54 THE ANALYST. INORGANIC ANALYSIS. Behaviour of Cacodylic Acid in the Marsh Apparatus. D. Vitali. (Boll. c7im. farm., 1903, xlii., 641; through Chem. Zeit. Rep., 1903, 300.)--Referring to his previous work in the same direction, Vitali again finds that cacodylic acid is not reduced in the Marsh apparatus if some platinic chloride has been added to assist in the evolution of the hydrogen. I€, however, the platinum is omitted, black flakesTHE ANALYST. of arsenic and the characteristic mirror are both produced. Presumably a com- pound between the cacodylic acid and the platinum salt is formed which is not decomposed in the conditions obtaining within the flask. F. H. L. Action of Saline Solutions upon Metallic Iron. P. N. Raikow and 0. Goworuchin-Georgiew. (Chem. Zeit., 1903, xxvii., 1192.)--All ammonium compounds are decomposed by iron at ordinary temperatures, a ferrous salt being produced, and ammonia liberated.At the boiling-point the action is much more rapid. No iron compound soluble in water is produced when the powdered metal is boiled with a nitrite, chlorate, bromate, iodate, or chromate; nor when the acid radicle of the salt forms a compound with iron which is not soluble in water. No iron compound is formed when the metal is boiled with any normal salt of the alkali or alkaline earth metals; nor with any of their hydrogen (acid) salts which have a distinctly alkaline* reaction. A soluble ferrous compound is produced when the metal is boiled with a normal salt of a heavy metal (Zn, Cd, Ag, Cu, etc.), and also when it is boiled with such hydrogen (acid) salts as exhibit an acid? reaction.Inasmuch as 0.01 milligramme of ferrous chloride can be detected by means of potassium ferricyanide, very minute traces of a compound of a heavy metal are easily recognised in the nominally pure salts of the alkalies or alkaline earths which do not act upon iron powder ; but it is necessary to observe that tihe presence of a nitrite, chlorste, or chromate may impede or prevent the success of the test. On a large scale, the rules laid down are worth considering in relation to the corrosion of steam boilers-that portion of the corrosion which is due to (‘ auto- oxidation ” being set aside. I t will be seen that the normal and alkaline salts of the “light ” metals should not attack the plates; while their acid salts, and all the soluble compounds of the c L heavy ” metals (among which Mg stands), should be detri- mental.This agrees with Ost’s work, except that he has found potassium chloride, potassium sulphate, and calcium chloride (though not the sodium salts) to be injurious. It is probable that the salts Ost examined were contaminated with compounds of the heavy metals, which gave them th4 solvent power he noticed. F. H. L. The Separation of Barium, Strontium, and Calcium. L. Robin. (Ann. de Chim. anal., 1903, viii., 445-447.)-The solution containing these metals in the form of chlorides or nitrates is rendered slightly alkaline with ammonia, and then treated with about 2 per cent. of ammonium chloride (free from sulphate), followed by acetic acid in slight excess. It is next heated to the boiling-point, treated with a consider- able excesg of potassium bichromate and cooled, and the barium chromate collected on a weighed filter, washed successively with a 0.5 per cent.solution of ammonium acetate (rendered alkaline with ammonia), and with dilute alcohol (95 per cent. alcohol, 10 parts; water, 90 parts), dried for two hours at 100” to 110” C., and weighed. * For example, K,HP04. t KHSO, NaHSO,, KH,PO,, NaH,PO,.56 THE ANALYST. The filtrate is made alkaline with ammonia, boiled, treated with 3 to 4 per cent. of pure crystalline ammonium sulphate, and kept at 100" C. for fifteen minutes, the evaporated ammonia being replaced from time to time. After cooling, the strontium sulphate is collected, washed with hot ammoniacal water containing 0.5 to 1 per cent.of ammonium sulphate, and with 10 per cent. alcohol, dried, ignited, and weighed. The filtrate is heated to about 80" C. and the calcium precipitated as calcium oxalate. A test series of determinations are given to show that the method yields very accurate results in a very short time. C. A. M. Solubility of Magnesium Ammonium Phosphate in Ammonium Citrate Solu- tion. A. Bolis. (Chem. Zeit., 1903, xxvii., 1151.)-By allowing some specially prepared MgNH,PO;6H20 to remain in contact for twenty-four hours at ordinary tempera- tures with a solution of ammonium citrate containing 400 grammes of citric acid per litre, then filtering off the undissolved portion, washing it with ammoniacal water, and igniting it, the author has found the degree of solubility of the precipitate to be from 0.30 to 0.59 (mean 0.457) gramme per 100 C.C.of the liquid. A t a temperature of 50°C. the solubility is from 0.575 to 0.60 gramme per 100 C.C. F. H. L. Behaviour of Normal Ammonium Salts towards Litmus. C. Reichard. (Chem. Zeit., 1903, xxvii., 1105.).-It is stated in many of the text-books that an aqueous solution of pure ammonium chloride reddens blue litmus paper or tincture. This, however, is an error. If a piece of blue litmus paper is immersed in a solution (of any degree of concentration) of ammonium chloride in previously boiled water, it retains its blue colour for more than twenty-€our hours; only when it is taken out and allowed to dry naturally in the air does it turn red.This reddening is evidently due to the action of the atmospheric oxygen upon the ammonium compound in presence of moisture, a certain amount of dissociation taking place, which leads to the production of free acid. It is interesting to notice that if the paper which has become red is immersed once again in the original solution it resumes its blue colour. The nitrate, sulphate, and oxdate of ammonium behave similarly towards blue litmus ; in the case of the bromide, reddening occurs almost immediately the paper is withdrawn from the liquid. F. H. L. Estimation of Selenium. G. Pellini and E. Spelta. Quantitative Separation of Selenium and Tellurium. (Gazx. chiin. ital., 1903, xxxiii., [2], 89 and [l}, 515 ; through Chem. Zeit. Rep., 1903, 298 and 281.)-For the determination of selenium Pellini and Spelta propose to make use of the reaction between selenious anhydride and hydrazine sulphate or hydrochloride- NgE4 + SeO, = Se + 2H,O + N2 ; but instead of weighing the reduced selenium, they measure the nitrogen, 201.83 C.C.G. Pellini.THE ANALYST. 57 of which are equivalent to 1 gramme of SeO,. The original solution, rendered faintly acid with hydrochloric acid if desired, is brought into a flask, diluted to 50 C.C. with water, and mixed with a few grammes of sodium chloride in order to prevent loss of selenious anhydride on boiling. The flask is fitted with a cork and leading tube, its contents are boiled till all the air is expelled, and a boiling solution containing about 2 grammes of hydrazine sulphate or .hydrochloride and some hydrochloric acid is introduced.The nitrogen is led through previously boiled water, and is collected preferably in the Schultze and Tiemann apparatus. Pellini's process for the separation of selenium from tellurium depends on the fact that if a solution of selenious and tellurous anhydrides containing a little free hydrochloric acid is mixed with saturated solutions of ammonium hydrogen tartrate, tartaric acid, and hydrazine sulphate (not the hydrochloride), the selenium is reduced and precipitated whiIe the tellurium remains dissolved. The reaction is complete in the cold, but it proceeds more smoothly at a temperature between 50" and 60" C., and is ended in one or two hours. The liquid is then warmed rather more strongly in order to cause the precipitate to cohere, a, little more hydrazine sulphate is added to see whether the selenium has been wholly thrown down, and finally it is collected on a tared paper or in a Gooch crucible, washed with warm water and absolute alcohol, dried at 105" C., and weighed.The filtrate is treated with sulphuretted hydrogen, and the preoipitate is brought on to a paper and washed with water. The filter with its contents are transferred to a crucible and mixed with fuming 1-52 nitric acid; the excess of acid is evaporated on the water-bath, and the residue is dissolved in hydrochloric acid. The sulphuric acid derived from the oxidation of the sulphur is next removed by mean8 of barium chloride, and from the filtrate the tellurium is precipitated with hydraaine hydrochloride.The tellurium is finally washed rapidly with water, and further treated as above described. The results are accurate; but if hydrazine hydrochloride is employed for the precipitation of the selenium, it will be found contaminated with tellurium. F. H. L. Determinationof Sulphur in Coal aad Coke. R. Nowicki. (Stahl und Eisen, 1903, xxiii., 1141; through Chm. Zed. Wep., 1903, 281.)-One gramme of the powdered sample is mixed with 2 grammes of a mixture of one part of sodium carbonate and two of m&gnesia. The mass is arranged in a platinum crucible in such a way that a vertical and cylindrical air-space is left in the centre. The lower part of the crucible is raised to dull redness, and a current of oxygen is intro- duced into the upper part by means of the lid of a Rose crucible.The material is stirred up every five minutes, the air-space being always restored ; and after twenty or thirty minutes' incineration will be found to be complete. The residue is treated in the usual manner. F. H. L. The Separation of Coal from Earthy Impurities. P. Nyssens. (Bull. de Z'Ass. belge, 1903, 317, 318.)-Since the specific graviby of coal ranges from 1.16 t o 1.60, whilst clay varies in density between 1-70 and 2-20, and the other earthy58 THE ANALYST. impurities are usually still more dense, it is possible to effect a rapid separation by treating the finely-divided powder with a solution of ferric sulphate of specific gravity 1.50. The coal rising to the surface is collected on a filter, washed with water, dried at 100" C., and analysed by the usual methods.C. A. M. Determination of Free Phosphorus. J. Katz. (Oesterr. Chem. Zezt., 1903, vi., 515.)-This is a modification of a method suggested by Straub, which the author has found to be very tedious. The liquid containing the phosphorus is shaken with a solution of copper nitrate until a permanent black emulsion of copper phosphide is formed, then ether is added, and the whole agitated again. A sufficient quantity of hydrogen peroxide is next introduced till, after much shaking, the black colour dis- appears. The ethereal liquid is separated and repeatedly extracted with water ; the combined aqueous liquids are treated with a few drops of hydrochloric acid, and concentrated on the water-bath to a volume of 10 or 20 C.C. After filtration ammonia is added until the precipitate which first forms is redissolved, and the phosphoric acid is finally estimated with magnesia mixture in the usual manner. F.H. L. A Method for the Estimation of Chlorides, Bromides, and Iodides. Stanley Benedict and J. F. Snell. (Journ. Amer. Chem. Soc., xxv., 1138.)-The author makes use of the difference in behaviour of chlorides, bromides, and iodides towards potassium iodate (ANALYST, xxviii., 305) to determine these salts quantitatively when present together. The total halogens present in the mixture are first determined by any suitable method; iodine and chlorine are next determined as described below, and bromine is found by difference. To determine iodine, a quantity of the sub- stance, containing not more than 0-5 gramme iodine and 0.15 gramme chlorine, is dissolved in 50 C.C.water in a, stoppered 100 C.C. cylinder, About twice the theoretical quantity of neutral potassium iodate is added, and the solution acidified with 4 or 5 C.C. of 30 per cent. acetic acid. The liberated iodine is removed from the aqueous liquid by shaking with 30 to 40 C.C. carbon bisulphide, the carbon bisulphide being then separated by filtration through a wet filter-paper. After washing the carbon bisulphide solution with cold water on the filter, it is transferred to a beaker and covered with 20 to 25 C.C. of 75 per cent. alcohol, the filter-paper also being washed with some alcohol of this strength. The iodine is then titrated with sodium thiosulphate without using starch. For the determination of chlorine, the aqueous filtrate from the carbon bisulphide is treated with 5 C.C.nitric acid (specific gravity 1.18) and boiled till the whole of the bromine set free has been expelled. The excess of iodate added is next destroyed by adding a small excess of potassium iodide and boiling the liquid till colourless, 2 or 3 C.C. more nitric acid being added if all the iodine has not been driven off after ten or fifteen minutes' boiling. The liquid is then exactly neutralized with sodium carbonate, a little calcium carbonate being added at first to act as indicator, and chlorine determined by titration with silver nitrate, using potassium chromate as indicator. From the test analyses quoted it appears that the method is satisfactory. A. G. L.THE ANALYST.59 The Estimation of Carbon Dioxide in the Presence of Chlorine. C. Offer- haus. (Zeit. f. angew. Chem., 1903, xliii., 1033.)-This problem is of some importance in view of the fact that electrolytic chlorine contains varying and con- siderable quantities of carbon dioxide. In the first the gaseous mixture to be examined is led through two Bunte burettes in succession until all the air has been displaced. In one burette chlorine is estimated by absorbing in potassium iodide solution and titrating in the usual way. In the other burette the carbon dioxide and chlorine are estimated by absorbing both in dilute caustic soda solution of one-fifth normal strength and observing the diminution in volume, making due correction for the moisture in the gas. From the difference between the results given by the two burettes the percentage of carbon dioxide can be calculated.In the second method use is made of a standard solution of caustic soda about one-fifth normal, in which the carbonic acid has been determined by adding barium chloride and titrating with oxalic acid in the presence of phenolphthalein, according to Winkler’s method. As before, a known volume of the gas is collected in a Bunte burette, 45 C.C. of the caustic soda solution are added and shaken with it ; then 5 or 10 C.C. of pure 3 per cent. hydrogen peroxide solution are introduced, and the burette is again shaken. The liquid is then made up to exactly 200 c.c., of which 50 are taken for the determination of carbon dioxide by Winkler’s method. It is claimed that both methods give satisfactory results.Two methods are recommended. A. M. The Gasometric Determination of Bromates. Max Schlotter. (Zeits. Anorg. Chm., xxxvii., 172.)-The author shows that alkali-metal bromates are quantitatively reduced by hydrazine sulphate according to the equation : 2NaBr0, + 3NH, - NH, = 2Na;Br + 6H,O + 6N. He consequently bases a method for the determination of bromates on the quantity of nitrogen evolved on treating them with hydrazine sulphate in the apparatus shown in the figure. The 200 C.C. flask A contains the alkali bromate, and the flask B three times its weight of hydrazine sulphate ; both flasks are filled to about two-thirds their capa- city with water, and are con- nected together as shown. At the commencement of a deter- mination the pinch-cock h, is closed, h opened, the exit of the tube r being immersed in water.and both flasks heated until only 30 or 40 C.C. water remain in each in order‘to expel the air completely. Towards the end of the boiling h is opened and h, closed in order to fill the tube r with steam; h is then also closed, the flasks are allowed to cool, and h, is opened, when the hydrasine solution will60 THE ANALYST. pass over into B from A ; h, is closed as soon as air begins to enter the tube rl ; A is then heated, h opened as soon as there is an excess pressure in the flask due to evolved nitrogen, and the heating continued until no more gas-bubbles pass into the measuring tube. The volume of the gas is then read off as usud. In carrying out determinatione by the above method, the author found that in every ctwe the quantity of nitrogen found exceeded the theoretioal by an amount which is independent of the quantities of bromste and hydrazine sulphate used, provided only that the latter is in excess.The excess volume found appears to depend only on the apparstus used, and if it is determined once for all and subtracted from the volumes found, good results are obtained. A. G. L. APPARATUS. A Nitrometer for the Estimation of Uric Acid and Urea in Urine. A. Jolles. (Oesterr. Chem. Zed., 1903, vi., 509.)-As shown by the accompanying illustration, the decomposing vessel in this apparatus is a flask having a long neck, the advantage of which is that it can be st lagged,” and the flask shaken without danger of communicating the warmth of the hand to the liquid inside. The inner recep- tacle for the bromine is fused into the neck of the flask proper, and it is bent t o one side at its lower end in order to render it a more efficient agitator of the liquid in the flask. This inner receptacle is charged with the bromine reagent-preferably by means of a long funnel-through the upper mouth of the flask, while the urine is introduced through the lateral aperture thereof. The lateral aperture of the inner receptacle is made somewhat large, so that the reagent can be added quickly when the flask is suitably tilted. For exact work, the flask is supported inside a vessel containing water in the manner shown, but a water- jacket round the measuring tube is not really required. The side cock on the tube leading from the flask serves for The apparatus is made by Gockel of a preliminary adjustment of the pressure. Berlin. F. H. L.