ii. 20 ABSTRACTS OF CHEMICAL PAPERS. Analytical Chemistry. New Form of Safety Pipette. A. K. REFIRMAN ( J . h l . Eng. Chem. 1917 9 1047).-A three-way tap is attached to the top of an ordinary pipette by means of a short length of rubber tubing and a rubber bulb fitted with suitable valves is connected with the upper limb of the tap. The pipette is filled by pressing and releasing the bulb. The tap is then turned so as t o admit air t o the pipette by which means tlie level of the liquid is lowered t o tlie Preparation of Ammonium Citrate Solution and the Estimation of Insoluble Phosphoric Acid. PHILIP McG. S'HUEY ( J . Ind. Eng. Chein. 1917 9 1045).-Neutrral ammonium citrate solution may be prepared by dissolving 1814.37 grams of citric acid in 6961 C.C. of water and 1760 C.C.of 28% ammonia the water and aniinonia being measured a t 2 3 O . The insoluble phos- phoric acid in acid phosphate may be estimated with practically identical results whether or not tlie weighed portion has been washed previously with water and preliminary washing of samples coil- taining cyanamide does not appear to be necessary. T t may be im- portant t o use a neutral ammonium citrate solution in the case cf ground tankage whale guano meat guano fish and similar mate- w. P. s. Micro-method for the Estimation of Inorganic Phos- phates in the Blood-serum. W. MCKIN MARRIOTT and F. H. HAESSLER ( J . Rid. Chem. 1917 32 241-243).-Full details of the method previously described (compare Howland Haessler and Marriott A. 1916 ii 269). The Perchlorate Method for the Estimation of the Alkali Metals.F. A. GOOCII and G. R. BLAKE ( A n z e ~ . J. Sci. 1917 [ivl 44 381-386. Compare A. 1917 ii 270).-Tn this method i t is unnecessary t o use an alcoholic liquid saturated with the substance to be precipitated since the volunie of the solution (97% alcohol containing 0.1% of percliloric acid) used for washing the precipitate may bel so restricted that the solubility of the precipitated per- clilorates is insignificant. A single evaporation with a moderate excem of perchloric acid (0.1 C.C. for each 0.1 gram of salt) is not sufficient to convert considerable quantities of alkali chlorides (for example 0.3 gram) completely into perchlorate ; the residue should b3 treated with a further quantity of perchloric acid and again evaporated.For the separation of large amounts of insoluble per- chlorates from sodium perchlorate the washed precipitate should be dissolved in a small volume of water and the evaporation with perchloric acid repeated. I n the case of rubidium perchlorate the precipitate should be digested for fifteen minutes with the alcoholic washing solution. It is to be noted that perchloric acid has a dis- tinct solvent action on glass. [See also J . SOC. Chem. Ind. 1918 mark a d then delivered. w. P. s. rials which have not been strongly acidified. H. W. B. 37 214.3 w. P. s.ANALYTICAL CHEMISTRY. ii. 21 Micro-method for the Estimation of Calcium and Mag- nesium in Blood-serum. W McKni MARRIOTT and JOHN HOWLAND (*7. R i d . ( ' h ~ m . 1917 32 233-239).-Full details of tlis niethtl previoiisly descrihtl (cortipare Rowlniicl Haeusler anil Marriott A .191G ii 269). €1. IV. 13. Estimation of the Metal Content of Powdered Metals. F. HODES (Zeitsch. nngezu. Chem. 1917 30 240).-The quantity of metallic tungst.en in powdered tungsten which has been prepared reduction with carbon may be estimated by igniting a portion o f the sample in an open crucible; the increase in weight is due t o oxidation of the metal to tungsten trioxide. As however the pow- dered metal usually contains small quantities of carbon and mois- ture i t is necessary to heat anotlier portion of the sample in a current of oxygen weigh the amounts of carbon dioxide and water given off and correct the first weight accordingly. w. P. s. Iodometric Estimation of Copper and Iron.HERM. LEY (Clmn. Zeit. 1917 41 763).-Soluble cupric and ferric salts both liberate iodine from potassium iodide in acetic acid solution and the iodine set free is a measure of the amount of cupric and ferric salts present. When it-is desired t o estimate copper in the p r e ~ n c ~ of iron the latter may be precipitated as ferric phosphate which is irisoluble in acetic acid and does not react with potassium iodide; ci:pric phosphate however is readily soluble in acetic acid. The solution containing the cupric and ferric salts is treated with sodium phosphate solution then acidified with acetic acid potassium iodide is added and the iodine is titrated with thiosulphate solution. The ferric salt is estimated by treating another portion of the original solution with potassium iodide and acetic acid and titrating the liberated iodine; the difference between the volumes of thiosulphate solution used in the two titrations is equivalent t o the quantity of ferric salt present.Aluminium and zinc do not interfere. IT. P. s. Sensitive Reaction of Hydrogen Peroxide depending on the Formation of Dihydroxytartaric Acid. G. DENIG~S (,4nn. C?iim. c1)iaZ. 1917 22 193).-A violet coloration is formed when a small quantity of dilute hydrogen peroxide solution is added t o a mixture of 2 C.C. of 5% tartaric acid solution and 2 drops o l 5% ammonium ferrous sulphate solution and the mixture then treated with 6 drops of sodium hydroxide solution. The reaction may be obtained with a quantity of hydrogen peroxide not exceed- w. P. s. ing 0.05 mg.Estimation of Glycerol by the Iodide Method using Small Quantities of Hydriodio Acid (Semimicro-iodide Method). R. NEUMANN (Zeitsclz. mzgezo. Chem. 191 7 30 234-237).-The author finds) that this method (compare A. 1902 ii 111 585; 1903 ii 515) yields trustworthy results when onlyii. 22 ABSTRACTS OF CHEMICAL PAPERS. about one-t,enth of t,lie usual quantities of sample and reagents are employed. Thp apparatus user1 has about one-eighth of the capa- city of t h e iodide apparatns d e s c r i h l hy Stritar ( A 1904 ii 95). w. P. s'. Oxidation of Organic Compounds with Chrorn ic Acid. A. WINDAUS ( Z e i t s c h . phgsiol. C'hem. 1917 100 lG7-1G9).- Certain organic compounds readily yield acetone and acetaldehyde when treated with chromic acid. The products can be separated by distillation and recogiiised by conversion into the comparatively iiisoluble p-nitrophenylhyrlrazones. In this way cholesterol coprost erol and sitosterol containing an isopropyl group can be readily clistinguished from the somewhat similar bile acids which do not yield any acetone or aldehyde when oxidised by chromic acid.Methylpentoses such as rhamnose are similarly easily detected in the presence of other pentoses and hexoses by means of the chromic acid oxidation test. The test must be carried out' under the prescribed conditions involving the use of a solution of chromic acid in glacial acetic acid. If an excess of sulphuric acid is sub- stituted for the glacial acetic acid lzvulose and even dextrose are found to yield acet,aldehyde and other volatile products which can form iodoform and iiisoluble 11-nitrophenylhydrazones (compare Engfeldt A.1917 ii 550). H. W. B. Formaldehyde as a Negative Catalyst in Sugar Reactions. H. MAGGI and G. WOKER (BPY. 1917 50 1331-1335).-Form- aldehyde hinders the reducing action of dextrose or maltose in the Moore-Heller and Rubner tests and especially the reduction of methylene-blue or picramic acid. Although formaldehyde has a more powerful reducing action on Nylander's solution than the sugars mixtures of them containing 1 C.C. of 40% formaldehyde to 0.5 C.C. of 1% dextrose or maltose do not affect this reagent. The aldehyde also hinders the reduction of Fehling and Pavy solu- tions by dextrose or maltose. J. C. W. Malic Acid and Citric Acid. T. C. N.BROEKSMIT (Phnm. TT'Pe?;bZnd 1917 54 1371-1373).-Both malic acid and citric acid answer to the iodoform test but can be distinguished by the fact that barium malate is not precipitated either in neutral solu- tion or in presence of acetic acid. A. J. W. Probable Accuracy in Whole Blood and Plasma of Colorimetric Estimations of Creatinine and Creatine. ANDREW HUNTER and WALTER R. CAMPBELL (1. R i d . Chen2. 1917 32 195-231. Compare Wilson and Plass A. 1917 i 360 and Gettler and Oppenheimer A. 1917 ii 184).-The authors have prepared a series of curves showing the rate a t which the colour utilised as the basis of a creatinine estimation develops not only in pure aqueous solutions but also in the various circumstancesii. 23 AXALYTICAL CHEMISTRY. in which it is applied to the analysis of blood.A coiupwisoii of these curves reveals the presence or absence of substaiices capable of producing a colour similar to that arising from creatinine. In plasma the only substance capable of simulating the reaction for creatinine is dextrose and its influence on the estimation of creatinine is shown to be too small to have much practical import- ance. I n whole blood however an unknown substance is present which although reacting more slowly than creatinine contributes in the Folin technique an appreciable fraction of the total colour developiiig within ten minutes. On laking the blood a much larger amount of this unknown substance is set free and passes into the protein-free filtrate. The filtrates from autocIaved blood or plasma also contain relatively large quantities of the substance which reacts similarly to creatinine.A consideration of these results indicates that although the Folin method estimates the preformed creatinine of plasma with a satisfactory approximation to accuracy the results obtained by the same method for the preformed creatinine of whole blood are on the average about 50% higher than the truth. The technique of Myers leads t o a still greater exaggeration of the preformed creatiiiine of blood. The Folin method for creatine whether applied t o whole blood or to plasma also gives decidedly erroneous results liable to be in the former case about twice and in the latter about four times as high as the amount actually present. H. W. B. Estimation of Faecal Indole. OLAF BERGEIM ( J .Biol. Chen~. 1917 32 17-22).-The fmes are mixed with potassiuni hydroxide solution and the indole distilled over in ail ordinary Kjeldahl dis- tillation apparatus Any ammonia in the distillate is removed by a second disttillation after acidifying with dilute sulphuric acid. The final distillate is treated with sodium B-naphthaquinone- sulphonate and alkali and the blue indole compound formed ex- tracted with chloroform and estimated colorimetrically (compare Herter and Foster A. 1906 ii 134 910). H. W. B. Improvement of the Volumetric Method of Estimating Albumin by means of Potassium Ferrocyanide. Presence of Rapidly Decomposable Proteins in certain Urines. ED. JUSTIN MUELLER (Bull. Sci. Plccrrrrincol. 1917 24 221-224 ; from C’kanz. Zentr. 1917 ii 325.Compare A . 1917 ii 555).-In the previous communication the author has described a method of estimating albumin in urine and pointed out that the immediate occurrence of the colour change with iroii alum (after addition of 3.5 C.C. of potassium ferrocyanide necessary for the saturation of the acidified water) is indicative of the absence of albumin. Sub- sequently however i t has been observed that the colour change can also occur in the presence of readily deconiposa,ble proteins. These differ from the albumins known in urology. They do not yield a precipitate when warmed or ~7he11 treated with acetic acid.ii. 24 ABSTRACTS O F CHEMICAL PAPERS. They are precipitated by warin trichloroacetic acid aud give pre- cipitates with cold trichloroacetic acid cold nitric acid and Esbach’s reagent which gradually dissolve when warmed.A cold solution of potassium ferrocyanide in acetic acid yields a pre- cipitate which is stable when warmed ; saturated ainmoiiium sulphate or sodium acetate solution gives a precipitate soluble in water whilst precipitates are not produced by sodium chloride or iiiagnesium sulphate. The prot eiiis are readily hydrolysed when warmed or in acetic acid solution. They are rendered more stable by addition of ammoiiiuin sulphate. For the volumetric estimation urine (50 c.c.) is treated with so much finely divided ammonium sulphate (8-9 grams) that a volume of 55 C.C. is attained; the solution is .filtered and 11 C.C. of the filtrate are used for the titration. H. W. Adaptation of Winkler’s Method to Biological Work.W. J. V. OSTERHOUT and A. R. C. HAAS ( J . B i o l . Chenz. 1917 32 141-146).-The apparatus consists of two glass tubes 30 cm. long and 27 mm. in diameter joined together by rubber tubing and furnished with rubber stoppers a t the free ends (all the rubber tubing and stoppers used in the apparatus should be coated with paraffin). One of the rubber stoppers is connected with a series of smaller glass tubes (13 mm. inside diameter) joined by rubber tubing with only enough space between them to permit the inser- tion of a clamp. The larger tubes are filled with the solution containing the organisins of which the oxygen metabolism is to be investigated. To estimate the oxygen in the solution a t any time the organisms are first allowed to gravitate to the bottom of the two tubes and then the lower one containing the organisms is clamped off and removed. The smaller tubes are then filled successively with alkaline potassinm iodide manganese chloride and concentrated hydrochloric acid each being separated from the other by means of the clamps. By releasing the clamps the reagents are successively introduced into the oxygenated solution and filially the contents are run into a beaker and titrated with thiosulphate as in Wiukler’s method for the estimation of oxygen i n solutions. The advantages claimed for the apparatus are that i t permits tlie reinoval of the organisms before adding the reagents and also the addition of the reagents without danger of contaminatioil by oxygen. Modifications are described whereby it is possible to take samples of the solution for analysis a t intervals during an experi- ment. H. W. B.