liquid is then electrolysed, using a platinum cone or cylinder as the cathode. The amount of copper being known by a previous determination, the electrolysis can be stopped when the solution contains the small amount of copper above specified. The anode is heated with the liquid until all deposit has dissolved, when the solution is evaporated down to 5 or 10 C.C. For quantities of arsenic up to 0.002 gramme, the timeataken for the experiment is two hours, and an additional hour for each 0.001 gramme. After the experiment the tuba containing the arsenical mirror is cut off, wiped clean, and weighed after being kept in a desiccator for twenty minutes. The arsenic is then dissolved out of the tube with nitric acid, and the clean and dry tube again weighed. A balance turning easily with 0.05 milligramme is required.A. R. T. Determination of Manganese in Drinking-Water. G. Baumert and p. Holdefleiss. (Zeit. Untersuch. Nahr. Genussmittel, 1904, viii., 177-181.)-The following method is proposed for the estimation of the small quantities of manganese occurring in some well waters (ANALYST, 1904, 201). The process is an iodometric one : From 250 to 1,000 C.C. of the water, according to the quantity of manganese present, are evaporated to a volume of 80 c.c., after adding 1 C.C. of concentrated hydrochloric acid. To separate any iron occurring in the water, a little zinc oxide or barium carbonate is added towards the end of the evaporation. The mixture is passed through a filter, and washed until the filtrate measures 100 c.c.; 5 C.C.of a 10 per cent. sodium hydroxide solution are then added to the filtrate, and the latter well shaken in a stoppered flask, the stopper being removed from time to time to admit air. Any manganous salts are oxidized. Five C.C. of a 10 per cent. potassium iodide solution are now added, and suficient concentrated hydrochloric acid, drop by drop, to completely dissolve the brownish precipitate, and leave a, clear yellow solution. After adding starch solution, the liberated iodine is titrated with a sodium thio-sulphate solution (0.30 gramme per litre) which has been pre- viously standardized on a solution containing a known amount of pure manganous chloride, and to which the above process has been applied.w. P. s. A New Separation of Thorium from Cerium, Lanthanum, and Didymium by Metanitrobenzoic Acid. (Journ. Amer. Chem. Soc., xxvi., 780.)-The author finds that metanitrobenzoic acid precipitates thorium quanti- tatively from neutral solutions 01 the nitrate, whilst cerium, lanthanum, and didymium are not precipitated. A quantitative separation of the thorium from these elements can be effected by two precipitations with this reagent from dilute solutions. The impure thorium metanitrobenzoate obtained by the first precipitation is best dissolved in dilute nitric acid, the thorium precipitated as hydroxide by an excess of potassium hydroxide, this precipitate dissolved in nitric acid, and the solution Arthur C. Neish.TRE ANALYST. 333: evaporated to remove free nitric acid, after which the precipitation with metanitro- benzoic acid is repeated.The reagent itself is readily prepared from benzoic acid ; as the ortho- and para-acids react similarly, though not so completely, the direct product of nitrating benzoic acid, containing about 75 per cent. of the meta-acid, may be used. The method gives good results in the analysis of monazite sands, and is more expeditious than Metzger's fumaric acid method, besides avoiding the use of alcohol. A. G. L. On the Separation of Calcium and Magnesium. Carl Stolberg. (Zeit. f- angew. Chern., xvii.., 741.) -According to the author, the separation of calcium from magnesium by means of the oxalate method is complete only if the calcium is precipitated twice from dilute slightly ammoniacal solutions containing sufficient ammonium chloride, a large excess of ammonium oxalate being used, and the pre- cipitate being allowed to stand for twelve hours in each case before filtering.As a, more rapid method he recommends conversion of the mixed calcium and magnesium salts into anhydrous sulphates by evaporation to dryness with sulphuric acid ; enough water is then added to the residue to form the hydrate MgSO, + 7H,O, after which the mass is treated with methyl alcohol containing 10 per cent. of absolute ethyl alcohol, which dissolves the magnesium sulphate and leaves the calcium sulphate undissolved. The residue is washed with methyl alcohol containing 5 per cent. of ethyl alcohol, and ignited and weighed as usual. The method gives excellent results. The addition of a slight excess of water does not appear to be injurious, but the whole should be heated on the water-bath for one minute with constant stirring to insure complete conversion of the magnesium salt into the hydrate. A.G. L. Chloride in Barium Sulphate precipitated by Barium Chloride. G. A. Hulett and L. H. Duschak. (Zeit. Anorg. Chern., XI., 196.)-The authors concur with the opinion of Richards that the chlorine present in barium sulphate precipi- tated by means of barium chloride should be determined, as it may amount to 1 per cent. of the precipitate. For this purpose the barium sulphate should be ignited at a temperature not exceeding 700" C. After weighing, it is dissolved in concentrated sulphuric acid, and the hydrogen chloride formed removed from the warmed solution by means of a current of dry air into a vessel containing silver nitrate.A. G. I;. The Determination of Sulphur in Iron. Allen P. Ford and Ogden G. Willey, (Journ. Amer. Clieuz. SOC., xxvi., 801.)-The authors believe that the different results often obtained by different analysts for the sulphur in the same sample of steel are due, not so much to differences in the methods used-provided that some form of the ordinary oxidation method is used-as to the want of skill of the operators. In their experience works chemists generally obtain correct values, whilst commercial chemists usually report low results. Too rapid solution appears to be one of the chief causes of error leading to loss of sulphur. The authors recommend covering the sample in a dish with a small inverted watch-glass before adding the acid.In322 THE ANALYST. this way the evolved gases are forced to remain in contact with the oxidizing liquid for some time; or else solution may be effected in an Erlenmeyer flask, covered with a small funnel. Instead of the ordinary oxidation method, they prefer to use Bamber’s method (Journ. Arner. Chem. Soc., xix., 114), in which the nitric acid solution of the steel is evaporated to dryness, the residue ignited after adding potassium nitrate, and the mass obtained extracted with dilute sodium carbonate. The solution is then evaporated to dryness with hydrochloric acid, the residue taken up’in dilute hydrochloric acid, and sulphuric acid precipitated as usual. The authors state that with this method accurate results are obtained even by analysts who have had no previous experience of the method.A. G. L. A Rapid Method for the Determination of Total Sulphur in Iron by Evolution. (Amer. Chem. Journ., xxxii., 84.)-The method, which is a modification of that of Walters and Miller, obviates the necessity for the expensive and unusual apparatus required in that method, and also materially reduces the time required for the operation. Two grammes of the sample are mixed with 1 gramme of the purest iron dust obtained by reducing the oxide in hydrogen, and the mixture is placed in a small porcelain crucible, and covered with 1 gramme of the iron dust. On top of this is placed a disc of ash-free filter-paper ; the lid is put on, and the crucible heated for ten minutes at the highest temperature of the blast-lamp.The contents of the crucible are then allowed to cool partially and placed in an ordinary evolution flask, after which the determination is finished as usual. The results obtained are said to be excellent, and a complete determination requires less than thirty minutes. The sulphur in the iron dust used must, of course, be determined. A. G. L. S. S. Knight. The Action of Acetic Acid on Portland Cement and Blast-furnace Slag. F. Hart. (Tonniizd. Zeit., xxviii., 809 ; through Chem. Zeit. Rep., xxviii., 241.)-Port- land cement and blast-furnace slag may be readily differentiated by means of alcoholic acetic acid solution as follows : 1 gramme of the finely-powdered sample is vigorously shaken for ten minutes in a closed flask with 100 C.C. of a 10 per cent.solution of acetic acid in absolute alcohol. After allowing the whole to settle for a few minutes the difficultly soluble slag may be recognised as light-coloured, glass- like particles ; any undecomposed Portland >cement remains as dark grains, whilst the precipitated silica is flocculent. I t is stated that 5 per cent. of slag may be recognised in Portland cement by this method. A. G. L. Determination of Nitrites in the Absence of Air. I. K. Phelps. (Amer. Journ. Science, xvii., No. 99; through Chenz. News. 1904, xc., 114, 115.)-The method proposed is based upon the interaction of nitrous and hydriodic acids. Nitric oxide and free iodine are liberated, the latter being subsequently indirectly titrated.The apparatus employed is the same as that described in the determination Qf nitric acid (ANALYST, 1903, 125). An amount of standard arsenious acid solutionTHE ANALYST. 323 slightly in excess of that required to take up the iodine set free later by the nitrous acid, and 25 C.C. of concentrated sodium carbonate solution, are placed in the flask, and boiled to expel air. The flame is then removed, the exit tube is lowered well into the mercury, and 7 C.C. of sulphuric acid (1 : 4) passed into the flask through the tapped funnel. This amount of acid is sufficient to liberate enough carbon dioxide to fill the flask, but still leave its contents alkaline. After the acid has been washed in and the mixture in the flask is quite cold, the nitrite solution to be analysed, con- taining 2 grammes of potassium iodide, is introduced, followed by sufficient sulphuric acid (1 : 4) to render the contents of the flask acid in reaction.A saturated potassium hydrogen carbonate solution is now added until the free iodine is taken up, and the mixture boiled for about five minutes to expel the nitrogen dioxide. After cooling, the solution is titrated with & iodine solution, using starch as indicator. In making the various additions to the flask, care must be taken that no air is allowed to, enter ; it is also necessary to expel the nitrogen dioxide before titrating. w. P. s. Analysis of Liquid Carbon Dioxide. R. Wog. (Zeit. ofeelztl. Chenz., 1904, x., 295-297.)-Anslyses are given of liquid carbon dioxide, the principal object of the investigation being a determination cjf the air in the carbon dioxide and also of the rate at which the air is liberated from the liquid. I t was found that the latter phenomenon takes place very gradually.For instance, the first 0-5 kilogramme taken from a cylinder gave 5.2 per cent. as the volume of air present. The next 0.5 kilogramme yielded 11.3 per cent. by volume of air. In the succeeding quantities taken from the cylinder the amount of air gradually diminished until, after with- drawing 5.0 kilogramrnes, only 1.1 per cent, was present. After standing overnight, the liquid remaining in the cylinder was again analysed, and found to contain 0-5 per cent. by volume of air. The average of all the results (fifteen) gave the amount of air as 5 per cent. by volume, a figure closely agreeing with that yielded by the first 0.5 kilogramme drawn from the cylinder.I t should be stated that the air mentioned above was the residue of gas unabsorbed by potassium hydroxide. I t contained only from 1-87 to 3-25 per cent. by volume of oxygen, and was therefore almost pure nitrogen, and not atmospheric air. w. P. s. Sources of Error in the Analysis of Coal. E. Goutal. (dim. d e Chint. anal. 1904, ix., 242-246.)-The calculation of the calorific power of coal by means of the author’s formula (ANALYST, xxviii., 128) may give erroneous results when the sample contains a high proportion of ash, owing to the influence of the volatile substances yielded by the schists themselves. This may be obviated by treating the powdered mineral with a mixture of methylene iodide with a third of its volume of benzene, and washing the separated fractions with pure benzene. Coal or anthracite thus treated is shown to give calculated calorific values in close agreement with those actaally determined. Another source of errm is introduced if oxygen prepared by electrolytic processes be used for the combastion in calorimeters. I n the author’s experience324 THE ANALYST. such oxygen sometimes contains as much as 2 per cent. and seldom less than 1 per cent. of hydrogen, with the result that the calorimetrical determinations are much too high. The oxygen should therefore be tested by calorimetrical determinations of naphthalene or other organic substance of known calorific value, purified if necessary by a preliminary combustion in Berthelot’s apparatus. C. A. M.