ANALYTICAL CHEMISTRY. 5 1 Analytical Chemistry. Measurement of Gases. By OTTO BLEIER (Be?*., 1898, 31, 2513).-The author finds the method recently described by him (Abstr., 1898, i, 183) for the absolute measurement of gases has been previously used by Hempel (Verhdlg. Pereins. Befard. Gewery., 1894, 460). A. H. Crystals in Blowpipe Beads. By WILHELM FLORENCE (Jahrb. f. Min., 1898, ii, 102--146),-A systematic account is given of the results obtained on heating various substances in borax and sodium phosphate beads with or without the presence of lead oxide ; numerous figures are given of the crystals characteristic for each element so treated. The method is recommended for the microchemical deter- mination of minerals in the dry way. Use of Ferrous Salts in Alkaline Solution as a Reducing Agent in Volumetric Analysis.By ANDRI~ JOB (Compt. rend., 1898, 127, 59-60).-A solution made by adding an excess of sodium pyrophosphate to a solution of ferrous ammonium sulphate, is recom- mended as an alkaline reducing agent in volumetric analysis. When titrated with iodine or potassium permanganate, an operation which should be carried out in a current of hydrogen, the liquid remains colourless up to the end point, which is extremely sharp. A solution of sodium pyrophosphate readily dissolves ferrous sulphide, and the green solution thus obtained is well adapted for the colorimetric estimation of small quantities of iron, The reaction is stated to be as delicate as that with potassium thiocyanate. The solu- tion to be tested for traces of iron is mixed with a few drops of sodium pyrophosphate solution and one bubble of hydrogen sulphide passed in.N. L. Estimation of Moisture in Invert-sugar. By LEONARD T. THORNE and E. H. JEFFERS (J. Soc. Cliern. Ind., 1898, 17, 114--116).-1n consequence of the difficulty of drying invert-sugar in the ordinary way, it is usual to take the specific gravity (at 60' F.) of a 10 per cent. solution of the sample, and divide the excess of gravity over that of water (taken as 1000) by 0.386. This gives the percentage of dry sugar in the solution and the moisture in the sample is obtained by L. J. 8. 4-252 ABSTRACTS OF CHEMICAL PAPERS. difference. When a notable amount of mineral matter is present this is estimated, the percentage is multiplied by 0.8, and deducted from the original excess gravity before calculating out the dry sugar (Heron).To obtain a direct method of estimating the moisture, the authors have devised an apparatus for drying the sugar a t 65-70' in a slow current of carbonic anhydride at a pressure of 50-80 mm. of mercury. For this purpose, the sugar is weighed on to the end of a coil of paper contained in a test tube (previously dried until constant in weight) and is then distributed over the paper by the addition of a few C.C. of water, with certain precautions. The test tube is then placed in a wider tube which is heated in the vapour of boiling methylated spirit, and whilst kept exhausted as completely as possible by a powerful and constantly working exhaust pump, a very slow current of carbonic anhydride is allowed to pass through the apparatus.The results agree very closely with those deduced from the specific gravity of the solutions provided that the ash is weighed after sulphating and GT deducted before multiplying by 0.8 ; also for invert-sugars, the factor 0.3888 given by Brown, Morris and Millar (Trans., 1857, SO), is employed instead of 0.386, which is the correct factor for saccharose. M. J. S. Use of a Solution of Phloroglucinol-vanillin for the Detec- tion of Halogens in Organic Compounds. By P. N. RAIKOW (Chem. Zeit., 1898, 22, 20--31).-An alcoholic solution of phloro- glucinol-vanillin has been successfully used for the detection of free hydrochloric acid in gastric juice; the author now states that it may be succesefully applied to test for halogens in the products of corn- bustion of organic substances.A porcelain dish is rinsed with some of the reagent and then dried; a little of the substance to be tested is burnt on a platinum wire in a spirit lamp, or, if it is a gas, it is either burnt from a jet, or a current of it is passed through the spirit flame. If the dish coated with the reagent is inverted over the flame, the least trace of halogen acid formed will be indicated By MAURICE LUCAS (J. Pharm., 1898, [vi], 8, 311-312).-By varying the details of t h e silver nitrate method of estimating cuprous oxide in copper, the autbor has determined the conditions which are necessary to give concordant results agreeing with those obtained on reduction by hydrogen. It appears, also, that when cuprous oxide alone is treated, the amount of copper in the residue corresponds with that required by Hampe's equation, and that the action of silver nitrate on mixed cuprous oxide and copper gives the same result as its action on the separate components of the mixture.On immersing a sheet of copper in a neutral solution of copper nitrate, no precipitate containing copper is obtained on filtering. Estimation of Sulphur in Coal. By GEORGE L. HEATH (J. Amr. Chenz. ~Soc., 1898, 20, 630-637).-The author iejects, for technical The safest way is to operate as follows. by the intensely red permanent coloration. L. DE K. Estimation of Oxygen in Copper. W. A. D.h~ALJ?TICA~, CHEMISTRY. 53 purposes, those processes which are based on fusing with an oxidising mixture, as this involves an evaporation in order t o render any silica insoluble.The best method is the one proposed by Eschka, provided that bromine is used as the oxidising agent; the details are as follows. One gram of the powdered coal is intimately mixed with 1 gram of pure light magnesia and half a gram of dry sodium carbonate, and heated in a 100 C.C. platinum dish with the flame of a large spirit lamp, at first very slowly ; the mixture is frequently stirred with a platinum wire, and the whole is finally heated for 15 minutes at a low, red heat. When cold, the contents are transferred t o a beaker, and boiled for 5 minutes with 100 C.C. of water and 15 C.C. of saturated bromine water; the filtrate is finally mixed with a slight excess of hydrochloric acid and then precipitated with barium chloride, with Estimation of Traces of Hydrogen Sulphide in the Air.By KARL B. LEHramN (C’hem. Centr., 1898, i, 139; from Arch. Hyg., 30, 262).-First method-Eight litres of air are drawn through 10 C.C. of N/100 iodine solution, at the rate of 6 litres in 30 minutes, by means of an aspirator, any traces of iodine which may be carried off by the current being caught in a trap containing an N/100 solu- tion of sodium thiosulphate. From the loss in iodine, the amount of hydrogen sulphide is calculated. Second method.-The same volume of air is passed through a glass tube 30 cm. long and 12 mm. wide containing a strip of filter paper 5 cm. long and 2 cm. wide which has been saturated with lead nitrate solution. I f the paper turns pale yellowish-brown, the air contains about 1 *4-2 parts of hydrogen sulphide per million ; if dark brown, the amount will be from 3-5 parts, and if brownish-black, 8 or more Titration of Combined Sulphuric Acid.By MAX REUTER (Chem. Zeit., 1898, 22, 357).-The author has tried Andrew’ process (Amer. Chem. J., 1880, 2, 567), and finds that it gives excellent results if the titration with thiosulphate is made in a liquid containing only 5 C.C. of hydrochloric acid instead of 20; as this necessitates a longer contact with potassium iodide, the beaker should be covered with a watch-glass, and the air kept off by a surface current of carbonic anhydride. The process, as will be remembered, consists in precipitating the sulphnric acid by means of a solution of barium chromate in hydro- chloric acid.The excess of barium chromate is then removed by means of calcium carbonate, and the filtrate which contains an amount of soluble chromate equivalent to the sulphate is acidified with hydro- chloric acid, mixed with potassium iodide, and titrated with sodium Estimation of Nitrogen by the Kjeldahl-Gunning Method. By K. WEDEMEYER (Chem. Zeit., 1898,22,21).-The author has proved by a number of experiments that the addition of potassium sulphate, as recommended by Gunning, to the usual mixture of sulphuric acid, the usual precautions. L. DE K. parts may be present. L. DE I(. thiosulphate. L. DE K.54 ABSTRACTS OF CHEMICAL PAPERS. phosphoric anhydride, and metallic mercury, considerably shortens the time of boiling without in the least interfering with the accuracy of the method.NoTE.-Those who have read Gunning's original Dutch paper know that his chief object was to obviate the addition of potassium perman- ganate, salts of heavy metals, phosphoric anhydride, &c. Detection of Nitrates in Milk by means of Formaldehyde. By E. FRITZMANN (Chem. Centr., 1898, i, 218-219; from Zeit. O$E. Chem., 3, 610--614).--Nitrates, which, if present in milk, are a sure sign of the presence of adulteration with water, may be detected by adding to the milk a small quantity of formaldehyde, such as one drop of a 40 per cent. solution to 250 C.C. of the sample. On adding strong sulphuric acid, which should be chemically pure, a blue ring is obtained, and, on shaking, the liquid turns reddish- or bluish-violet. Too much formaldehyde interferes with the reaction ; if, therefore, the reaction still shows when more of it is added, the amount of nitrates is L.DE K. considerable. L. DE K. Estimation of Phosphoric Acid. By HENRI LASNE (Compt. rend., 1898, 12'7, 62--64).-According to Vignon (Abstr., 1898, ii, 639), the sum of the amounts of phosphoric acid precipitated as ammonium magnesium phosphate from the solutions obtained by treating superphosphate, (1) with water, (2) with ammonium citrate solution, is always considerably less than the amount precipitated from the mixture of the two solutions. The author, in his large experience of superphosphate analysis, has never observed such a discrepancy, and quotes the results of three experiments in support of his opinion of the exactitude of the precipitation of phosphoric acid in presence of ammonium citrate.It is also maintained that ammonium magnesium phosphate is absolutely insoluble in a liquid containing one-third of its volume of ammonia of 22O, and experi- ments are described which show that even as little as a milligram of phosphoric acid in a litre of liquid is aompletely precipitated as the double salt. N. L. Estimation of Phosphoric Acid, By LBO VIGNON (Compt. rend., 1898, 127, 191).-A reply to Lasne (preceding abstract). The author does not dispute the fact that the phosphoric acid soluble in citrate solution may be determined with accuracy directly, but maintains that his method of estimating it by difference offers the combined Estimation of Phosphoric Acid for Physiological Pur- poses.By ALBERT NEUMANN (Chem. Centr., 1898, i, 219-223); from Dzc Bois Reymond's Arch., 1897, 552-553).-The destruction of the organic matter is effected by means of strong sulphnric acid and ammonium nitrate, 1 gram for every C.C. of the acid, the operation being conducted in a Kjeldahl flask and the substance introduced in two or three separate portions. To destroy the organic matter in 25 C.C. of urine, 10 C.C. of acid is required ; for 25 C.C. of milk or 8 grams of fresh meat, 15 C.C. ; and for 5 grams of dry faxes, 20 c,c. of acid is needed. advantages of rapidity and accuracy. N. L.ANALYTICAL CHEMISTRY. 55 The phosphoric acid is then estimated graviine tricdly or, if possible, Method of Preparing a Strictly Neutral Ammonium Citrate Solution. By A. D.COOK (J. A n z e ~ . Chew,, S'oc., 1898, 20, 585-586).-Seven hundred and forty grams of citric acid is placed in a 4-litre measure containing 1900 C.C. of 10 per cent. ammonium hydroxide, the mixture is vigorously stirred until it is all dissolved. and the solution made up to the mark ; after being again well stirred, it is transferred to a large porcelain evaporating dish and allowed to remain overnight; in the morning i t will be found to be perfectly neutral both to corallin and cochineal. It is then Estimation of Citrate Soluble Phosphoric Acid in Bone Meals, Superphosphates, &c. By 0. BOTTCHER (Chew.. Zeit., 1898,22, 201-202).-Five grams of the sample is repeatedly ground up with a solution of ammonium citrate containing free citric acid, and the whole introduced into a Stohmann's half-litre fiask, which is then filled up to themark with the citrate solution a t 17.5".After closing the flask with an indiarubber cork, it is rotated for 30 minutes in a suitable apparatus which turns 30-40 times per minute round its axis. The liquid is then filtered and 50-100 C.C. is used for the estimation of the phosphoric acid either by the molybdic or the magnesia method. Preparation, qf the Citrate XoZutiorn.-One hundred and fifty grams of citric acid is put into a litre flask, dissolved in water, and exactly neutralised with ammonia; 10 grams of citric acid is then added, and the whole is made up to 1 litre. Before use, 1 volume of the By HUGO MASTBAUM (Zeit. anal. Chena., 1898, 37, 581-582).-With reference to the rapid method of igniting magnesium ammonium phosphate described by Schmoger (Abstr., 1898, ii, 455), the author states that even the preliminary drying of the precipitate is unnecessary, but that the wet precipitate, wrapped in its paper filter, may be placed in the platinum crucible and the latter at once exposed to the full flame of a Bunsen burner. The water assumes almost instantly the spheroidal condition, and no loss of precipitate takes place.After 15-20 minutes, the grey residue is moistened with strong nitric acid, which is then carefully evaporated. Exposure t o the blowpipe flame for half a minute completes the ignition. Volumetric Analysis of a Mixture of Alkyl Phosphates with Phosphoric Acid, By JACQUES CAVALIER (Compt. rend., 1898, 127, 60--61).-The composition of a mixture of phosphoric acid with tho two alkyl phosphates, RH,PO, and R2HP0,, where R is methyl, ethyl, or allyl, may be determined volumetrically by a process based on the behaviour of these compounds towards diff erent indicators and on the different solubilities of the barium salts.The dialkyl hydrogen phosphate is acid to methyl-orange and to phenolphthalein, and is rendered neutral to both indicators by combination with one equivalent of alkali. The monalkyl dihydrogen phosphate is neutralised by one by the volumetric uranium process. L. DE K. diluted with water to a sp. gr. = 1.09 at 20'. L. DE K. solution is diluted with 4 volumes of water. L. DE K. Ignition of Magnesium Ammonium Phosphate. M. J. S.56 ABSTRACTS OF CHEMICAL PAPERS. equivalent of alkali if methyl-orange is used as the indicator, and by two equivalents if phenolphthalein is employed ; the barium salt is soluble in water.Phosphoric acid resembles the preceding compound in its behaviour with indicators, but the barium salt, BaHPO,, is insoluble in water. The mixture of the three acids is first titrated with baryta water in the cold, using methyl-orange as indicator, and the liquid is then warmed to about 60° and the titration completed to the end point indicated by phenolphthalein. The first titration gives the sum of the three acids and the second the two polybasic acids, RH,PO, and H,PO,. The liquid is now filtered from the precipitated barium hydrogen phosphate, and both precipitate and filtrate are mixed with excess of standard hydro- chloric acid.The phosphoric acid and the monalkyl dihydrogen phos- phate respectively contained in the two solutions thus obtained are then titrated with baryta water, the volume of alkali required to pass from the end point indicated by methyl-orange to that indicated by phenolphthalein being determined. The sum of the two results should bo equal to that given by the second titration of the original mixture, N. L. By JOHN PATTINSON and HUGH S. PATTINSON (J. XOC. Cl~rn. Ind., 1898, 17, 211-214).- Arsenious sulphide requires for dissolution not more than 400,000 parts of hydrochloric acid of sp, gr. = 1 *17 saturated with hydrogen sulphide. I n the absence of hydrogen sulphide, the solubility is somewhat greater, although still very minute. Lead, cadmium, tin, antimony, and bismuth are not precipitated at all by hydrogen sulphide from acid of this strength, and the separation of arsenic from these metals is quantitatively complete.The precipitate may be washed with a mixture of 5 parts of hydrochloric acid and 2 parts of hydrogen sul- phide solution without loss, and without risk of precipitation of the other metals. If the arsenic be in the form of arsenic acid, it may be reduced by stannous chloride before passing the gas. The arsenious sulphide is best collected on an asbestos (Gooch) filter, After washing with acid, and then completely with water, it may either be weighed, or, if the quantity amounts to only a few centigrams, i t may be heated with 10-15 C.C. of strong sulphuric acid in an uncovered beaker at a temperature at which the acid gives off white fumes.No loss of arsenic is to be feared if chlorides are absent. Hydrogen sulphide is first given off, then sulphurous anhydride j the acid darkens at first but becomes colourless when a11 the sulphurous anhydride is expelled. A clear solution of arsenious acid is thus obtained. The solution is then nearly neutralised with sodium hydroxide, rendered alkaline with excess of sodium hydrogen carbonate, and titrated with iodine. A blank experiment with the same sulphuric acid is desirable, Very accurate results are obtained. Volumetric Estimation of Boric Acid. By VADAM (J. Phrm., 1898, [vi], 8, 109-1ll).-In presence of an excess of mannitol, the quantity of boric acid present in a solution can be determined by adding decinormal caustic soda until an alkaline reaction with phenol- Separation and Estimation of Arsenic, M.J. S.ANALYTICAT, CHEMISTRY. 57 phthalein or litmus is obtained ; the end point of the reaction is more sharply defined than in Barthe’s method, using glycerol (Abstr., 1896, ii, 337). In both cases, however, 1 C.C. of decinormal alkali is equivalent to 0.0062 gram of boric acid. The above method is avail- able for determining boric acid in butter ; 1.5 grams of the latter is extracted with warm water i(20 c.c.), the free fatty acids present neutralised with decinormal caustic soda in presence of litmus, 1-2 grams of mannitol added, and the boric acid determined as above described. W. A. D. Estimation of Potaasium. By JULIUS DIAMANT (Chem. Zed., 1898, 22, 99).-The author, finding a great difficulty in making potassium platinochloride perfectly anhydrous, now proposes the following process.The double salt, which should weigh about 0.5 gram, is dissolved in boiling water in a 500 C.C. flask, and when cold is diluted to the mark, and 1 gram of zinc dust is added; this rapidly decomposes the compound, forming zinc and potassium chlorides. Two hundred and fifty C.C. is then pipetted off and titrated with N/10 silver nitrate; from the chlorine thus found, the amount of potassium chloride may Test for the Presence of Sodium Hydrogen Carbonate in Sodium Carbonate. By MELCHIOR KIJBLI (Arch. Phurm., 1898,236, 321-324).-The test for the presence of sodium hydrogen carbonate in sodium carbonate, as given in the Pharmucop. German., III., and in the Phurmcop.Boss., IV., is not sufficiently delicate, as, according to the author, it mill not show the presence of 2 per cent. of the acid salt, and according to Thummel (Pharm. CentraLH., 1891, 86) as much as 4-5 per cent. may escape detection. The author therefore proposes the following test, which depends on the fact that an aqueous solu- tion of a soluble quinine salt of a certain concentration is not pre- cipitated by normal sodium carbonate, if the latter does not contain more than 2 per cent. of sodium hydrogen carbonate. Ten C.C. of a solution of quinine hydrochloride, containing 0.4 gram of the salt in 100 C.C. water, is added to 10 C.C. of the sodium car- bonate solution (3 grams in 50 C.C. water) ; if there is not more than 2 per cent.of sodium hydrogen carbonate present, no permanent cloudiness is produced. The quinine solution should be preserved from air and light, Estimation of Perchlorate in Chili Saltpetre. By OTTO FOERSTER (Chem. Zeit., 1898, a2, 357).-Ten grams of the sample is mixed with 10 grams of pure dry sodium carbonate and heated in a covered platinum or porcelain dish over the full flame until the fused mass flows quietly; this takes about 10 minutes. When cold, the mass, which can be readily detached from the dish, is dissolved in dilute nitric acid, and the total chlorine titrated by Volhard’s method ; afterallowing for chlorine present in other forms, the difference is Estimation of Lead in Ores. By FR. MOLDENHAUER (Chem. Zeit., 1898, 22, 256-257).-Two grams of the finely powdered ore is boiled be readily calculated.L. DE K. A. W. C. calculated to perchlorate. L. DE I(.5s ABSTRACTS OF CHEMICAL PAPERS. with strong hydrochloric acid until hydrogen sulphide is no longer evolved; if much antimony is present, this may be practically ex- pelled by evaporating to dryness two or three times with strong hydrochloric acid; the addition of a little nitric acid is advisable. The dry residue is now moistened with hydrochloric acid, 100 C.C. of water is added, the liquid boiled to dissolve the lead chloride, and the insoluble matter repeatedly washed with boiling water until quite free from lead ; 300 C.C. generally suffices, The filtrate, heated to redissolve any deposited lead chloride, is nearly neutralised with ammonia, preci- pitated with hydrogen sulphide, and the lead sulphide collected, washed, and digested with ammonium sulphide. It is then treated, together with the paper, with nitric acid, evaporated to dryness, and after being heated to expel any sulphur and organic matter, is con- verted into sulphate by means of sulphuric acid; finally, the sulphate is freed from other metals by washing, dried, and weighed.L. DE I(. Estimation of Traces of Lead in Water. By A. LIEBRICH (Chem. Zeit., 1898, 22, 325)-One, or more, litres of water is con- centrated, acidified with acetic acid, precipitated with hydrogen sul- phide, and the precipitate ignited and converted into sulphate by heating with a drop of sulphuric and nitric acid. The lead sulphate is then dissolved in a few C.C. of 10 per cent.aqueous potash, diluted to 20 c.c., mixed with 2 C.C. of freshly prepared ammonium sulphide, and the tint compared with that given by a standard solution of lead sulphate in aqueous potash containing 1 milligram of lead per c.c., and also diluted to 20 C.C. The experiment is then repeated, using more, or less, of the standard lead. By way of control, the lead sulphide New Volumetric Method for the Estimation of Copper. By RICHARD K. MEADE ( J . Amer. Chm. Xoc,, 1898,20, 610-613)- The solution of the copper ore is evaporated with sulphuric acid, the residue is dissolved in water, the greater part of the free acid neutra- lised with ammonia, and sulphurous acid added in excess ; the copper is then precipitated by adding a slight excess of ammonium thio- cyanate.The precipitate, collected on asbestos, is well washed and then put back into the beaker, where it is heated with aqueous (3 per cent.) potash; the suboxide thus produced is again collected on an asbestos filter, well washed with hot water, and then heated in a beaker with solution of pure ferric sulphate until it is entirely dissolved. The cuprous oxide sets f ree a corresponding equivalent of ferrous oxide, which is then titrated with permanganate, the iron equivalent of the perman- ganate multiplied by 1.125 giving the weight of the copper in the sample. Instead of sulphurous acid, a mixture of equal weights of sodium hydrogen sulphite and potassium thiocyanate may be used to precipi- tate the copper. Ferric chloride may be used instead of ferric sul- phate, but the titration with permanganate is then less safe.The process is not interfered with by the presence of arsenic, antimony, bismuth, or zinc. may be again converted into sulphate and weighed. L. DE K. L. DE K.ANALYTICAL CHEMISTRY. 59 Estimation of Copper in Vegetables. By VICTOR VEUROUI (Chert&. Zeit., 1838, 22, 103--IU4).-This is a controversy with Lehmann on the subject of the occurrence of copper in vegetable substances. The author (Abstr., 1897, ii, 602) has found in the same samples a very much larger amount of copper than Lehmann did, and attributes this to the inaccuracy of the latter’s colorimetric process. The author also expresses his view that the presence of even 0.8 gram of copper per kilogram of wheat does not constitute a danger to health, as the metal is there in the form of a harmless compound.L. DE I(. Estimation of Copper in Vegetable Substances. By KARL B. LEHMANN (Chem. Zeit., 1898, 22, 296--297).-A reply to Vedrodi (preceding abstract). The author cannot understand the great dif- ferences in the amount of copper found by Yedradi and himself in the ash of the same cereals, but he upholds his own views. L. DE I(. Estimation of Cuprous Oxide by Permangante. By ROBERT M. CAVEN and ALFRED HILL (J. Xoc. Chem. Ind., 1897,16, 981-983 ; 17, 124).-Cuprous oxide dissolves completely and rapidly in a mixture of potassium permanganate and sulphuric acid, with formation of cupric sulphate and corresponding reduction of the permanganate. One part of sulphuric acid is diluted with 3 parts of water, and 4 volumes of this acid is added to 1 vol.of standard permanganate. The cuprous oxide, collected and washed on an asbestos filter, is introduced into a measured excess of this mixture, and after stirring for a minute or two the unreduced permanganate is titrated with oxalic acid. Highly satis- factory results are obtained with great rapidity. M. J. S. Quantitative Separation of Metals by Hydroxylamine and Hydraeine. By PAUL JANNASCH (Ber., 1898, 31, 2377-2394).- [With a. D~v~~.]-Mercury may be separated from copper, bismuth, lead, cadmium, arsenic, antimony, or tin, in solutions of their salts, by the employment of an ammoniacal solution of hydroxylamine. A mixture of mercuric chloride and a salt of one of these metals is dissolved in water, or in hydrochloric acid if necessary, tartaric acid, ammonia, and hydroxylamine hydrochloride are added in excess, and the solution heated until the precipitation of mercury is complete; the latter is then collected, redissolved in fuming nitric acid, the solution evaporated to dryness, the residue taken up with hydrochloric acid and water, and the mercury precipitated as sulphide.From the ammoniacal filtrate, the remaining metal may be separated by suitable means, the copper as cuprous thiocyanate, the bismuth, lead, cadmium, antimony, and tin as sulphides, and the arsenic as magnesium ammonium arsenate. The copper, cadmium, bismuth, and tin are finally weighed as oxides, the lead as sulphate or peroxide, and the antimony as sulphide or oxide. [With F. ALFFERs.1-In a similar manner, mercury is separated from molybdenum or tungsten by ammoniacal hydroxylamine in the presence of tartaric or citric acids ; the precipitated metal is treated as described above, whilst from the filtrate the molybdenum is obtained as sulphide and weighed as trioxide.GO ABRTRAC'L'S OF CHEMICAT, PAPERS. r i 1 he separation of mercury from aluminium, chromium, or manganese is effected by the same reducing agent in the presence of oxalic acid ; aluminium and chromium are obtained as oxides, whilst the manganese is precipitated from the ammoniacal filtrate by hydrogen peroxide and weighed as the protosesquioxide. Tartaric or citric acid may be substituted for oxalic acid in the case of these metals, but its use is less convenient.Mercury and cobalt are separated by the addition of ammoniacal hydroxylamine to the slightly acid solution of their salts, the cobalt being precipitated as sulphide in the filtrate from the mercury.Mercury and nickel are similarly separated in the presence of an organic acid such as tartaric or citric. Either of these acids or oxalic acid may be employed in the separation of mercury from uranium by ammoniacal hrdroxylamine, the latter element being determined by evaporating the filtrate to dryness, igniting, and weighing both as U,08 and UO,. [With WILHELM H~InlANN.]-selenium, either in the form of selenious or selenic acid, is separated from sulphuric and phosphoric acids by boiling their solutions with excess of hydroxylamine hydrochloride, the selenium being precipitated and weighed, whilst the sulphuric and phosphoric acids in the filtrate are determined in the usual way. The separation of tellurous or telluric acids from sulphuric or phosphoric acids is effected in the same way.[With M. MULLER.]-A~ ammoniacal solution of hydroxylamine reduces tellurous and telluric acids, causing the complete precipitation of the tellurium, whilst an acid solution of this reagent is without action on these acids ; hydroxylamine may therefore be employed in separat- ing selenium and tellurium. The selenium is completely precipitated when a hydrochloric acid solution of these elements is boiled with hydroxylamine, and the tellurium separates quantitatively when the filtrate is rendered ammoniacal and boiled with more of the reducing agent ; four molecules of hydroxylamine are required for the reduction of one molecule of selenious acid, water and nitrogen being produced at the same time.Tellurium is separated from palladium by distilling an alloy of these elements in a current of bromine, the tellurium bromide is volatilised and received in hydrochloric acid, and the solution reduced with ammoniacal hydroxylamine ; the residue containing the palladium is dissolved in aqua regia, the metal precipitated as sulphide, and finally weighed as PdO, and also in the metallic state. Selenium is most conveniently separated from barium by means of hydroxylamine. Barium selenite dissolved in hydrochloric acid is boiled with excess of the reducing agent, and the selenium is collected and weighed; the barium remains in the filtrate and is estimated as sulphate. [With M.MU~~~~.]-Hydrazine sulphate reduces selenium compounds even more energetically than hydroxylamine and precipitates selenium both from hydrochloric and nitric acid solutions. Tellurium is also precipitated from its hydrochloric acid solution by this reagent but not from the solution in nitric acid ; from a hot ammoniacal solution, hjdrazine precipitates it quantitatively, and theANALYTiCAL CHEMISTRY. 61 reduction of this element from its salts by hydrazine and by hydroxyl. amine is being employed in the determination of its atomic weight. G. T. M. Detection of Cobalt in the Presence of Nickel. By ADAM JAWOROWSKI ( C l ~ m . Cerctr., 1898, i, 144; from Pharm. Zeit. Rus~., 36, 632-634). -The liquid, if containing free acid, is neutralised with sodium carbonate, and solid sodium pyrophosphate is added until the tur- bidity a t first produced has again disappeared ; the liquid, after being poured off from any undissolved pyrophosphate, is, if necessary, diluted with water until nearly colourless, and 6-43 C.C.of this solution is shaken with 1-1 -5 grams of sodium carbonate and 5-8 drops of brom- ine water; if cobalt be present, even in traces, the liquid will be of a By ANDRE JOB (Compt. rend., 1898, 127, 100-103. Compare McConnell and Hanes, Trans., 1897, 584).-It is well known that when hydrogen peroxide or bromine is added to a solution of potassium hydrogen carbonate mixed with a cobalt salt, the pink colour of the solution changes to green, but the composition of the green compound has not hitherto been ascertained. By titration with the special ferrous reagent proposed by the author (this vol., ii, 51), he finds that the ratio of cobalt oxide to active oxygen is 2Co : 0.Hence the green compound is a de- rivative of the oxide Co,O, ; its exact formula is Co2(OH),(HC0,), The oxidation of the cobalt in presence of an alkali followed by the titra- tion of the green compound with the ferrous reagent and iodine affords a rapid and very accurate means of estimating cobalt in presence of iron or nickel. Any excess of hydrogen peroxide soon decomposes or may be expelled by boiling the liquid for a few minutes. The best plan is to add the solution containing the cobalt salt t o a mixed solution of potassium hydrogen carbonate and hydrogen peroxide, heat gently for a few minutes and then titrate.beautiful green when observed in reflected light. L. DE I(. Oxidation of Cobalt Salts in Presence of Alkalis. C. H. B. Analysis of Swedish Magnetic Iron Ores. By TH. WETZKE (Chem. Centr., 1898, i, 144; from Zeit. O$z. Chem., 3, 575-577).- The solution in hydrochloric acid is effected by micro-burners so as to avoid boiling, and the vessel is covered; the iron is finally titrated with permanganate, any sulphurous insoluble residue being examined by Hauffe’s method. Phosphorus is best estimated by the process recommended by Meineke (Abstr., 1897, ii, 157), namely, ignition of the phosphomolybdate precipitate in a Gooch crucible. L . D E I(. Estimation of Tin in Tin Plate. By ROBERT JOB (J SOC. Chem. Ind., 1898, 17, 325-326).-1n a 200 C.C.flask, about 100 C.C. of pure hydrochloric acid of sp. gr. = 1.2 is placed, a piece of marble weighing about 5 grams is dropped in, and the flask is closed with a perforated cork fitted with a Bunsen valve. After about 5 minutes, 4 square inches of the tin plate, from the middle of a sheet, cut into strips, is dropped into the acid, and the flask is heated until the action becomes vigorous. When the plate is nearly dissolved, another fragment of marble is added, and immediately the plate is entirely dissolved, and while62 ABSTRACTS OF CHEMICAL PAPERS. effervescence from the marble still continues, the flask is cooled as rapidly as possible to 0'. I t is then filled to the mark with ice cold water, the contents mixed, and 50 C.C.of the solution titrated with iodine (1 C.C. = 0°00056 gram of tin). A fresh quantity of the iodine solution equal to that required in the first; titration is then placed in a flask, and 50 C.C. of the tin solution delivered into it, keeping the point of the pipette just below the surface; in this way, oxidation by air is completely avoided. A little starch is added and the titration finished with the iodine. Electrolytic Estimation of Tin in Tin Ores. By EDWARD D. CAMPBELL and E. C. CHAMPION (J. Amer. Chem. Soc., 1898, 20, 687--690).-0ne gram of the finely ground ore is fused with 3 grams of sulphur and 3 grams of sodium carbonate in a covered porcelain crucible inserted in a larger one; this is then placed on a triangle and covered with a Hessian crucible, the bottom of which has been removed.After being heated for an hour to a full red heat, it is allowed to cool, the inner crucible is placed in a beaker with water to dissolve the soluble tin compound, and the insoluble portion is again fused with soda and sulphur and the twofiltrates are united. Hydro- chloric acid is now added until all the tin is precipitated, and the liquid is evaporated to about 75 c.c,, mixed with 10 C.C. of hydrochloric acid, and fully oxidised by means of sodium peroxide. After filtering off from any sulphur, ammonia is added to the filtrate until the liquid becomes slightly turbid, and then 50 C.C. of a 10 per cent. solution of ammonium hydrogen oxalate is added, which will cause the solution to become clear, The liquid is now electrolysed during a night with a current N.D.l O O = O * l O amphre with an E. M. F. of 4 volts. This will cause the metal to be completely precipitated as a pure white, firmly adhering coating upon the platinum dish, and after being washed first with warm water, and then with alcohol, it is dried and weighed. The dish, cleaned by dissolving off the tin with hydrochloric acid, is Estimation of Tin in Commercial Antimony. By JOHN PATTINSON AND HUGH S. PATTINSON (J. SOC. Chem. Tzd., 1898, 17, 214--215).-1n this process, advantage is taken of the very sparing solubility of antimony pentasulphate in cold strong sulphuric acid. About 6.5 grams of the pulverised metal is heated with 75 C.C. of sulphuric acid of sp. gr. = 1.83 until completely dissolved, which re- quires about 10 minutes, and then allowed to cool completely, when the greater part of the antimony sulphate crystallises out, whilst the tin, arsenic, copper, iron, and other impurities (except lead) remain in solution, The whole is thrown on an asbestos filter, the precipitate washed several times with sulphuric acid (sp. gr.= 1-83), and the filtrate diluted with 3 times its volume of water, cooled, nearly neutralised with ammonia, and treated with an excess of strongly yellow ammonium sulphide. All the metals except tin, arsenic, and antimony are precipi- tated, and are filtered off and washed with ammonium sulphide. The filtrate is then feebly acidified with dilute sulphuric acid and the sulphides of tin, arsenic, and antimony collected and washed. The precipitate is next boiled with a solution of bromine in strong hydro- M.J. S. washed, dried, and weighed. L. DE I(.ANALYTICAL CH EMISTBY. 63 chloric acid, any sulphur which separates is oxidised by fuming nitric acid, and the united solntions are boiled until free from bromine. The greater part of the acid is now neutralised by ammonia, and the tin is separated from the other metals by Clarke’s oxalic acid process, for which purpose oxalic acid is added to saturation and hydrogen sulphide passed through the hot solution; this throws down the arsenic and antimony, whilst the tin remains in solution. Finally, the filtrate is nearly neutralised with ammonia, an excess of ammonium sulphide added, and the stannic sulphide thrown down by acetic acid. M. J. S . Assay of Telluride Ores, By CHARLES H.FULTON ( J . Amer. Chem. Xoc., 1898, 20, 586--597).-As telluride ores, when assayed for gold and silver, give less accurate results than other auriferous minerals, the author has made a laborious investigation into the cause of this, and has now greatly improved the process. I f the ore is very poor, 1 assay-ton of the powder is mixed with the same amount of dry soda, and half its amount of silica; 1-6 assay- tons of litharge, 10 grams of borax glass, and 2 grams of argol are added, and the whole well mixed and covered with salt. If traces of silver are likely to be present, an extra 0.02 gram of this metal is added, and also 4 iron nails to assist in desulphurising the ore. I n dealing with rich ores, only one-fifth of an assay-ton is used with a t least 2 assay-tons of litharge, 1 assay-ton of soda, and 1 gram of argol.The fire should be moderately hot, and the fusion should last from 40-50 minutes; the lead button obtained should weigh about 20-28 grams, and be, if possible, submitted a t once to cupellation without previous scorification. The traces of gold absorbed by the cupel may be disregarded, but the slag should always be powdered and remelted with another assay- ton of litharge and 2 grams of argol ; the button formed is then also By 9. N. RAIKOW (Chem. Zeit., 1898, 22, 149--150).-Dietze has proposed the following standard tests for the purity of Otto of roses. (1) The sp. gr, a t 1 5 O should not exceed 0.870. (2) The solidification point should not be below 15--20°. (3) The rotatory power observed in a 100 mm.tube at 20° should not exceed -lo 30’. (4) The saponification number should not exceed 9.5-10, and the relation between the acid and ether numbers not be higher than 7. The author having tested the process with two samples of Otto of roses of undoubted genuineness, states that the tests do not safeguard Technical Analysis of Asphaltum and Asphalts. By STEPHEN F. PECKHAM (J. SOC. Chem. Izd., 1898, 17’,438--439).-For technical purposes, a more or less empirical examination is of greater value than a scientific analysis, and it isof primary importance to ascertain what proportion of the mineral is soluble in light petroleum, boiling oil of tur- pentine, and chloroform respectively. If the ‘‘ organic matter not bitumen ” is large in amount, its nature should be ascertained.The cupelled. L. DE K. Examination of Otto of Roses. against an adulteration with oil of geranium. L. DE K.64 ABSTRACTS OF CHEMICAL PAPERS. amount and form of combination of the sulphur is next in importance, then the amount of the nitrogen and, lastly, that of the iron and aluminium. The solvents used should be carefully freed from water. Quantities of the mineral containing about 0.5 gram of bitumen aro weighed into two 11 cm. extracted filters of known weight, placed in stop-cock funnels ; the filters are filled with light petroleum which is allowed to run off slowly until the colour has perceptibly diminished, then the stop-cock i8 kept closed for 15-30 minutes after each addi- tion of solvent and the last digestion should extend to 12 hours.The filters are dried a t 6 6 O and weighed. They are then exhausted with boiling oil of turpentine which is displaced by light petroleum before drying ; the same filters are then exhausted with chloroform. They may next be extracted with 10 per cent. hydrochloric acid which re- moves calcium and magnesium carbonates and calcium sulphate ; and lastly, they are deflagrated with sodium carbonate and potassium nitrate, and the sulphur present as pyrites, silica, iron, and aluminium are estimated in the fused mass. For total sulphur, a fresh portion of the mineral is deflagrated in the sameway. Another portion is dissolved in benzene, the solution washed with dilute (10 per cent.) sulphuric acid, and the acid solution is treated with sodium nitrate when any nitro- genous basic oils present will be recognised by their odour.Another portion is exhausted with chloroform, and the residue extracted with ammonia for peat acids. Free sulphur will be found in the petroleum extract, organic aluminium compounds in the turpentine solution. Nitrogen may be determined by any method. The interpretation of the analytical results is reserved for a later communication. By FREDEBIC WILLIAM RICHARDSON and ADOLF J A F F ~ (J. Soc. Chem. Ind., 1898, 17, 330--333).-The only trustworthy methods for the estimation of glycerol are the dichromate and the acetin methods (Abotr., 1890, 425), and of these the dichrom- ate process is the simpler and more satisfactory. The authors depart slightly from Hehner’s instructions, relying on basic lead acetate for the removal of chlorides as well as of organic impurities, removing the excess of lead by diluto sulphuric acid and filtration before adding the dichromate.They recommend somewhat different proportions, namely, 20-25 C.C. of the glycerol solution, 25 C.C. of dichromate (Hehner’s solution, of 74-86 grams per litre : 1 c.c.=O-Ol gram of glycerol), and 25 C.C. of concentrated sulphuric acid, and find that 20 minutes heating in the water-bath completes the reaction. It is possible to estimate the glycerol in commercial samples by using a table of specific gravities, but i t is necessary to determine the ash and to deduct certain constants from the specific gravity, in order to correct for the organic impurities usually present ; these constants require to be varied for different percentages.Estimation of Phenols in Disinfectants in the Presence of Soap. By WERNER SPALTEHOLZ (Chm. Zed., 1898,22, 58)-The following process is recommended for testing compounds which contain the phenols in the free state; the sample is heated in an iron retort to 200°, and steam is passed through i t until oily drops no longer come over with the distillate. The temperature during the distillation should M. J. S. Estimation of Glycerol. M. J. S.ANALYTICAL CHEMISTRY, 65 not exceed 210°, as otherwise any olein soap which may be present will undergo decomposition ; rosin soap, however, may be heated with- out decomposition to 220". The distillate is shaken with benzene to dissolve the phenols, and the benzene is then agitated with aqueous soda and the phenols precipitated from the alkaline solution by means Estimation of Small Amounts of a-Naphthol in Commercial P-Naphthol.By JOHN PROCHAZKA and H. N. HERMAN (J. Xoc. Chem. Ind., 1897, 16, 894--895).-The method is based on the fact that a-naphthol combines with diazo-compounds more readily than ,&naphthol does, and forms more soluble products. The diazo-compound employed is thatof 1 : 4-naphthylaminesulphonic acid (naphthionic acid) : 27.5 grams of (90 per cent.) sodium naphthionate is dissolved in 185 C.C. of water, 6.2 C.C. of concentrated sulphuric acid is diluted to 186 C.C. and slowly added with stirring, the mixture is cooled below 5", and 7 grams of (98 percent.)sodium nitrite ina 10 per cent. solution is slowlyadded ; the paste obtained in this way is made up t o 750 c,c.The P-naphthol solution is prepared as follows : 1.5 grams of P-naphthol, 200 grams of a 25 per cent, solution of sodium hydroxide, and 140 grams of a 10 per cent. solution of sodium carbonate are heated together until the P-naphthol is dissolved, and the solution is made up to 250 C.C. and cooled below 5" before use. To this P-naphthol solution, 37.5 C.C. of the diazo-solution (being GG of the theoretical amount) is slotvly addedwith stirring. If the P-naphthol solution is pure, the coloiir produced is entirely insoluble; but the presence of less than 10 per cent. of a-naphthol is indicated by a peculiar bluish tint and the production of a coloured supernatant liquid. The amount of a-naphthol can be very fairly estimated by dipping strips of filter paper into this liquid, and comparing the depth of t i n t with that produced by a solution of known composition.By ADOLF LIEBMANN (J. Xoc. Chem. Ind., 1897, 16, 294-296).-Quantities of a-naphthol greater than 0.1 per cent. present in the P-naphthol used for preparing paranitraniline red, are deleterious as regards the colour of the resulting dye. P-Naphthol can readily be freed from a-naphthol by crystallising it from toluene, washing the crystals first with a mixture of toluene and light petroleum, then with the latter alone, steaming, and crystallising the steamed product several times from boiling water. For determining the quantity of a-naphthol in P-naphthol, the following process gives good results even when 0.01 per cent. only of the former is present ; 0.144 gram of the P-naphthol is dissolved in 5 c,c.of absolute alcohol contained in a graduated test-tube, and the solu- tion made up to 15 C.C. with pure toluene ; 0.14 gram of paranitraniline dissolved in 9 C.C. of dilute hydrochloric acid is then cooled and diazo- tised with 1 C.C. of normal sodium nitrite, and 1 C.C. of the diazo-solution added to the tube containing the &naphthol. Water is then added, the toluene solution separated and shaken with 5 C.C. of normal caustic soda, and the colour of the alkaline solution compared with that of the alkaline solution obtained in exactly the same way from P-naphthol containing a known quantity of a-naphthol. The test is based on the fact that the hydroxyazo-derivative obtained from of an acid.L. DE K. M. J. S. Paranitraniline Red. VOL. LXXVI ii. 566 ABSTRACTS OF CHEMICAL PAPERS. a-naphthol and paranitraniline is soluble in alkalis, whereas the corresponding P-naphthol derivative is insoluble. The following method of estimating metanitraniline in paranitr- aniline is available when more than 0.1 per cent. of the former is present. 0.25 gram of the paranitraniline is reduced to paraphenylene- diamine by zinc dust and hydrochloric acid in a flask which permits gas t o escape and yet does not allow air to enter; the colourless solution is then quickly filtered and diluted t o 250 C.C. 0.25 gram of metanitraniline is reduced in the same way, and made up to 250 C.C. Ten C.C. of the paraphenylenediamine solution is then diluted to 50 c.c., and a drop or two of dilute sodium nitrite added ; if the paranitraniline is pure, only a very slight change in colour occurs.If, however, metanitraniline is present, a light brown coloration (Bismarck brown) is produced ; by comparing the latter with that obtained from a known quantity of the standard metaphenylenediamine solution, the amount of metanitraniline can be determined. W. A. D. Estimation of Sugars as Osaeones. By CARL J. LINTNER and KROBER (Ann. Agrron., 1898, 24,448; from Zeit. Bvauwesen, 1896,153). -Twenty C.C. of t h e solution, which should not contain more than 1 per cent, of sugar, is heated for one hour and a half on a water bath with phenylhgdrazine (1 gram) and 50 per cent. acetic acid (1 gram). I n presence of dextrin, the heating must be prolonged to 2 hours, Boiling water (20 c.c ) is then added, and the osazone collected on a tared tilter, dried a t 110' for 3 hours, and weighed, Saccharose, which is not completely inverted by acetic acid, is first inverted with hydrochloric acid ; sodium acetate is then added, and the process continued as described.The method is suitable for dextrose, levulose, and saccharose both alone and in presence of maltose and dextrins; the latter increase the weight of the osazone slightly. Detection and Estimation of Glucose in Diabetic Urine, By A. CARPENI~ (L'Orosi, 1697, 20, 157--160).-A quantity of urine containing not more than 0.2 gram of glucose is, if acid, neutralised with potash or, if alkaline, boiled to expel ammonia, mixed with a slight excess of lead acetate solution, filtered, and the precipitate on the filter washed with water.Five or six grams of glycerol are added to the filtrate and washings, and the liquid is then mixed with six times i t d volume of alcohol of 95-96', filtered again if necessary, and about 10 C.C. of baryta water added. If glucose is present, a precipitate is produced, which may be collected, washed with alcohol of 95-96', and converted into barium sulphate in the usual manner. Each decigram of the sulphate corresponds with 0.0772 gram of glucose. The test analyses given are very satisfactory. The glycerol in this process is added to prevent the precipitation of barium hydroxide by the alcohol. According to the author, the precipitate which glucose forms with baryta has the composition BaC6HIoo6, and not Ba(C6H1,0,),, as By ARTHUR R.LING ( J . Soc. Chem. Ind., 1898, 17, 1 10-lll).-The author gives N. H. J. 11. usually stated. N. L. Clerget's Method of Estimating Saccharose.ANALYTICAL CHEMISTRY. 67 lOOIl 142.66 - 0.58’ results confirmatory of Herzfeld’s well-known formula, S = which is used when K stands for the difference between the readings of the Soleil-Ventzke-Scheibler polariscope, before and after the inversion of 26.048 grams of saccharose per 100 C.C. It is customary t o make the solution after inversion of only half this strength, and t o multiply the observed reading, i, by 2 before deducting it from the original reading. If any quantity other than 13.024 grams of sugar is inverted and diluted to 100 c.c., the formula becomes X= 141.84 + o.05i - 0,5t.Modiflcation of Clerget’s Saccharimetric Process, applicable to After Products, and Molasses. By ARTHUR R. LING and JULIAN L. BAKER (J. Xoc. Chern. Ind., 1898, 17, lll-l14).-The authors recommend the use of invertase (compare O’Sullivan, Trans., 1886, 58; 1891, 46) instead of hydrochloric acid for the inversion of saccharose, on the ground that, in highly impure substances, its hydrolytic action is u o r e likely to be confined to saccharose than that of a mineral acid, and they show by experiments on pure sugar that identical results are obtained with the two inverting reagents. Satisfactory agreement is shown between duplicate determinations made by this process on various dark commercial products, 10011 M. J. 5. M. J.5. Estimation of Sugar in Chocolate. By P. CARLES (J. Phwm., 1898, [vi], 8, 245--257).-The optical method of determining sugar in chocolate is inaccurate, because it neglects to take into account the volume occupied in the solution by dissolved cocoa, and the rotation often caused by the latter. The values obtained by cupric reduction are also high, since, after inversion, substances, other than dextrose, which reduce Fehling’s solution are often present. The following process, however, gives good results : 16.20 grams of the chocolate in thin shavings is heated with 80 C.C. of water during 1+ hours on the water-bath, 2-3 C.C. of a solution of basic lead acetate added, the mixture well shaken, and, after being diluted to 100 C.C. with water, filtered, and the amount of sugar determined by the saccharimeter.I f N denotes the value thus obtained, the true percentage of sugar =0*919 N. W. A. D. By GEORG LEBBIN (Ann. Agron., 1898, 24, 352 ; from Arch. Hyg., 1897, 28, 213).-The finely powdered substance (3 t o 5 grams) is heated t o boiling with 100 C.C. of water, for half an hour, to gelatinise the starch; 20 per cent. hydrogen peroxide (50 c.c.) is then added, and the boiling continued for 20 minutes, during which time 5 per cent. ammonia (15 c.c.) is added, 1 c.c at a time. The mixture is then boiled for 20 minutes longer, filtered through a tared filter, the precipitate washed with boiling water, dried, and weighed ; the weight of the ash is afterwards deducted. In some cases, the nitrogen is determined and the corresponding amount of protein also deducted.Concordant results, obtnined with rye and wheat bran and with wheat grain, are given. Estimation of Cellulose. 5-268 ABSTRACTS OF CHEMICAL PAPERS. Filter paper treated with hydrogen peroxide and ammonia lost 2-72 to 5.65 per cent. Estimation of Crude Fibre in Fodders and Foods, By JOSEF KONIG (Bied. Centr., 1598,2'7, 706-711 ; from Zeits. Unters. NcdiTungs. u. Genussmittel, 1898, 3).-The air-dried substance (3 grams), contained in a dry 500 C.C. porcelain dish, is stirred with nearly 200 C.C. of glycerol (sp. gr.=1*230) containing 30 C.C. of strong sulphuric acid per litre, the rest of the 200 C.C. being used to rinse the glass rod, when done with, into the dish; the dish is then placed in an autoclave and heated for an hour at 137" (3 atmospheres pressure).When cooled to SO-loo", the contents of the dish are diluted with 200-250 C.C. of boiling water, and at once filtered through asbestos; and as soon as the whole of the residue is on the filter, it is washed successively with 300-400 C.C. of boiling water, about 50 C.C. of alcohol (about 93 per cent.), and a warmed mixture of ether and alcohol, until the filtrate is colourless. The filter is dried until the weight is constant, weighed, ignited, and again weighed ; the difference between the two weights is the amount of crude fibre free from ash. Gooch crucibles of 90 C.C. capacity and 35 mm. in diameter at the bottom were employed. An alternative method, not involving the use of an autoclave, is also given. The substance is boiled with glycerol and sulphuric acid (the quantities are as given above) in a reflux apparatus for 1 hour ; when the temperature reaches 120-1 30") the contents generally froth, and the flask has to be shaken a few times.The process is afterwards completed as already described. By the above method, the pentosans are removed either completely or nearly completely (to 6.62 per cent.). Leguminous fodders seem to be freed from pentosans with greater difficulty than gramineous fodders. The method is simpler, as well as more efficacious, than the older ones. Separation and Estimation of Acids of the Saturated Fatty Series. By SIGMUND ~ L Z M A N N ( A ~ L . Pharm., 1898,236,409-441). -The author has prepared and studied the properties of the barium salts of the following acids : formic, acetic, propionic, butyric, iso- butyric, isovaleric, hexoic, octoic, decoic, lauric, myristic, palmitic, stearic, and cerotic.Experiments on the separation of the volatile acids are described, based on the supposition that, if a mixture of the alkali salts of several fatty acids be partially acidified with sulphuric acid and distilled, the acid of highest molecular weight passes over first, and on further acidification, the next lower acid, &c. (compare Wechsler, Abstr., 1894, i, 12, and Crossley, Trans., 1897, 580). Formic and acetic acids cannot be separated in this way, but with the higher members a partial separation is possible. The solubility of the alkali salts of the volatile fatty acids in methylic and ethylic alcohols increases with the molecular weight, whereas amylic alcohol dissolves more sodium propionate than sodium butyrate. The salts of the lower members are not readily soluble in acetone, whereas sodium isovalerate is very soluble, and on cooling N.H. J. M. N. H. J. M.ANALYTICAL CHEMISTRY. 69 the solution no milkiness is caused, but a separation of long, glistening, prismatic needles ; by this means, isovaleric can be separated from propionic, butyric, or hexoic acid. Ethylic palmitate, prepared by heating the acid with a 3 per cent. solution of hydrogen chloride in alcohol (yield 80 per cent. of theory) melts at 24O and boils a t 1S4.5-185-5° a t 10 mm. Ethylic stearate, produced in a similar manner (yield, the same), melts at 33.5’ and boils at 199-201’ at 10 mm. The melting and boiling points of mixtures of these two ethylic salts are given.A. W. C. Sulphuric Acid as a Reagent in the Analysis of Fatty Acids By ERNST TWITCHELL (J. SOC. Chem. Ind., 1897, 16,1002-1004).- The saturated fatty acids (palmitic, stearic, &c.) dissolve in, but are not sulphonated by, concentrated sulphuric acid, even at 100’; on the other hand, oleic acid seems to be quantitatively sulphonated by acid of even S5 per cent., the sulphonated acid thus produced, C,,H,,* OH,* CH(SO,H)*COOH, being practically insoluble in light petroleum, whilst the unattacked fatty acids can be dissolved out by petroleum, if acid not stronger than 85 per cent. has been employed. From 95 per cent. sulphuric acid, petroleum fails to dissolve out stearic and palmitic acids.For the examination of a mixture of saturated and unsaturated fatty acids, about 1 gram is solidified in a thin layer on the bottom of a stoppered conical flask, about 3 C.C. of S5 per cent. sulphuric acid is added, and the flask slightly warmed ; as soon as a clear solution has been obtained, the flask is quickly cooled, 50 C.C. of light petroleum is added, the stopper inserted, and the flask violently shaken. The petroleum extract is then decanted, the flask twice rinsed with 10 C.C. of petroleum, and the united extracts washed with water and evaporated. Lard yielded 42.3 per cent., two samples of cotton-seed oil 32.6 and 23.9 per cent. respectively, of saturated acids, having melting points of 53-53.5’. M. J. S. Estimation of Tartaric Acid. By JOHN MOSZCZEKSKI (J.Xoc. Chem. Ind., 1898, 17, 215--216).-Goldenberg’s (‘ acid method,” which, a t the present time, is almost exclusively employed, does not provide for the separation of iron, aluminium, or the pectinic substances which are present in argols and lees, and which probably interfere greatly with the final titration. The following method excludes these substances, and is shorter than Goldenberg’s. Five grams of the finely pulverised substance is treated with 26 C.C. of 13 per cent. sulphuric acid; this provides a sufficient excess of sulphuric acid for all ordinary materials, a large excess should be avoided. After stirring for a few minutes, the mixture is made up to 250 C.C. with alcohol of at least 90 per cent., well shaken, im- mediately thrown on a ribbed filter, and 200 C.C.of the filtrate measured into a basin ; any long delay a t this stage is objectionable, as there is a risk of ethylic tartrate being formed. An alcoholic solution of potassium acetate is now added as long as it produces a precipitate of potassium hydrogen tartrate; also 5 C.C. of strong potassium chloride solution. After 6 hours, the precipitate is col-70 ABSTRACTS OF CHEMICAL PAPERS. lected on a filter, washed with strong alcohol, and titrated as .usual. A correction of 1.2 C.C. should be made for the bulk of the insoluble matter in the 250 C.C. flask, and 0.320 ( 1 gram or per cent.) of tartaric acid should be added to the result to compensate for the solubility of potassium hydrogen tartrate in alcohol. The method is interfered with by the presence of potassium oxalate, as also is the case with Goldenberg’s ; phosphoric acid also raises the results, apparently to the extent of 1 molecule of tartaric acid for 1 molecule of P,O,.M. J. 8. Estimation of Potassium Hydrogen Tartrate in Wines. By LOUIS MAGNIEB DE LA SOURCE (Chem. Centr., 1898, i, 149; from Rev. Intern. Falszjic., 10, 195--196).-First method.-Ten C.C. of the wine is evaporated over sulphuric acid to dryness, and meanwhile a solution is made by dissolving 1 gram of potassium hydrogen tartrate in 100 C.C. of water containing 10-12 C.C. of alcohol, and adding an excess of pure potassium sulphate; 10 C.C. of the filtrate is then added to the wine residue, so as to dissolve all matters except the cream of tartar and any potassium sulphate which may be present.After washing the insoluble matter with the same liquid, the residue is dissolved in hot water and titrated with standard baryta water. Second method.-One hundred C.C. OE wine is mixed with 1.5 gram of potassium sulphate, and evaporated down to 15 C.C. ; after the lapse of 48 hours, the crystals are collected, washed with the alcoholic liquid already described, and finally titrated with baryta. L. DE I(. Analysis of Raw Materials containing Tartaric Acid. By JOH. SCHAFER (Chem. Zeit., 1898, 22, 255-256 ; 269 ; and 404-405). -The first paper deals with the estimation of tartaric acid in argols, &c. To obtain correct results, Goldenberg’s method, slightly modified, should be used. The original process, which is all but universally adopted, consists in heating the sample with dilute hydrochloric acid, neutralising the filtrate with excess of potassium carbonate, evapo- rating to a small bulk, adding acetic acid and alcohol, and finally titrating the potassium hydrogen tartrate with standard alkali.The author recommends first extracting any fatty matter by means of a mixture of alcohol and ether ; after adding the acetic acid and alcohol, the mixture should not be allowed to stand overnight, but should be at once filtered. I n titrating, feebly red litmus paper should be used as indicator. The second paper is a reply to Eckstein’s criticism (following abstract). L. DE K. Analysis of Raw Materials containing Tartaric Acid. By F. ECHSTEIN (Chem Zeit., 1898, 22, 351-352).-This is a criticism on Schafer’s article (preceding abstract).The chief point of interest is the author’s condemnation of the use of red litmus paper. Violet litmus paper or azolithmin paper should be used when titrating the precipitated potassium hydrogen tartrate with standard soda. The rest of the paper is not adapted for abstraction. L. DE K.ANALYTICAL CHEMISTRY. 71 Test for Uric Acid, and its Volumetric Estimation. By TORQUATO GIULI (Chem. Zeit., 1898, 22, 330-331).-The paper is a preliminary communication on a new test for, and a volumetric estimation of, uric acid ; 10 C.C. of the urine, or a solution of uric acid containing about the same quantity of the acid as occurs in urine, is mixed with 10 C.C. of a 7-5 per cent, solution of ammonium molybdate, and then with 10 C.C. of normal sulphuric acid.After a few minutes, 10 C.C. of normal soda is added, which causes the precipitate to redissolve; in the presence of uric acid, a bluish liquid is obtained. N/100 solution of potassium permanganate is now added, 1 C.C. at a time, until the liquid, which first becomes green and then yellow, changes to a reddish- yellow. This is the end reaction, and if the permanganate has been checked in the same manner with a known quantity of pure uric acid, the amount in the sample is readily calculated. The author admits that urine probably contains matters which also reduce molybdic acid, and thus count as uric acid; the chief consti- tuent of urine, the urea, does not, however, reduce the reagent. 1,. DE K. Estimation of the Iodine Number. By C. ASCHMAN (Chem.Zeit., 1898, 22, 59 ; 71-’72).-This is a modification of Hubl’s process, the fats being measured instead of weighed, and a solution of iodine chloride used instead of the mercurial solution; 30 grams of potassium iodide is dissolved in 100 C.C. of water contained in a flask furnished with a side tube, which serves to carry off the excess of chlorine. Washed chlorine is now passed into the solution until the liberated iodine has redissolved, forming a dark yellow liquid. After being left for 6 hours in a cold place, potassium iodate and iodic acid separate. The clear liquid is then poured off, the crystals are washed four times with water, and the whole is made up to one litre. The solution is further diluted to such an extent that 10 C.C. of it, on adding 10 C.C.of a 20 per cent. solution of potassium iodide, liberates as much iodine as will correspond with 40 C.C. of N/10 sodium thiosul- phate. The fatty solution is made by measuring exactly 10 C.C. of the oil a t 15O, or 10 C.C. of a melted fat at 50°, and dissolving this in exactly 10 C.C. of chloroform. Of this solution, 0.4 C.C. is carefully measured into a 250 C.C. flask, and after adding 10 C.C. of chloroform, 20 C.C. of the iodine chloride is introduced, the mixture is well shaken, and left for 24 hours in the dark with occasional shaking, when it is titrated with thiosulphate in the usual way, after adding 10 C.C. of 20 per cent. solution of potassium iodide. A list of test results is given, the iodine figures being decidedly lower than those obtained by the original Hub1 method. It is claimed, however, that they are more trustworthy and concordant.L. DE I(. The solution will keep unaltered for a year. Detection of Sesame Oil in Butter or Margarines. By PAUL SOLTSIEN (Chem. Centr., 1898, i, 224 ; from Pharm. Zeit., 1897,42,83’7 ; 846)-In applying the furfuraldehyde test to butter or margarine, care should be taken to use a hydrochloric acid of sp. gr. = 1.125, as the stronger acid of sp. gr. = 1.19 causes a distinct red coloration with furfuraldehyde alone, and also with turmeric, which is often used t o72 ABSTRACTS OF CKEMICAL PAPERS. colour butters. Fats containing sesame oil give a permanent, fine red colour when heated on a water bath with stannous chloride; the colour is not discharged by moderately diluting with water.Turmeric also gives a reddish coloration with strong hydrochloric acid or stannous chloride, but the colour is not stable and disappears on Detection of Rape Oil. By A. PALAS (L’Orosi, 1897, 20, 49-50).-Thirty cubic centimetres of a 0.1 per cent. solution of rosaniline are mixed with 20 C.C. of sodium hydrogen sulphite solution of 34O, 200 C.C. of water, and 5 C.C. of sulphuric acid of 65’. The reagent thus obtained, which should be perfectly colourless, is agitated in the cold with an equal volume of the oil to be tested, when a rose tint is developed if rape oil is present. Two per cent. of rape oil can be detected in olive oil by this means. The reaction is not given by other oils, neither is it given by the fatty acids of rape oil itself.adding water. L. D E K . N. L. Apparatus for the Rapid Estimation of Fat in Soap, and of Hehner’s Number. By GIOVANNI POSSETTO (L’Orosi, 1897, 20, 73--76).-The apparatus consists of a flask of about 200 C.C. capacity, into the ground neck of which is fitted a glass tube graduated from 0 to 40 C.C. and closed a t the top with a stopper, I n order to determine the fatty acids in soap, about 1 gram of the sample is introduced into the flask and dissolved in hot water. The solution is then cooled, acidified with dilute sulphuric acid, the graduated tube inserted into the neck of the flask, and sufficient water added to bring the level of the liquid to the zero mark or thereabouts. Ether is now added, and the contents of the apparatus well mixed and allowed to settle ; the volume of the ethereal layer is then read, and an aliquot part of it is pipetted off and evaporated, the residue being afterwards dried and weighed.For the determination of Hehner’s number (percentage of insoluble fatty acids), 1-1.5 grams of the fat or oil is placed in the flask and dis- solved in alcoholic potash, the alcohol being subsequently expelled by evaporation and the residual soap treated as above described. The apparatus is also adapted for the estimation of the amount of Use of Lead Dioxide as a means of Clarifying Urine for Analysis. By A. LOUBIOU ( J . Phawn., 1898, [vi], 8,270-272 ; from Bull. Xoc. Phavm. Bordeaux, July, 1898).-Turbid or fermented urine can be readily clarified by the following method ; to 10 C.C. of it caustic soda is added until the solution becomes alkaline to phenolphthalein, 1.0-1.5 grams of lead dioxide is then added, and the mixture well shakenduring a minute, and filtered.In the clear solution thus obtained, the albumin can be determined by Tanret’s reagent; experiment shows that no albumin is removed by the lead dioxide during clarification. The process described is especially useful in the case of turbid biliary urine which cannot be obtained clear by other met hods. W. A. D. Modification of Morner and Sj oquist’s Method of Estima- ting Urea, By HENRI MOREIGNE (J. Pharm., 1898, [vi], 8, 193-197. Compare Abstr., 1891, 758 and 1561).-The process named oil in paints, and for other purposes. N. L.ANALYTICAL CHEMISTRY. 73 is more rapidly carried out by determining the nitrogen of the urea by means of sodium hypobromite, instead of by Kjeldahl’s method, as originally recommended.After the process has been carried out as usualuntil the addition of magnesia, the solution is evaporated to 8 or 10 c.c., a little distilled water added, and the solution decanted from the magnesia; the latter is washed by decantation, and the washings added to the solution of urea, which is diluted to 50 c.c., and the urea present in 10 c.c., determined by the hypobromite method. The above method gives lower results with urine than are obtained on purifying the latter by basic lead aoetate, before estimating the urea; the difference is due to the ammonium salts not being removed in the latter process. Estimation of Urea in Urine by means of Sodium Hypobro- mite.By HENRI MOREIGNE (J. Pharm., 1898, [vi], 8,197-200, and 241-245).-The results obtained by the ordinary method, using basic lead acetate to purify the urine, are too high, owing probably t o the incomplete removal of ammonium salts (compare preceding abstract) and small quantities of guanine and creatinine. All the substances present in urine, including ammonium salts, which are capable of yielding nitrogen when acted on by sodium hypobromite, are, however, completely precipitated on adding phosphotungstic acid. It is im- portant that the latter should be prepared as follows : 20 grams of sodium tungstate and 2.34 grams of crystallised phosphoric acid are boiled with 100 C.C. of water during 20 minutes, the solution slightly acidified with hydrochloric acid, and, after standing, filtered.To 10 C.C. of the urine, a small quantity of water is added, together with 4 C.C. of hydrochloric acid, and the amount of phosphotungstic acid (15-20 c.c.) necessary for accurate precipitation, and the whole diluted to 50 c.c.; after standing 24 hours, the solution is filtered, and 25 C.C. of the filtrate, rendered slightly alkaline by caustic soda, diluted to 50 C.C. The urea is estimated in 10 C.C. of the solution thus obtained by the hypobromite method. In the case of certain pathological urines, it is advisable to subject the latter to a preliminary precipitation with basic lead acetate. After adding this and diluting the solution to a known volume, it is left for some time, filtered, and the lead precipitated in a n aliquot portion by dilute sulphuric acid ; any excess of the latter is neutralised with caustic soda, the solution acidified with hydrochloric acid, and the lead sulphate allowed to subside.One-half of the total volume is then filtered, precipitated with phosphotungstic acid, and the urea determined as above described. W. A. D. Relation between the Total Nitrogen of Urine and the Nitrogen Present as Urea. By HENRI MOREIGNE (J. Phcmn., 1898, [vi], 8, 293-302).-Although the method of determining the total nitrogen of urine by Kjeldahl’s process, generally adopted, is satis- factory, the methods for determining the nitrogen of the urea are by no means accurate (compare preceding abstract). If, for example, the solution to which sodium hypobromite is added, in the hypobromite method, contains more than 0.5 gram of urea in 100 c.c., an error as W.A. D.74 ABSTRACTS OF CHEMICAL PAPERS. great as 2.5-3 per cent. may arise ; the composition of the solution of hypobromite also exercises great influence on the volume of nitrogen obtained. In the author’s opinion, little value can be attached to the ratio of total nitrogen of urine to urea nitrogen determined by other investigators. W. A. D. Volumetric Estimation of Alkaloids as Higher Periodides. By ALBERT €3. PRESCOTT and HARRY M. GORDIN ( J . Amer. Chem. Xoc., 1898, 20, 706-728). Volumetric Method for the Estimation of the Total Alkaloids in ‘‘ Cortex china succirubr.” By H. EKROOS (Arch. Yhccrm., 1898, 236, 328-334).-The finely powdered bark is extracted with a mixture of chloroform, ether, and aqueous soda for 3 hours, and a certain portion of the clear ethereal solution is then shaken with normal sulphuric acid, the excess of acid being titrated back with decinormal caustic potash solution, using a freshly prepared alcoholic solution of hsemntoxylin as indicator.The end of the reaction is denoted by a pale yellow colour, which, on vigorous shaking, changes to a bluish-violet. The method can also be used for the analysis of quinine extracts. A. W. C. Indigo Testing by Permanganate. By J. GROSSYANN (J. XOC. Chem. Ind., 1S97, 16, 974-975).-1t has been stated by Rawson (J. Xoc. Chem. had., 1885, 489) that, by direct titration with perman- ganate after sulphonation, a fairly approximate idea OF the value of a sampleof indigocan be formed ; the author shows, however, that,although this may be true as regards high class indigoes (Bengal, Java), the method, when applied to low class samples (Kurpahs, Oiides), gives very variable results, which, in some cases, may err to the extent of 34 per cent.Results sufficiently trustworthy for practical purposes are, however, obtained if the sulphindigotic acid is precipitated by sodium chloride, and after washing is dissolved in dilute sulphuric acid and titrated. The instructions given by Rawson (not here reproduced) must be exactly followed, and they should be combined with the author’s dye test, which is not described in this place, See this vol., i, 89. M. J. S. Indigotin and Nitrobenzene. By B. WILLIAM GERLAND (J. SOC. Chem. fnd., 1897, 16, 108-109.Compare Abstr., 1898, ii, 102).- Carefully purified cold nitrobenzene dissolves indigotin very slowly ; on warming, however, a blue solution is obtained, which becomes tinged with red a t 65O, and pure blood-red a t 95’. On cooling, the solution remains red during several hours, but if a glass rod is introduced immediately becomes blue. Owing to the tendency to supersaturation which nitrobenzene solutions of indigotin exhibit, the solubility of the latter a t the boiling point of the solvent could not be accurately determined ; the numbers vary from 0.5-1 gram in 100 C.C. After indigotin has separated from its nitrobenzene solution, 25 C.C. of the filtrate contains only 0.000225 gram of indigotin; in the author’s process of determining indigotin (Zoc. cit.), this amount is quite negligible.The indigotin crystals, before weighing, should beANA1,YTICAL CHEMISTRY. 7.5 extracted on the filter with acid and alcohol, dried, washed with benzene, and digested with concentrated hydrochloric acid ; after boiling with water and with aqueous sodium carbonate, the crystals are finally thoroughly washed with water, dried, and weighed. To determine indigotin in woollen and cotton fabrics? it is best to destroy the fibre by treatment with dilute sulphuric acid, and after filtering, washing, and drying to extract the residue with nitrobenzene. The author has confirmed the accnracy of his method by comparing the results given by the latter with those obtained by the Muller- Bernthsen method ; in addition, accurate results were obtained when known weights of indigotin were suitably adulterated, and then determined by the nitrobenzene method.Detection of Turmeric in Rhubarb Powder, By ADAM JAWOROWSKI (J. Phavm., 1898, [vi], 8, 303-304; from Ann. Chim. Anal., 1898, 102)-If the powder contains turmeric, on agitating it with chloroform (10 c.c.) for several minutes, and filtering, a yellowish-brown solution is obtained with a greenish fluoresc- ence; and on adding light petroleum (50 c.c.), a yellow, flocculent precipitate is formed, although the solution remains yellow and fluorescent. This is now divided into two portions; to the first, sulphuric acid (2-3 c.c.) is added, and to the second, a saturated solution of borax (1-1.5 c.c.). I n presence of turmeric, the former becomes violet, whilst the acid layer, at first bright red, becomes succes- sively reddish-brown and yellow ; the borax solution becomes coloured violet, without, however, any colour-change taking place in the petrol- eum layer. W, A.D. Standard Methods for the Sampling and Analysis of Tan- ning Materials. By HENRY R. PROCTER and J. GORDON PAHKER (J: Xoc. Chem. Ind., 1898, 17, 6--lO).-This communication makes public the resolutions on the above subject adopted by an Inter- national Conference of the Chemists connected with the leather industry, held in London on September 28-30, 1897. The “hide powder” method was adopted for the estimation of tannin. The resolutions came into force on January lst, 1898, and all leather trades’ chemists who are members of the Association have agreed to work by the Conference methods until these are altered at a future Conference. M. J. S. Chemical Method of Ascertaining whether Milk or Cream has been Heated to at least 80”. By V. STORCH (Bied. Centr., 1898, 2’7, 711-714; from 40de Beret. kgl. Vet. LandbohCjskoZes Lab. Zindokon. POTS. Copenhagen, 1898, 1-1 6).-Not only mixed milk, but the milk of single cows during the whole period of lactation, shows a bluish-grey coloration when treated with a drop of dilute hydrogen peroxide and a few drops of a 2 per cent. solution of praphenylene- diamine. The action seems to be rather less in the case of cream and rather greater with butter milk, probably owing to the varying amount of fat globules present. Whey produces a dark violet-brown colour. It is supposed that the blue coloration obtained with milk is the result of a special interaction of the casein and the diphenylenediamine. W. A. D.76 ABSTRACTS OF CHEMICAL PAPERS. Butter milk from sour cream gave the coloration only when the free acid was partly neutralised with lime water. Milk fat globules free from milk serum, and pure casein solutions, gave no coloration. The '' active substance" could not be obtained quite pure. It is especially abundant in the deposit which forms on the inner side of a centrifugal separator. When obtained as free as possible from proteids, the aqueous solution gives the same violet-brown coloration as is obtained with whey, but when added to boiled milk (which gives no coloration with hydrogen peroxide and paraphenylenediamiue) the dark blue colour is at once obtained. Nilk which has been heated to 79-80' may still produce the blue coloration, provided that it is cooled as soon as that temperature is reached. Heating for 2 minutes a t 75" is sufficient to destroy the active substance " if the cooling is slow. As regards the effect of acidity, it was found that, whilst with milk which required 17 C.C. of N/lO alkali to neutralise 100 c.c., the limit of temperature was 79", butter milk which required 90 C.C. of alkali was rendered inactive by heating to 71". The reaction is not without importance in the case of Pasteurisation. It sometimes happens that the whole of the milk has not been heated to 79". Milk heated to 8Z0, mixed with 10 per cent. of milk heated to 7S0, showed the blue coloration distinctly. The presence of 1 per cent. of milk heated at 75" could also be detected. By P. ONFROY (J. Pharm., 1898, [vi], 8, 7-9).-1n presence of 5 per cent. of gelatin, as much as 10 per cent. of water can be added to chocolate without altering the latter's appearance. When a considerable proportion of gelatin is present it can be detected by boiling the chocolate (5 grams) with water (50 c.c.), adding 5 C.C. of 10 per cent. solution of lead acetate, filtering and adding a few drops of a saturated solution of picric acid; the latter produces an immediate, light yellow, amorphous precipitate with solutions containing more than 0.01 per cent. of gelatin. When only a small quantity of gelatin is present in a chocolate, it is ren- dered insoluble by the tannin of the latter, and the following method bas t o be adopted for its detection. After removing the fat from 10 grams of the sample by means of ether, 100 C.C. of warm water is added, then 5-10 C.C. of a 10 per cent. solution of potassium carbonate, and 10 C.C. of 10 per cent. lead acetate; on filtering and carefully neutralising, a solution is obtained which can be tested for gelatin by the ordinary methods. N. H. J. M. Gelatin in Chocolate. W. A. D.