ii. 72 ABSTRACTS OF CHEMICAL PAPERS. An alg tic a1 Chemistry . The Estimation of W a t e r and of Carbon Dioxide i n Minerals and Rocks. MAX DITTRICH and W. EITEL (Zeitsch. anorg. Chem. 1912 77 365-376. Compare A. 1912 ii 804).-It has not been found practicable to estimate water and carbon dioxide in minerals and rocks by fusion with sodium metaphosphate or borax as a constant weight is not obtained in blank experiments. On the other hand heating the material in a stream of dry air in a silica tube is found to be satisfactory. The air supply tube is also of silica and is ground in and held by springs. A spiral of silver gauze is used to retain sulpbur. I n the case of minerals which lose water with difficulty the final heating is performed by means of the blow-pipe. Rocks contain- ing epidote mica or hornblende often do not lose the whole of their water through these minerals becoming enclosed.Much better results are obtained by the use of a platinum electric furnace sur- rounding the silica tube. The water is then driven off completely except in the case of materials containing much fluorine. Calcite does not lose its carbon dioxide completely unless first mixed with powdered quartz with a blowpipe but is completely decomposed in the electric furnace a t 1100O. The silica tubes last well a t 1000° but devitrifg at higher temperatures By the use of a platinum-iridium tube and a boat of the same material the most refractory rocks lose their water and carbon dioxide completely at 1250-1300'. C. H. D. New Reagent for Chlorine and Bromine Free and- Com- bined.GEORGES DENIQES and L. CHELLE (Compt. relad. 1912 155 1010-1012. Compare A. 1912 ii 1208 ; Guareschi ibid. 989).-The reagent is prepared by the addition of 10 C.C. of a solution of magentaANALYTICAL CHEMISTRY. ii. 73 (1 in 1000) to 100 C.C. of sulphuric acid (5% by volume). This is then mixed with an equal volume of acetic acid and 4 C.C. of strong sulphuric acid are added. This solution is coloured yellow by chlorine and reddish-violet by bromine the colour being best shown by shaking with chloroform. If the halogens are combined as metallic salta the solution is acidified the iodine removed by a ferric salt the bromine liberated by the addition of potassium chromate and tested for and the chlorine then liberated by potassium permanganate and separately tested for.The vapours are absorbed in dilute alkali before applying the test. The method can be applied quantitatively for dealing with small quantities of bromine as in natural waters and is more delicate than the methods at present in use. The Iodic Acid Process for the Estimation of Bromine in Halogen Salts. FRANK A. GOOCH and P. L. BLUMENTHAL (Amer. J . Xci. 1912 [iv] 469-474).-A criticism of Bugarszky’s process (boiling the solution containing bromides and chlorides with potassium iodate and dilute sulphuric acid and determining the loss in iodine A 1896 ii 216). The authors find that although the process was thought to be an ideal one it is vitiated by secondary effects. The latter may be reasonably attributed to the action of small amounts of iodine monochloride or monobromide formed in the interaction of iodic acid and free iodine with hydrochloric acid and hydrobromic acid Substitution of chloric acid for iodic acid cannot be recommended.L. DE I(. W. G. Oxidisable Substances in the Air. L. SCHWARZ and GEoRa MUNCHMEYER (Chern. Zentr. 1912 ii 1695 ; from Zeitsch. Hyg. Imfekt.-Krankh. 1912 72 371-384).-Henriet and Bouyssy’s method (A. 1911 ii 532) whereby the titration of condensed water is taken as a measure of the degree of vitiation of an atmosphere is criticised. Such condensed water gives different readings at different tem- peratures and would only give a correct value if all the moisture in the atmosphere could be condensed or if it contained oxidisable substances to the same extent as the collected water which is not the case.J. C. W. An Apparatus for Microanalysis of Blood Gases and Micro-respirometry. HANS W INTERSTEIN (Biochem. Zeitsch. 19 12 46 440-449).-The apparatus (which is figured) is a combination of the Barcroft-Haldane and Petterson systems. It consists essentially of two pear-shaped flasks one of which is a compensation flask and the other serves for the actual analysis. They communicate by means of three-way stopcocks either with the outer air or with one another. The latter communication is through an etched capillary tube containing an oil drop. The analysis flask communicates furthermore with a manometer of which the narrower limb which is nearer the flask is a capillary graduated in millimetres. The outer limb is broader and is closed by a piece of indiarubber with a screw-clamp and a piece of solid glass rod.The flasks communicate ; the meniscus of the oil drop in the capillary connecting the twoii. 7’4 ABSTRACTS OF CHEMICAL PAPERS. flasks is brought to a certain point by turning the screw-clamp on the indiarubber and the height of the mercury in the graduated limb of the manometer is then read. This takes place with the same materials in each flask and the whole apparatus is kept in a water- bath. If the measurement to be carried out consists in the deter- mination of the oxygen content Haldane’s ferricyanide method is employed the reagent being kept in spoons fused on to the stoppers. The apparatus after the first reading and adjustment is then removed from the water-bath and shaken so as to bring the ferri- cyanide into the blood and ammonia mixtures whilst the two flasks are kept closed and out of communication with one another.They are then brought back to the water-bath brought into communication the position of the oil drop in the communicating flask is readjusted and the height of the manometer is read. The difference between this and theGrigina1 reading gives the amount of oxygen evolved. 3. B. 8. Estimation of Tellurium by means of Hydrazine Hydrate. J. B. MENKE (Zeitsch. anorg. Chenz. 1912 77 282-288).-See this VOI. ii 41. Detection of Nitrogen in Organic Substances H. ZELLNER (Pharm. Zeit. 1912 57 979-980).-For the detection of small quantities of nitrogen or when the amount of the substance to be tested is not large the following modification of the usual test may be employed.The substance is fused with potassium the fused mass is dissolved in water and the solution is warmed after the addition of ferrous sulphate. The solution is then filtered and the filtrate is poured on the surface of ferric chloride solution acidified with hydro- chloric acid. If nitrogen is present in the substance a blue zone appears at the junction of the two liquids. w. P. s. A New Gas-analytical Method for the Estimation of Nitric Oxide. I. OSKAR BAUDISCH and GABRIEL KLINGER (Ber. 1912 45 3231-3236).-The method depends on the fact that when air is passed into nitric oxide standing in contact with solid potassium hydroxide nitrogen trioxide (N,O,) is formed and immediately converted by the hydroxide into potassium nitrite there being no formation of nitrogen dioxide.Four-fifths of the contraction thereby caused is due to nitric oxide as expressed by the equation 4N0 + 0 + 4KOH = The analysis is carried out in Fuller’s gas-analysis apparatus ; the pipette contains slightly moistened stick potassium hydroxide with which the nitric oxide is in contact before air is passed in and all measurements are carried over mercury. Before passing the nitric oxide into hhe pipette the latter is completely filled with mercury. The method gives accurate results even in the presence of nitrous oxide or hydrogen. The reverse method of passing the nitric oxide into air standing over potassium hydroxide cannot be used since a mixture of N20 and NO is formed 4KN0 + 2H20.T. S. P.ANALYTICAL CHEMISTRY. ii. 7 5 Detection of Nitrous Acid in Water. PRIMOT (Chevz. Zentr. 1912 ii 1546-1847; from Bull. Sci. PharntacoZ. 1912 19,546-547). -To 10 C.C. of the water 4 or 5 drops of a 1-1*5% solution of benzidine o-tolidine or dianisidine in 30 -40% alcohol are added. The mixture is then acidified with 5 or 6 drops of acetic acid and shaken when the development of a yellow cclour which deepens in time shows the presence of a nitrite. Benzidine is the least sensitive reagent. The limit of sensitiveness is below 0.01 mg. of nitrous acid per litre. J. C. W. LUIGI ER~IANNO CAVAZZA (C?hem. Zentr. 1912 11 1061-1062; from Atti 11. Congr. Nax. Chifiz. uppZic. 1912j.-Ten grams of powdered pure indigo are intro- duced (with cooling) in small quantities into 50 C.C.of fuming sulphuric acid and after forty-eight hours the solution is diluted t o a litre. Fifty C.C. of the liquid are then diluted so that 10 C.C. corres- pond with 0.001 gram of nitrogen pentoxide. When dealing with nitrates a stronger solution may be employed. Organic matters if present should be removed by adding to 25 C.C. of the solution the requisite amount of potassium permanganate (previously determined) ; 30 C.C. of sulphuric acid are added and the whole heated over a small flame and titrated. When the liquid turns bright yellow the heating is discontinued and the titration rapidly brought to an end. I n testing ordinary water concentration by evaporation is not necessary as the GEORG LoCIiEMANN (Chem. Zed. 1912 36 1465-1 466).-Historical. The original Marsh apparatus and Estimation of Nitrates with Indigo. end reaction is so very sensitive.1,. DE K. The Marsh Arsenic Test. modifications of it are described. L. DE I(. Estimation of Small Quantities of Carbon Monoxide. OTTO BRUNCK (Zeitsch. angew. Chum. 1912 25 2479-2481).-A suitable flask of known capacity (say a litre flask) is fitted with a doubly perforated rubber cork closed by means of rods. After filling the flask with the air to be examined 50 C.C. of sodium palladium chloride (1 C.C. =5 0.004762 gram of palladium) are introduced from a pipette (while gently loosening the second rod) followed by 2.5 C.C. of 5% solu- tion of sodium acetate; the volume of liquid ip of course deducted from the volume of the gas tested. After an hour’& action with frequent shaking the palladium which has deposited owing t o the reducing action of the carbon monoxide is collected and ignited with the usual precautions; 1 gram of metallic palladium = 0.2624 gram or 210 C.C.of carbon monoxide. Hydrogeu and the unsaturated hydrocarbons should be absent as Simple Apparatus for the Estimation of Carbon Dioxide. W. R. FORBES (CIhem. News 1912 106 284).-The apparatus consists essentially of an inverted V-tube closed at either end and provided in the middle with a ground stopper carrying a delivery tube and tap. In use the acid is placed in one limb and the carbonate in the other. Beaction is started by suitably inclining the apparatus. they also exert a reducing action. L. DE K. H. W.ii. 76 ABSTRACTS OF CHEMICAL PAPERS. Cobaltinitrite Method of Estimating Potassium.F. H. MACDOUGALL (J. Amer. Chern. Soc. 1912 34 1684-1686).-1n the estimation of potassium by the cobaltinitrite method i t is usual to boil the yellow precipitate with excess of N/10-potassium permanganate then add dilute sulphuric and oxalic acids and finally to titrate with N/10-permanganate. The oxidation of the nitrite thus takes place in alkaline solution and the permanganate is reduced to manganese dioxide. The latter reacts but slowly with the oxalic acid added subsequently since the mixture is so dilute. The author has therefore studied the effect of adding dilute sulphuric acid to the yellow precipi- tate before oxidising it with perrnanganate. It has been found that accurate results can be obtained in this way but that a different factor must be eniployed in the calculation ; accord- ing to the usual method 1 C.C.of lY/lO-permanganate is equivalent to 0.0007111 gram K or 0,0008564 gram K,O whilst with the modified method 1 C.C. of N/lO-permanganate corresponds with 0.0006518 gram K or 0.0007850 gram K,O. IE. G. Determination of the Sensitiveness of the Hydroxide Re- actions for the Common Metals. LOUIS J. CURTMAN and A. D. ST. JOHN (L Amer. Chern. Soc. 1912 34 1679-1684).-A study has been made of the sensitiveness of the hydroxide test for metals. The experiments consisted in treating a solution of a salt of the metal with 10% ammonia or sodinm hydroxide the final volume being 5 C.C. I n the following cases ammonia was used as the precipitant and the limiting concentrations of the metals a t which a precipitate was visible were lead 1-6500 ; mercury (in the mercuric state) 1-13,000 ; bismuth 1-10,000 ; iron 1-80,000 ; aluminium 1-100,000 ; chro- mium 1-170,000 ; zinc 1-80,000.For the following metals sodium hydroxide was employed as the reagent and the limits found were silver 1-40,000 ; mercury (in the mercurous state) 1-200,000; copper 1-10,000 ; cadmium 1-40,000 ; nickel 1-80,000 ; cobalt 1-80,000 ; magnesium 1-10,000. I n the case of manganese either ammonia or sodium hydroxide can be used the limiting concentration in each case being 1-170,000. Estimation of Lead in Lead Paints. FRANZ UTZ (Chern. Zentr. 1912 ii 1788 ; from Fas.benxeit. 18 18-20).-The nitric acid solution of the lead compound is electrolysed at 60-65' by a current of 0.05-1.5 amperes using gauze electrodes and the lead dioxide is washed without breaking the current and then dried a t 200O.E. G. J. C. W. Cuprous Iodide. MORITZ KOHN and ARTHUR KLEIN (Zeitsch. anorg. Chem. 1912 77 252-254).-Cuprous iodide may be analysed by the method of Baubigny and Chavanne for iodine in organic substances (A. 1903 ii 510; 1904 ii 203). One to 1.5 gram of silver nitrate is dissolved in 40 C.C. of concentrated sulphuric acid in tl flask and after cooling 5 grams of pure potassium dichromate are added. This is then dissolved by warming and the whole is again cooled. After adding 0.3 gram of the substance the flask is warmedANALYTICAL CHEMISTRY. ii. 77 and shaken until the svolution of oxygen begins. The silver iodide formed a t first dissolves as iodate.After cooling the contents of the flask are poured into 80 C.C. of water reduced by sodium sulphite and allowed to remain until t h e silver iodide has settled. This is collected on a Gooch filter after the addition of nitric acid dried a t 130° and weighed. Copper is estimated by heating 0-5 gram of the substance with 15-20 C.C. of nitric acid (1 l) converting into sulphate and precipitating as sulphide. Like potassium bromide (Kohn A. 1909 ii 891) ammonium bromide solutions dissolve cuprous iodide the solubility increasing rapidly with the concentration. Sensitiveness of the Bead and Lead Dioxide Tests f o r Manganese with Special Reference to the Interference of Iron. LOUIS J. CURTMAN and A. D. ST. JOHN ( J .Amer. Chern. Soc. 1912 34 1675-1679).-A study has been made of the sensitiveness of the bead and lead dioxide tests for manganese and the influence of iron on these tests. The results show that the lead dioxide test is not trustworthy in presence of 300 or 400 parts of iron to 1 part of man- ganese unless as much as 0.2 mg. of manganese is present but that the bead test is efficient with larger quantities of iron provided that as much as 0.005 mg. of manganese is introduced into the bead and certain specified precautions are observed. F. KONIG (Cliern. Zentr. 1912 ii 1064 ; from Apoth. Zeit. 2’7 536-’537).-Klut’s nitric acid process (ibid. 1909 ii 1076) for the colorimetric estimation of iron in waters gives good results with samples absorbing not more than 80 mg.of potassium permaaganate per litre. When this amount is exceeded the iron should be estimated colorimetrically in the residue after Reagent for D e t e c t i n g Small Quantities of Nickel. V. FORTINI (Chem. Zed 1912 36 1461).-The following solution is recommended for the detection of nickel in plated wares and nickel alloys. 0.5 Gram of dimethylglyoxime is dissolved in 5 C.C. of 98% alcohol and 5 C.C. of ammonia are added. After rinsing the surface with ether a drop of the reagent is applied when a rose-coloured spot mill form should nickel be present. Previous heating in the oxidation flame mill render the test still more delicate. The presence of other metals usually present does not interfere. The spot formed may be Test for Methane. I. OTTO HAUSER and H.HERZFELD (Ber. 19 12 45 351 5 -35 16).-Methane and ozonised oxygen react with each other with the formation of formaldehyde which reaction may be used as a very delicate test for methane. The gas t o be tested is mixed with ozonised oxygen and the mixture passed through a tube containing moistened glass wool which is afterwards extracted with water and the solution tested for formaldehyde with morphine and sulphuric acid C. H. D. E. G. Estimation of Iron in Water. ignition. L. DE K. easily removed with a towel. L. DE K.ii. 78 ABSTRACTS OF CHEMICAL PAPERS. Methane is the only hydrocarbon which gives formaldehyde. Ethane gives acetaldehyde or acetic acid but the reaction is much slower than with methane whilst acetylene is oxidised explosively to carbon dioxide.The reaction with acetylene takes place also i n acetone solution and is accompanied by a beautiful green luminescence. Toluene is oxidised by ozone with formation of formic acid. T. S. P. Analysis of Hydrocarbon Mixtures. OSCAR ROUTALA (Chem. Zelztr. 1912 ii 638; from Ann. AcacZ. Sci. FennicaB A. 2 No. 13 19 pp.).-The iodine absorption process for the testing of petroleums is too much affected by the personal equation and reports founded on it should be accompanied by a careful description of details of analysis including time of action temperature exposure to light and volume of iodine solution added. Frank's bromine process (Chem. I d 1901 24 263) may be successfully applied when small quantities of the substance are t o be tested also when the refining degree of colourless or pale petroleum products has to be determined or when the degree of purity of olefines has to be ascertained.From the bromine number the percentage of the olefine is calculated directly by the formula A = [K+ (.n - 5)c].G in which A = % of the olefine K= a constant n the mean number of carbon atoms for the olefine boiling at between the limits of temperature (at least 5) c = a constant ( =Om8750) and G the bromine number. For amylene K = 0,4375 ; for hexylene 0,5950 ; for heptylene 0.6075 ; for each additional CH,-group the constant increases on the average by 0*@€375 To show the accuracy of the formula a bromine number was taken of a mixture of amylene and isopentane and the result agreed with the sulphuric acid process and the density. L. DE K.Methoxyl Estimation with Hydriodic Acid and Phenol. FRITZ WEISHUT (Monatsh. 1912 33 1165-11~2).-1t is found that many substances which fail to give satisfactory results for the estima- tion of the methoxyl group even in the presence of acetic anhydride can be made to do so by using a mixture of hydriodic acid with approximately one-quarter its volume of phenol. I n most cases the reduction is complete in about one and a-half hours but methyl nitroanisate required five and a-half hours. D. F. T. Estimation of Higher Alcohols in Wine. THEODORE VON FELLENBERG (Chem. Zentr. 1912 ii 1851 ; from Mitt. Lebensmitt. Hyg. 3 231-235).-The colour reaction which higher alcohols give with salicylaldehyde and concentrated sulphuric acid (A. 191 1 ii 667) has been applied to wine. Natural wines are found to contain about three parts per 1000 of higher alcohols which suggests that the accepted value for cognac (1 per 1000) is very low. THEODORE VON FET'LENBERG ( C h m .Zentr. 1912 ii 1850-1851; from Mitt. Lebensmitt. Hyg. 3 228-231).-A colour reaction which phloroglucinol gives with vanillin and hydrochloric acid is also obtained with wine which J. C. W. A Colour Reaction of Wine.ANALYTICAL CHEMISTRY. ii. 79 probably contains therefore an unsaturated alcohol. may be matched against a standard solution of acid magenta. The red colour J. C. W. Examination of Lead Rrecipitates in Wine; Pentose and Methylpentose Estimations in Grapes and Wine. THEODORE VON FECLENBERG (Chem. Zentr. 1912 ii 1850 ; from Mitt. Lebensmitt. Hyg. 3 213-217).-After precipitation with just the required amount of lead acetate (found by a series of preliminary tests) there are still some substances in natural wines although scarcely at all in sophisticated wines which may be precipitated by basic lead nitrate.The precipitate contains succinic acid inositol pectin pentosan and methylpentosan. The pectin of grapes contains pentose and methyl- pentose but does not account for the whole of these sugars which are present both in the free state unprecipitable by basic lead nitrate as well as in the form of pentosans. J. C. W. A Simple Method of Characterising Acraldehyd 8. WILLIAM OECHSNER DE CONINCK (BUZZ. h a d . roy. Belg. 1912 534).-The acraldehyde vapour is condensed in distilled water through which a slow stream of oxygen is then passed until the solution is strongly acid.On dissolving freshly precipitated silver oxide in this liquid and concentrating the clear solution silver acrylate crystallises out in the pure state. W. G. A Bromometric Estimation of Formic Acid. H. MADEH (Chem. Zentr. 1912 ii 1401 ; from Apoth. Zeit. 1912 27 746-747). -A method for the estimation of formic acid is given; it depends on the oxidation of the acid by bromine. The solution is left for twelve to fifteen hours with bromide and bromate of potassium and phosphoric acid in a stoppered bottle when the unused bromate is decomposed by potassium iodide and estimated with thiosulphate. J. C. W. The Detection of Acetoacetic Acid by Sodium Nitroprusside and Ammonia. VICTOR JOHN HARDING and ROBERT FULFORD RUTTAN (Bio-Chem.J. 1912 6 445-45O).-Le Nobel’s test for acetone in the urine in which the urine is treated with acetic acid sodium nitro- prusside and ammonia mas found to give a less distinct reaction with urine to which acetone had been added than in natural acetonuric urines ; furthermore in the latter cases the reaction was less distinct in distillates containing relatively large amounts of acetone than in the original urine. The authors show now that the Le Nobel test is really a reaction for acetoacetic acid and not for acetone and as such it is capable of detecting 1 part of this acid in 30,000 whereas the ordinary ferric chloride test fails when the dilution is only 1 in 7000. s. B. s. Detection of Benzoic Compounds. OTTO SCHNATOLLA (Pharna. Zeit.1912 57 947).-Twenty C.C. of a solution t o be tested for the presence of benzoic acid are mixed with 5 C.C. of hydrogen peroxideii. 80 ABSTRACTS OF CHEMICAL PAPERS. and a solution prepared by dissolving 5 grams each of ferrous sulphate and boric acid in 100 C.C. of water is added until the mixture no longer becomes darker in colour. Should benzoic acid be present a dark bluish-green coloration develops with a few seconds. The test will detect the presence of 1 part of benzoic acid in 15,000 parts of solution. A similar reaction is given by all benzoic compounds hippuric acid etc. Benzoic acid however may be separated from hippuric acid in mixtures of the same by extracting the former acid with light petroleum. By means of suitable solvents benzoic acid may be separated from salicylic acid and then identified by the above reaction.w. P. s. Nature of the Reactions of Naphthenic Acida. I(. W CHARITSCHKOV (Chem. Zeit. 1912 36 1378. Compare A. 1910 ii 549 ; 1911 ii 543)-The reactions OF naphthenic acid described previously with copper sulphate ferrous salts cobalt nitrate and hydrogen peroxide are characteristic of cycEopentanemonocarboxylic acids and are not shown by acids which contain the cyclohexane ring such as cyclohexylacetic acid. H. N7. A New Method f o r the Colorimetric Estimation of Uric Acid in Urine. OTTO FOLIN and A. B. MACALLUM jun. (J. Bid. Chenz. 1912 13 363-370).-The method is based on the colour produced by the action of phosphotungstic acid. W. D. H. Detection and Estimation of Arachie Oil. NORMAN EVERS (Analyst 1912 37 487-492).-After trying various methods already proposed the author gives the following process 5 grams of the sample (say olive oil) are saponified with 25 C.C.of alcoholic potassium hydroxide (80 grams in 80 C.C. of water and diluted to a litre with 90% alcohol) in a reflux apparatus. To the hot soap solution are added 7.5 C.C. of dilute glacial acetic acid (1:2) and 100 C.C. of 70% alcohol containing 1 vol.% of hydrochloric acid and the liquid is then cooled to 12-14' for a n hour. The deposit is collected and washed with the acid alcohol (temp. 17-19") until the filtrate gives no turbidity with water the washings being measured. The precipi- tate according t o its bulk is dissolvod in 25 to 70 C.C. of hot 90% alcohol and cooled tc 15-20'.If abundant crystals appear then after one to three hours they are collected and washed first with 90% alcohol (about half the volume used for crystallisation) and finally with 50 C.C. of 70% alcohol. The crystals are now dissolved in hot ether and this is then distilled off in a weighed flask; the residue is dried at 100' and weighed. Should the m. p. be lower than 71° the acid should be recrystallised from 80% alcohol. A correction for solubility should be made for which a table is given. If no or but few crystals are obtained from the 90% alcohol a sufficient quantity of water is added to reduce this to 70% strength. Any deposit then formed within an hour at 17-19' is collected washed with 70% alcohol and weighed as before applying the correction for solubility.It must be re- crystallised if the m. p. is below 71'. The following 4.54 Grams of crystals= 100 grams of arachis oil.ANALYTICAL CHEMISTRY. ii. 81 oils gave no crystals olive oils including three superfine Maloga and eight of unknown origin almond poppy and rape oils. 5 L. DE K. E s t i m a t i o n of Cyanogen Compounds in Gas-Liquors. FRANZ WEISSEB (Chem. Zeit. 1912 36 1285-1287).-Estimatiorr 01 ammonium cyanide.-This is found by difference from the total cyanogen and that present as ferrocyanide. Totul cyanogen.-500 C.C. of the liquor are mixed with 50 C.C. aqueous potassium hydroxide (1 1) and 50 C.C. 10% ferrous sulphate and heated for one to two hours on the water-bath. After cooling and diluting to a litre 100 C.C. of the filtrate are acidified with hydrochloric acid and 5 C.C.of 10% ferric chloride are added The Prussian-blue formed is collected and then again decomposed with dilute aqueous potassium hydroxide. The filtrate which contains all the cyanogen as ferrocyanide is evaporated with excess of sulphuric acid and the residue is gently ignited. The residual iron oxide is then estimated as usual by titration with pernianganate. One C.C. of N/lOO-permanganate = 0.00156 gram of cyan og en. Ammonium ferrocynnide.-l00-500 C.C. of the liquor are evaporated and the residue is gently ignited. The iron is then titrated as usual; 1 C.C. of N/100 permanganate = 0.00334 gram of the ferrocyanide. Ammonium thiocyanak-The author has slightly modified Pfeiffer’s colorimetric process. Fifty C.C. of the sample are diluted with 50 C.C.of water and evaporated on the water-bath to 30 C.C. so as to expel the free ammonia. After diluting to 100 c.c. 20 C.C. of the solution ( = l o C.C. of sample) are mixed with 10 C.C. of ferric chloride solution (5 parts of the salt made up to 100 parts with 10% hydro- chloric acid) and diluted to 500 C.C. An aliquot part of the liquid is then matched as usual with a solution of ammonium thiocyanate of known strength. L. DE K. Estimation of Scatole and Indole in Faeces. CURT MOEWES (Chem. Zentr. 1912 ii 1702 ; from Zeitsch. expt. Path. Ther. 11 555). -The scatole and indole obtained by extracting the steam distillate with light petroleum are precipitated by dirnethylaminobenzaldehyde and the dye is dissolved in water and compared in a Plesch chromo- photometer with a test solution of 1 100,000 made from equal parts of indole and scatole.Values of 0*008-0-024 gram per 100 grams of fsecal matter were obtained J. C . W. Action of Activated Aluminium on Alkaloidal Extracts Its Use in Toxicology. XMILE KOHN-ABREST (Comnpt. rend. 19 12’ 155 1179-1181. Compare A. 1912 ii 645-768).-A study of the behaviour of the more common vegetable alkaloids pure or mixed with viscera in the presence of activated aluminium. The hydrated oxide formed from the activated aluminium only fixes a certain small proportion of alkaloids such as strychnine puinine and cocaine in alcoholic solution acidified with tartaric acid whilst on the other hand it retains almost all the nicotine. Data are given for some fourteen alkaloids.Alcoholic extracts of viscera containing morphineii. 52 ABSTRACTS OF CHEMICAL PAPERS. and strychnine after treatment with activated aluminium are obtained in a state of great purity wherein the alkaloids can be readily detected and estimated. W. G. Distinction between Cocaine and its Substitutes. D. SCHERBATSCHEV (Chem. Zentr. 1912 [ii] 386 ; from Apoth. Zait. i912 27 441. Compare Hankin A. 1911 ii 162).-The appearance under the microscope of solutions of /3-eucaine nirvanine alypine holocaine and novocaine when solutions of ammonia (lo%) potassium hydroxide (lo%) or sodium hydrogen carbonate (saturated) are added to them is described T. A. H. Gravimetric E s t i m a t i o n of Quinine as Nitroprusside. P. J. KRUYSSE (Pharm. Weekblad 1912 49 11 17-1120).-Quinine nitro- prusside is obtained as a crystalline precipitate when a neutral solution of a quinine aalt is treated with sodium nitroprusside; its solubility is less than that of the tartrate or oxalate 100,000 parts of water dissolving 39 parts of the salt.For the estimation of quinine in powdered cinchona bark 5 grams of the latter are mixed with 3.5 C.C. of water 1 C.C. of ammonia and 2.5 grams of slaked lime and the mixture is extracted in a Soxhlet apparatus for two hours with acetone. After the acetone has been evaporated from the extract the residue is dissolved in 25 C.C. of 2% hydrochloric acid the solution is filtered rendered alkaline and shaken with ether. The ethereal solution is then shaken with dilute hydrochloric acid the separated acid solution is neutralised diluted to 50 c.c.heated t o 90° and treated with 0.5 gram of ammonium oxalate. After cooling the pre- cipitate is collected on a filter washed with a smallquantity of water and then dissolved in 5 C.C. of 2% hydrochloric acid. This solution is nearly neutralised and the excess of oxalate is removed as calcium oxalate. The solution is now neutralised with ammonia diluted to 70 c.c. heated to boiling and treated with 0.5 gram of sodium nitro- prusside. After cooling the precipitate is collected washed with water dried at looo and weighed. The treatment with ammonium oxalate is for the purpose of separating cinchonine from the quinine. w. P. 8. Estimation of Quinine Sulphate in Cinchona Bark. P. J. KRUYSSE (Pharm. Weekblad 1912 49 1135-1136).-The author (preceding abstract) regrets he was not acquainted with the work of Kramers (A.1897 ii 83) who however did not succeed in applying the nitroprusside reaction quantitatively. L. DE K. Hyposulphite in Volumetric Analysis. I. WILHELM SIEGMUND (Monatsh. 1912 33 1431-1445).-The hyposulphites give stable compounds with the ketones which in contradistinction to the form- aldehyde-hyposulphites act as reducing agents in the cold. Solutions of these compounds may therefore be used as standard reducing agents in volumetric analysis. The author finds that a convenient solution has the following composition 10 grams of sodium hyposulphite 15 C.C. of acetone and 35 C.C. of ammonia (D = 0.91) to one litre ofANALYTICAL CHEMISTRY. ii. 89 water. The solution is kept under a layer of paraffin oil in a stock bottle from which a burette can be filled automatically; the top of the liquid in the burette is covered with toluene to prevent contact with the air.The tip of the burette is connected with the titration flask by means of a piece of glass tubing which is bent twice a t right angles and passes through the rubber stopper of the flask. The stopper also contains inlet and exit tubes for a current of oxygen-free carbon dioxide and a thermometer. The solution of the hyposulphite is standardised by means of iron alum or an oxidised solution of ferrous ammonium sulphate using ammonium thiocyanate as indicator and in the presence of sulphuric acid. The titration is carried out in the cold with vigorous shaking; the sulphite formed has no action on the ferric salt.The solution is fairly stable but should be standardised daily. &Experiments mitb a large number of azo-dyes dyes which form leuco-compounds and nitro-colouring matters show that they are reduced quantitatively by the above hyposulphite solution. T. s. P. Methods of Estimating Saponins. (Mlle.) MARIE KORSAKOV (Compt. rend. 1912 155 844-846).-A critical study of the methods at present employed for estimating saponins. Christophson's method based on the precipitation of the saponin as a barium compound is inaccurate since the whole of the saponin does not separate and further the barium hydroxide precipitates from plant extracts other substances than saponins which on hydrolysis yield dextrose. A modification of Kruskul's method gives good results.The plant is dried finely powdered and extracted with boiling alcohol (60%). The extract is filtered the alcohol distilled off and the residue evaporated on the water-bath with calcined magnesia. The resulting paste is extracted with boiling alcohol (80%) the extract filtered and precipitated with ether. The precipitate is dissolved in 3% sulphuric acid and hydrolysed in an autoclave a t 105' for one hour the sapogenin liberated being washed with water dissolved in absolute alcohol and weighed after distilling off the alcohol. W. G. A Delicate Test for the Detection of Albumin in Urine. ADOLF JOLLES (Zeitsch. physiol. Chm. 1912 81 205-206. Compare A 1896 ii 344).-The reagent is composed of mercuric chloride 10 parts citric acid 20 parts sodium chlorate 20 parts water 500 parts.Five C.C. of the filtered urine are added to each of three test-tubes. To (1) and (2) 1 C.C. of 30% acetic acid is added 5 C.C. of the reagent are put in (l) and the other tubes are filled with water t o the same level. The tubes are observed against a black background (3) being placed in the middle. The reaction is sensitive to 1 part of albumin in 120,000 E. F. A. Chemistry and E s t i m a t i o n of Gelatin. M. BERRAR (Biochent. Zeitsch. 1912 47 189-214).-Gelatin forms saturated solutions in water in the following strengths 0.62% a t 21° 0.60% at lS" and 0.56% a t 17". If the concentrations only slightly exceed these pro-ii. 84 ABSTRACTS OF CHEMICAL PAPERS portions a gel is formed. Gelatin cannot be precipitated quantitatively by alcohol or by metaphosphoric acid as it completely dissolves in excess of the latter reagent the proteins being only partly soluble in excess.A t room temperature it is not precipitated quantitatively by picric acid but at 8' precipitation is complete when a gelatin solution is mixed with a saturated aqueoussolution of the acid. The acid com- bining power of gelatin was estimated by precipitation with potassium- mercuric iodide and picric acid solutions. Hydrocbloric acid could not be used for this purpose as the precipitate dissolves in excess of the acid. By this means i t W A S found that 1 gram of gelatin combines with the equivalent of 11.3 C.P. N/lO-acid which corresponds with a molecular weight of 823. The picric acid precipitate of gelatin dissolves in alcohol and in 2% urea solution containing sodium chloride. If a mixture of 1 part of saturated aqueous solution of picric acid and 4 parts of alcohol are added to a gelatin solution the gelatin is not precipitated whereas other proteins are.The gelatin is however pre- cipitated quantitatively from this gelatin-picric acid solution in alcohol by the addition of excess of aqueous picric acid solution in the cold. The precipitate again dissolves on warming to 40° or treating again with absolute alcohol. On these reactions a method is based for estimating quantitatively gelatin in the presence of other proteins. It was incidetltally observed that the nitrogen in picric acid can be quantitat'ively estimated by Kjeldahl's process if it is first reduced by iron and glacial acetic acid. S. B. S. The Rideal- WGlker Phenol Control. A Possible Discordant Factor in the Standardisation of Disinfectants. J. T. AINSLIE WALKER and JOHN 31. WEISS (J. Frank& Inst. 1912 101-112).- The accnracy of the Itideal-Walker process depends on the purity of the phenol which is employed as the standard. Ordinary crystallised phenol is usually contaminated by cresols to such a n extent as t o render it unsuitable for the purposes of bactericidal control ; four specimens of phenol examined by the authors were found to contain from 1 * 1 to lams% of cresols. As cresol has approximately three times t h e bactericidal e6ciency of phenol the error from this cause may be very considerable. The purity of the phenol should be ascertained by a determination of its solidifying point and phenol having a lower solidifying point than 40' should be rejecked. Synthetic phenol prepared by the authors showed a solidifying point of 40.5'. The bromine titration method does not yield trustworthy results as regards the purity of the phenol but i t may be employed for checking the strength of the 5% stock solution prepared from pure phenol. w. P. s.