ANALYTLCAL C'HEMiS'f HY. Analytical Chemistry, 45 Microcheinical Test for Alkalis and Acids ; Detection of small Qantities of Ozone and Water. By FRIEDRICH ERIICH (Xonatsh., 1901, 22, 670-678).-As a microchemical test for alkalis and acids, the author uses silk dyed either with red or blue litmus. A drop (0.95 mg.) of the liquid the reaction of which is to be tested is placed on a glass slide and observed under a microscope magnifying46 ABSTRACTS OF CHEMICAL PAPERS. 200 times, a condenser being used. Into this drop, a silk thread coloured with litmus is introduced. It is found that a perceptible colour change is given by the following quantities of alkalis, stated as millionths of a milligram ; 0.3 of sodium or potassium hydroxide, lithium, or czesiixm carbonate ; 0.5 of sodium, potassium, or rubidium carbonate; 30 of barium hydroxide and 10 of calcium hydroxide; and of acids 0.5 of sulphuric, hydrochloric, or nitric acid ; 1 of oxalic and 3 of acetic acid.Solutions of lithium, potassium, or rubidium carbonate, which have no effect on the colour of the flame, will yet give a marked reaction with the above reagent. Silk coloured with litmus can also be used for the detection of ozone, in the presence of potassium iodide or potassium ferrocyanide when red litmus is used, or of sulphur or potassium thiocyanate with blue litmus. Methods of Standardising Acid Solutions. By CYRIL G. HOPKINS (J. Amer. Chem. Xoc., 1901, 23, 727--740).-The author has studied the respective merits of the following irtethods for standard- king volumetric acids : the silver chloride method, conversion into ammonium sulphate, use of metallic sodium, use of pure crystallised borax, electrolysis of copper sulphate, and standardising oxalic acid by means of metallic iron and potassium permanganate. Of these, the first two methods gave the best results.I n the silver chloridemethod, a carefully measured quantity of approxi- mately correct hydrochloric acid is precipitated with silver nitrate and the resulting silver chloride is collected in a Gooch crucible, washed, dried at 130-150°, and weighed. The ammonium sul- phate method consists in neutralising a definite volume of approxi- mately correct sulphuric acid with ammonia and drying the residue at Normal Alkalis and Indicators in Acidimetry. By C. A. JUNGCLAUSSEN (Chew.Centr., 1901, ii, 896--S97 ; from Apotl~. Zed., 16, 664--666).--Normal potassium hydroxide may be conveniently made from fresh ordinary ‘( caustic potash purified by alcohol.” The solution should then be standnrdised with normal hydrochloric or oxalic acid, phenolphthalein being used as indicator. Although the solution may not be free from carbon dioxide, this does not matter in the least whether titrating from acidity to alkalinity or the reverse way, pro- vided the liquid is cold and t h a t phenolphthalein is employed. A deci- or centi-normal solution may be prepared from the above normal solution and used for the titration from acid to alkaline reac- tion only, using iodeosin as indicator ; the solutions should, however, be checked with N/10 o r N/100 hydrochloric acid and the necessary correction applied.In cases where haematoxylin is used as indicator, the alkali should be completely freed from carbon dioxide by cautious addition of barium By LUDWIG W. WINKLER (Zeit. anal. Clhern., 1901, 40, 596 -600).-In titrating chlorides by Mohr’s method, the rerjnlts are seriously too high when K. J. P. 0. 120’ (compare Weinig, Abstr., 1893, ii, 245). L. DE K. hydroxide. L. DE K. Estimation of Chlorine in Natural Waters.ANALYTICAL CHEMISTRY. 47 the chlorine present is less than 25 mg. per litre, because a certain amount of silver nitrate is needed to produce a visible precipitate of silver chromate. Using a silver nitrate solution, of which 1 C.C. equals 0.001 gram of chlorine, the amount to be subtracted, when working in the manner here prescribed, is shown in the following table : Solution used, c.c.....0'2 0.3 0'4 0.5 0-6 0.7 0.8 0.9 1.0 2.0 Correction, C.C. ....... 0'2 0.25 0'3 0.33 0-36 0'38 0.39 0'40 0'41 0'4-2 From 2 C.C. up to 10 c.c., the correction increases regularly by 0.02 C.C. for each additional C.C. of solution used. I n each of two bottles of 150 C.C. capacity, there is placed 1 C.C. of a 1 per cent. solution of potassium chromate. To the first bottle is added about 90 C.C. of the water to be titrated and just enough silver solution to produce a red colour, which is then removed by another 10 C.C. of the water. This mixture then serves as a standard for com- parison. To the second bottle, 100 C.C. of the water are added and then silver nitrate until a just visible red coloiir persists for 5-10 minutes.The bottles must be protected from light as'much as possible. The re- sults of some titrations of very weak chloride solutions show that even when the correction is three times the amount to be estimated, very close agreement with gravimetric determinations can be obtained. M. J. 8. Estimation of Sulphur and Phosphorus in Iron and Steel. By UEALDO ANTONY (Gaxxetta, 1901, 31, ii, 274-277).-To rapidly determine sulphur and phosphorus in iron, the author recommends the use of an oxidising mixture consisting of 4 parts of manganese dioxide, 1 of potassium permanganate, and 2 of dry sodium carbonate, the procedure being as follows : 5 grams of the finely powdered sample are well mixed in a platinum crucible with 40 grams of this oxidising mixture, a layer of the latter being also laid on the surface and the whole heated gradually at first, afterwards more strongly, and finally to a bright red heat by means of a blowpipe flame, the mass being meanwhile kept well stirred with a platinum wire.When cool, i t is extracted with boiling water, the filtrate acidified with nitric acid and evaporated to a volume of about 30 c.c., t o which is added a little ferric chloride, then ammonium chloride and ammonia, the liquid being then heated and filtered ; by this means, silica, phosphates, and arsenates are removed. The filtrate is used for the estimation of the sulphur as barium sulphate, whilst the precipitate is dissolved in hydro- chloric acid, the solution evaporated to dryness, and the residue main- tained for some time a t 120-130°, t,o insure the insolubility of the silica.The mass is then dissolved in dilute hydrochloric acid and any arsenic removed by means of hydrogen sulphide, the excess of whichis then boiled off. Ammonium molybdate is then added to precipitate the phosphoric acid, which is afterwards reprecipitated as ammonium magnesium phosphate and weighed as magnesium pyrophosphste. Other elements, such as tungsten, chromium, sto., often present in steel, can also be detected and determined by this method. T. H. P.48 ABSTRACTS OF CHEMICAL PAPERS, New Nitrometer for use with the Sprengel Pump. By GIUSEPPE ODDO (Gazxettcc, 1901, 31, ii, 215-217),-This nitrometer, which is especially adapted for collecting and measuring nitrogen in its estimation in organic compounds, is a modification of Schiff’s and is fitted near the bottom with three apertures, one for running off the mercury, the second for connecting with the Sprengel pump, and the last, which is slightly above the other two, for the entry of the potash solution.A sketch of the apparatus is given. T. H. P. Action of Ammonium Carbonate on the Arsenic Sulphides By LUDWIG VANINO and C. GRIEBEL (Zeit. anal. Chem., 1901, 40, 589--591).-Solutions of arsenious or arsenic sulphide in ammonium carbonate should, when acidified, yield in the form of sulphide the whole of the arsenic they contain, without the necessity for adding hydrogen sulphide. Practically, however, some hydrogen sulphide is always given oft’ on adding acid, and the precipitation is consequently incomplete.If, however, the solution is largely diluted and acidified in a bottle, which is then closed for 24 hours, the hydrogen sulphide is re- absorbed and every trace of arsenic is precipitated. If a n open vessel is used, or a concentrated solution is acidified, addition of hydrogen sulphide is indispensable. I n separating the arsenic sulphides from those of tin and antimony by concentrated hydrochloric acid, boiling must be avoided, or a loss of arsenic by volatilisation will occur. AT. J. S. Gravimetric Estimation of Boric Acid. By ALFRED PARTHEIL and J. A. KOSE (Bey., 1901, 34, 3611--3612).--The distribution ratio obtained by shaking N/10 boric acid solution witb ether at 26’ is 34.2 : 1. The boric acid solution, acidified with hydrochloric acid, is extracted with ether for some 18 hours in a specially constructed extractor or ‘ perforator,’ in which the ether is made to work i t s way continually t.hrough the aqueous solution contained in a spiral tube, The flask containing the ethereal solution is afterwards placed in a vacuum desiccator over sulphuric acid and the residue weighed.The ether must not be distilled off at the atmospheric pressure, as boric acid volatilises with ether vapour. The method gives good results and may be employed for estimating the acid in various minerals. Sulphuric, phosphoric, and nitric acids, or appreciable amounts of iron, must not be present. J. J. 8. Separation and Estimation of small amounts of Potassium in Saline Mixtures. By FREDEEIK H. VAN LEENT (Zeit.anal. ChenL., 1901,40, 569--5’73).-From solutions containing small amounts of potassium with large quantities of sodium chloride, magnesium and calcium salts, the potassium is best separated a s potassium cobalti- nitrite, after which it may be weighed as perchlorate or platinichloride. The calcium and most of the magnesium should be removed by sodium carbonate. The filtrate is then slightly acidified with acetic acid and treated with ths .cobalt reagent, which is prepared by mixing, just heFore use, equal volumes of sodium nitrite solution (180 grams per litre) and cobalt solution containing 19.16 grams of crystallised cobaltANALYTICAT, CHEMISTRY. 49 chloride, and 50 C.C. of glacial acetic acid in a litre. The precipitate is allowed to subside for 6-7 hours a t 40-50°, and then all night in the cold, and after collecting on a filter is washed once with the re- agent and then thoroughly with 80 per cent.alcohol. The dried pre- cipitate is decomposed by evaporating with hydrochloric acid ; per- chloric acid is then added, and the mixture evaporated until white fumes are given off. The potassium perchlorate is triturated and washed with 96 per cent. alcohol, to which 0.2 per cent. of perchloric acid has been added, and then on the filter with ether-alcohol. I t is dried on the filter at 120-130' and weighed. For weighing as platinichloride, the yellow prezipitate should be decomposed by gentle ignition, the potassium nitrite dissolved out by a weak sodium chloride solution (to prevent the cobalt oxide from passing through the filter), and evaporated with hydrochloric acid to convert it into potassium chloride.34. J. S. Ekdima&ion of Byd,r,nlrlde in the P*r&%Rce of AJkaJi Carbonate. By W. E. RIDENOUR (Chem. News, l901,84,202).-The author has tested the process of titration of alkali hydroxide in the presence of carbonate, using first phenolphthalein and then methyl- orange; he finds that phenolphthalein does not indicate half the carbonate, either alone or in presence of the hydroxide. To ascertain the number of C.C. of normal acid corresponding with the carbonate in a mixture of alkali hydroxide and carbonate he multiplies by 3 the number of c,c. of normal acid indicated by methyl-orange, using phenol- phthalein, then methyl-0ran.e. and divides Qv 104.5726.D. A. L. Estimation of Cadmium. By EDRIUND H. MILLER and ROBERT W. PAGX (Zeit. an0s.g. Chepnz., 1901, 28, 233-241).-The electrolytic method of estimating cadmium is convenient and gives trustworthy results if care is taken t o avoid a large excess of potassium cyanide and the presence of other salts. A current of 0.1 to 0.15 ampere is employed and the estimation takes about 16 hours, The estimation by precipitation with sodium carbonate gives very unsatisfactory results. A convenient and accurate method is to precipitate the cold neutral cadmium solution with an excess of diammonium phosphate. The precipitate must be allowed to remain f o r some time and is then transferred to a weighed filter aiid dried a t 1 0 5 O , or is converted into pyrophosphate and then weighed.The solution must not be heated, since the precipitate of ammonium cadmium phosphate gives off ammonia on boiling and is partially converted into cadmium ortho- phosphate. E. C. R. Estimation of Mercury in Antiseptic Solutions containing Mercuric Chloride, Iodide, or Cyanide. By G. MEILLERE (J. Pharm. Chim., 1901,[vi],14,356--359).-Attention is drawn to the fact that mercurial antiseptic solutions contain, as a rule, other substances besides the mercury salt, so that mere evaporation of the solution and weighing the residue gives erroneous results. With solutions con- taining mercuric chloride or iodide, accurate results may be obtained by extracting them with ethyl acetatme and veighing the residue VOL. LXSXIJ. ii. 450 ABSTRACTS OF CHEMICAL PAPERS.left on evaporation of the ethereal solution. Mercuric cyanide may be estimated by means of a standard solution of iodine in the presence of an excess of an alkali hydrogen carbonate. It is necessary t o add a n excess of the iodine solution, which excess may be estimated by means of a standard solution of sodium thiosnlphate. H. R. LE S. Gravimetric and Volumetric Estimation of Mercury, Copper, and Zinc. By ROBERT COHN (Bey., 1901, 34, 3502--3508).- Volumetric estimation of mercury.--An excess of ~11/10 ammonium thiocyanate is added to the solution of mercury and the excess determined after the addition of nitric acid and ferric alum by means of N/lO silver nitrate; it is best in the titration to add more silver nitrate than is necessary for decolorisation, and to titrate back with the thiocyanate as in Volhard’s method.The method depends on the formation of sparingly soluble, non-dissociating mercury thiocyanate, Volumetric estimation of copper and zinc.-A solution of mercuric chloride (0.1 mol.) and ammonium thiocyanate (0.4 mol.) in water (1 litre) is standardised by means of N/10 silver nitrate according t o Volhard’s method ; on adding t o an excess of this a known volume of the copper or zinc solution, a precipitate of the type M”Hg(SCN), is formed, and, after filtering, an aliquot portion of the liquid is titrated againpt the silver nitrate solution. From the difference in the values for the titration of the same quantity of mercuric thiocyanate solution before and after precipitation, the amount of copper or zinc follows from the relationship Cu(or Zn) : SCN = 1 : 2.The equations involved are, for example, (1) HgCI, + 4NH,SCN + 4AgN0, = Hg(SCN), + 2AgSCN Hg(SCN),,Cu(SCN), + 2AgC1+ K,SO, + SNH,NO,. Gravirnetric estimation of the three metals.-In the case of copper and of zinc, the precipitate of the type MHg(SCN), is left for two hours i n the cold, filtered, and ignited ; with copper, the filter paper and precipitate are burned together, and when a large quantity has t o be weighed, the product is converted into cuprous sulphide by reduction in a stream of hydrogen and the addition of sulphur. With smaller amounts, the product is ignited with pure mercuric oxide, and weighed as copper oxide. I n the case of zinc, the procedure is similar, but the filter paper has to be initially removed.To estimate mercury, it is precipitated b y a solution of zinc sulphate (1 rriol.) and alkali thiocyanate ( 4 mols.), and the zinc oxide, ultimately obtained, weighed. H!dCNS),. + 2AgCl+ 4NH,NO, ; (2) HgCI, + 4NH4SCN + CUSO, + 2AgN0, = W. A. D. Volumetric Estimation of Manganese. By HUGH RAMABE (C”1~e.m.. News. 1901, 84, 209-210).-To correct for various incidental reactions, the method for titrating manganese described by the author and Reddrop (Trans., 1895, 47, 268) has been modified so far as wrought iron, steel, and pig-iron are concerned:-1.1 gram of the sample is dissolved by boiling with 30 C.C. of dilute nitric acid, the solu- tion is cooled, then boiled for 3 minutes with 0.5 gram, or more, of sodium bismuthate, cooled again, treated with a slight excess of sulphurous acid, reoxidised with 1.5 gram of sodium bismuthate, and filtered.Hydrogen peroxide is run in at first until the reddish colour givesANALYTICAL CHEMISTRY. 51 place to a clear yellow solution, then 0.6 to 1.0 in excess is added. The solution is titrated with N/10 permanganate. D. A. L. Detection of Chromic Acid by Hydrogen Peroxide in presence of Vanadic Acid. By C. REICHARD (Zeit. ccntcl. Chem., 1901, 40, 577--586).--The blue colour attributed to perchromic acid, produced by the action of hydrogen peroxide on chromic acid, is destroyed or prevented by the addition of vanadic acid with production of a brown colour. The blue ethereal solution of perchromic acid is decolorised by ammonium metavanadate. Addition of sodium phosphate or arsenate annuls the action of vanadic acid ; a nitrate has no such effect.Molybdates and tungstates also destroy perchromic acid, but not so energetically as vanadates. Estimation of Uranium. By EDWARD F. KERN (J. Amer. Chem. Xoc., 1901, 23, 685-’726).-Uranium solutions may be freed from the- metals of the fifth or sixth group by a current of hydrogen sulphide providing 50 parts of the liquid contain no more than 1 part of free hydrochloric or nitric acid. From metals-of the third or fourth group, it may be isolated by boiling for 15 minutes with a large excess of sodium carbonate. Ferric iron may be completely separated from uranium by shaking the hydrochloric acid solution three times in succession with pure ether; it is, however, essential that the acid should have a sp.gr. of 1-1 and that the ether be previously saturated with the acid. From an acetic acid solution containing alkalis or alkaline earths, uranium may be precipitated as hydrated oxide, U30$,3H,0, by electrolysis, or i t may be separated by a thrice repeated precipitation in a hot solution with ammonia in the presence of ammon- ium chloride ; the ammonium uranate, at first slimy and yellow, becomes darker and crystalline after 20 minutes’ boiling and is readily con- verted into U,O, by ignition over the blast with free access of air. Uranium may be also separated from alkalis by precipitating the boiling liquid with ammonium phosphate in the presence of ammonium acetate. After boiling for 15 minutes, the precipitate becomes crys- talline, and, like the ammonium uranate, it is washed with a 2.5 per cent.solution of ammonium chloride and then ignited i n a porcelain crucible. It is then moistened with nitric acid, reignited, and weighed as uranyl pyrophosphate. Several methods are given for the separa- tion of uranium from phosphoric acid, the most convenient process being the treatment of the nitric acid solution with metallic tin a t the boiling temperature. The most rapid estimation of uracium is accomplished by reducing the sulphuric acid solution with metallic zinc and titrating the resulting uranous sulphate with standard per- manganate in an atmosphere of carbon dioxide. Full particulars will be found in the original paper. Reduction of a hydrochloric acid solution either by zinc or stannous chloride gives unsatisfactory results.The assay of uraninite (pitchblende) is best made by the ether method, which is briefly as follows : The mineral is dissolved in nitric acid and repeatedly evaporated with addition of hydrochloric acid. After removing lead, copper, &c., by hydrogen sulphide, the filtrate is M. J. S.52 ABSTRACTS OF CHEMICAL PAPERS. boiled and oxidised with nitric acid and precipitated whilst boiling with ammonia. The precipitate which also contains ammonium uranabe, is, after washing, dissolved in hydrochloric acid and agitated with ether to remove the ferric chloride. The aqueous acid solution is nearly neutralised with ammonia and then boiled with a large excess of ammonium carbonate. The filtrate which contains the uranium is then boiled down to a small bulk, the precipitate redissolved by addition of hydrochloric acid, and the solution again boiled t o expel carbon dioxide.The uranium is then precipitated as directed by ammonia, or by ammonium phosphate, or estimated volumetrically after expelling the chlorine by means of sulphuric acid, or precipitated by electrolysis. Some ores of uranium such as carnotite contain vanadium. This may be got rid of by simply evaporating the nitric acid solution to dryness, and dissolving the residue in a hot solution of ammonium nitrate which dissolves the uranium and leaves the vanadium un- dissolved; no phosphates should be present. I t may also be re- moved by excess of sodium hydroxide or by neutralising the nitric acid solution with mercuric oxide and then adding mercurous nitrate.L. DE K. Detection and Estimation of Traces of Antimony in presence of large quantities of Arsenic. By GEORGIES DENIGBS (Compt. rend., 1901, 133, 688-689).-1f tin is used instead of zinc in contact with platinum for the detection of antimony by deposition on the platinum in the form of a dark stain, the reaction becomes much more delicate and will detect 0.002 mg. of antimony in 0.05 C.C. of hydro- chloric acid (1 : 4) provided that the quantity of arsenic present does not exceed 5 mg. per C.C. The rapidity with which the stain appears is a function of the quantity of antimony present and for a given time the depth of the stain is also a function of that quantity. The reaction can be used quantitatively by making precisely similar tests with solutions containing known quantities of antimony.A still more sensitive reaction is obtained with antimony cssium iodide. The antimony compound is dissolved in dilute hydrochloric acid (1 : 4) or dilute sulphuric acid (1 : lo), and the reagent is made by dissolving 1 gram of potassium iodide and 3 grams of czesium chloride in 10 C.C. of water. A drop of the antimony solution and one of the reagent are mixed and examined uader the microscope, when the antimony cmium iodide is seen to form in yellow or garnet-red hexagonal lamells, often grouped in stellate macles. The reaction will detect 0*0001 mg. of antimony in presence of 500 times the quantity of arsenic. I t is important that the quantity of arsenic does not exceed 50 mg. per c.c., as with a larger quantity iodine will probably be liberated.Qnantitative results can be obtained by the method of comparison. c. 13. B. Analysis of White Metal Alloys, By FRED IBBOTSON and HARRY BREARLEY (Clhem. News, 1901, 84, 167--169).-Antimony as precipitated reduces solutions of stannic chloride in the cold very slowly, whereas pulverised smelted antimony is inactive. The authors thereforeANALYTICAL CHEMISTRY. 53 employ the latter a t the boiling point in place of iron for the estima- tion of tin, With finely powdered metal, the reduction takes place rapidly, and when complete the solution is allowed to cool in an atmosphere of carbon dioxide and titrated with iodine. One-fifth the volume oE strong hydrochloric acid is usually employed, but in the presence of copper one-third the volume is used, and in the presence of lead there must be plenty of acid to prevent the formation of cuprous iodide in the one case, or lead iodide in the other.Iron, chromium, nickel, zinc, manganese, aluminium, bismuth, tungsten, phosphorus, sulphur, mercury, molybdenum, and cobalt are inactive or exert a negligible influence in the cold. Arsenic, however, is precipitated and carries some tin with it; and in the case of antimony, when iron is used for the reduction, the tin is all reduced before the appearance of the antimony, a result which serves as an indication that all the stannic chloride is reduced. D. A. L. Estimation of Chloroform. By WILLIAM A. PUCKNER (Pl~cwnz. Arch,, 1901, 4, 124-128).-The following process is recommended as being particularly suitable for the estimation of chloroform in a mixture of chloroform and ether.A quantity of the mixture equal to 0.05-0.2 gram of chloroform is put intoa strong flask containing 10 C.C. of N alcoholic potassium hydroxide free from chlorine, the flask is closed with a sound cork, covered with cloth and tied down firmly. After gently mixing the liquids, the flask is put in boiling water for 3 hours. When cold, the contents arecarefully neutralised with N sulphuric acid, using pheriolphthalein as indicator, and the chlor- ine is then estimated by means of silver nitrate and potassium chromate. Or the cold liquid may be acidified with nitric acid and titrated by Volhard’s thiocpanate process. From the amount of chlorine thus found, the percentage of chloroform in the mixture is readily calculated. If the percentage of chloroform in the sample is quite unknown, it may be determined approximately by digesting 1 C.C.for an hour with 25 C.C. of N alcoholic potassium hydroxide and titrating the excess of alkali, One C.C. of Nalkali consumed x 0,02977 equals the amount of chloroform per C.C. L. D E K . Estimation of Cyanide in the presence of a Chloride. By FRANK B. GATEHOUSE (Chem News, 1901, 84, 197).-When silver nitrate is added to a solution of potassium cyanide, the soluble double cyanide, KCN,AgCN, is first formed, pcnd no precipitate is obtained so long as there is cyanide in solution. Therefore, titration with N/lO silver nitrate until a permanent turbidity appears may be used to estimate the cyanide; each C.C.used =0*013036 gram of potassium cyanide. An equal volume of the N/lO solution is then added, the burette read, potassium chromate introduced as indicator, and the chloride estimated in the usual way. D. A. L. Estimation of Alcohol in Ether. By FRANZ FREYER (Chenz. Centr., 1901, ii, 900 ; from Zeit. Zandw. Verszcclk-Wes. Oesterr., 4, 955--959).-The amount of alcohol and water is estimated by agitat-54 ABSTBACTS OF CHEMICAL PAPERS. ing 20 C.C. of the sample with a saturated solution of calcium chlor- ide and noticing the diminution in volume. Twenty-five C.C. of the sample (which should not contain more than 1 gram of alcohol and water, otherwise it should be diluted with anhydrous ether) are put into an Erlenmeyer flask, iiiixed with 50 C.O.of a 10 per cent. solu- tion of acetyl chloride in chloroform and the flask closed with an india- rubber cork, through which passes a separating funnel containing 100 C.C. of water. After the lapse of an hour, a little of the water is in- troduced, and the whole well shaken and then titrated with 2 N alkali, phenolphthalein being used as indicator. A blank experiment is then made with 50 C.C. of the acetyl chloride solution, mixed with anhydrous ether, which is titrated in the same manner, The difference in the two titrations represents the alcohol, inasmuch as by the action of alcohol on acetyl chloride only 1 mol. of free acid is formed, whilst water liber- ates 2 mols. One C.C. of 2 N alkali = 0.092 gramor 0.1157 C.C. of alcohol. L. DE K. Quantitative Esterification and Estimation of Alcohols and Phenols.By ALBERT VERLEY and FR. BOLSING (Ber., 1901, 34, 3354--3358).-Esterifcation takes place rapidly and completely whena mixture of acetic anhydride and pyridine is used in p1ace:of acetic anhy- dride. For quantitative esterification, a mixture of 120 grams of acetic anhydride and 880 grams of pyridine is used ; this is titrated with stand- ard alkali, and again after heating with a known weight of the alcohol or phenol. The method has been successfully used with ethyl alcohol, amyl alcohol, cinnamic alcohol, menthol, phenylglycol, glycerol, phenol, P-naphthol, guaiacol, saligenin, thymol, eugenol, carvacrol, and santalol, but geraniol, terpineol, vanillin, benzyl alcohol, and linalool could not be satisfactorily determined by this method.Estimation of Eugenol in Oil of Cloves. By ALBERT VERLEY and FR. BGLSING (Ber., 1901, 34, 3359-3362).--Eugenol can be satisfactorily estimated in oil of cloves by quantitative esterification with a mixture of acetic acid and pyridine, provided that other phenols and alcohols are absent. Umney’s method (Pharm,J., 1895,25, [iii],950), in which the oil is extracted with 10 per cent, alkali and the residue weighed, is liable to large errors; Thorns’ method (Abstr., 1892, 250) gives small values with oils rich in terpenes. T. M. L. T. M. L. Nickel Salts as Reagents for Reducing Sugars. By MAURICE DUYK (Avzrz. Chim. ancd. appZ., 1901, 6, 364).-The reagent is best prepared by adding to 25 C.C. of a 20 per cent. solution of nickel sulphate 20 C.C. of aqueous sodium hydroxide of sp.gr. 1.33 and 3 grams of tartaric acid dissolved in 50 C.C. of water. A clear, slightly green liquid is thus obtained which does not change on boiling, but is a t once reduced on adding a reducing sugar, with separation of a lower oxide having a brown, or even intense black, colour. The re- agent is likely to be of great service in urine analysis, as it is not in the least affected by normal urine. Sollmann (Abstr., 1901, ii, 535) has also applied nickel salts, but his conclusions differ somewhat from those of the author. L. DE K.ANAT,YTICAL CHEMISTRY. 5 5 Detection of Sucrose in Plants by means of Invertin, and of Glucosides by means of Emulsin. By EMILE BOURQIJELOT (Compt. rend., 1901, 133, 690-692).-The presence of sucrose in vege- table extracts is most conveniently detected, and its quantity esti- mated, by the action of invsrtase, the rotatory power of the liquid being determined before and after the action of the ferment.Tnver- tase also hydrolyses gentianose and raffinose, but these substances are comparatively rare, and the products of hydrolysis are readily dis- tinguished from those of sucrose. Emulsin may be employed with advantage in the same manner for the detection and estimation of glucosides. For example, the pericarp of COCOS Yatai and the seed of asparagus were found to contain a considerable proportion of sucrose but no glucosides, whilst the rhizome of Xcrophularia nodosa contained a considerable quantity of a lzvorotatory glucoside. C. H. B. Estimation of Starch in the Grain of Cereals. By L ~ O N LINDET (J.Phama. C/’&L, 1901, [ vi], 14, 397-400. Compare Abstr., 1897, ii, 525).-The process only differs from the one previously de- scribed in that, instead of collecting the starch on a tared filter, the amylaceous liquid is siphoned off, and the starch which is left behind is washed by decantation, and then hydrolysed with dilute sulphuric acid. The resulting glucose and dextrin are then estimated by means of Fehling’s solution and the polarimeter. H. R. LE: 8. Estimation of Formaldehyde. By LUDWIG VANINO and E. SEITTER (Zeit. rind, Chew,., 1901, 40, 587-589).-An excess of potass- ium permanganate strongly acidified with sulphuric acid oxidises formaldehyde quantitatively to carbon dioxide and water. The pro- portions recommended are 30 grams of concentrated sulphuric acid, 50 C.C.of water (mixed and cooled), 35 C.C. of N/5 permanganate, and 5 C.C. of a 1 per cent. solution of formalin. After 10 minutes, the excessof permanganate is titrated by an empirical solution of hydrogen per- oxide. The results agiee well with those obtained by Romijn’s method, which is adopted by the Verein fur Chemische Industrie in Mainz [Abstr., 1900, ii, 326). M. J. 8. Some sources of Error in the Estimation of the Volatile Acidity of Wines. By CURTEL (Ann. Chim. ~ 6 d . appl., 1901, 6, 361--364).-The author states that the total acidity of a wine as found by direct experiment is always less than the sum of acids exist- ing in it, and has investigated the cause of this phenomenon. The volatile acids mill be found too high if the wine should contain neutral acetates or similar salts, as, on boiling, part of thoacid is liber- ated by the action of potassium hydrogen tartrate, or even tannic acid.Several experiments are given showing that the volatile acids thus formed may be equivalent to as much as 0.06 per cent. of acetic acid. Another serious source of error is the introduction of carbon dioxide by means of the steam which is passed through the wine when esti- mating the volatile acids. The steam should be generated from recently boiled distilled water. The error may amount to 0.029 per cent, expressed as acetic acid. A third, although less important, source5 G ABSTRACTS OF CHEMICAT, PAPERS. of error is the presence of ethyl acetate, which, when boiled with water, yields a faintly acid distillate, which may account for 0.009 per cent.of volatile acid. I n orcler to avoid these errors, the author first estimates the total acidity, less carbon dioxide. Twenty-five C.C. of the mine and 25 C.C. of water are introduced into a small flask furnished with a doubly per- forated cork and placed on a heated sand-bath, and steam is passed through at such n rate that the ?liquid does not undergo any sensible alteration in volume. After the lapse of an hour, the acidity is again determined, the difference between the two experiments being the volatile acidity. Supposed Use of Oxalic Acid for the Preparation of Hydro- ~ I I f i - ~ ~ ~ i d t ~ By AO-GVSTE ATiLvLGE (C~C[~G. Ccfltc, .i!&n~, ii, 834-835 ; from Xon. scient., [iv], 15, ii, 576. Compare Abstr., 1901, ii, 622).-The following process is proposed for the detection of oxalic acid in commercial hydrogen peroxide. Five hundred C.C. of the sample are evaporated to dryness on the water-bath with addition of pure sodium hydroxide, the well-dried residue is dissolved in dilute nitric acid, about 50 c.c.of water are added, together with calcium nitrate or chloride, and then excess of ammonia. The liquid is heated, the precipitate washed free from soluble salts by decantation, and then heated a t 50" with 30 C.C. of sulphuric acid of sp. gr. 1.2-1.4. The filtrate and washings are evaporated on the water-bath i n a vacuum until about 30 C.C. are left ; on cooling, any oxalic acid will crgstallise, and may then be further identified. The author has not as yet succeeded in detect- Probable Errors of Analysis resulting from the Presence of Salicylic Acid in Strawberries. By L. PORTES and A. DES- MOULI~RES (J. Pliaym., 1901, [ vi], 14, 342--351).-Compare this vol., ii, 40. L. DE K. ing oxalic acid in commercial hydrogen peroxide. L. DE K. New Process for the Detection and Estimation of Salicylic Acid. By HENR~ PELLET (Ann. Chirn. ancd fippZ., 1901,6, 364-365). -Twenty C.C. of the liquid to be tested are acidified and boiled down in a beaker and from time to time a cold stirrer is held over the liquid so as to condense the vapour, which is then tested for salicylic acid by dropping it on a slightly greased porcelain slab on which are placed minute' drops of ferric chloride. The author has noticed that no reaction will be obtained until the salicylic acid reaches the con- centration of 0*06-0*07 gram per litre. Supposing, therefore, that the liquid has to be boiled down to 6 C.C. before the test is obtained, then the 20 C.C. contain 10 : 6 =0*0006-7 or about 0*0004 gram of salicylic acid. Liquids containing much salicylic acid must first be suitably diluted. When dealing with wines, the process should be applied to the product resulting from the extraction of the wine with benzene. L. DE K