70 Analg t i c a l Chemistry. New Forms of Gas Burettes. By 0. BLEIER (Bw., 1895, 28, 2423-2427).--Tn order to avoid inaccuracy in measuring the original volume of gas taken, the burette is made to contain exactly 100 C.C. under the conditions of experiment. The burette is filled with water, and the gas introduced through a tap at the top by allowing the water to run out by a tap at the bottom of the burette. If the gas is readily soluble in water, i t must be passed through the burette for a certain time, in no case more than 30 seconds, in order to saturate the moisture adhering to the walls. A. H. Inertness of Oxidising and Reducing Agents in Analyses in the Wet Way. By XARCELLIN BERTHELOT (Ann. Chim. Phys., 1895, [7], 4, 429432).-The author calls attention to the fact that in the titrstion of persulphuric acid with ferrous sulphate and sul- phuric acid, a discoloration of the liquid often occurs, and the final point is difficult to detect.This leads to erroneous results, and the muse is, probably, that an intermediate compound of persulphuric acid with ferrous sulphate is formed ; this has a most decided colour, is unstable, but is not instantaueouslg reduced. I n cases of tit,ration of persulphuric acid, it is, therefore, better to at once add an excess of ferrous sulphate, and then to titrate back again with permsn- gttnate. Titrations of oxalic acid with potassium permanganate in similar manner, proceed very slowly in the cold, but, in order to complete the reduction quickly, it is merely necessary to add a trace of manganese sulphate.A similar observation has also been mado by Engel (Abstr., 1892, 277). These resultts may be due to the temporary formation of a manganic salt, or to the union of permanganic acid with a manga- iiese salt. J. J. S. Estimation of Water in Commercial Ammonium Sulphate. By JOHN HUGHES (C'hem. News, 1895, 72, 6).-The author calls attention to the importance of always estimating the moisture, and also the free sulphuric acid, in the samples. The water should be estimated by drying at looo, and this shouId be done first in the original sample, and then be repeated after the salt has been finely ground for the purpose of analysis, as the moisture lost during the preparation of the sample may occasionally amount tc! 1 per cent. The analytical results are afterwards calculated into the state of the sample as received.L. DE K. Detection of Chlorine, Bromine, and Iodine in Organic Compounds. By P. N. RAIKOW (Ohem. Zeit., 1895, 19, 902-903). -A fragment of silver nitrate is for it moment gently heated with a little sulphuric acid in a teat tube. After cooling, the substance to be tested is added, and the mixture is again gently heated. If iodine isANA LTTICAL OHEhllSTRY. ‘I1 present, part of it escaps, but some of i t forms a yellow precipitate of silver iodide, which, however, gradually disappears on boiling. Chlorine will be detected by a temporarv white precipitate ; bro- mino by a temporary pale yellow deposit. The process is applicable t o all organic compounds containing halojids. Electrolytic Estimation of the Halogens.By GEORG VORT- MANN (Monatsh., 1895, 16, 674-683) .-The author has previously described a method for the electrolytic estimation of iodine (Abstr., 1894, ii, d26), and now calls attention to several important improve- ments in the process. An anode of pure silver, shaped like a clock- glass, a cathode of platinum or‘ copper, aiid a current which does not exceed 2 volts in the case of cold solutions containing no alka.li t.artrate, or 1.3 volts when $he Bolution is warm and contains alkali tartrates, are employed. The electrolysis is continued until the solu- tion no longer gives the iodine reaction or until a new anode placed i n the solution does not gain in weight. Undor these conditions, no silver is dissolved from the anode, which retains on its surface the whole of the iodine in the form of silver iodide.L. DE K. G. T. M. Estimation of Iodine in Organic Substances. By M. C. SCHUYTEN (Chem. Zeit., 1895, l9,1143).-The author’s method is applicable to members of the fatty series only. The substancc is introduced into a 15-cm. test tube, and mixed with finely powdered, previously fused, potassium dichromste. After covering the mixture with anothey 5-6 cm. layer of the chromate, the tube is drawn out t o a bent capillary. While the latter is kept cool with a wet piece of cloth, the contents of the tube are heated, commencing at the top, until no further sublimation of iodine takes place. The tube is cut in two, the part containing the iodine is rinsed with cz solution of potassium iodide, and the iodine is then titrizted as usual; or the iodine may be weighed.I n the latter case, the cupil- lary part containing the iodine is connected with a calcium chloride t,ube containing also a fewlpieces of soda-lime ; when a magnifying glass no longer shows the presence of water, the tube is weighed. After gently heating to volatilise the iodine, the glass is re-weighed. Estimation of Sulphur in Pyrites. By ALEXAXDEE vos ASB~TH (Chern. Zeit., 1895, 19, 598--599).-The author .has investi- gated the processes recommended by Johuson (reduction of the ferric salt by means of sodium hypophosphite before precipitating with barinm chloride), Xoehnel and Glaser (fusion with sodium carbonate and sodium peroxide), and Fresenius (fusion with sodium carbonate and potassium nitrate).The second method is, in the author’s opinion, the best, but it is necessary to evaporate the aqueous solution to dryness with addition of bromine and hydrochloric acid so as to ensure the complete oxida- Detection of Sulphates, Sulphites, and Thiosulphates in Presence of each other. By R. GREW SMITH (Cheitz. ATe~~s, 1895, L. DE K. tion of the sulphur and the removal of any silica. L. DE I(.72 ABSTRACTS OF CHEMICAL PAPERS. 72, 39-40) .-Excess of barium chloride and pleni,y of ammonium chloi-ide are added to the solution, which is made dilute when thio- sulphates are present; hydrochloric acid is then dropped in until barium sulphate alone remains undiasolved. The filtered solution is next rendered permanently yellow with iodine, when a turbidity or precipitate indicates the presence of sulphite, which has been oxidised to sulphate, whilst any t.hiosulphate is converted into tetra- thionate ; this in its turn is oxidised to sulphate on adding bromine water to the clear solution.Hydrogen sulphide mould interfere with these reactions, and is removed at the start by means oE a current of carbonic anhydride. D. A. L. Qualitative Analysis of a Mixture of Hydric Sulphide, Polysulphid e, Thiosnlphate, Sulphite, and Sulphate. Bg W. POPPLEWELL BLOXAM (Chem. News, lb95,72,63-64).-The solation is precipitated with an animoniacal solution of zinc, cadmium, and am- monium chlorides. The filtrate is divided into two portions; one of these is tested for sulphites and thiosulphates by Fresenius’s pro- cess, which consists in neut.ralising part of the liquid with acetic acid and adding a trace of sodium nitroprusside and, if necessary, some potassiumferrocyanide, which gives a red coloration with sulphites.Thiosulphates ape tested for by adding hydrochloric acid. The other half is tested for sulphates as follows. After adding a little sodium hydrogen carbonate, the liquid is placed in a flask fitted with a trebly perforated cork ; through an inlet tube, a current, of carbonic anhydride is admitted, and an outlet tube is provided dipping below the surface of water, whilst through the third hole, a small stoppered separating funnel is passed, the tube reaching nearly to the bottom of the flask. While the gas is passing, the liquid is gradually raised to boiling, and, after all the air is expelled, hydrochloric acid is slowly admitted through the funnel until i t is in excess; the liqnid is then boiled down to one-fift)h of its original bulk, t,he current of carbonic anhy- dride still being maintained. After filtering off the separated mlphur, the liquid is tested for snlphates as usual.L. DE K. Estimation of Organic Nitrogen by the Kjeldahl Process. By HENRI E. CAUSSE (J. Pharm., 189!5, [ 6 ] , 1, 543-549; compare Dyer, Trans., 1895, 81 l).-The ordinary modification of the Kjeldahl process yields satisfactory results with substances, such as guano, con- taining less than 5 per cent. of nitrogen, but with more highly nitro- genous substances, such as wool and leather, it gives results as much as 2 per cent. too low. The addition of mercuric oxide, recommended by the American Committee, is either unnecessary or objectionable, for, although the time required to effect the decoloriaation of the liquid may be considerably diminished by using sufficient oxide, the clear product then contains unreduced nitrogeu.A sample of dried wool, for instance, which yielded about 12.5 per cent. of nitrogen by the soda-lime method, and also by the American method when a small amount of the oxide (0.2-0-3 gram to 0*3-0*35 gram of wool) was used, gave only 10.5 per cent., or even less, wberi the amountANALP TICAL CHEMISTRY. 73 of the oxide (2-4 grams) was such as to effect decolorisation in an hour. Neither is the result improved by the addition o€ sodium snlphic?e to the alkaline liquid before distillation, and t.he deficiency is there- fore not due to the retention of ammonia in mercuric combination ; the addition of permnnganate is equally inefficacious, as nitrogen ring compounds are stable towards that agent.The author recommends the substitution of copper sulphate for the mercury or mercuric oxide. The substance (0-3-0.8 gram) is boiled in a round-bottomed flask with concentrated sulphuric acid (20 c.c.), and a small amount (0.3 gram) of the sulphat,e, until n clear solution is obtained ; the operation requires about three hours, but does not need attention if the flask is surrounded by a metal jacket. The product is diluted (to 300 c.c.), rendered alkaline with caustic soda, and distilled into standard sulphuric acid in the usual way. The copper remains in the solution, which is at first blue, and finally colonr- less, provided the caustic soda is free from carbonate, b u t if carbonate is present Eopper is precipitated, and a portion of the ammonia re- tained.The specimen analyses quoted agree well with duplicates made by the soda-lime method. J-r. W. Estimation Gf Phosphoric acid by the Molybdenum Method. By HUGO NEUBAUER (Zeit. anorg. Chem., 1895, 10, 60-65).-1n the estimation of phosphoric acid by means of molybdic acid, the precipi- tate obtained is usually not more than 0.27 gram Mg.,P207. %'he author has determined the correction necessary to compensate for the volatilisation of phosphoric acid. With 0.07 gram Mg2P207, the volatile P,O, = 0 ; with 0.35 gram Mg2P207 the volatile P20a cal- culated as Mg2P,0, = 0.006 gram.Whence the following equation is obtained for the correction number 2. When n = the MgzP207 found in milligrams, z = (n - 70) . 0.021. The author recommends the use of a crucible lid coated with mag- nesium oxide to prevent loss of phosphoric acid, especially when larger quantities of phosphoric acid are being estimated. The follow- ing directions must be closely followed in order to obtain trustworthy resnlts by the molybdic acid method. The yellow precipitate is dis- solved in 100 c . ~ . of tL$ per cent. ammonia. The precipitation with the magnesia mixture must be performed slowly, with stirring. When the volatile phosphoric anhydride is determined directly by means of a crucible lid coated with magnesia, the filter is burned at the lowest possible temperature, and the lid put on directly the teni- perature is raised, the precipitate beiiig heated for one hour over R strong Terquem- or Teclu-burner, in such a way that the whole crucible is at a full red heat.Care must be taken to determine if the magnesia coated lid gains in weight under the influence of the burning gas. The author recommends a spirit burner instead of gas, especially when the gas coutains any quantity of sulphur compounds. E. C. R.74 ABSTRACTS OF CHEMICAL PAPERS. Analysis of Artificial Manures. By VON GRUEBER (Zeit. angw. Chein., 1895, 504--516).-An elaborate article, dealing with the practical analysis of manures in the way proposed by the society of G winan manure manuf~cturers. The methods present, on the whole, no novel features, but., by strictly adhering to them, analytical differ- ences will, no doubt, be to a great extent obviated.L. DE K. Gravimetric Estimation of Arsenic. By CARL FRIEDHEIM and I’AUL MICHAELIS (Zeit. anal. Chem., 1895, 34, 505-545) .-I. The am- monium magnesium arseiiate method.-The authors, combining the most advantageons features of the proposals of Puller, Wood, Fresenius, and others, proceed as follows :-About 0.3 gram of arsenic acid in 100 C.C. is mixed with 10 C.C. of ammonia (sp. gr. 0*96), 20 C.C. of magnesia mixture (made from magnesium chloride), and about 45 C.C. of alcohol; after.48 hours, the precipitate is collected and washed with a, mixture of 2 vols. alcohol, 1 vol. ammonia, and 3 vols. water until frce from chlorine. The filtrate contains only unweighable trac2s of arsenic, so that Puller’s cori%cction of 1 milligram per 16 C.C.is not required. The use of the Gooch crucible, with a filter-bed of asbestos, is strongly recommended, since, with paper filters, a loss of ai-senic is inevitable. The precipitate is ignited in oxygen, a t first by ;L single burner, then by a multiple one for 15-30 miiiutes, until a constant weight is attained ; the blowpipe should not be used, for, a t the temperature which it gives, decomposition of the pyroarsenate and loss of arsenic occur, moreover, the crucible should stand in a small porcelain capsule to protect its contents from the flame. The precipitate undergoes partial decomposit.ion when heated with water, hence attempts to ascertain its weight by rinsing it into a crucible and eyapcrating the water results in serious loss.Bunsen’s method of dissolving in nitric acid and evaporating is also unsatisfactory, but cxperiments conducted as above did not in any case deviate more than 0.0007 gram of As,03 from the amount taken. 11. Estimation as Trisu@hide.-Here, also, the employment of the Gooch crucible avoids the inconveniences attending the use of dried paper filters. The chief dif€iculty in this process is the tendency of the precipitate to contain more sulphur than corresponds with t>he formula AS& ; part, a t least, of this excess, is shown to exist in the form of arsenious hydrosulphide, which compound is not decomposed by expelling the dissolved hydrogen sulphide by a current of carbonic anhydride, and is only very slowly decomposed, with partial oxidation and separation of sulphur, when the precipitate is dried at 100-103”.Hence, all attempts to dissolve out the excess of sulphur by carbon disulphide, fail to remove it completely. For the same reason, also, the method of Classen and Ludwig (Abstr., 1885, 932) does not give correct results. Mobr’s method of dissolving in ammonia and evapo- rating t’he solution is also untrustworthy, as, in consequence of the large surhce exposed, oxidation of sulphur and loss of weight occur dui ing the drying. Rose’s method of oxidising the precipitate, determining the total sulphur as barium sulphate, a n d estimating the arsenic from the difference, seems to be free from sources of error. A customary niethod is to oxidise the precipitate by nitric acid andANALYTICAL CHEMISTRY.75 thyow down the arsenic acid as ammonium magnesium arsenate, but the presence of the sulphuric acid resnl ting from the oxidation, entails thc co-precipitation of some basic magnesium sulphate, and renders a double precipitation necessary. In this second precipitation, too much dilution must be avoided ; 30 C.C. of solution for 0.1 gram of Asz& is an appropriate proportion; a little magnesia mixtiire should be added, and 3 vol. of alcohol, or the precipitation will be incomplete. Backstrom has proposed (Abstr., 1893, ii, 299) t o separate the sul- phuric acid from ihe arsenic acid in the oxidised precipitate by simply heating until the former is expelled, but the authors show that how- ever carefully the heat is applied, a loss of arsenic acid occurs before :dl the sulphuric acid is driven off.Arsenious acid, intended for precipitation as sulphide, is kest dis- solved in potassium hydrogen carbonate, as its solution in an alkali hydroxide oxidises gradually to arsenate on exposure to air. Before precipitating, it muot be acidified with a quarter to one-half its volume of hydrochloric acid of 1.12 sp. gr. ; passing hydrogen sul- phide for half an hour suffices for compIete precipitation, but one hour is recommended. Subsequent expulsion of the dissolved hydro- gen sulphide, either by warming or by carbonic anhydride, is neither necessary nor desirable, and since arsenious sulphide is not soluble in hydrogen sulphide water, moreover, the passage of carbonic anhydride does not diminish the excess of snlphiir in the precipitate if filti+aticn is commenced within 10 minutes of stopping the current of hydrogen sulphide, but, on the other hand, dissolves traces of arsenic from the precipitate, in consequence of the reaction As,S3 + 3H,0 = As,03 + 3H2S.One and a-half hours drying a t 105-110° suffices to give a constant weight, ; in an atmosphere of carbonic anhydride, no loss of weight takes place a t 140". Boric acid. By A. VILLIERS and M. FAYOLLE (J. Pharm., 1895, [6J, 2, 241--244).-The ordinary tests f o r boric acid leare much to be desired. Those founded on the turmeric reaction a1.e unsatisfac- kory, because the colour change is not sufficiently characteristic. Those based on the flame coloration, on the one hand, if applied directly t o t'he solid substance, are liable to give misleading indications if copper be present, as is often the case in wine ash, and to be masked by the more intense colorations due to sodium, &c.; and, on the oiher hand, if applied to the alcoholic distillate, they are open to the objec- tion that the delicacy 01 the test is very variable, the intensity of the coloration depending on the stage of the distillation at which it is observed, as well as on the amount of boric acid present. I n testing for boric acid in wines, the authors recommend the repeated distillation of the ash (from 25 c.c.) with small quantities of mefhylic alcohol (3 c.c., three times), after moistening i t with cowen- trated sulphuric acid (1 c.c.). The whole of the boric acid passes over, and its amount may be estimated colorirnetrically with fair accuracy, by comparison of the flame with those given by standard solutions of the acid.Quantities from 0.5 to 0.1 milligram may be detected in this way. A number of natural wines which werc examined by this process M. J. S.76 ABSTRACTS OF CHEMICAL PAPERS. did not contain a tlrace of boric acid. The marked quantities found by other authors must, therefore, either be aktributed to the mesence 01 copper, or to the iutrodaction of boric acid i n fining tde wines. JN. W. Estimation of Boric acid. By HENRI JAY and DCPASQUIER {Compt. reitd., 1895, 121, 260--262).-1n order to estimate boric acid by converting it into methylic borate, the substance is very slightly acidified with hydrochloric or sulphnric acid, mixed with 25 to 30 C.C.of methylic alcohol, and placed in a flask, which is connected with another flask containing a normal alkali hydroxide solution free from carbonates, and also with a condenser, i n such a way that the methylic borate formed in the first flask passes through the alkali solntion in the second flask, whilst the methylic alcohol, which is again set frec, is condensed, and falls back into the first flask. After boiling for about an hour and a half, the alkali solution is heated to expel methylic aleohol, very slightly acidified with dilute hydrochloric acid, again heated to expel carbonic anhydride, and titrated with decinormal alkali until neutral to litmus paper. Two drops of an aqueous solution of the blue C.4.B. are added, and the addition of alkali is continued until the first colour-change begins.The second quantity of alkali, less 0.2 to 0.3 c.c., according to the volume of the liquid, gives the quantity of boric acid present. I n order to obtain accurate results, the temperature and volume of the liquid operated on must be constant, and methylic alcohol and carbonic anhydride must be expelled. Pure wines contain from 0.0105 to 0.022 gram of boric anhydride per Iitre; cider and perry 0.011 t o 0.017; urine, 0.008 to 0.017. The acid was not f o n d in the bone or flesh of an ox. The accuracy of the estimation is slightly affected by the presence of fluorides, bnt not by that of other salts. C. H. B. Estimation of Carbonic Oxide in Air. By JOHN S. HALDAXE ( J . Physiol., 1895, 18,463-469).-The method described depends on fhe fact that when a hmmoglobin solution is shaken with air con- taining carbonic oxide, the proportion of the pigment which finally combines with the gas varies with the percentage of carbonic oxide in the air.This proportion is determined by a colorimetric method, standard carmine solution being added to a standard solution of diluted blood till it becomes as pink as that which has been shaken with the air containing the poisonous gas. The details of the method, which yields delicate rather than accurate results, are described in full. W. D. H. Gold and Silver in Copper and Copper Matte. By ERNEST A. SMITH (Chem. News, 1895, 72, 76--77).-Cornmenting on the discrepancies that often occur in the results obtained by different assayers working on the same sample, it is suggested that neglect to take notice of the silver in the test-lead, use of one acid only, and that probably strong, for parting, and bad balances are the most frequent causes of the differences alluded to.D. A. L.ANALYTICAL CBEMISTRY. ‘77 Detection and Estimation of Mercury in Urine. B y ADOLI’ ,JOI,LES (Monafsh., 1895, 16, 684--692).--From 100 to 300 C.C. of urine, according to the amount of mercur.7 supposed to be present,, is treated with stannous chloride and free hydrochloric acid i n presence of about 2 grams of granular gold. The precipitated mercury amal- gamates with the gold, which is afterwards washed by decantation and treated with hot coiicentrated nitric acid, whereby the mercury is dissolved. The solution thus obtained is diluted with water, and, 011 the addition of stannous chloride, yields a precipitate even when only a minute trace of mercary is present.To determine the mercury quantitatively, the amalgam of gold obtained as above is washed with water, alcohol, and ether, weighed in a hard glass tube, and the mercury distilled off. The loss in xeight gives directly the amount of mercury present. G. T. M. Volumetric Estimation of Platinochlorides ; Estimation of Potassium, Ammonium, Nitrogen, and Platinum. By LUCIEX L. DE KONIKCK (Chem. Zeit., 1895, 19, 901-902).-The platino- chlorides of potassium or ammonium obtained during an analysis, rimy, instead of being weighed, be volumetrically treated by apply- ing t h e reduction principle of Correnwinder atid Contamina.The precipitate is dissolred in boiling water and heated for some time with calcium formate, which soon removes the platinurn, leaving potassium 01- ammonigm and calcium chloyides, also free hydrochloric acid in solution. The mixture is neutralised by means of calcium carbonate suspended in water, and, after filtering, the chlorine is estimated as usual by rneans of silver nitrate. Six atoms of chlorine represent 2 atoms of l~otassium, ammonium, or nitrogen, and 1 atom of platinum. The calcium carbonate is made from calcium nitlaate and sodium carbon- ate. L. ~ l i : K. Quantitative Separation of Benzene from Light Petroleum. By HOBERT HFNRIQUYS (Chem. Zeit., 1895,18, 958-959).-5-7 C.C. of the mixture is introduced into a 25-C.C. glass stoppered cylinder, divided to 0.2 c.c., and shaken with twice its volume of sulphuric acid containing 5 per cent.of added sulphurio anhydride, uiitil notliing more is dissolved. Ti Le benzenes soon become sulphonated and dissolve, whilst the light petrolenm is scarcely attacked, and floats on the surface of the acid ; its volume may then be read off. Analysis of the Cyanide Solutions used in the Extraction of Gold. By GEORGE A. GOYDER (Chew. News, 1895, 72, 80-82).--- S?e this vol., ii, 28). Estimation of Glycerol in Fermented Liquors. By J. LLABORDE ( J . Yhccrm., 1895, [6], 1, 568-57O).-The liquor (50 c.c.) is boiled neai-ly to dryness (1 c.c.) in a flask containing zt quantity of lead shot (100 grams). The glycerol, which is retained in combination with the acid of the wine, is liberated from the residue by the addi- tion of finely powdered, slaked lime, with which i t is incorporated L.DE I(.i 8 ABSTRACTS OF OHEMIUAL PAPERS. into a stiff paste by agitation with the shot. If sugar is present,, the lime is previously moistened with alcohol, and a larger quantity is used. The liberated glycerol is extracted from the paste by repeatedly shaking with alcohol-ether, and the extract having been a9idifieci with sulphuric acid (10-15 drops), the solvent is as far as possible distilled off, and the residual alcohol removed by repeatedly boiling with water; the acid aqueous solution thus obtained is evaporated to dryness and carbonised, and the amount of glycerol calculated from that of the carbon obtained. Trial analyses of wine and beer, to which various amounts of glycerol were added, are stated to have been satisfactory.The ratio of alcohol to glycerol in red wines appears to vary within the limits 1 : 10 and 1 : 16. JN. W. Estimation of Glycerol in Wine and Beer. By ALFRED PAI~THE~L ( A ~ c h . Pliarm., 1895, 233, 391-398).-The liquid under examination (50 c.c.) is mixed with calcium carbonate and concen- trated to 10-15 C.C. ; it is then filtered into a tubulated retort, a, of 100 C.C. capacity, fitted at the tubulus with a bored cork, which is closed with a glass rod; the retort is heated in an air bath, b, thc bottom being of sheeh iron, the sides and top of asbestos paper, fastened by means of water-glass. The bath is heated at 120°, and when the contents oE the retort have been distilled almost to dryness, the bath is cooled to 60’, the apparatus exhausted, and the distillatiou coii- tinued at 180’ (25-30 mm.) ; this requires 1.5 hours.The apparatus is allowed to cool, the pressure restored, water (10 c.c.) added to the retort, and the distillation repeated at 120’. All the glycerol will be found in the receiver, d ; should i t be coloured, it must be distilled Again under similar conditions. The glycerol is determined by Uaumert and Schaumann’a method (Abstr., 1892, 1529). The con- tents of the receiver and condenser, e, are washed into an ErlenmeyerANALYTICAL CHEJIISTRY. 79 flask, diluted t o about 200 c.c., sodium hydroxide (8-10 grams) added, and then potassium pernianganate solution ( 5 per cent.), until the liquid remains bluish-black ; it is next heated on the water bath for an hour, decolorised by means of sulphurous anhydride, acidified with glacial a c e h acid (20 c.c.), and evaporated until free from sulphurous anhydride.The residue is diluted to about 200 c.c., and calcium chloride added in excess, the mixed precipihate of calcium oxalate and calcium sulphate being collected on an asbestos filter, and washed until, the liquid ceases to affect potassium permanganate solution ; it is then dissolved in dilute sulphiiric acid, and the oxalic acid estimated in t'he usual manner by means of standard potassium permangnnate solution (about 5 : 1000). Under the above conditions, 1 mol. glycerol yields 1 mol. oxalic acid, and the analyses given of wine, beer, and aqueous glycerol solutions agree closely with thc theoretical.In tbc sketch, In is the manometer, and r a tube and valve to prevent the back flow of water from the pump. If the pressure in the apparatus is reduced without previously cooling, in the manner described by von Torring, loss of glycerol always occurs. Estimation of Pentoses and Pentosans in Diffusion Cuttings, Sugar Beet, and some Food Stuffs. By A. Smw (Chenz. Cenfr., 1895, i, 4 4 8 4 4 9 ; from 0sterr.-ung. Zeit. Zucker-Ind., 23,925-933). -The substance under examination (2-5-5.0 grams), together with hydrochloric acid, ~ p . gr. = 1.06 (100 c.c.), is placed in a fl:~& (300 C.C. capacity), which is fitked with a funnel and heated b y means of a bath of fusible metal; after distilling during lO-1h minutes, when 30 C.C.should have passed over, hydrochloric acid (30 c.c.) is added, the distillation continiied, and these operations repeated until the production of furfuraldehyde ceases. The distil- late is treated exactly in the manner described by Tollens. The whole analysis requires 5i-6 hours ; it presents no special difficul- ties, and afYord3 valuable information as to the relative worth of the most varied food stuffs. J. B. T. Relative Proportion of Glucose and Levulose in Sweet Wines. By F. JOSEF KONIG (C'henz. Zeit., 1895, 19, 999-lOOO).- The relative proportion of glucose and levulose in sweet wines may be calculated from the copper-reducing power? coupled with a polan- metric ohservation. The colouring and tannin matters are best removed by means of basic lead acetate, as the use of animal charcoal causes a slight loss of sugar.If there is a large excess of levulose present, it may be concluded that the wine has been prepared by fermenting a concentrated juice, and then suddenly stopping the fermentation by adding alcohol or by other means ; whilst, if the dextrose is largely i n excess, its wilful addition may be suspected. If both are pretlent in about equal amounts, with perhaps a slight excess of levulose, no particular con- clusion can be drawn; the relation of the dextrose to the levulose will not decide the question whether the grape-juice has, from the commencement, been fortified with cane or starch sugar. On the J. B. T.80 ABSTRACTS OF OEIlhlICAL PAPERS. whole tlie analysis of sweet wines is stiil in a somewhat nnsatisfac- Estimation of Glycogen in Liver and Muscle.By WLADIMIR A. I<ISTIAKOFFSKY (C/iem. Centr., 1895, i, 449 ; from Pharnz. Zd. EZLSS., 34, 25).-The use of 0.1-0.3 per cent. solutions of alkali instead of 2 per cent., for the extraction of glycogen, is recommended by Brucke ; but glycogen may also be obtained by extracting the finely divided substance 5-6 times with boiling water, and pressing the residue. In addition to the glycogen, the liquid contains alkali albuminates, glutin, and traces of peptone; these are removed by precipitation with potassio- mercuric iodide after the liquid has been concentrated and acidified with hydrochloric acid ; the glycogen is precipitated from the filtrate by means of alcohol (compare Huizinga, this vol., i, 6). Estimation of Glycogen.Detection of Albumoses in Presence of Glycogen. By D. HUIZINGA (PJliiger’s Archiv, 1895, 61, 32--38).--See this vol., i, 6 . Estimation of Formic acid. By FRANZ FREYER (Chew. Zed.. 1595, 19, 1184--1185).-The author’s process has been invented for the purpose of estimating calcium formate in the presence of acetate. The mixture having been distilled with dilute sulphuric acid i n a current of steam until the distillate is no longer acid, the total acidity is estimated in an aliquot part of the distillate by means of standard soda, and another portion of the liquid is neutralised with aqueous hoda and concentratled to a small bulk. I t is then boiled with a mix- tnre of dilute salphuric acid and potaesium dichromate, which is withoufi any action on the acetic acid, but rapidly and completely oxidises the formic acid to carbonic anhydride and water.The strength of the dichromate must, of course, be accurately known, and t h e excess is afterwards determined as usual by means of potassium iodide and sodium thiosulphate. To better see the change in colour, The Resorcinol Test for Tartaric acid. By GEORGES DENIG~S ( J . Yharnz., 1895, [6], 1, 556-589 ; compare Mohler, Abstr., 1891, 867).-Mohler’s test, which consists in the addition of tlie substance to be tested to a dilute solution of resorcinol in concentrated sulphuric acid at 120°, with the consequent production, if a tartrate is present, of a violet coloration, is open to the objections t h a t the reagent is perish- able, that the test is applicable only t o solid substnrices, thus necessi- tating the evaporation of so!utions to dryness, and that the violet coloration, or one closely resembling it, is produced by various oxidising agents.The author has modified the process in a manner to obviate these defects. A solution of pure, white reeorcinol (2 grams) in very dilute sul- phuric acid (-$ C.C. to 100 C.C. water) is quite stable, and, by adding a portion of this to about 20 vols. of concentrated acid a t the time of testing (2 or 3 drops to 2 c.c.), a solution is obtained equivalent to Mohler’s reagent. A small quantity of the liquid to be tested (1 or 2 drops) is added to the reagent, and the mixture gradually tory condition. L. DE K. J. B. T. a little metaphosphoric acid may be added. L. DE K.ANALYTICAL OEEMISTRP.81 warmed to 115"-140°, when, if tartaric acid is present, the charac- teristic piolet-red colour is developed. It is due to t,he preRence of a broad absorption band in the spectrum, extending from X = 510 to x = 545, and best seen with solutions diluted with conceut'rated snlphcric or acetic acids. If the coloration appears before heating, i t is due to the presence of an oxidising agent in the liquid, and this must be previously removed by reduction with copper-zinc and dilute sulphuric acid ( 5 C.C. of solution to be tested, 5-6 drops copper sulyhate solution, 2 grams of zinc, and 1 C.C. of acid), Tartaric acid can in this way be detected in a few drops of a 1 per cent. solution containing, in addition, 10 per cent. each of sodium nitrite? and potassium nitrate, chromate, and chlorate.Substances suvli as cane sugar, which are blackened by sulphuric acid, are best eliminated by Mohler's method of precipitating the acid ae lead salt, and washing the latter with dilute nitric acid before decomposing it and making the test. By VIKTOR VEDRODI (Chem. Zeit., 1895, 19,600-601).-The autjhor calls attention to the enormous difference in composition of various commercial brands of fish oil, particularly as regards their acidity and percentage of unsaponifiable matter. The acidity was fcund to yary from 0.74 to 38.9 per cent., and the nn- saponifiable matter from 0.6 to 82 per cent. No sample should be used for tanning purposes unless tho acidity and the unsaponifiable matter do not exceed 15 and 4 per cent.Examination of Fats by Means of the Refractometer. By HhINKlCH B~crtUrtI's arid H. HEILER (Arch. Yharm., 1895, 233, 423--428).--The objects of this invest.igat>ion were the determin- ation (1) of the influence of temperature on the refractive power of various fats and oils, and (2) the relationship, if any, between the refractive power, the percentage of volatile acids, and tho iodine additive capacity : the experiments were conducted with a Zeiss' refractometer. The increase in the refractive power in scale divi- bions, for a rise of lo, is as follows :-Butter = 0.54-0.58 ; olive oil = 0% ; cotton seed oil = 0.5-0.6 ; ground nut oil = 0-6-0.7 ; oil o€ apricots = 0.4-0.6 ; sun-flower seed oil = 0.5-0 6 ; sesame oil = 0.6-0.7 ; almond oil = 0.5-0.6 ; oil of peaches = 0%.No relation- ship was observed between the refractive power of butter and of deer fat and their iodine absorption value, hut high i*efractive power is accompanied by a relatively great additive power for iodine in the case of tho above oils, and also in the following : poppy oil, cocoa nut oil, palm oil, butter (?), tallow, lard, margarine. Nothing is said directly of the volatile fatty acids. By HEINRICH BECKURTS and F. OELZE (Arch. Pharm., 1895, 233,429-430).-Deer fat melts a t 49-49.5" and solidifies at 48", the corresponding values for tallow and beef fat are 44-45.5'and 32-36', and 43-44.5" and 37" respeztively. The fatty acids of JN. W. Analysis of Fish Oils. re spec tively . L. DE K. J. B. T. Deer Fat. VOL. I Y Y . ii. 782 ABSTRACTS OF CHEAIIOAL PAPERS.deer fat melt a t 49*5O, of tallow at 45--47O of beef f;tt a t 44*546° ; the iodine absoi*ption values are 19.8-21, 32.7-46-2, and 85.4-44 respectively. The refractive powers (Zeise' refractometer) at 40" are 44 5, 46, and 45. J. B. T. Lard Analysis. By A. GOSEE (Chem. Zeit., 1895,19,1043-1045). -The author recommends crystallising the sample from ether a t temperature of 12-13'. With a little practice, it will be found com- paratively easy to distinguish microscopically between tallow-stearin and any lard-stearin which, however, does not readily crystallise. When applying the silver test, the author uses the original Bechi test as approved by the Italian C'omiuittee. The test gains in delicacy if the lard is first pressed a t a temperature of 26-30', aud the reaction applied to the expressed oil.L. DE K. Examination of Pepper. By WALTER BUSSE (Zeit. anat. Chem., 1895, 34, 638-643 ; from Arbeiten I<. Geszcndlzeitsanzte, 9, 509).- The methods of estimating moisture, ash, sand, and total alcoholic extract are described. The estimation of the colouring mat (em, which occur only in the huak, i R considered important; these are extracted from the residue insoluble in alcohol by digesting it with hot sodium hydroxide solution, and are precipitated by lead acetate from the extract a,fter acidifying with acetic acid. The amount of lead so precipitated is called the "lead number," and seldom exceeds 0.122 gram per gram of pepper. M. J. S. Nutmeg. By WALTER BUSSF; (Zed. anal. Chem., 1895, 34, 643- 644 ; froin Arbeiten K.Gesuirdhei/samte, 11, 390j.--The total ash should not exceed 5 per cent., nor the amoiint insoluble in hydro- chloric acid (sa.nd) 0.5 per cent. For estimating the fat, 2 grams of the grated powder is extracted with ether for eight hours, then dried, rubbed down with qumtx sand, and again extracted for four hours; after evaporating the ether, the f a t is absorbed by 8 grams of ignited sand, arid dried a t 100" for five liours only, to avoid oxidation. The amount varies from 30 to 40 per cent. M. J. S. Margarine Cheese and its Analysis. By M. KUHX (Erper. Stat. Becord, 1895, 7, 158 ; from Molk. Zeit., 1895,9, 185--187).--The percentage ccjmposition of the cheese is said to be valuelew in dis- tinguishing it from natural cheese. The following determinatims are recommended to be made in the ether extract: sp.gr. at 100" by means of a Westphal balance and Konig's butter areometer; in- soluble fatky acids (Hehner's method) ; rolatile fatty acids [Reichert-Meissl-Wollnyj Kottesdorfer saponification equivalent,; angle of refraction in the Zeisa-Wollny butter refractometer. A number of results obtained with difterent cheeses are given in the original paper. Composition of Meat Extract. By F. JOSEF KijNIG and A. BOMER ( Z e i t . a n d Chern., 1895, 34, 548-562).-1t has hitherto N. H. J. M.ANALYTICAL CHEMISTRY. 83 been accepted that mcat extmcts contain gelatin and other protei’ds. Thus, C. Karmrodt found 10.4 per cent. of gelatin in Liebig’s extract, Kemmerich (Sbstr., 1894, ii, 150) 33.23 per cent. of proteids, and Stutzer (Abstr., 1853,146) 20.5 to 22.6 per cent.of peptone. The num- bers obtained by Kemmerich were deduced from the results of fractional precipitation by alcohol, on the assumption that. gelatin is precipitated by 60 per cent. alcohol, albumoses by 80 per cent., and peptones only by 90 per cent. The authors have repeated Kernruerich’s work, but employing Kjeldahl’s process, instead of weighing the precipitates, as was done by Keinmerich (a metliotl to wliicli they take exc:eption), have obtained numbers for gela t’in and rllbu- moses which are less than one-third of Kernmerich’s. 111 Liebig’s extract, the mode of manufacture would seem to exclude the pos- sibility of more than traces of gelatin being present. The usual mode of estimating peptone, namely, precipitation by phospho- molybdic acid, is erroneous, since this reagent gradually (in the course of 5-7 days) throws down the flesh bases also, about 90 per cent.of the total nitrogeu being precipitable. The absence of peptone may be shown qualitatively, as after precipitating with ammo~iinm sulphate to ensure the absence of albumoses, the tilt~ate does not give the biuret, reaction. In thecases of Kemmerich’s meat peptone, and Cibils’ meat extract obtained by means of the digestive ferment of Ca~ica Papaya, the filtrate is pale enough f o r the detec- tion of traces. The filtrates from Liebig’s and Kemmerich’s meat extracts are darker, but the colour is incompetent to mask the reac- tion if as much as 2-3 per cent. of peptone is present (see, how- ever, Stutzsr, loc.c i t . ) . Stutzer’s statement as to the presence of ammonia is confirmed, but no evidence could be obtained of the presence of amido- or acid amido-compounds. A1 bumose (prccipit- able by ammonium sulphate or zinc sulphate; see next abstract) is the oiily proteiid present in notable quantity, but the flesh bases constitute by far the largest portion of the total nitrogenous con- stituen ts. M. J. S. Precipitation of Albumoses by Zinc Sulphate. By A. ROMER (Zeit. anal. Chem., 1895, 34, 362--567).-Zinc sulphate possesses many advantages over ammonium sulphate for the separation of albuniose from peptone. The precilitation is equally complete, comparative estimations by means of the two reagents haviiig given identical numbers in the four meat extracts examined by the autlior and J.Koniq (preceding abstract), whilst the tiltrates in all cases gave no indication of the presence of a prote‘id by the biuret reaction. The presence of zinc gulphate in no way disturbs tlre Kjeldahl process, so that the nitrogem in the precipitate can be estimated without the need of applying any coi~ection. The filtrate is ac once suitable for precipitation by phosphomolybdic acid, eit’her for the purpose of testing the completeness of the washing or (nf’ter strongly acidifying with an equal volume of dilute sulphuric acid, 1 : 4) for t’he estimation of peptone, flesh bases, $c. Since phos- phates give a precipitate with zinc sulphate, it is desirable to slightly acidify the solution Kith sulphuric acid; 1 C.C. of acid (1 : 4) is84 ABSTRAOTS OF CHEMICAL PAPERS.therefwe to be added to 50 C.C. of solution containing 1-2 grams of the extract, and previously freed from insoliible a d coagulable substances; the liquid is tlien saturated in the cold with ;I small excess of finely powdered zinc salphatc, and the precipitate washed with a cold saturated solution of the same salt. Although meat extracts contain ammonia, and ammonia forms with zinc sulpha'te a sparingly soluble double salt, in no case was ammonia found in the albumose precipitate, but, on the contrary, the whole of the ammonia of the origiiial substance was obtained when the filtrate from the zinc precipitate was distilled with sufficient magnesia to precipitate all the zinc present and leave the liquid strongly alkaline. Estimation of Gelatin in Meat Extracts and Commercial Peptones. By ALBERT S w t z E R (Zeit. anal. (Jltem., 1895, 34, 568- 370).--The following is the exact method of carrying out the estima- tion ot wbich the outline was previously given (Abstr., 1895, ii, 543). Sand, previously ignited and freed from fine dust by a sieve, is used instsend of asbestos for absorbing the solution in the tinfoil capsule. After complete drJ-ing at looo, the contents of the capsule are pow- dered, and, with the cut up capsule, placed in a beaker, where they are extracted four times with absolute alcohol, filtering the extracts through an asbestos filter, but leaving as far as possible the insoluble mather in the beaker. The beaker (marked 4) together. with four others (t, c, (a, e) are then plunged into crushed ice, as also a flask contailling ;t niixturc of 100 C.C. of alcohol, 300 grams of ice, arid cold water to 1 kilogram. Of this mixture, the temperature of which must not exceed 5O, about 100 C.C. is poured on the sand, stirred therewith for two minutes, aid decanted into beaker b ; the second ducantate is poured into c, and so on, until the last washing is colourless, a frag- ment of ice being added to each as soon as it is poured off. Three funnels are then arranged with filter beds of long-fibred asbestos sup- ported by perforated porcelain plates, and connected with a pump by which getitlc and gradually increasing suction can be applied. The contents of beaker a are tiltered through the first., finally transferring ttie sand to the fuiinel, b is poured into the second, and c and d iIito the third. The three filters, as well as that through which the absolute alcohol extract had been filtered, are then thoroughly washed with the ice cold dilute alcohol, transferred to a basin, and repeatedly extracted by boiling with water. The aqueous extract, alter fiitration, is concentrated and submitted to Kjeldahl's process for the estimation of the gelatin. ln the process given in Abstr., 1893. 146, the albumose (line 2 from bottom of page) must be cor- rected for any coagulable albumin present in the meat extract. M. J. S. M. J. S.