ii. 30 ABSTRACTS OF CHEMICAL PAPERS. Analytical Chemistry. Quantitative Analysis of Small Quantities of Gases. H. M. R Y D ~ ( J . Amer. Chem. Soc. 1918 40 1656-1662).-A somewhat complicated apparatus is described by means of which a quantitative analysis of 5 cu. mm.-1 C.C. of a gaseous mixture of water vapour carbon dioxide carbon monoxide oxygen hydrogen nitrogen and methane may be carried out with an accuracy of about 5% for each constituent The apparatus makes use of a McLeod gauge and an optical lever gauge but for details the original should be consulted. OLIVER D. BURKE (Chenz. News 1918 117 368-369).-The glass tube through which the J. F. S . Gas Bubbler for Gas Analysis.ANALYTICAL CHEMISTRY. ii. 32 gas enters is drawn out into a fine capillary. Another glass tube 1s fused on to the side of the first tube and bent so that the end extends just below the end of the capillary where it is flattened and the top surface ground.The capillary is made to fit tightly on the ground surface so that the gas escaping from the capillary is broken up into very fine bubbles. [See further J . SOC. Chem. I d . January 1919.1 w. P. s. Oxidising Action of Potassium Dichromate as Com- pared with that of Pure Iodine. CARL’ R. MC~ROSKY (J. Amr. Chenz. SOC. 1918 40 1662-1674).-The author has studied the trustworthiness of potassium dichromate as a standardising agent in iodometric determinations. It is shown that potassium di- chromate always liberates more than the theoretical quantity of iodine from hydriodic acid. This excess of iodine may be reduced but not wholly removed by removing dissolved air from the solu- tions by drying and fusing the dichromate in the absence of air and by recrystallising repeatedly to remove oxidising impurities.It is also shown that the excess of iodine is not due in any way to the catalytic action of chromium chloride. J. F. S. Estimation of Active Oxygen in Sodium Peroxide. JAROSLAV MXLBAUER ( J . p. C’hem. 1918 [ii] 98 l-S).-The following methods have been proposed for the estimation of active oxygen in sodium peroxide (1) liberation of hydrogen peroxide by water followed by titration with potassium permanganate (2) treatment of sodium peroxide with potassium iodide and potassium hydrogen carbonate and titration of the iodine liberated with sodium arsenite and (3) measurement of the oxygen liberated by water in the presence of cobalt nitrate.The results obtained by the first method are very low those by the second are somewhat better but still low whilst the third method gives high results. The methods have been critically examined and improved; the following processes yield accurate results (1) Water (100 c.c.) is mixed with concentrated sulphuric acid (5 c.c.) and chemically pure boric acid (5 grams); sodium peroxide (0.5 gram) is gradu- ally added to the mixture which is kept briskly shaken and the liberated hydrogen peroxide is titraked with potassium perman- ganate. The usual permanganate method gives low results since a portion of the hydrogen peroxide is catalytically decomposed by the manganese sulphate formed during the process. (2) Sodium peroxide is gradually introduced into a solution of potassium iodide (2 grams) in dilute sulphuric acid (1 in 20; 200 c.c.); the iodine is titrated with standard thiosulphate.The results agree fully with those obtained by the permanganate method. (3) Sodium peroxide (0-2-0.3 gram) is mixed with about 10 C.C. of copper sulphate solution (0.05%) in a small flask connected to a nitro- meter; the flask is shaken and decomposition is complete within a minute when the liberated oxygen is measured. The gas evolved contains about 0.32% of carbon dioxide and 0.08% of5. 32 ABSTRAWS OF CHEMICAL PAPERS. hydrogen. With cobalt nitrate as catalyst the results are in- variably high; the authors consider this may indicate the presence of an oxide higher than t,he peroxide. The action of the atmosphere on sodium peroxide has also been investigated; moisture appears to be more active than carbon dioxide in causing decomposition. H.IV. Pregl's Micro-estimation of Nitrogen. HANS FISCHER (Ber. 1918 51 1322-1325) .-The estimation of nitrogen in difficultly combustible substances by the micro-Dumas method (compare Dubsky A. 1918 ii 130) gives untrustworthy results and the author now employs only the original Pregl method. The causes of the inaccuracies in the former method are discussed. The author uses the micro-Pregl method in preference t o the Lassaigne test for tihe detection of nitrogen in rare or very valuable organic compounds. c. s. Detection and Estimation of Hydrogen Phosphide in Hydrogen. J .SOYEB (Ann. C k m . anal. 1918 23 221-225). -Hydrogen prepared by the action of sodium hydroxide on ferro- silicon always contains traces of hydrogen phosphide. The presence of the latter may be detected by burning the hydrogen from a platinum jet and directing the flame on to the edge of a porcelain basin; the flame has a green coloration. When examined with the spectroscope the flame exhibits the phosphorus spectrum. If tt drop of water suspended on a glass rod is held in the flame for fifteen seconds and then tested with molybdic acid reagent a yellow precipitate is obtained. The amount of hydrogen phosphide present is estimated by passing a definite volume (from 2 t o 30 litres) of the gas together with a large excess of air through a platinum jet arranged in a silicon tube heated to bright redness; this tube is inclined slightly and its lower end is connected with absorption vessels containing water.When the desired quantity of the gas has been burned the tube and the contents of the absorption vessels are rinsed into a basin treated with 5 grams of ammonium nitrate concentrated t o about 40 c.c. and the phos- phoric acid precipitated with molybdic acid reagent. [See further w. P. s. J . Soc. C"hem. Id. 1918 765a.I Estimation of Minute Quantities of Arsenic. 0. BILLETER (Helv. Chim. Acta 1918 1 475-498).-A more extended and somewhat modified account of the method previously published (A. 1915 ii 578). The estimation of arsenic in organic substances is effected in the following manner. The organic matter is destroyed either by treatment with a mixture of nitric and concentrated sulphuric acids or in the case of urines by rendering the latter alkaline with sodium carbonate evaporation to a syrup admixture with potassium perchlorate (2 grams) and potassium sulphate (4 grams) for each 100 C.C. of urine desiccation of the mixture at 120° andANALYTICAL CHEMISTRY.ii. 33 gradual introduction of the latter into a platinum crucible heated to dull redness followed by more intense ignition until tranquil fusion is attained The arsenic is separated from other metals by distillation with a mixture of sodium chloride (2 grams) and potassium bromide (0.2 gram) for each 20 C.C. of sulphuric acid (the addition of bydrazine sulphate previously recommended is found t-o be unnecessary) or if destruction of organi matter has been effected with potassium perchlorate by treatment with potassium bromide (0.1-0.2 gram) and sulphuric acid (go% 6 c.c.).Mercury is completely retained by one distillation but if antimony is present it is necessary to redistil after addition of 5-6 grams of sulphuric acid (90:L). Hydrochloric acid is eliminated from the distillate by breatment with hypochlorous acid and the solu- tion is evaporated t o dryness. The residue is dissolved in sulphuric acid (lZ% 1 c.c.) and evaporated on the water-bath to destroy any chloric acid that may be presentl; after addition of water (0.85 c.c.) it is transferred to a Marsh’s apparatus. By this method 0.01 mg. of arsenic may be detected The precautions necessary for ensuring the requisite purity of the reagents are fully described.The commercial pure sulphuric acid is diluted to 85(;. and heated with sodium chloride (3%) and potassium bromide (0.3:4,) in a quartz flask; the treatment is twice repeated with smaller quantities of salts and the acid is finally distilled from a quartz retort and collected in a quartz flask. The nitric acid and the salts are purified by Lockemarm’s method. Zinc is conveniently obtained by the electrolysis of an aqueous solution of zinc sulphate using a zinc anode and copper cathode; the metal is obtained in the pulverulent state and dissolves readily in dilute sulphuric acid without being activated. It is free from arsenic. Hypochlorous acid is best prepared by the solution of chlorine monoxide in water. The sensitiveness and constancy of the arsenic mirrors depend considerably on the quality of the calcium chloride used in desic- cating the gas.Commercial fused calcium chloride is unsuitable as it is always strongly alkaline; a neutral product can be obtained by dehydrating the crystalline substance a t ZOOo and subsequently gradually heating the finely divided product to its melting point in a quartz tube in a slow current of dry hydrogen chloride. The best results are obtained by starting from metallic calcium. H. W. Amount of Amorphous Silica in the Soil. BELA VON HORVATH (Bied. Zentr. 1918 47 97-98).-For the estimation of the amorphous silica 5 grams of the soil are extracted with 100 C.C. of a 1% sodium carbonate solution for fifteen minutes a t looo. Solutions of sodium carbonate of greater concentration than 1% dissolve silicates and quartz besides amorphous silica the results obtained being consequently too high.When estimated by the author’s method soil is found to contain only a few milligrams per cent. of amorphous silica. El. w. B.ii. 34 ABSTIEACTS OF CHBMICBL PAPERS. Estimation of Metals by Electrolytic Deposition without using an External Supply of Electricity. NAURXCE E’RAN~OIS (Compt. rend. 1918 167 725-727).-A strip of nickel is placed across the top of a platinum crucible and a zinc rod 5 m. in diameter is suspended from the strip; the rod is notched so as to fit on to the nickel strip and the lower end of the rod extends nearly to the bottom of the crucible. The zinc rod is amalgamated at least twenty-four hours before being used and is wrapped in filtier-paper which is tied on to the rod with ordinary cotton thread.The electrolyte used for the deposition of silver or gold consists of 9 C.C. of 10% potassium cyanide solution 5 C.C. of potassium hydroxide solution (D 1*332) and 2 C.C. of ammonia; for the depositaon of mercury the solution should consist of 20 C.C. of 10% sulphuric acid containing 0-5 gram of potassium iodide. I n all cases the deposition requires twenty-four hours for com- pletion. [See further J . SOC. Chem. Ind. 1918 784a.1 w. P. s. Quantitative Estimation of Ions by Microanalytical Methods. I. ROBERT STREBINGER (&sterr. Chem. Zeit. Lii] 21 71-73; from Chem. Zentr. 1918 ii 471).-The author has extended Pregl’s method of quantitative organic micro- analysis to inorganic substances and describes the estimation of silver nickel arsenic iron chromium and copper and the separations of silver from copper and lead from tin.Precipita- tion of nickel with a-benzildioxime is unsuitable for micro- analytical purposes the results being too high. H. W. Gravimetric Analysis. VI. Estimation of Calcium. VII. Separation of Calcium from Magnesium. L. W. WINKLER (Zeitsch. angew. Chem. 1918 31 187-188 203 214-216).-Estimation of calcium as oxalate or carbonate was submitted to critical examination. It is recommended that the precipitation as oxalate should be made from an acetic acid solu- tion in the presence of ammonium chloride; the calcium oxalate should be weighed as such since ignition to oxide is less trust- worthy especially in the presence of sulphates.If the latter are present the oxalate always contains sulphate but the weight is not affected since the two have practically identical molecular weights. For the separation of calcium from magnesium the calcium is precipitated as described from an acetic acid solution; the magnesium is subsequently precipitated as ammonium mag- nesium phosphate and weighed in this form. [See further J . SOC.. Estimation of Calcium and Magnesium in different Saline Solutions. E. CANALS (Bull. SOC. chim. 1918 [iv] 23 4 2 2 4 3 0 ) .-The simplest and most satisfactory method of estim- ating calcium is t o precipitate it as oxalate in ammoniacal solu- tion and weigh it as oxide. The method of precipitation as Chem. Ind. 1919 29A.I w.P. s.ANALYTICAL CHEMISTRY. ii. 35 sulphate and weighing as such is also very exact providing that numeroas precautions are taken. [See further J. SOC. Chem. Ind. January 19191. W. G. Sensitiveness of the most usual Tests for Copper. A. WOBEB (Osterr. Chenz. Z e i t . [ii] 21 105-107; from Ch-em. Zentr. 1918 ii 560).--The usual tests for the detection of copper have been investigated with respect t a sensitiveness with a view to their application in microanalysis; the results are given in the form of a table. H. w. Separation of Hydroxides in the Ammonium Sulphide Group. W. D. TREADWELL (Schzuezz. Chem. Zeit. 1918 2 59-61 71-74; from Chem. Zenltr. 1918 ii 663-664).-The separation of bivalent f ro,m tervalent metals of the ammonium sulphide group by simple precipitation with dilute ammonia from solutions containing ammonium salts is generally not quantitative.Transitory local excess of the reagent cannot be avoided and the precipitate then carries down varying amounts of the bivalent metal. The necessary asymptotic approach to the neutral point can be easily att'ained by leading a current of dilute ammonia (obtained by blowing air through a saturated solution of ammonium chloride in the presence of calcined magnesia) through the solution. Precipitation by the gas has the advantage that it is possible to approach the neutral point with the necessary caution without correspondingly diluting the solution. I n the presence of manganese ammonia should not be added until neutralisation is complete since the manganese ion is readily oxidised by air in neutral solution and passes into the precipitate; neutrality to litmus is the farthest possible limit.I n the separation of chromium and manganese the latter is always adsorbed by the chromium hydroxide. The hydroxides of the tervalent metals carry down nickel and cobalt more readily than manganese and zinc. The error can be somewhat diminished by precipitating and filtering the main portion of the sesquioxides in distinctly acid solut.ion and subsequently precipitating the remainder from the filtrate after fresh addition of ammonium salts. The precipitates contain much more basic sulphate than chloride. Precipitates from solutions containing sulphates are more sandy in character and do not adhere to glass; when washed they lose the sulphate ion and become slimy.Mercuric oxide is not suitable for the precipitatioii of the ter- d e n t metals of the ammonium sulphide group as hydroxides. When warmed with dilute ammonium sulphate solution it only liberates an insufficient quantity of ammonia ; addition of ammonium chloride effects an improvement due to the formation of complex mercury salts. The precipitates are very easily filtered; they are ignited wet; when the admixed mercury compounds are quantitatively volatilised. For the precipitation of aluminiumt chrominm ancl iron afi. 36 ABSTRACTS OF CHEMICAL PAPERS. freshly prepared mixture of sodium or pobassium nitrite and ammonium chloride is to be preferred to the very unstable ammonium nitrite. The precipitates obtained from solutions of ferric salts by this process are very difficult to filter and are greatly contaminated with basic salta.The difficulty may be avoided by displacing the iiitrous acid from the solution by a current of inert gas at the temperature of the water-bath. 'The utility of the method is limited oming to the oxidising action of the nitrous fumes. A portion of the aluminium remailis in solution when a dilute solution of aluminium chloride is precipitated by sodium thio- sulphatei and boiling is continued until sulphur dioxide is com- pletely expelled; precipitation is still less complete in the case of chromium. H. VV. Arsenite Titrations of Perrnangaaate Solutions. ALOKE BOSE (Chem. News 1918 117 369-370).-Results of experiments are recorded showing that free nitric acid is not the cause of the abnormally high reducing value of sodium arsenite solution when this is used for the titration of permanganate (compare Ibbotson A.1918 ii 175). The formation of manganic compounds does iiot seem to be possible. Some complicated reactions may take place during the titration but- i t is quite clear from the results of titratious with ammonium ferrous sulphate solution thatl all the manganese is present as permanganate before the titration with arsenite is commenced. [See further J . SOC. Chem Znd 1919.1 w. P. s. Separation of Germanium from Arsenic by the Dis- tillation of the Chloride in the Presence of a Chromate. PHILIP E. BROWNING and SEWELL E. SCOTT (,4nzer. J. Sci. 1918 [iv] 46 663-665).-A modification of a method described previously (A. 1917 ii 546); chromic acid is used to oxidise the arsenic and t,he germanium chloride is then distilled from the hydrochIoric acid solution. Five C.C. of 10% potassium dichromate solution are sufficient to oxidise 0.25 gram of arsenious acid. A current of carbon dioxide inay be passed thrmgh the apparatus during the distillation t o facilitate the removal of the germanium chloride. If as little as 0.0005 gram of germanium oxide is pre- sent the distillate yields a white precipitate of germanium sulphide 011 the addition of hydrogen sulphide. [See further ,7. Snc. The Estimation of Phenol and the Three Isomeric Cresols irm Mixtures of these Substances. HARRY MEDFORTH DAWSON and CHRISTOPHER ARCHIBALD MOUNTFORD (T. 1918 113 C h e s n . J u d . 1918 7 8 5 ~ 1 w. P. s. 935-944)