ANALYTICAL CHEMISTRY. 107 A n a l y t i c a l Chemistry. Estimation of Hydrogen Chloride in Solutions of Hydroxyl- amine Hydrochloride. By J. A. MULLEK. (Rzill. Xoc. Chim. [ 3 ] , 3, 605).-Phenolphthalei'n is unaffected by solutions of hydroxylamine hydrochloride, and the amount of acid present may be estimated by means of a standard solution of sodium hydroxide, free from carbon- ate, using phenolphthalein as the indicator. Pyridine, picolines, and lutidines behave similarly. T. G. N. Estimation of Sulphur in Inorganic Sulphides. By L. BLUM (Zeit. anal. Gem., 29, 411412).--The method publiahed by Jannasch (Abstr., 1889, 1244, and 1890, 1187) is not new, having been already brought forward by Sauer, in 1873 (Abatr., 1873, 939). M. J. S. Estimation of Nitrogen by the Schultze-Tiemmaam (Schloesing's) Method.By F. COGHIUS and T. MOELLER (Chern. Zeit., 14, 3%).-Low results are obtained by this method, especially in the examination of explosives. This is attributed by the authors to the addition of too much water, and to the want of proper relation- ship between the size of the apparatus used and the quantity of mate- rial employed. In some test experiments they used a long-necked 350 C.C. flask, a measuring tube of 150 C.C. capacity, ferrous chloride solution containing 70 grams in 100 grams of water, hydrochloric acid of about 37 per cent., and employed 0-3-0.4 gram of potassium nitrate, 5-15 C.C. of the ferrous chloride solution, and twice the quantity of the hydrochloric acid. The ordinary course of operation was followed, avoiding unuecessary boiling to drive out the air.When 25 to 50 C.C. of water was added, the analysis lasted 30 to 40 minutes, and the results varied between 13.76 and 13.86, mean 13.81, whilst with 80 to 150 C.C. of water the variation in the results was from 13.05 to 1339, mean 13.21, and the analysis lasted 70 to 90 minutes. D. A. L. Estimation of Nitrogen in Sodium Nitrate. By 0. FOERSTER (Chem. Zeit., 14, 509-510 ; compare Ahstr., 1889, 547, 746).-Two or three grams of the nitrate is dried at 150" or by heating to inci- pient fusion, weighed, and repeatedly evaporated to dryness on a water- bath in a tared crucible, with 25 C.C. of about 19 per cent. hydro- chloric acid. After about the third evaporation, the nitrate is com- pletely converted into chloride, which is dried at 150°, ignited slightly, and weighed, and the nitrogen calculated from the difference. The method yields satisfactory results, but only in the absence of other substances, which would be attacked by hydrochloric acid.D. A. L. Estimation of Nitric Nitrogen as Nitric Oxide. By F. SCHEIDING (Chem Zeit., 14, 635-637).-For estimating nitric nitrogen as nitric oxide, the author has devised and employs the apparatus shown in the drawing, which is provided with a measuring tube108 ABSTRAOTS OF CHEMIOAL PAPERS. having a globular expansion, a glass tap H with a small funnel at the top, and a special arrangement intimately attached by india-rubber tubing or fusion to the bottom. In operation, tube L is connected in a suithble manner with a, movable reservoir containing sodium hydroxide, sp.gr. 1.25, with which the apparatus is charged to the level of 4 by raising the reservoir, and clip 4 is closed. The substance is placed, along with a little water, in a 200-250 C.C. flask, to which the stopper and tubes are fitted, connections made, and to expel the air through tubes 2 and 3, the water in the flask is boiled until the water into which tube 3 dips is caused to boil by the issuing steam, clip 5 is then closed, and the air still in tube 1 driven into the mea- suring tube by opening clip 4, which is again closed, and the flame removed from below the flask. The measuring tube is filled to the top with sodium hydroxide, and tap H is closed. 20-25 C.C. of cold saturated ferrous chloride, and then 8-10 C.C. of concentrated hydro- chloric acid are carefully drawn into the flask through tubes 2 and 3, which are then washed with water in the same way; the flask, suspended a few cm.above the wire gauze, is now heated, and asANALYTIGA L OHEMISTRY. 109 soon as a, pressure i R indicated in the india-rubber tube at 4, that clip is opened, t,be nitric oxide passes into the measuring tube, and by the time the liquid has yolatilised in the flask, all the nitric oxide is con- chided to be in the measuring tube. The temperature in the jacket tube surrounding the measuring tube is made to correspond with that i n the vicinity of the bulb, and the level in the reservoir being adjusted to that in the measuring tube, the volume is read off, and after the necessary corrections are made, the percentage of nitrogen is. calculated therefrom.The saucer under the measuring apparatus is filled with water to keep the tubes immersed in it cool. For substances which might be deromposed by boiling with water, a tap funnel is fitted to the flask, and is used for charging it. D. A. L. Estimation of Nitrogen in Organic Substances by means of Alkaline Permangmate. By R. L. WAGNER (Chem. Zeit., 14. 269). -The autbor some years ago recognised the possibility of oxidising nitrogenous organic substances by means of alkaline permanganate, without the formation of ammonia. In his experimcnts he mixed 0.5 to 1 gram of substance with 25-30 times its weight of potassium permanganate, and 5 C.C. of 25 per cent. potassium hydroxide, placed the mixture in a tube closed at one end, terminating at the other in a capillary for the escape of oxygen, warmed in a water-bath to aid admixture, and theu heated at 150-170" in an air-bath for two to two and a half hours. The conQents of tbe tube were turned into a porcelain basin, the excess of manganate reduced with manganese sulphate and sodium carbonate, and the nitric acid deteimined in the clear liquid by a modification of Eder's method ; but irregularity of combustion and breaking of tubes rendered the method practically useless, except, perhaps, for substances soluble in alkalis.Non-vola- tile nitro-derivatives and ethereal nitrates can be safely oxidised by alkaline permanganate in a porcelain dish, excess of permanganate being subsequently reduced with alcohol, and the diluted filtrate treated with ferrous sulphate, zinc-dust, and hydrochloric acid ; the nitrogen is then estimated as ammonia, by any of the usual distilla- tion methods.Carbon bisulphide and thiophen can be oxidised by similar treatment, and the sulphur estimated in them; they are enclosed in thin glass bulbs, and placed in tubes containing the alkaline permanganate ; the tube is sealed up, the bulb broken, and tbe digestion proceeded with. D. A. L. Detection of Foreign Raw Phosphates in Powdered Basic Slag. By L. BLUM (Zeit. anal. Chem., 29, 408--411).-The relative superiority of basic slag as a fertiliser over natural phosphatic minerals, owing to its ready absorbability, and the high price which i t has in consequence atkained, have led to its falsification with other raw phosphates.Only such are likely to be used as, from their low percentage of phosphoric acid, cannot profitably be worked a p as superphosphate, and these in most cases contain much calcium carb- onate. Fresh basic slag is almost absolutely free from carbonates, and even ou long exposure to air, absorbs very little carbonic acid110 ABSTRAOTS OF OHEMIOAL PAPERS. (2.47 per cent. was found in an extreme case), so that the presence of much carbonate i n a specimen would be enough to throw suspicion on it. A low percentage of iron and manganese might furnish an additional indication, since tlhese metals are rarely present in natural phosphates. In estimating the carbonic acid by decomposition with an acid, some chromic acid should be added, to prevent evolution of hydrogen aulphide from the sulphides present, but a simple estima- tion of the loss on ignition would generally allow an opinion to be formed.M. J. S. Estimation of Water in Superphosphates. By J. STOKLASA (Zeit. anal. Chem., 29, 390--397).-Pure monocalcium tetrahydrogen phos- phate, CaH4(P04)2 + HzO, loses its water of crystallisation at loo", but only completely after 40 hours. It may be kept at 105" for 20 hours with but little change, but on longer heating at the same temperature begins to show decomposition. At higher temperatures, the amount of change is dependent not alone on the temperature, but also on the time of drying. The statement of Drewsen (Abstr., 1881, 465) that drying even at 300" does not diminish the proportion of soluble phosphate, but merely reduces it to a soluble pyroyhosphate, cannot be confirmed for pure or nearly pure monocalcium phosphate.I t might be true for a superphosphstc in which free phosphoric acid constituted 80 per cent. of the total soluble phosphoric acid. On drying for one hour at ZOO", one-half of the monocalcium phosphate undergoes decomposition, thus :-4Cs€&( PO,), = Ca?,P207 + Ca(P03), + CaH2P207 + 2H3POa + 4H20. At lower tempera- tures for the same length of time, the proportion decomposed is smaller, but if the time is prolonged, a further decomposit'ion takes place ecen at 150", and less free phosphoric acid is fouud in the soluble part. A temperature of 200" sufficiently prolonged results in the following decomposition : 4 C a H 4 ( PO& = 3Ca(P03)2 + CaH2P207 + 7H20, whilst at 210", there remains nothing but insoluble, glassy calcium metaphosphate.I n presence of free phosphoric: acid, the contrary action may on heating take place, thus: ChP,O, + 2H,PO, = 2CaH2P207 + H,O, and thus the soluble phosphoric acid actually undergo increase. M. J. S. By R. FRESENIUS (Zeit. anid. Chem., 2 9 , 4 1 3 4 3 0 ; see Abstr., l890,924).-A11 attempts to Qbtain complete separation by niems of chvomic acid in a single pre- cipitation resulted in failures. The seemingly satisfactory separation obtained by Frerizhs and by Russmann (next abstract) resuited from the accidental compensation of opposite errors, since they washed the barium chromate wibh acetic acid, in which it is distinctly soluble, and weighed it after drying a t 110", at, which temperature it still retains some moisture.In a solution containing alkaline acetate and dichromate, barium chromate is, however, quite insoluble. It can also be rendered anhydrous without decomposition by ignition at a dull red heat, even the portion adhering to the filter reoxidising after temporary reduction. By double precipitation of the barium, a com- plete separation can be effected even when the proportion of Separation of Barium from Strontium.ANALYTICAL CHEMISTRY. 111 strontium is large. The solution of the chlorides is feebly acidified with acetic acid, and diluted until it contains not more than 0.5 per cent. of the bases, then precipitated hot with a n excess of ammonium chromate, which has been carefully neutrahed with ammonia, After cooling €or an hour, the precipitate is washed by decantation with very di1ut.e ammonium chromate until the washings no longer give a precipitate with ammonium carbonate, and then further with warm water until the washings are scarcely coloured by silver nitrate.The precipitate is then dissolved in the smallest possible quantity of nitric acid, and the solution again diluted and heated. Ammonium metate is added in sufficient quantit-v to displace the free nitric acid by acetic acid, and then ammonium chromate until the odour of acetic acid bas completely disappeared. After an hour, the liquid is poured through a filter, the precipitate is digested with hot water, cooled, filtered, and washed thoroughly with cold water. It is then free from strontium, whilst the filtrates contain no barium.Double precipita- tion from neutral or alkaline solutions has not been successful. M. J. S. Separation of Barium, Strontium, and Calcium. By A. RUSSMANN (Zeit. anal. Chent., 29, 447-454; from Tnaug. Diss. Berlin, 1887).-Barium cannot be satisfactorily estimated by Frericbs’ method (precipitation from an acetic acid solution by normal potassium chromate), since the filtrate always contains traces of barium, and some potassium chromate is carried down by the preci- pitate. The precipitate will also contain strontium, if the proportion of the strontium in the solution exceeds 30 parts per 100 of barium. Calcium is not so precipitated. The simplest way to ascertain the weight of the barium chromate, is to dissolve it in dilute hydrochloric acid, add potassium iodide, and immediately tit rate with thiosulphate.Diehl’s method for separating barium and calcium by digesting the sulphates with sodium thiosulphate solution is complicated by so many sources of error that it cannot be recommended. Fresenius’ method of separating barium and calcium by dilute sulphuric acid in a solution acidified with hydrochloric acid is thoroughly satisfactory. The method of Sidersky (Abstr., 1883, 509) for separating strontium and calcium only yields approximate results. For separating barium and calcium, it is, however, serviceable. Bloxam’s method (Abstr., 1886, 920) is not suitable for quantitative separations, as the strontium sulphate carries down with it considerable quantities of ccLlcium, and the calcium ammonium arsenate cannot readily be brought into a, form for weighing in which i t contains a constant proportion of calcium.Fleischer’s method for separating barium and calcium by digestion with 3 parts of potassium sulphate and 1 part of carbonate, followed by titration of the calcium carbonate in the weighed precipitate gives good results. Lastly, Leison’s met.bod for the estimation of the individual alkaline earths, by precipitation with oxalic acid and alcohol, and titration of the oxalic acid in the precipitate by permanganate, is accurate. The barium oxalate must be dissolved by hydrochloric acid, as it is not completely decomposed by. sulphuric acid. Strontium and calcium oxalates can be decomposed by sulphuric acid. The solutions must not be filtered throngh paper,112 ABSTRACTS OF CHEMIOAL PAPERS.and must be highly dilute. Ignition of the oxalates is, however, as a rule, the quicker process. Estimation of Cadmium in the Products of Zinc Manu- facture and in Calamine. By W. MINOR (Chem. Zeit., 14, 4, 34, and 348-349) .-The material is dissolved in hydrochloric acid, treated with hydrogen sulphide, and the precipitate washed with hot, water, dissolved in hydrochloric acid, heated to boiling, and poured into dilute sodium hydroxide likewise heated to boiling. This preci- pitate, after washing with hot water, is ignited in a current of oxygen, and weighed as cadmium oxide. Material containing but little iron, such as “ pure cadmium,” is dissolved in hydrochloric acid, and preci- pitated directly with the sodium hydroxide.This method of precipitation may also be used to separate zinc and cadmium i n t,he ordinary method of examining calamine ; the ammoniacal solution containing the zinc and cadmium is rendered slightly acid and poured hot into the hot hydroxide, &c. In the method described in the last of the three papers, the material is dissolved in hydrochloric acid, filtered from undissolved lead, precipitated with hydrogen sulphide, the precipitate, containing zinc and an inconsiderable amount of arsenic, is washed, dried, weighed, dissolved in dilute hydrochloric acid, and treated with sodium hydroxide in excess. The cadmium hydroxide is filtered off, and the zinc titrated in the filtrate with sodium sulphide, calculated . to zinc sulphide, and deducted from the weight of the cadmium sulphide precipitate.I n another method (requiring the absence of other metah precipitated by sodium hydroxide) after removal of iron with ammonia, the solution of zinc and cadmium is nearly neutralised with hydrochloric acid, and then treated with sodium hydroxide. The precipitate of cadmium hydroxide i R dissolved in dilute hydrochloric acid, evaporated to dryness, dissolved in water, and titrated with standard sodium hydroxide, using litmus or sodium sulphide papers as indicators. Good results have been obtained by both methods, the first being the more suitable in the presence of much zinc and vice versc2. D. A. L. M. J. S. Estimation of Cadmium as Sulphide by Precipitation with Sodium Sulphide Solution. By W. MINOR (Chem. Zeit., 14, 439-440) .-The material is dissolved in hydrochloric or nitric acid, and the lead separated by sulphuric acid ; the solution is then treated with soda, and the precipitate digested with ammonia.The ammoniacal solution is free from lead, zinc, and iron, but contains all the cadmium, which can then be determined by means of sodium sulphide solution, either volumetrically by tit ration, nsing ferric hydroxide as indicator, or gravimetrically by precipitating, and weighing the precipitate after drying for some hours at 140-145O. D. A. L. Volumetric Estimation of Zinc and Copper. By E. DONATR and G. HATTENSAUR (Chenz. Zeit., 14, 323--325).--Various experi- ments have been made by the authors. They find that for titrating zinc by Schaffner’s method, it is better to use sodium hydrosulphideANALYTICAL CHEMISTRY.113 (prepared by adding a known volume of dilute sodium hydroxide to an equal volume of the same solution previously saturated with hydrogen sulphide) than a solution of the crystalline sulphide of commerce ; howerer, in using this reagent in solutions containing tartaric acid and ammonia, the iron commences to precipitate before all the zinc is converted into sulphide. The estimation of zinc by using excess of ferrocyanide, after the removal of the iron, and titrating back with permanganate does not answer, since in the cold a clear solution cannot be obtained, whilst if warm, decompositions occur which cause irregularities. It is noticed that ferrocyanide precipitates zinc bu h not iron in the presence of tartaric acid and ammonia, and that the excess of either of these substances does not seriously disturb the relative quantity of 1 mol.ferrocyanide to 2 atoms of zinc ; therefore 1 C.C. of a solution containing 33.5 grams of potassium ferrocyanide per litre corresponds with 0.010 gram of zinc. As small an excess of ammonia as possible, and a hot solution, are favourable to the preci- pitation. The zinc precipitate is not decomposed by acetic acid; therefore, by placing drops of this acid and the solution under exami- nation in contact, in the presence of iron, a coloration indicates complete precipitation of' the zinc. The following method it; based on these considerations :--3 t o 4 grams of material is dissolved in hydro- chloric acid with some nitric acid, diluted to a definite volume with water, an aliquot part filtered, treated with 20-25 C.C.of concen- trated tartaric acid solution, a slight excess of ammonia added, and the liquid warmed to about 80". The ferrocyanide is now run in until the precipitation of the zinc is complete, as indicated in the manner described above. The proportion of iron to zinc ili the solution under examination should be the same as that present in the solution used for standardising the ferrocyanide. Under similar circumstances, copper is precipitated in a like manner, but the precipitation is greatly influenced by ammonia ; fherefore the solutio~r for titration should be neutral or nearly so. The ferrocyanide is standardised from a solution of copper of known strength, and cannot, be approximated to by the weight of fcrrocyauide employed, inasmuch as the composition of the copper precipitate is uncertain.Copper and zinc may be estimated in the same solution by this method ; first both are titrated, then the copper is precipitated out of another portion of solution, and the zinc alone titrated, &c. D. A. L. Estimation of Lead by Phosphomolybdic Acid. By H. BEUF (Bull.. Soc. Chinz. [3], 3, 852--855).-To the boiling neutral solution of the metal, an aqueous solution of phosphomolybdic acid is added until the supernatant liquid is coloured yellow by the excess of reagent used. After washing, the precipitate is dried at 90-100° and weighed. It forms a dense, white powder which is insoluble in water (1 in 500,000) and aqueous ammonia, but dissolves in nitric and in acetic acids ; it contaiiis 54.8 per cent.of lead, and corresponds with the formula Mo,Pb2,P2H,,0,,,; at a high temperature it loses 7 mols. H20. By decomposition of the precipitate with dilute sulphuric acid and einc at a gentle heat, a brown liquid is obtained, wbich may be YOL, LX. i114 ABSTRACTS OF CHEMIUAL PAPERS. titrated for lead by a solution of permanganate which has been previously standardised against a solution resulting from tho similar treatment of a known weight of a lead salt. The phosphomolybdic acid is made by evaporating to dryness a solution of ammonium phosphomoly bdate in nitric acid. Iron is eliminated by a previous treatment with sodium hydroxide, copper, potassium, and ammonium by washing the mixed pbosphomolybdates with ammonia-water, but the presence of zinc or arsenic vitiates the estimation.T. G. N. Separation of Copper from Arsenic by the Electric Current. By L. W. MCCAY (C‘laem. Zcib., 14, 509).-Under the influence of the current from four to six Meidinger elements, alkaline arsenates remain in solution, whereas copper is completely and quantitatively precipitated, and has been estimated wikh good results. Moreover, t h e copper is quite free from arsenic, and the solution may be safely employed for the determination of the original amount of the latter metal. D. A. L. Estimation of Aluminium in Commercial Aluminium. By 0. KLEMP (Zeit. anal. Chsm., 29, 388--390).-The prccess employed for zinc (Abst , 2890, 1190) cannot be applied to aluminium since, always evolved, but bv dissolving the aluminium in potash, and burn- ing the hydrogen in Fi=eserlins’ apparatus, a very accurate estimation cam be made.About 1 gram of the metal in filings is placed in a 150 c c. flask with a little vaselin to prevent frothing, and the potash solution (35 grams of KOH in ‘LOO c.c.) is added gradually, with warming towards the close. Estimation of Alumina in Bread, &c.; Solubility of Alu- minium Phosphate in Acetic Acid. By W. C. YOUNG (dnalyst, 15, 61-63 ; 83-S4).-1n Duprk’s process, the aluminium is precipi- tated as phosphate from an acid solution containing ammonium chloride and acetate, and is collected after remaining all night in the cold. Test analyses, with weighed quantities of alum, show that under these conditions the resalts are much below the truth.The best result (from a mixture of alum, sodium phosphate, and acetic acid) was obtained by boiling the mixture both before and after the addition of ammonium acetate, and filtering immediately. The amount of ammonium acetate must not be too small, nor that of acetic acid too large. For 0.1 gram of potash alum, there mas used 1 gram of ammonium acetate and 5 C.C. of ordinary acetic acid. The presence of ammoriiuni chloride has little effect when the liquid is filtered immediately after boiling, but lowers the result i E the pre- cipitation is performed iu the cold, or the mixture is allowed to cool before filtering. M. J. S. Estimation of Iron Oxide and Alumina in Phosphates. By R. JONES ( C h e w Zeit., 14, 269--271).-The author criticises the methods employed until recently, and recommends as the best the following Combination of Glaser’s method, slightly modified, with even with big ;i; ly dilute alkali and a large excess of iodate, hydrogen is The operation takes about 45 minutes.M. J. S.ANALYTICAL CHEMISTRY. 115 part of Stutzer's method. The phosphate is dissolved in hydrochloric and nitric acids, made up to a definite volume, an nliquot part taken and treated with ft quarter of its volume of sulpluric acid of sp. gr. 1.84, and its om11 volume of 95 per cent. alcohol, making up to definite volume with the alcohol ; after 12 honrs, the gypsum is collected, and when weighed, gives good results for calcium i n absence of much magnesium. The alcohol is driven off from the filtrate, which is rendered alkaline with ammonia and boiled to completely drive off the ammonia.The precipitate may be weighed, and half the weight taken as iron oxide and alumina, which gives good results, or it may be treated with molybdic solution, the phosphoric acid 'separated in the usual way, the iron and alumina precipitated wlth dilute ammonia, redissolved in hydrochloric acid, reprecipitated, &c., and weighed. Titration of Chromates, Barium Salts, and Sulphates. By P. SOLTPIEN (Clteni. Centr., 1890, ii, 217-218 ; from Pharm. Zeit., 35, 372) .-The titration of solutions of barium salts with potassium &chromate, and inversely the titration of chromates with barium salts, may be readily performed with either hzematoxylin or logwood extract as indicator. A solution of barium chloride is prepared equal to one of potassium dichromate, and for the determination of barium salts, potassium dichromate is run in from the burette until a drop placed on a warmed porcelain plate with a drop of haematoxylin just shows the formation of a blue-black coloration. The solution to be titrated must be neutral, and may not contain more than the merest trace of either acetic acid or ammonia.Chlorides and nitrates do not in- terfere with the reaction, nor does rosolic acid, which latter may be used a s an indicatlor for the titration of soli~tions of salts of barium. If the solution of a chromate contains sulphat,es, the titration with barhim chloride gives the total qnaritity of the two salts, from which must be deducted the amount of the latter as determined gravimetrically.For the determination of the combined snlphuric acid, an excess of barium chloride solution is added, and the excess determined by titration with pot.assium dichromate. Salts of aluminium, copper, and iron must be removed from the solutions. J. W. L. Estimation of Antimony by Marsh's Method. By A. VAN BYLERT (Ber., 23, 2968-2971 ; compare Kuhn and Saeger, Abstr., ISYO, 1187).-For the estimation of antimony in alloys of tin, silver, and antimony, the author recommends the following process :- A three-necked Woulffe's bottle is connected on the one hand with an npparatus for evolving carbonic anhydride, and on the other with thu usual calcium chloride tube and hard glass tube. The central neck of the Woulffe's bottle is fitted with a wide tube reaching to the bottom of the flask.About 0.5 gram of the alloy is dissolved in 20 C.C. of mercury at 60", and poured, after cooling, into the fiask. 100 C.C. of 10 per cent. sulphuric: acid is then added through one oLc the side tubes, and the air expelled from the apparatus by carbonic anhydride. A freshly prepared sodium amalgam, obtained by dis- solving 5.5 grams of sodium in 25 c . ~ . of mcrcury, is then aclded dro]~ D. A. IJ. i;L116 ABSTRACTS OF CHEMICAL PAPERS. by drop through the wide tube. The sublimate of antimony quickly appears in the heated tube ; the apparatus is then periodically shaken, care being taken that no liquid is allowed to pass into the central tube. After the evolution of gas has ceased, carbonic anhydride is again passed tbrough the apparatus t o expel all the hydrogen and hydrogen antimonide. The mercury solution is then poured off, dried with filter-paper, and divided into two equal portions, one of which is returned t,o the cleaned and dried apparatus, and covered with 75 C.C.of a 10 per cent. sulphuric acid, whilst the other is mixed with 3 grams of sodium, and added drop by drop through the central tube as before. The results obtained are fairly accurate, but might possibly be improved by employing lirdrogen in place of carbonic anhydride for driving out the air. Another soiirce of error is the oxidation which takes place in the manufacture of the alloy, and during its solution in the mercury. H. G. C. Estimation of Hardness of Natural Waters. By E. JJ. NEUGE- JMJER.(Zeit. anal. Chem., 29, 399-401 ).-The author proposes the following modifications of Clark's test. The standard water is a mixture of 8 vols. of calcium sulphate solution of 12" of hardness with 2 vols. of a 12" magnesium sulphate solution. Of this mixture, 100 C.C. is used. The soap solution is of such strength that 12 C.C. is required for the 100 C.C. of standard water. The following new table has been drawn up from titrations of the standard water diluted to the required degrees :- Hardness.. .......... 0" lo 2" 3" 4" 5" 6" C.c of soap solution . . 0.6 1.7 2.8 3.9 4.9 5.9 6.9 Hardness.. .......... I 8" 9" 10' 11" 12" C.c of soap solution ,. 7.8 8.7 9.6 10.4 11.2 1200 F O and a, special burette (titanometer) constructed, the readings of which give a t once the degrees of hardness.The Analysis of Sulphurous Waters. By D. VITALI (Chem. Cen.fr., 1890, ii, 166 ; from L'Orosi, 13, 73-778).-Thiosulphates may be detected by the addition OE potassium nit,rite and a mineral acid or acetic acid. Nitric oxide is liberated, and the solution is thereby coloared yellow ; later, sulphur is precipitated, and the liquid becomes milky. 'Ibis reaction is extremely delicate, 0.0001 per cent. of thio- sulphate being detectable. I n testing for nitrites in presence of thio- sulphates with potassium iodide and starch, this reaction of the tbio- sulphate may prevent the formation of the blue iodide of starch, a yellow coloration being produced instead. In testing for iodine, in presence of thiosulphates, with potassium nitrite and an acid, it is better to evaporate the water to dryness after neutralising with sodium carbonate, and then to extract the residne with absolute alcohol, which dissolves the iodide, leaving the thiosulphate undis- solved.In the case of waters containing large quantities of calcium snlphnte, it is recommended to first precipitate the calcium as calcium carbonate before determining the silicic anhydride. Xu order to prevent tho precipitation of sulphur during the evaporatior of water M. J. S.ANALYTICAL UHEMISTRY. 117 for determination of the total solid residue. the author recommends that a current of hydrogen should be passed through the water. J. W. L. Examination of Water for Contamination by Gas Works, By F. DICKMAEN (Zeit. miat. Cirem., 29, 398--399).-1n a specimen of water from a brook which had suffered contamination from a neigh- bouring gas works, and by which poultry had been poisoned, tho author detected traces of a substance giving the reactions of di- pbenylamine.Owing to the stability of this compound and the sensitiveness of its reaction with nitric acid, its presence might be used as a proof of contamination by tar-water, if i t should be found to be a constant constituent of that liquid. Detection and Estimation of Organic and Inorganic Poisons in Corpses. By A. SEYDA (Chem. Zeit., 14,31--32, 51-53,128 -129, 181-184, and 198-200) .-The author describes his system of examining corpses. The chemical examination proper is preceded by a preliminary examination of the blood, urine, and contents of the stomach and accessories.The blood, when not too decomposed to show the absorption bands, is examined spectroscopically ; if dry, it is dissolved in water, and made very slightly alkaline with sodium hydroxide ; the two oxyhaemoglobin bands amd the intermediate haemoglobin band merge into one con- tinuous band in partially decomposed blood; the faint band in the red is attributable to methaemoglobin in alkaline material or to hmmatin in acid blood. In the former case, fiwther examination is only made under exceptional circumstances, such 8s the presence of hydrogen sulphide not emanating from ordinary putrefactive pro- cesses ; but, in the latter case, the red band being due to hsematin, is regarded as indicating the presence of other reducing agents, or of acids, potassium chlorate, ferricyanides, nitroglycerol, or nitro- benzene, which are tested for in the urine and in parts of the body.Other isolated lines in the red are carefully noted, and carbonic oxide is sought for ; the presence of hsmatin, already reduced or otherwise, indicates the absence of carbonic oxide poisoning ; carbonic oxide haemo- globia is recognised in presence of oxyhEmoglobin and hEmoglobin by the mere displacement of the absorption band towards the red part of the spectrum when ammonium sulphide is added ; chemical tests with sodium hydroxide with or without calcium chloride are con- sidered of little value. In the urine :-Notes are taken of the quantity, colour, odour, of re- actions as to the presence of blood, albumin, and sugar, of the action in alkaline copper solutions, of the behaviour toyvards barium chloride before and after treatment with hydrochloric acid, and of the occur- rence of balsams, alkaloids, soluble metallic poisons, and of such salts as potassium chlorate, iodide, or bromide.The contents of tlie stomach are examined in the dark for phosphor- escence, which is only due to phosphorus in acid mixtures, otherwise fungoid growth may cause it. The, udour of the contents of the stomach may be acid, alkaline, like decayed cheese, or they may have a specific odour, or be putrid ; sometimes they are well preserved, and M. J. S.118 ABSTRACTS OF OHEMICAL PAPERS. have a sweetish, repulsive odour, observed by the author in cases of ar.ienical poisoning. Food iizagmas and vomits are carefully sampled, treated with alcohol and then with ether, and examined macro- and micro-scopically.The alcoholic extract is examined for oxalases and foreign bases and acids ; matters found in the folds are also examined. The arsenic test is applied, and an aqueous extract of the magma is examined f o r soluble poisons. For the chemical examination proper, parts of organs, finely divided, arc heated with water in a boiling water-bath for several hours, acidified q-ith tartaric acid, and distilled with steam. Two fractions are made: the first will contain the more volatile matters, such as alcohol, aldehyde, acetone, chloroform, nitrobenzene, ethereal oils, turpentine, camphor, amines, and their volatile corn binations ; the other, the remains of these more volatile products, and any less volatile substances, such as fatty acids, phenol, hydrocyanic acid, &c.A plain distillation of alkalilie material is required at times ; if, how- ever, the presence of chloral bydratc or hydrocyanic acid and ferro- cyanides is suspected, the finely divided orgaris are first moistened with potash or with hydrogen sodium carbonate respectively. A steam distillation from alkaline solution, when necessary, is best ejfected, not directly, but with the liquid obtained by extraction with water and tartaric acid. The residue from the acid distillat'ion is employed in testing for alkaloids. The odour, colour, opdesceiice, quantity, &c., of the first fraction of the acid distillate are noted. I t is tested with silver nitrate, both in nitric acid and in ammoniacal solution, with sodium nitroprusside, potash, and acetic acid ; with alkaline permanganate ; with iodine and potash (iodoform reaction) ; with zinc-dust and hydrochloric acid for nitrobenzene ; with hydrochloric acid and alcoholic phloroglucinol for otheyeal oils (a reaction frequently takes place, but too much depend- ance is not to be placed on it; it is better to examine the urine for some of these substances, also for altered camphor). The reagents for tur- pentine oil, in tangible quantities, are a mixture of fresh guaiacum alcohol and fresh citronella oil.Resorcinol and potash are better reagents for detecting chloroform than the isonitrile test. A quantita- tive examination for these substances is seldom possible, with the exception of alcohol, but this cannot be estimated by distillation in the presence of amines ; therefore it is oxidised to acetic acid as follows.A portion of the distillate is redistilled, the first portions collected are dried with potassium carbonate, distilled again, treated with sodium dichromate and sulphuric acid, rendered alkaline with potash, boiled t o eliminate amine bases, then acidified, steam distilled, and the distillate, containing the acetic acid, titrated ; any sulphuric acid, if accidentally present, being estimabed and allowed for. The author at,taches great importance to the estimation of alcohol, especially in the case of children. Various parts of corpses of persons addicted t c alcohol yield distillates containing alcohol, but not pure ethyl alcohol ; therefore such distillates reduce alkaline permanganate, &c., give the iodoform reaction, and yield an inflammable distillate on redistillation from potassium carbonate.Alcohols readily evaporate from corpses, the more volatile disappearing first.ANALYTICAL CHEMISTRY. 119 Passing on to the second fraction from the acid distillation, phenol, as n normal product of the decomposition of nlbuminoids, is frequently detected by Millon's reagent, less readily by bromine-water. Large quantities of phenol are estimated by filtering the fraction containing it from the fatty acids, extracting with ether, drying the ethereal residue over sulphuric acid, and weighing, taking precautions against the phenol creeping over the edge of the evaporating dish. As regards hydrocyanic acid, the distillate is tested with copper sulphate and guaiacum, and if the reaction is noticed, i t is confirmed by some other test, and the acid determined as silver cyanide.Phosphorus is generally recognised by the phosphorescence, but if this does not occur phosphorous acid must be tested f o r and estimated in the residue. I n examining the distillate for phosphorus, it is redistilled, using an upright bulbed tube, the distillation being continued for an hour after phosphorescence has ceased to appear; the new distillate is treated with silver nitrate, &c., and also is tested for phosphorus by oxidising with chlorine-water and adding ammonium molybdate. I n examining for alkaloids, special care must be taken not to mistake ptomaines for other alkalo'ids, and to allow for the impure form of the latter.The material is treated with alcohol containing tartaric acid, the extract filtered, evaporated, dissolved in water, filtered, neutra- lised with potash, concentrated, treated with alcohol, separated from the potassium tartrate, and the neutral aqueous fiolution is tested with alkaloid reagents (a preliminary test with tmtaric acid and iodic acid being made for morphine) and examined systematically if required. A portion is rendered alkaline, and steam-distilled for nicotine, coniine, miline, &c. I n the absence of these volatile bases, another portion is extracted with ether successively when neutral, acid, and alkaline, ehen with chloroform while still alkaline, and finally is made ammoniacal and extracted with aniyl alcohol. The t8hird portion is reserved for the direct confirmatory examination of any alkaloid icdicated in the other portions.The aqueous residue is tested for narceine and curarine, whilst the various extracts are examined separately. The residue from the dkaline-ether extract is tested with phosphoric acid for aconitine. Vitali's atropine reaction is liable to be hidden by xaiithoprotcin colour reactions. For strych- nine, a double test is made: first., a drop of vanadic acid solution is mixed with the Gesidue, dissolved in sulphuric acid, which is subse- quently diluted with concentrated sulphuric acid, and then solid ammonium vanadate is dusted over another portion of the same solution. Attention is called to the fact that colocynthine gives reactions with both vanadic and chromic acids, which resemble those of strychnine with the same reagents. The atuyl alcohol extract, after purifying, serves for confirming the presence of morphine.Another portion of the original extract is examined for metallic poisons soluble in alcohol. The examination for metallic poisons generally is made with the residue from the first distillation or that, from the alcoholic extraction ; in the latter case the alcohol is expelled by warming. The residue is heated with hot water, potassium chlorate, and hydrochloric acid until the organic tissue is destroyed a r d all chlorine driven off, the magma is treated with tartaric .acid, then120 ABSTRAOTS OF OHEMIOAL PAPERS. largely diluted with water and, after 24 hours, filtered.Tho insoluble portion is treated (if required with more chlorate and hydrochloric acid, then) with alcohol, and with ether to extract fat, and is ignited, The ash is treated with very dilute hydrochloric acid, and the residue dried, ignited, weighed, and proved conclusively to be silica by fusion with sodium carbonate ; or is examined for silver, lead, barium, and strontium. The soluble portion is marde up to a definite volume, and must be free from chlorine and chloric acid. For mercury, a portion is nearly nentralised by means of potash, and is digested with brass wool €or 15 minutes a t 70" ; if the brass is visibly amnlgnmatecl, the presence of mercury is confirmed by heating in a test-tube, &c. ; but when the presence of niercury is not so evident, the brass wool is burnt with copper oxide in a current of air in a tabe 25 cm.long, drawn out to a doubly bent capillary, in which any mercury is accumulated, and identified with iodine. To estimate mercury, the hot hydrochloric solution is treated with hydrogen sulphide, the precipitate collected on an asbestos filter (or if arsenic is present, i t is first digested with yellow ammonium sulphide) washed with hydrochloric acid, dissolved in nitric acid, filtered through some asbestos, and the washings and filtrate diluted and treated with phos- phorous acid. After 24 hours, the calomel is filtered off, washed with, water, alcohol, and ether, and weighed on a tared filter. Tbi*oughout the estimation, high temperatures and contact with organic matter are to be avoided.To detect antimony, some of t,he liquid partly neutralised with ammonia is placed in a bright platinum dish with a piece of zinc for six hours ; the brownish-black antimony flakes obtained in this manner being more trustworthy than Marsh'R test. This test does not answei- in the presence of tin ; but by fusing the mixed oxides wit11 sodium hydroxide, most of the tin can be separated. To estimate antimony, hydrogen sulphide is passed through the slightly acid, and at first boiling, solution until it has cooled down ; after three days, most of the hydrogen sulphide is driven off by carbonic anhydride, the preci- pitate washed with an acetic acid solution of ammonium acetate, treated and washed with a solution of sodium sulphide, containing hydrogen sulphide, the solution treated carefully with hydrochloric acid, warmed, and then boiled.After 24 hours, the precipitate is col- lected, treated with an acetic acid solution of ammonium acetate, then carefully with nitric acid, evaporated, the residue moistened with sodium hydroxide, intimately mixed with dry sodium carbonate, dried, introduced by small quantities at a time into fused sodium nitrate in a, silver crucible, and the mass, when cold, treated with water. After 24 hours, the precipitate is washed with 45 per cent. alcohol, contain- i n g soda, digested with il hot solution of tartaric and hydrochloric acids for half an hour, filtered, and washed with a dilute solution of tartaric a d hydrochloric acids. The filtrate and wasbings combined are concentrated in a water-bath, the excess of acid reduced with ammonia, and the antimony precipitated by hydrogen sulphide as a pure orange-coloured sulphide which is converted into oxide by Bunsen's method.The purity of the antimony is ultimately confirmed by its volatility.ANALYTICAL CHEMISTRY. 121 To detect arsenic, mercury and antimony being absent, various obvious precautions are observed in applying the Marsh test to some of the liquid, and when a mirror is obtained, the tube containing i t is divided by a diamond into four parts, of which one is used for the odour test, another for solubility in freshly-prepared sodium hypo- chlorite, a third for dissolving in nitric acid arid testing with silver nitrate, whilst the fourth is dissolved in nitric acid, and converted into arsenic sulphide by colourless ammoniuin sulphide.To estimate arsenic, hydrogen sulphide is passed through the warm hydrochloric solution for 12 hours, and after remaining three to five days in a closed flask, most of the hydrogen sulphide is driven off by a current of carbonic anhydride ; the procedure then resembles that described in the antimony estimation, but various points are to be observed :- 1. Ammonia or ammonium carbonate are the only solvents used for arsenic sulphide on the filter. 2. The fusion is conducted in a porcelain crucible with fusion mixture and potassium nitrate. 3. The arsenic is always weighed as magnesium pyroarsenate. 4. The alkaline solution is not precipitated directly with magnesia mixture, but is first Rub- mitted to the following treatment :-Neutralisation with nitric acid, expulsion of carbonic anhydride and nitrous acid, precipitation with hydrogen sulphide, and conversion into arsenic acid.5. The ammo- nium magnesium arsenate is redissolved in hydrochloric acid and re- precipitated by ammonia. 6. Small quantities of magnesium arsenate are converted into pyrosrsenate by dissolving in very dilute nitric acid, evaporating in a porcelain crucible over a water-bath, and carefully and gradually igniting the residue ; the method is susceptible of great accuracy ; as little as 0.0093 to 0.0077 gram of arsenic in a portion of a dead body has been estimated. 7. Large quantities of pyroarsenate are preserved as such for reference ; small quantities aro converted into metallic arsenic in a Marsh’s apparatus, and are preserved in EL sealed tube.In the absence of mercury, antimony, and arsenic, the hydro- chloric acid solution is made alkaline with soda, acidified with acetic acid, hydrogen sulphide passed in at the boiling point of the liquid and until cold, sodium carbonate added to distinct alkalinity, and the whole allowed to remain corked up until clear. The solution serves for de- tecting and estimating tin. The precipitate is washed with sodium sulphide containing hydrogen sulphide, oxidised with nitric acid, evapo- rated, moistened with sodium hydroxide, mixed with fusion mixture, dropped into molten nitre in a silver crucible, extracted with water, supersaturated with hydrochloric acid, filtered, and the hydrochloric acid solutions submitted to the ordinary methods of analysis, weighing any metal isolated in a definite form. The separation of iron, alumina, and zinc in the presence of calcium and magnesium phosphates is effected in the following manner :-The hydrochloric acid filtrate from the hydrogen sulphide precipitate is concentrated on the water-bath, treated with chlorine-water, evaporated, the residue dissolved in very dilute hydrochloric acid, filtered, the solution supersaturated with am- monia, excess of the latter nearly expelled on the water-bath, the pre- cipitate removed, the solution acidified with acetic acid, and hydrogen sulphide passed into the boiling hot liquid until it gets cold.The zinc122 ABSTRACTS OF CHEMICAL PAPERS. sulphide is weighed. The ammonia precipitate is dissolved in nitric acid, treated in a platinum dish with tin, which by repeated and careful treatment with nitric acid is converted into stannic oxide, and with i t the phosphoric acid into insoluble stannic phosphate, from which the alumina is washed out by yery dilute nitric acid, and estimated in the solution by precipitating with ammonia, igniting the precipitate, fusing the ignition residue with sodium carbonate, extracting with water, filtering, &c.The alkaline filtrate from the hydrogen sulphide precipitate is examined for tin :-It is acidified with hydro- chloric acid, boiled, hydrogen sulphide passed through until it is cold ; after remaining for 24 hours in a warm place the precipitate is collected, washed with an acetic acid solution of ammcinium acetate, and ignited with the filter.The residue is moistened with nitric acid, evapo- raked, ignited, and to get rid of the iron present it is washed into a silver crucible, dried, and treated for I.~df-nn.hour with molten sodium hydroxide, extracted with water, filtered, the filtrate acidified with hydrochloric acid, and the tin obtained in the usual manner. I n concliision, it is pointed out that not only are potassium, sodium, calcium, magnesinm, iron, and niaiiganese noimally present in the human body, but that aluminium, copper, and zinc are always met with, and less frequently tin and lead. These extraneous metals are derived from food, cooking utensils, medicine, &c. ; aluminium comes from various sources, and even after death may be introduced iii the dust, when the post-mortem takes place i n the countr-y.The authoi* in advo- cating his employment of a solution prepayed directly froin the corpse inaterial for the detection of arsenic, points out that consideriiig the sensitiveness of the arsenic reaction a concentrrttion of the solution is not necessary, that the “brown speck” on the porcelain lid referred to by Otto cannot interfepe in his method, which, more- over, obviates any chance of vitiation through arsenical hydrogen sulphide. It is also shown that the presenceof chlorides and nitrates does not stop the formation of gaseous hydrogen arsenide, provided that the zinc and hydrochloric acid are in excess and the evolution of hydrogen is allowed to proceed snfficiently long ; but the presence of free nitric acid stops the evolution temporarily. It is still doubtful whether solid hydrogen areenide is converted into the gaseous modi- fication by zinc and hydrochloric acid.Ordinarily only one poison is found in a corpse, but nevertheless i t should he borne in mind that there is the possibility of more than one being present. Detection of Paraffin in Beeswax. By H. HAGEI~ (Zeit. anat. Cliem., 29, 480481 ; from Pharrn. Centi-allmlle, 30, 565).-A few grams of the substancs in fine, air-dried shavings is gradually heated in a small, porcelain capsule, until fumes begin to rise. A half-litre wide-mouthed bottle is then inverted upon the capsule, and when filled with white vapours is closed and set aside until the fumes bave condensed upon its walls. The sublimate is then dissolved in 3 C.C.of chloroform, the chloroform evaporated in a test-tube, and the residue boiled with 4 C.C. of soda solution. If paraffin was present, it, will after cooling be found floating on the clear solution. A drop of D. A. L.ANALYTICAL CHEMISTRY. 123 the chlorofoi*m solution niay also be evaporated on a slip of glass and examined microscopically. The fumes from pure beeswax are not so whiteas from paraffin, and are only obtained at a higher teinpe~ature (300-320"). The sub- limate gives a coloured solution with chloroform, and a coloured and turbid solution with soda. The residue from the chloroEorm solution is a dnll film ; paraffin on the contrary gives separate grains in a clear field. M. J. S. Condition of the Sulphuric Acid in Plastered Wines, and a Method of Distinguishing between Plastered Wines and Wines mixed with Sulphuric Acid.By L. Roos and E. THOMAS (Compt. reitd., 111, 573--577).-Wines which have been mixed with calcium sulphate do not contain potassium hydrogen sulphate. The liberated tartaric acid interacts with the organic potassium compounds in the wine, and forms a new quantity of potassium hydrogen tartrate. Direct experiment shows that when calcium sulphate is added to a solution of potassium hydrogen tartrate, and an acetate, malate, citrate, or succiaate, the liquid contains no free sulphuric or tartaric acid, but acetic, citric, malic, or succinic acid is liberated. No potas- sium hydrogen sulphate could be detected i n plastered wines by the following method, which will detect the addition of 0-25 gram of sulphuric acid per litre. 'l'he proportion of chlorine and the total sulphuric acid in the wine are estimated.50 C.C. of the wine is mixed with a small quantity of ammonium acetate, and exactly procipitated with a standaiad solution of barium chloride. The filtrate is evaporated to dryness, heated gently, and the chlorine in the residue is estimated. If only normal potassiuni sulphate is present, the reaction is &SO, + BaC'I, = BaSOl + 2KC1, and the chlorine in the residue should be equal to the chlorine of the barium chloride, plus the chlorine originally present in the wine ; i f the acid sulphate i s present, the reaction ia KHSO, + BaCl, = BaSOp + HC1 t KC1, and the free hydrogen chloride is cxpelled i n the process of evaporation, the loss increasing with the quantity of hydrogen-sulphate present.Estimation of Dissolved Solids in Wine. By E. L ~ S Z L ~ (Chem. Zeit., 14, 438, 455).--Results are quoted, showing t,he unsatisf'nctsry character of estimations of " extractives " made by drying residues for 2$ hours. The author suggests determining the alcohol both by distillation and by an alcoholometerat 15" ; the difference between t.he two observations being due to the "extractives" present may be utilised as a measure for them, and he finds that miiltiplying this difference by 0.32 gives numbers for the quantity of dissolved solid matter in 100 C.C. of wine concordant with actual determinations. 'l'he alcoholometer should not have a greater range than lo", or at the outside 12", and f o r wines of sp.gr. greater than 1.000, a saccharo- ueter showing volume percentages is emplcyed. Detection of Methylated Nitrous Ether. By J. MUTER (Analyst, 15, 48).-Much of the sweet spirits of nitre in commerce is prepared from- methylated spirit instead of from pure ethyl alcohol C . H. B. D. A. L.124 AliSTRAOTS OF CHEMICAL PAPERS. as prescribed in the pharmacopceia. The two may be discriminated by dissolving a fragment of solid potash in a sample. The methylated ether darkens, the colour varying from deep-yellow to orange-red, while the odour of methylated spirit becomes very distinct. The ether from pure spirit loses its odour of ethyl nitrite, and retains only that of ethyl alcohol, and it does not dnrkeu beyond the faintest straw colour. On treating with Hubl’s reagent the dis- tillate obtained after digestion with potash, tho methylated sample will absorb 0.4 to 0.7 per cent.of iodine, but that from pure spirit none. Analysis of Carbolic and Sulphurous Disinfecting Powders, By J. MUTER (AnaZyst, 15, 63--68).-The author calls attention to tbe ambiguities in the usual forms of specification for disinfecting powders. Whilst the contracts are nominally for “ carbolic acid,” i t is commonly understood that the powder may contain chiefly cresol and other high-boiling tar phenols. There have, however, been cases where the supply of the more costlyabsolute phenol has been insisted on. The omiesion of the word “ available ” before “ sulphurous acid” some- times renders a literal compliance with a specification impossible.For the estimation of the phenols, the author still employs his own process (Abstr., 1888, 92), with the single modification that 150 C.C. of a 10 per cent. solution of sodium hydroxide is now used instead of 200 C.C. of a 5 per cent. solution. The cresol, measured in contact with brine, retains about 5 per cent. of water. Since anhydi-ous cresol increases in volume by about 5 per cent. when shaken with 3 volumes of brine, whilst that containing water does not increase o r may even diminish, this furnishes a rough but ready test for t h e presence of water. For more accurate work the water must be dis- tilled out. Naphthalene, which is usually present in commercial cresol, may be estimated as follows :-50 C.C. is shaken with 200 C.C. of a 10 per cent.solution of sodium hydroxide. The phenols dissolve, leaving the naphthalene floating. The solution is removed, the naphthalene washed with a 5 per cent. soda, solution, then rapidly filtered off. It is rinsed from the filter with water and again collected on a pair of filters. After drying as far as possible by pressing between blotting paper, the filters are separated and the inner one with its contents is weighed, using the outer one as a tare. For estimating the available sulphurous acid, 2 grams of the powder is washed 011 a filter with dry ether until the phenols and tarry matters are removed. As soon as the ether has evaporated, the contents of the filter are thrown into a bottle containing 50 C.C. of N/10 iodine solution, and after half an hour the residual iodirie is titrated by thiosulphate.This method is unsatisfactory when the basis of the powder is lime. I n sulphurous powders which have undergone osidation, the amount of original sulphurous acid cannot be ascertained if the mixture had consisted of gypsum and calcium sulphite, but where the basis is silica, the sulphates present may be regarded as oxidised sulphites, and where sodium hydrogen sulphite has been mixed with gypsum, the estimatiou of calcium, sulphuric, and sulphurous acids in an aqueous extract will give the necessary data. M. J. S. M. J. S.ANALYTIOAL OHEMISTRY. 125 Arabinose. Milligrams. 17 -0 18 *6 20 '3 21 -9 23.5 25 -1 26 *7 28.3 29.9 31 *5 33.1 --- Detection of Diresorcinol as an Impurity in Synthetically Prepared Phloroglucinol.By J. HERZIG and S. ZEISEL ( J h a t s h . , 11, 421-423) .-The presence of diresorcinol, as an impurity in phloro- glucinol, scarcely affects its melting point, or the nnmbere obtained on estimating carbon and hydrogen. It may be best detected by dissolving a few milligrams of the sample in about 1 C.C. of concen- trated snlphuric acid, adding 1-2 C.C. of acetic anhydride, and warming the mixture for a few minutes i n a water-bath. If diresor- cinol-or its tetrethyl ether or tetracetyl derivative-be present, a bluish-violet colour, which disappears on the addition of much water or of an excess of alkali, will be produced. By this means, the presence of 0.4 per cent. of diresorcinol may be clearly shown, and the delicacy of the test is probably much greater.G. T. M. Estimation of Sugars by means of Copper Potassium Carbonate Solution. By H. OST (Bey., 23, 3003-3011 ; compare Abstr., 1890, 103l>.-A solution cmtaining 23.5 grams of crystallised copper sulphate, 250 grams of potassium carbonate, and 100 grams of hydrogen potassium carbonate per litre has the following advantages over Fehling's solution for the gravimetric determination of sugars :- (1) It is unchanged by keeping. (2) Its action on cane-sugar is relativcly slight. (3) After 10 minutes boiling, the precipitation of cuprous oxide is practically complete, and thus more concordant results are obtained. (4) The monosaccharoses precipitate almost twice asmuch cuprous oxide from this solution as from Pehling's solution. ( 5 ) The quantity of precipitate obtained from different kinds of sugars varies considerably, thus rendering it possible to determine the composition of mixtures.The solution may also be employed for Folumetric estimations, as the end reaction is sharp ; the time required for boiling, is, however, longer than with Fehling'H solution. For gravimetric determinations, 30 C.C. of the copper solu- tion is mixed with 25 C.C. of the sugar solution, water is added and the liquid boiled for 10 minutes, filtered through an asbestos filter, and the cuprous oxide reduced in a stream of hydrogen. The follow- ing table shows the quantity of copper precipitated by different sugars :- Copper. Xilligmins. 50 55 60 65 70 75 80 85 90 95 100 ---- Invert-sugar. Milligrams. 15 *2 16 *6 18'0 19 *4 20.8 22-3 23-7 85 *2 26 *6 28 *1 29 ' 5 -- Dextrose.Milligrams. 15 *6 17 *O 18 '5 19 '9 21 '4 22 *9 24 -4 25.8 27 -3 28 68 30 -3 --- Levulose. Milligrams. 14 *7 16 '1 17 * 5 18 *9 20 '3 21 -7 23.0 24 -3 25 -7 27 -1 28 -5 &lac tose . Milligrams. 17 '4 19 *1 20 *8 22'5 24 *2 25'9 27 -6 29 -3 81 '1 32 -8 34 -5 --ABSTRACTS OF CHEMICAL PAPERS. Copper. Milligrams. 105 110 115 120 125 130 135 140 115 150 155 160 165 170 175 180 186 190 195 200 2c5 210 215 220 225 230 235 240 245 250 255 260 265 270 275 280 285 290 295 298 -7 CI- v Invert-sugar. Milligrams. 31 '0 32 *4 33 -9 35.3 36 -8 38 *2 39 *7 41 '1 42 *6 44-0 45 -5 47.0 48 -5 50 '0 51 -5 53 -0 54.5 56 -0 57 -5 59-1 60 *7 62 '41 84 *1 65 -8 67 *5 69 *3 71 -1 7s -9 74 -8 76.7 78 *6 80 -5 82 -5 84.7 87 *I 89.7 92 -3 95 -1 98 '0 103 '0 -- Dextrose.Milligrains. 31'8 33 -3 34 *8 36 '3 37'8 39 '3 40 *s 42.3 43-s 45 -3 46.8 48.3 4g9 * 8 51 -4 52 *9 54'5 56.0 57 *c; a9 -2 60 *8 62 *4 64 '1 65.8 67.8 69 *2 70-9 72'7 74 *5 76 *4 78 -4 80 -5 82.8 85.1 87'5 89.9 92.4 94 -9 97.6 100 '4 102 *8 -- Levulose. Milligrams. 29 -8 31 -2 32.6 34.0 35.4 36 *8 38 *2 3'3.6 41 -0 42.5 43 .9 45-3 46.7 48'1 49.5 61 -0 52 -5 54 -0 55 *5 57 -0 58 fi 60.2 61 .S 63.5 65.2 6G 9 68 -7 70 *G 72 -5 74 -4 76 -5 78 *8 81 '1 83 -5 85 -9 88 6 91 '3 94 '2 97 -2 99 '0 ~ Galactose. I'Jilligrams. 36 '2 38 -0 39 -7 41 '4 43 -1 44 *8 46 5 48 -3 50.0 51 *8 53.6 55 *4 57.2 59 *o 60.8 62 *7 64 -5 66 '4 68 -3 70 *3 72 '3 74 *3 76 *3 78 -3 80 *3 82 -4 84 -5 86 -6 88.9 91.2 93.5 95 -9 98.3 1GO * 7 103 -3 106 *l 109 -0 112 -0 115'1 117 *O -- Arabinose.Milligrains. 34 -7 36.3 37 *9 39 *5 41 1 42 -8 44.3 46 '0 47'6 49 '3 50 '9 52 *G 54 '3 55 .9 57 *5 59 -2 60 9 62 -7 61 '4 66 *2 68.0 69.8 71.6 73 -5 75 4 77 -3 79 *3 81 -3 83.4 85 -5 87 -6 89 -8 92 -2 94 -6 97 *I 99 '6 102 -3 105 *1 107 -9 109 -5 -- In the case of lactose, the factor copper/l~ctose = 1.31 to 1.57 for solutions containing from 125 to 198 milligrams of sugar. For volumetric work, a n indicator must be employed; after 20 minutes boiling, 198 milligrams of lactose precipitate 190 milligrams of copper, Ra5ose, C,H,Ola + 5H20, does not affect the copper solution,ANALYTICAL CHEMISTRY. 127 but after hydrolysis, it has the highest reducing power, 50 milligrams precipitating 150 milligrams of copper. Estimation of Sugar in Milk.By M. KUHN (Bied. Centr., 19, 628 ; from 3IiZc*hzeit., 18, 926) .-Results obtained by Tollens’ method agree better with those obtained by Soxhlet’s method, when only so much serum solution is employed that the colour is bluish after boiling. If so much sugar solution is used that the liquid is greenish after reduction, results will be obtained which are 0.1 to 0-15 per cent. too low. The phosphotungstic acid method is not rccornmended. If Soxhlet’s method is not used, the lead acetate method should be employed. Estimation of Ash in Raw Sugar. By W. MINOR (Chenz. Zeit., 14,51O).-Stammer objects to the use of oxygen, and recommends air for the incineration of raw sugay in estiniations of ash. The author has investigated the point, and sees no reason for disqualifying oxygen nor any special virtue in the atmospheric nitrogen, and as burning the charred sugar in oxygen takes 25 minutes, whilst corn- biistion in air, with the aid of mechanical agitation, requires from 6 to 15 hours, he considers the oxygen method is distinctly to be recommended.11. A. L. J. B. T. N. H. J. M. Estimation of Starch. By 0. REINKE (Zeit. anal. Cl2etn., 29. 472-475 ; froin Zed. XpiritiiLdusf.) .-The author divides the processes hitherto proposed into those with and without high pressuw, and recommends the following as the best of the respective methods :- With high pressure : 3 gi-arns of the finely ground substance is stirred with 25 C.C. of a 1 per cent. solution of lactic acid and 30 C.C. of water in a metallic beaker, then covered and heated for 24 hours in a digester (Soxhlet’s or Lintiier’s) a t 3-& atmospheres pressure, then mixed with 50 C.C.of hot water and, after cooling, made up to 250 C.C. and filtered. 200 C.C. is then inverted by cohobating with 15 C.C. of hydrochloric acid (1.125 ~ p . gr.) for Z& hours, then neutralised with soda, made np to 500 c.c., and 25 C.C. of it titrated with Febling’s solution. Without high pressure : 3 grams of the substance is boiled with 50 C.C. of water, and then digested for au hour a t 62.3” with 0.05 gram of Lintnep’s diastase. It is then cooled, mads up to 250 c.c., and 200 C.C. inverted with acid 8s above. For the preparation of Lintner’s wade diastase, I part of green malt is extracted for 24 hours with 2 to 4 parts of 20 per cent. alcohol. The extract, filtered by suction, is precipitated with twice, or at most 2$ times, its volume of absolute alcohol.The upper liquor is poured off and the precipitate thrown upon a pressure filter, then rubbed down with absolute alcohol in a mortar, again filtered and washed with absolute alcohol and then with ether, and finally dried in a vacuum over sulphuric acid. For the purification of this raw product, the precipi- tation and digestion with alcohol, washing with ether, and drying are repeated. By this means, fclbuminoiid impurities are rendered insoluble and dextrino’id ex tractive matters removed. The dried product is a loose, yellowish-white powder, which has no action on128 ABSTRACTS OF CmMICAL PAPERS. Fehling’s solution either before or after boiling with hydrochloric acid, and which does not turn brown when its solution is evaporated on the water-bath. M.J. S. . A New Application of Molisch’s Reactions. By G, COLASANTI (Gazzetta, 20, 2C39--%05) .-Molisch (Abstr., 1886, 923) found that, the merest traces of sugar or glucosides (O*OOOOl per cent.) could be detected by the addition of one or two drops of an alkaline solu- tion of a-naphthol or thymol (15 to 20 per cent.), together with an excess of coiicentrated sulphuric acid. Molisch further derived from this reaction a confirmation of the alleged presence of sugar in normal urine. The author finds that extremely dilute solutions of potassium or sodium thiocyanate, treated ih the same manner, show first a green band and, on agitation, an iiiiense violet coloration resembling in all respects that obtained from solutions of sugar.On cooling the liquid, a, compound containing the naphthalene nucleus and the sulphonic group separates in a mass of long, slender needles. The solution of thiocyanate or thiocynnic acid must be very dilute, or on addition of sulphuric acid a brown coloration is produced, and hydrogen sulphide is evolved. Urine must similarly be diluted before treatment with a-naphthol, and altogether fails to give the thymol reaction. As urine has been found to contain thiocyanic acid, Molisch’s reaction affords no confirmation of the presence of sugar in khat fluid. Reaction of Thiocyanic Acid. By G. COLASANTI (Gazzetta, 20, 306--308).-1f a few drops of a solution of auric chloride (&th per cent.) made alkaline with a saturated solution of sodium carbonate or a 5 per cent..solution of potash are added to a few C.C. of a dilute solution of a thiocyanate (0.01 per cent.), a deep violet coloration is obtained, and a precipitate of metallic gold gradually separates. The thiocyanic acid in urine does not give the reaction, the liquid merely acquiring a reddish coloration. Schneider’s Method €or the Estimation of Malic Acid in Wine. Uy IS. NIEDERHAIJSER (C’hem. Centr., 1890, ii, 172; from Pharm. CentraZhaZZe, 31, 378-379) .-lo0 C.C. of t3he wine is neutral- ised with decinormal alkali, evaporated, incinerated, and the carbonic anhydride in the ash determined. From t,his amount, the quantity of carbonic anhydride equivalent to the total tartaric acid present is deducted, the difference being then calculated into malic acid.Since, however, wines usually coiitain other substances (tannic, succinic, acetic acids), all of which rieutralise alkalis, and would when incinerated produce carbonates, the author considers the method valueless. J. W. L. It exhibits great hydrolytic activity. S. B. A. A . S. B. A. A. comparison between Methods for Estimating Tartaric Acid. By J. T ~ T H (Chem. Beit., 14, 63-64).---To compare the three rival methods for the estimation of tartaric acid, the “ original Goldenberg ’ 9 method, the “ Ilorenz-Goldenberg” method, and the ‘( modified Goldenberg ” method, the author made simultaneous and duplicateANALYTICAL CEEMISTttY. 129 estimations in crystalline calcium tartrate, in wine lees, in argol, and in tartaric acid, following rigidly the directions laid down in each method ; the numbers obtained are tabulated, and from the results i t is concluded thstt the Lorenz modification of the Goldenberg method is the best method, and is applicable in all cases, a specially valuable factor about it being the introduction of one-third normal soda for the titration. With regwd to Roessneck's suggested method, the author shows that the amount oE antimonious oxide taken up by the calcium tartrate is not + a mol.for 1 mol., but is a variable quantity, which seems to depend on the amount of free tartrate in solution. D. A. L. Estimation of Tartaric Acid. By J. WOLFMANN (Ohem. Zeit., 14, 320 ; compare T6th, preceding abstract).-The author considera the use of litmus tincture unsatisfactory in deeply colonred tartaric solutions ; he has noticed neutralisation of alkali by humus in such solutions, and does not regard the question of the estimation oE tartaric acid as solved bay the Lorenz method, in fact, looks with greater favour on the Goldenberg-Geromont results.He himself endeavoured unsuccessfully to determine tartaric acid by titration with permangnnate. 1). A. L. Estimation of Tartaric Acid in the Crude Products of Tartaric Acid Factories. By J. TELRISZ (Crhem. Zeit., 14, :347).- In consideration of results recently published by T6th (see above), the author has made several estimations of tartaric acid in various samples of calcium tartrate and dried wine lees, applying, with much precision, both the " original '' and " modi6ed " Goldenberg-Geromont, and also the Lorenz method; the results are tabulated, and in his hands the latter method yielded undoubtedly higher results than the first two methods, and he agrees with Wolfmann (preceding abstract) in considering tbe modified Goldenberg-Geromont method the most trustworthy, up to the present time.Variations as great as 7-10 per cent., noted by Tdth, in different estimations of the same sample by this method, have not been obaerved in the preseut experiments. D. A. L. Estimation of Citric Acid in Parts of Plants. By E. Cr,AAssEN ( Z e i t . anal. Chern., 29, 468--469).-The plant is extracted with very dilute ammonia and ammonium csrbonate, the liquid somewhat con- centrated, precipitated with lead acetate, and filtered. The dried precipitate is boiled out with strong alcohol, then suspended in water, and decomposed by hydrogen sulphide.The filtrate is evaporated to zt thin ~yrup, mixed with ammonium chloride, excess of ammonia, and calcium chloride, and 3 volumes of alcohol added. The precipi- tate is filtered, washed with 75 per cent. alcohol, dried, and dissolved in hot dilute hydrochloric acid. After cooling, it is filtered, treated with excess of ammonia, and again filtered, and evaporated on the water-bath to dryness. The residue is %taken up with boiling ammoniacal water, and the insolkble calcium citrate collected on a weighed filter. Traces of citrate in the filtrate may be recovered by repeating the evaporation. M. J. S. VOL. r,x. k130 ABSTRACTS OF CHEMICAL PAPERS. Amount of Volatile Fatty Acids in Rancid Butter.By P. C o m E w A (Chern. Zed., 14, 406).--8amplen of fresh butter were taken and examiried on the 16th of Febriiary for volatile fatty acids ; they were then exposed in vessels covered with psper, and again examined on April 3rd, when, in all cases, a reduction in tbe quantity of volatile fatty acids was observed ; in a subsequent examination on April 30th, no further change was noted, but a final test, on AuguHt gth, indicated a still further falling off in these acids. The disappear- ance of T-olatilc fatty acids in the rancid buttJer, although progressive in these experiments, was in no instance very considerable, and in no case could volatile fatty acids be washed from the rancid butter either by water or sodium hydrogen carbonate.Butter and Margarine. By C. VIOLETTE (Compt. rend., 111, 345-347).--'l'he acids resulting from the saponification of 50 grams of pure, dry butter by aqueons potash are distilled in a current of ateam, and the successive portions of the aqueous distiliste (the total volume of which should not be less than 10 litzes) are titrated with normal sodium hydroxide, using phenolphthalein as an indicator. The volatile acids, which solidify, and tho non-volatile acids also, are weighed, after being dried in a vacuum and melted. A table i s given showing the results obtained with various butters and with margarine. It is assumed, on t,he evidence of Duclaux's results, that the ratio between butyric and caproic acids in genuine butter remains constant, and equal to 1.645. In ordinary butters the mean proportion of volatile acids is 7.6 per cent., wit,h a minimum of 7.0, and the proportion of non-volatile acids is 84.0 per cent., with a, maximum of 84-6.I n the case of a butter of high quality, the addition of about 20 per cent. of margarine would lower the proportion of volatile acids from 8.5 to the minimum of 7 per cent., and would raise the non-volatile acids from 82.63 to 84-76. In the case of ordinary butter, the addition of 9 per cent. of margarine would reduce the volatile acids to 7 per cent. Optical Analyses of Butters. By C. VIOLRTTE (Compt. rend., 111, 348).-From his observations, the author concludes that butter and margarine have different indices of refraction, the deviations in the oleorefractometer being -35" to -27" for bucters, and - 15" to -so for margarines.The indications of the oleoref ractometer are suffi - ciently exact when the instrument is applied to mixtures of con- stituents giving known deviations. It is necessary to ascertain, by means of a large number of observations, the minimum deviation below which a butter may be regarded as adulterat4ed with margarine. The olcorefractomet,er may be used for the analysis of commerci81 butters, but its iudicidions will not be very exact, because these butters will give deviations below the minimum f o r good butters, and the proportion of margarine deduced from the results will be too low. Analysis of Lard, Cotton Oil, and Tallow. By J. MUTER and L. DE KONINGH (Analyst, 15, 48--50).--Employing the method described by themselves (Arkalyst, April, 1889) for estimating the 1).A. L. C. H. B. C. H. B.ANALYTICAL CHEMISTRY. 131 liquid fatty acids in fats, and for treating them with iodine without. exposure to air, the ttiithors have obtained the following results. They regard tallow as the best material for the preparation of olei'c acid, and for this acid they find the iodine absorption to be 90 per cent., and to vary at most 0.2 per cent. from theory. The olei'c acid from lard never gives so low a number, the average being about 93 per cent., whilst that from cotton oil is found to be 135, with very little variation. In consequence of this wide difference, the per- centzge of cotton oil in a sample of adulterated lard can be indirectly estimated with considerable accuracy.M. J. S. Beeswax. By A. BUISINE and P. BUISINE (Bull. SOC. Chim. [3], 3, 867-8731 --The authors confirm the results previously obtained by Hub1 and Hebner with respect to the free, total, and combined acids of beeswax ; further, they have det,ermined the iodine numbers for this substance, m d describe a process for estimating the alcohols present. This consists in the fusion of the wax with potaasium hydroxide and pot,a,sh lime a t 250°, which causes the evolution of Iijdrogen proportionally to the amount of alcohols acted on, and from the residue of this experiment the hydrocarbons existing in the wax are determined by extraction with a suitable solvent. Their results for pure, dry, washed beeswax are summarised :- M. p. 63-64'. Entirely soluble in hot chloroform.Wax Acids. Free acids correRponding with 19--21 m i l l i p m s KHO per gram. 7, 7, 97 13.5-15.5 per cent. cerotic acid. Total acids 9 ) 91-97 milligrams KHO per gram. Combined a c i h ,, 72-76 ,, 7, 7 9 9 , 9 , 9 7 32-85-34.67 per cent. palmitic acid Ratio of free io combined acid 3.5 to 3.8. Iodine Numbers. 100 parts of wa.x absorb 8-3-11 parts iodine ; which corresponds to 9-12 per ccnt. olejic wid. Wax Alcohols. Hydrogen liberated by fusion with KHO, 53.5-57.5 C.C. per gram. Wax Hydrocadbons. M. p. 49.5. Iodine fixed by 100 parts of Estimation of Resin in Soap. By R. WIT,T,TAMS (AnaZyst, 15, 81--82).-Gladding's method (Abstr., 1883, 603) yields vcry good remits. The author prefers to work on the soap itself rHther than on che acids separated from it.Estimation of Csmpkor. By F. FOERSTER ( B e y . , 23, 2981-- 2989).-A number of substances now occur in commerce, consisting Percentage 12-5-14. 11 y drocar-bon 22-03. T. G. N. M. J. S.232 ABSTRACTS OF OREMIOAL PAPERS of nitrocellulose and camphor, and up to the present no method is known for estimating the amount of camphor which they contain. The author proposes to carry out the estimation by distilling the substances with soda solution, when the camphor readily passes over. This may be then extracted with benzene, and the specific rotatory power of the benzene sol ution ascertained. Detailed instructions for carrying out the reaction, and tables of the rotation of camphor in benzene solution at different concentlrations and temperatures are given in the original.The results obtained are about 0.7-1.0 per cent. too low, probably owing to the difficulty of driving out the last portions of camphor. The autbor finds that sublimed camphor contains a small quantity of impurity, and that for the determination of its rotatory power, high temperatures must be avoided in its preparation, and the camphor finally twice recrystallised from 50 per cent. alcohol. It then melted at 174+3-173.3", and after six crystallisations a t 1 'i6*3-176-5", and after 10 crystallisations the solidifying point was found by Landolt's method (Abstr., 1890,l) to be 15'8.7" (con-.). The boiling point of the purified camphor was 209.1" under 739 mm. Estimation of Tannin in Tea. By P. MALTSCHEFFSKY (.I. P~CWWL. [ 5 ] , 22,270-271 ; from Pharm.Zed. f. RUSS., 29,12i).-The tanuiri is precipitated by meitus of normal copper acetate, and the excess of copper is titrated by the aid of potassium ferrocpnide solution, The copper solution contains 7.657 grams of copper oxide per litre (1 C.C. = 0.01 tannin), and its strength is controlled by evaporating a measured volume t o dryness, moistening with nitric acid, heating to redness, and weighing tbe oxide. The ferrocyanide solution is prepared by making up to 1 litre, 100 C.C. of a saiurated solution. To standardise this solution, it is added, 1 C.C. at R time, to 5 C.C. of the copper solu- tion diluted to 100 c.c., until a drop of the mixed liquids gives a blue colonr with a solution (1 : 100) of ferric chloride. A second assay, in which the additions of ferrocyanide solution are made by tenths of a C.C.towards the end, gives the exact strength of the solution. 2 grams of tea, dried at 100-107" is extracted four times with 100 C.C. of boiling water each t h e ; the filtrates are united, made up to 400 C.C. ; 100 C.C. of this solution is boiled and treated with 10 C.C. of copper solution. The precipitate is filtered oft', washed with hot water, and the filtrate and washiip are made up to 200 C.C. ; half of this is taken, and the excess of copper is determined approximately by means of the ferrocyenide solution ; the second half of tbe solution then serves for the exact determination of the copper. In 14 samples, the amount of tannin yaried from 6.10 to 11.08 per cent. The water varied from 5.59 to 12 48 per cent.; ash, 3.14 to 9.25 ; aqueous extract, 17.3 to 39.4; caffeine, 1.09 to 2.88 per cent. J. T. Estimation of Urea. By P. M~QUEL (Compt. rend., 111, 501- 502).-Many of the nrophagic microbes, and especially micrococci and earcinae, can develop in a neutral and even in tt slightly acid cnltivation fluid. Several grow solely a t the bottom of the vessels and produce more or leks granular deposits without rendering the H. G. C. .ANALYTICAL CHEYISTRP. 133 liquid turbid, whilst at the same time tbey produce a large quantity of the soluble ferment (this vol., p. 100 ). These clear liquids should be used for the estimation of urea. An aqueous solution of urea is simply mixed with the cultivation fluid containing the ferment; the alkalinity i s at once estimated by titration, and the liquid is heated at 50" for two hours in a well- closed vessel, which it nearly fills.The alkalinity is again deter- mined? and from the quantity of ammonium carbonate formed the amount of urea present is calculated. Urine and ohher organic liquids are previously heated with a sligbt excess of ammonium carbonate, filtered if necessary, and then mixed with the feiment, the object of this treatment being to prevent loss of ammonia from formation of double salts, neutralisation of any acid present, &c. A quantity of urea exceeding 10 per cent. interferes with the activity of the ferment, and in solutions o€ 30 per cent. the ferment is it mctive. Concentrated solutions must, therefore, be diluted. Am- monium carbonate, sodium chloride in small proportion, uric acid, arnmoniacal and alkaline salts, extractive matters, albumin, and sugar in large quantity do not interfere with the results.C. H. B, Simple Mode of Estimating Urea. By C. W. HEATON and S. A. VASEY (Analyst, 15, 106--107).-The method, which does not aim at great accuracy, is suggested for the use of medical men in cases where none of the special forms of apparatus is available. An &-ounce bottle is fitted with a thistle funnel and gas delivery tube which dips under water in a basin. In the bottle is placed 1 fluid drachm of bromine and 10 fluid drachms of a 40 per cent. solution of caustic soda. A bottle full of water is inverted over the delivery tube to receive the gas ; 2 fluid drachms of urine is then poured into the generator and rinsed in by 1 fluid dirtchm of water, and +,be bottle is shaken until gas ceases to be evolved. The receiver is then closed by the thumb, removed from the basin, placed in an upright position, and filled up with water, the volume required being noted. Deducting 200 minims for the volume of air dis- placed by tho urine and water introduced into the generator, the remainder is equal in volume to the nitrogen evolved, and each 100 minims corresponds with 0.25 per cent of urea.M. J. S. Rapid Method of Estimating Urea in Urine. By C. J. H. WARDEN (Analyst, 15, 201--203).-l'he apparatus iR a modified Crum's nitrometer, 630 mm. long and of about 75 C.C. capacity. Into its lower end is ground a stopper, on which 10 narrow grooves have been filed. The cup above the stopcock i s of 5 C.C.capacity, and is accurately marked a t 2.5 C.C. The tube is graduated to show percentages of urea a t once, asmmirig that 1 per cent. of urea in 2.5 C.C. of urine will yield 9.27 C.C. of gas. The hypo- brornite solution is stated to be made by dissolving 100 '' grains" (? grams) of caustic soda in 750 C.C. of water and adding 25 C.C. of bromine. The inverted tube is filled with this solution and the stopper inserted. Ifs cxterior and the cup are then rinsed and134 AUSTHACTS OF CHEhUCAL PAPERS. dried. It is stood in a vessel of brine and the stopper is removed. 2-5 C.C. of urine is then placed in the cup and there mixed with its own volume of saturated brine to increase its densify, and this mixture is allowed to enter the tube in small portions.The last traces are rinsed in by brine. The tube is then grasped by the right hand, the thumb being tightly pressed against tbe open end, aiid the contents thoroughly agitated. It is then transferred to a vessel of water, where the heavy sollitions flow away, and the volume of the nitrogen is read in the usufil manner. M-. J. S. Estimation of The'ine in Tea. By G. L. SPENCER (CZcex2, Centr., 1890, ii, 172 ; from J. Arner. Chern. Soc., 4, 158).-2 to 3 grams of the finely ground tea is extracted i n a small beaker seven times with boiling water, the extract being each time decanted off, and the residue finally transferred to a filter, and washed with a few C.C. of boiling water. Ha.sic lea*d acetate is added to the extract, about 8 C.C. usually being snfficient; the precipitate is filtervd, washed with hot water, and the lead sepayated as sulphide, after wbjch the filtrate is concentrated to about 50 c.c., with addition of about 5 grams of calcium hSdroxide or mngnesium oxide.The liquid is again filtered, the insoluble portion extracted with hot water, and the tiltrate is extracted with chloroform seyen times. 'I'he chloro- form extract is distilled from a tared flask, and the weight of the residual theine recorded after drying at 75". The method has been in use in the Department of Agriculture. J. W, L. Estimation of Quinine. By SEATON and H. D. RICHMOND (AnaZyst, 15, 42- 43) .-In soh tions containing quinine bisulphate dissolved in an acid, and free from salts whose base is pyecipitable by baryta, the quinine may be estimated by titration.Quinine bi- sulpbate is neutrai to methyl-o~ange, whilst the base itself has no action on phenolphthalein. To 25 C.C. of the solution there are added 2 drops of methyl-orange solution (0.25 gram iii a litre of water), and 2 drops of phenolphthalein solution (0.5 gram in a litre of 50 per cent. alcohol). Baryta solution (N/10) is then run in until the red colour changes to a brown, at which point all the free acid is neutralised. The addition of baryta is then continued until the pink colour of the phenolphthalein appears. As the pink colour develop8 slowly, care must be taken not to overstep this point. The number of cubic centimetres required for this second stage, multiplied by 0.0218, gives the weight of the hepta-hydrated quinine sulphate present.M. J. S. Reaction for Cocai'ne. By F. DA SILVA (Compt. rend., 111, 54$-349).-A small quantity of cocaine, or one of its salts, or of thc residue obtained by evaporating a solution, is mixed with a few drops of fuming nitric acid of sp. gr. 1.4, evaporated to dryness on the water-bath, and the residue mixed with 2 or 3 drops of concentrated alcoholic potasli. A distinct, and peculiar (;dour, recalling that of peppermint, iLi developed. In lhageudorfl's systematic scheme ofASALTTICAL CHEMISTRY. 135 analysis, cocaine is found among the alkaloidu extracted by benzene from an aqueous ammoniscal solution. Of the other alkaloids of the same group, atropine, hyoscyamine, strychnine, codeine, and eserine give colortttions wlien treated in the same way,and eseriiie also develops a disagreeable odour resembling that of phcn.ylcarbylamine.Del- phinine, brucine, and veratrine give only indistinct odours, which cannot be confonnded with that from cocaine. Sabadilline and narcotine can be recognised in the same way, but the other alkaloids give no sensitive reactions of this order. The reaction will detect 0.5 milligram of cocaine hydrochloride. C. H. B. Detection of Colchicine in Corpses. By N. OBOLONSEI (&it. a n d Chem., 29, 493).-The finely divided viscera are rubbed up with glass powder treated with oxalic acid, and digested for 2‘2 hours with alcohol. The liquid is squeezed out, and the dry residue twice mashed wit.h alcohol. The extract is concentrated at a temperature not exceeding SO”, and the cooled residue made up to the original volume with alcohol.The filtered liqriid is evaporated as before, nad this operation repeated until no clots separate on the addition of alcohol. The residue is then dissolved in water, the solution purified by shaking with light petroleum, and the colchicine finally extracted with chloroform as usual. Tbe alkaloid is best identified by means o€ the violet colour pro- duced by nitric acid ; by Erdmann’s reagent (nitrosnlphuric acid), which gives in succession green, dark-blue, violet, and yellow colourti, turning to raspberry-red on adding alkali ; also by Mandelin’s reagent (1 gram of ammonium vanadate in ‘200 grams of siilphuric acid) which gives a green colour. Colchicine is with difficulty destroyed by putrefaction of animal matter.The kidneys, bladder, and urine are best suited for forensic examination. M. J. S. Detection of Bile Constituents in Urine. By A. JOLLUS (Zeit. a n d Chenz., 29, 402 -406).--Of the various tests proposed for detecting bile pigments in mine (Grnelin’s, Huppert’s, Vitali’s, Rosenbach’s, Ultzmann’s, Hoppe-Seyler’s, Dragendorff ’s), those of Rosenbach and Huppert, with the following modifications, give the best results :- Rosenbach’s Test.-A large quantity of the urine is filtered through clean, white filter-paper, the interior of the filter is touched with a drop of strong nitric acid containing nitrous acid, and the funnel is gently warmed over a flame. After a few minutes a green ring is formed round the spot moistened by the nitric acid. Huppert’s Test.-About 10 C.C.of the urine is shaken with an equal volume of milk of lime containing 10 grams of calcium oxide in the litre. The success of the test depends on the proper concentratioti of the milk of lime. The precipitate is filtered off and washed into a test tube with alcohol and dilute hydrochloric acid, then filtered, and the filtrate boiled, With only traces of bile pigments, the liquid becomes green to blue. An estimate of the amount of bile con- stituents can be obtained from the iodine number oE the urine. If g136 ABSTRACTS OF CHEMICAL PAPERS. is the number of grams of iodine absorbed by 10 C.C. of the urine, and s the specific gravity, the iodine number is - The number s-1 for normal urine, filtered after cooling, is 6.3 to 8.1, though even in specimens rich in uric acid it rarely exceeds 7.8.The presence of even traces of bile pigments raises the uumber to 9.6, and values as high as 17.4 have been observed. New Test for Albumin. By A. JOLLES (Zeit. ai2aE. Chem., 29, 406--407).-About 8 or 10 C.C. of albuminous urine is mixed with an equal volume of concentrated hydrochloric acid, and then 2 or 3 drops of a saturated solution of bleaching powder deposited quietly on the surface. If a8 little as 0.01 gram of albumin per 100 C.C. is present, a white turbidity appears at the surface of contact. This test, being less sensitive than that with nitric acid, which latter will detect 0*0015 gram per 100 c.c., may be used to find approximately the proportion of albumin present, since by diluting the urine until the one test gives an iiidication but the other none, the percentage may be known to lie between the above minimum limits.M. J. S. M. J. S. Detection of Albumin in Bacterial Urines. By A. JOLLES (Zeit. anal. Chem., 29, 407--408).-The most sensitive test for albumin in urine is that with acetic acid and potassium ferrocyanide, the lower limit of which is 0.0008 gram in 100 C.C. It is, however, necessary to filter the urine to obtain a standard with which to com- pare the turbidity produced by the test. When bacteria are present, ti clear filtrate is best obtained by shaking with infnsorial earth before filtering. In the case of purulent., slimy urines, rich in leuco- cytes, traces of albumin may adhere to the precipitate ; but by wash- ing this with warm potash, and testing the filtrate, the smallest traces of albumin may be detected.M. J. S.ANALYTICAL CHEMISTRY. 107A n a l y t i c a l Chemistry.Estimation of Hydrogen Chloride in Solutions of Hydroxyl-amine Hydrochloride. By J. A. MULLEK. (Rzill. Xoc. Chim. [ 3 ] , 3,605).-Phenolphthalei'n is unaffected by solutions of hydroxylaminehydrochloride, and the amount of acid present may be estimated bymeans of a standard solution of sodium hydroxide, free from carbon-ate, using phenolphthalein as the indicator. Pyridine, picolines, andlutidines behave similarly. T. G. N.Estimation of Sulphur in Inorganic Sulphides. By L. BLUM(Zeit. anal. Gem., 29, 411412).--The method publiahed byJannasch (Abstr., 1889, 1244, and 1890, 1187) is not new, havingbeen already brought forward by Sauer, in 1873 (Abatr., 1873, 939).M.J. S.Estimation of Nitrogen by the Schultze-Tiemmaam(Schloesing's) Method. By F. COGHIUS and T. MOELLER (Chern.Zeit., 14, 3%).-Low results are obtained by this method, especially inthe examination of explosives. This is attributed by the authors tothe addition of too much water, and to the want of proper relation-ship between the size of the apparatus used and the quantity of mate-rial employed. In some test experiments they used a long-necked350 C.C. flask, a measuring tube of 150 C.C. capacity, ferrous chloridesolution containing 70 grams in 100 grams of water, hydrochloric acidof about 37 per cent., and employed 0-3-0.4 gram of potassium nitrate,5-15 C.C. of the ferrous chloride solution, and twice the quantity ofthe hydrochloric acid.The ordinary course of operation was followed,avoiding unuecessary boiling to drive out the air. When 25 to 50 C.C.of water was added, the analysis lasted 30 to 40 minutes, and theresults varied between 13.76 and 13.86, mean 13.81, whilst with 80to 150 C.C. of water the variation in the results was from 13.05 to1339, mean 13.21, and the analysis lasted 70 to 90 minutes.D. A. L.Estimation of Nitrogen in Sodium Nitrate. By 0. FOERSTER(Chem. Zeit., 14, 509-510 ; compare Ahstr., 1889, 547, 746).-Twoor three grams of the nitrate is dried at 150" or by heating to inci-pient fusion, weighed, and repeatedly evaporated to dryness on a water-bath in a tared crucible, with 25 C.C. of about 19 per cent.hydro-chloric acid. After about the third evaporation, the nitrate is com-pletely converted into chloride, which is dried at 150°, ignited slightly,and weighed, and the nitrogen calculated from the difference. Themethod yields satisfactory results, but only in the absence of othersubstances, which would be attacked by hydrochloric acid.D. A. L.Estimation of Nitric Nitrogen as Nitric Oxide. By F.SCHEIDING (Chem Zeit., 14, 635-637).-For estimating nitric nitrogenas nitric oxide, the author has devised and employs the apparatusshown in the drawing, which is provided with a measuring tub108 ABSTRAOTS OF CHEMIOAL PAPERS.having a globular expansion, a glass tap H with a small funnel at thetop, and a special arrangement intimately attached by india-rubbertubing or fusion to the bottom.In operation, tube L is connectedin a suithble manner with a, movable reservoir containing sodiumhydroxide, sp. gr. 1.25, with which the apparatus is charged to thelevel of 4 by raising the reservoir, and clip 4 is closed. The substanceis placed, along with a little water, in a 200-250 C.C. flask, to whichthe stopper and tubes are fitted, connections made, and to expel the airthrough tubes 2 and 3, the water in the flask is boiled until thewater into which tube 3 dips is caused to boil by the issuing steam,clip 5 is then closed, and the air still in tube 1 driven into the mea-suring tube by opening clip 4, which is again closed, and the flameremoved from below the flask. The measuring tube is filled to thetop with sodium hydroxide, and tap H is closed.20-25 C.C. of coldsaturated ferrous chloride, and then 8-10 C.C. of concentrated hydro-chloric acid are carefully drawn into the flask through tubes 2 and3, which are then washed with water in the same way; the flask,suspended a few cm. above the wire gauze, is now heated, and aANALYTIGA L OHEMISTRY. 109soon as a, pressure i R indicated in the india-rubber tube at 4, that clipis opened, t,be nitric oxide passes into the measuring tube, and by thetime the liquid has yolatilised in the flask, all the nitric oxide is con-chided to be in the measuring tube. The temperature in the jackettube surrounding the measuring tube is made to correspond with thati n the vicinity of the bulb, and the level in the reservoir beingadjusted to that in the measuring tube, the volume is read off, andafter the necessary corrections are made, the percentage of nitrogen is.calculated therefrom.The saucer under the measuring apparatus isfilled with water to keep the tubes immersed in it cool. For substanceswhich might be deromposed by boiling with water, a tap funnel isfitted to the flask, and is used for charging it. D. A. L.Estimation of Nitrogen in Organic Substances by means ofAlkaline Permangmate. By R. L. WAGNER (Chem. Zeit., 14. 269).-The autbor some years ago recognised the possibility of oxidisingnitrogenous organic substances by means of alkaline permanganate,without the formation of ammonia. In his experimcnts he mixed0.5 to 1 gram of substance with 25-30 times its weight of potassiumpermanganate, and 5 C.C.of 25 per cent. potassium hydroxide, placedthe mixture in a tube closed at one end, terminating at the other in acapillary for the escape of oxygen, warmed in a water-bath to aidadmixture, and theu heated at 150-170" in an air-bath for two totwo and a half hours. The conQents of tbe tube were turned into aporcelain basin, the excess of manganate reduced with manganesesulphate and sodium carbonate, and the nitric acid deteimined inthe clear liquid by a modification of Eder's method ; but irregularityof combustion and breaking of tubes rendered the method practicallyuseless, except, perhaps, for substances soluble in alkalis. Non-vola-tile nitro-derivatives and ethereal nitrates can be safely oxidised byalkaline permanganate in a porcelain dish, excess of permanganatebeing subsequently reduced with alcohol, and the diluted filtratetreated with ferrous sulphate, zinc-dust, and hydrochloric acid ; thenitrogen is then estimated as ammonia, by any of the usual distilla-tion methods.Carbon bisulphide and thiophen can be oxidised by similar treatment,and the sulphur estimated in them; they are enclosed in thin glassbulbs, and placed in tubes containing the alkaline permanganate ; thetube is sealed up, the bulb broken, and tbe digestion proceededwith. D.A. L.Detection of Foreign Raw Phosphates in Powdered BasicSlag. By L. BLUM (Zeit. anal. Chem., 29, 408--411).-The relativesuperiority of basic slag as a fertiliser over natural phosphaticminerals, owing to its ready absorbability, and the high price whichi t has in consequence atkained, have led to its falsification with otherraw phosphates.Only such are likely to be used as, from their lowpercentage of phosphoric acid, cannot profitably be worked a p assuperphosphate, and these in most cases contain much calcium carb-onate. Fresh basic slag is almost absolutely free from carbonates,and even ou long exposure to air, absorbs very little carbonic aci110 ABSTRAOTS OF OHEMIOAL PAPERS.(2.47 per cent. was found in an extreme case), so that the presence ofmuch carbonate i n a specimen would be enough to throw suspicionon it. A low percentage of iron and manganese might furnish anadditional indication, since tlhese metals are rarely present in naturalphosphates.In estimating the carbonic acid by decomposition withan acid, some chromic acid should be added, to prevent evolution ofhydrogen aulphide from the sulphides present, but a simple estima-tion of the loss on ignition would generally allow an opinion to beformed. M. J. S.Estimation of Water in Superphosphates. By J. STOKLASA (Zeit.anal. Chem., 29, 390--397).-Pure monocalcium tetrahydrogen phos-phate, CaH4(P04)2 + HzO, loses its water of crystallisation at loo",but only completely after 40 hours. It may be kept at 105" for20 hours with but little change, but on longer heating at the sametemperature begins to show decomposition. At higher temperatures,the amount of change is dependent not alone on the temperature,but also on the time of drying.The statement of Drewsen (Abstr.,1881, 465) that drying even at 300" does not diminish the proportionof soluble phosphate, but merely reduces it to a soluble pyroyhosphate,cannot be confirmed for pure or nearly pure monocalcium phosphate.I t might be true for a superphosphstc in which free phosphoric acidconstituted 80 per cent. of the total soluble phosphoric acid.On drying for one hour at ZOO", one-half of the monocalciumphosphate undergoes decomposition, thus :-4Cs€&( PO,), = Ca?,P207 + Ca(P03), + CaH2P207 + 2H3POa + 4H20. At lower tempera-tures for the same length of time, the proportion decomposed issmaller, but if the time is prolonged, a further decomposit'ion takesplace ecen at 150", and less free phosphoric acid is fouud in the solublepart.A temperature of 200" sufficiently prolonged results in thefollowing decomposition : 4 C a H 4 ( PO& = 3Ca(P03)2 + CaH2P207 + 7H20, whilst at 210", there remains nothing but insoluble, glassycalcium metaphosphate. I n presence of free phosphoric: acid, thecontrary action may on heating take place, thus: ChP,O, + 2H,PO,= 2CaH2P207 + H,O, and thus the soluble phosphoric acid actuallyundergo increase. M. J. S.By R. FRESENIUS(Zeit. anid. Chem., 2 9 , 4 1 3 4 3 0 ; see Abstr., l890,924).-A11 attemptsto Qbtain complete separation by niems of chvomic acid in a single pre-cipitation resulted in failures. The seemingly satisfactory separationobtained by Frerizhs and by Russmann (next abstract) resuited fromthe accidental compensation of opposite errors, since they washed thebarium chromate wibh acetic acid, in which it is distinctly soluble,and weighed it after drying a t 110", at, which temperature it stillretains some moisture.In a solution containing alkaline acetate anddichromate, barium chromate is, however, quite insoluble. It canalso be rendered anhydrous without decomposition by ignition at adull red heat, even the portion adhering to the filter reoxidising aftertemporary reduction. By double precipitation of the barium, a com-plete separation can be effected even when the proportion ofSeparation of Barium from StrontiumANALYTICAL CHEMISTRY. 111strontium is large. The solution of the chlorides is feebly acidifiedwith acetic acid, and diluted until it contains not more than 0.5 percent.of the bases, then precipitated hot with a n excess of ammoniumchromate, which has been carefully neutrahed with ammonia, Aftercooling €or an hour, the precipitate is washed by decantation withvery di1ut.e ammonium chromate until the washings no longer give aprecipitate with ammonium carbonate, and then further with warmwater until the washings are scarcely coloured by silver nitrate. Theprecipitate is then dissolved in the smallest possible quantity ofnitric acid, and the solution again diluted and heated. Ammoniummetate is added in sufficient quantit-v to displace the free nitric acidby acetic acid, and then ammonium chromate until the odour of aceticacid bas completely disappeared. After an hour, the liquid is pouredthrough a filter, the precipitate is digested with hot water, cooled,filtered, and washed thoroughly with cold water.It is then free fromstrontium, whilst the filtrates contain no barium. Double precipita-tion from neutral or alkaline solutions has not been successful.M. J. S.Separation of Barium, Strontium, and Calcium. By A.RUSSMANN (Zeit. anal. Chent., 29, 447-454; from Tnaug. Diss.Berlin, 1887).-Barium cannot be satisfactorily estimated by Frericbs’method (precipitation from an acetic acid solution by normalpotassium chromate), since the filtrate always contains traces ofbarium, and some potassium chromate is carried down by the preci-pitate. The precipitate will also contain strontium, if the proportionof the strontium in the solution exceeds 30 parts per 100 of barium.Calcium is not so precipitated.The simplest way to ascertain theweight of the barium chromate, is to dissolve it in dilute hydrochloricacid, add potassium iodide, and immediately tit rate with thiosulphate.Diehl’s method for separating barium and calcium by digesting thesulphates with sodium thiosulphate solution is complicated by somany sources of error that it cannot be recommended. Fresenius’method of separating barium and calcium by dilute sulphuric acid ina solution acidified with hydrochloric acid is thoroughly satisfactory.The method of Sidersky (Abstr., 1883, 509) for separating strontiumand calcium only yields approximate results.For separating bariumand calcium, it is, however, serviceable. Bloxam’s method (Abstr.,1886, 920) is not suitable for quantitative separations, as thestrontium sulphate carries down with it considerable quantities ofccLlcium, and the calcium ammonium arsenate cannot readily bebrought into a, form for weighing in which i t contains a constantproportion of calcium. Fleischer’s method for separating bariumand calcium by digestion with 3 parts of potassium sulphate and1 part of carbonate, followed by titration of the calcium carbonate inthe weighed precipitate gives good results. Lastly, Leison’s met.bodfor the estimation of the individual alkaline earths, by precipitationwith oxalic acid and alcohol, and titration of the oxalic acid in theprecipitate by permanganate, is accurate.The barium oxalate mustbe dissolved by hydrochloric acid, as it is not completely decomposed by.sulphuric acid. Strontium and calcium oxalates can be decomposedby sulphuric acid. The solutions must not be filtered throngh paper112 ABSTRACTS OF CHEMIOAL PAPERS.and must be highly dilute. Ignition of the oxalates is, however, asa rule, the quicker process.Estimation of Cadmium in the Products of Zinc Manu-facture and in Calamine. By W. MINOR (Chem. Zeit., 14, 4, 34,and 348-349) .-The material is dissolved in hydrochloric acid,treated with hydrogen sulphide, and the precipitate washed with hot,water, dissolved in hydrochloric acid, heated to boiling, and pouredinto dilute sodium hydroxide likewise heated to boiling.This preci-pitate, after washing with hot water, is ignited in a current of oxygen,and weighed as cadmium oxide. Material containing but little iron,such as “ pure cadmium,” is dissolved in hydrochloric acid, and preci-pitated directly with the sodium hydroxide. This method ofprecipitation may also be used to separate zinc and cadmium i n t,heordinary method of examining calamine ; the ammoniacal solutioncontaining the zinc and cadmium is rendered slightly acid andpoured hot into the hot hydroxide, &c.In the method described in the last of the three papers, the materialis dissolved in hydrochloric acid, filtered from undissolved lead,precipitated with hydrogen sulphide, the precipitate, containingzinc and an inconsiderable amount of arsenic, is washed, dried,weighed, dissolved in dilute hydrochloric acid, and treated withsodium hydroxide in excess.The cadmium hydroxide is filtered off,and the zinc titrated in the filtrate with sodium sulphide, calculated .to zinc sulphide, and deducted from the weight of the cadmiumsulphide precipitate. I n another method (requiring the absence ofother metah precipitated by sodium hydroxide) after removal ofiron with ammonia, the solution of zinc and cadmium is nearlyneutralised with hydrochloric acid, and then treated with sodiumhydroxide. The precipitate of cadmium hydroxide i R dissolved indilute hydrochloric acid, evaporated to dryness, dissolved in water,and titrated with standard sodium hydroxide, using litmus or sodiumsulphide papers as indicators.Good results have been obtained byboth methods, the first being the more suitable in the presence ofmuch zinc and vice versc2. D. A. L.M. J. S.Estimation of Cadmium as Sulphide by Precipitation withSodium Sulphide Solution. By W. MINOR (Chem. Zeit., 14,439-440) .-The material is dissolved in hydrochloric or nitric acid,and the lead separated by sulphuric acid ; the solution is then treatedwith soda, and the precipitate digested with ammonia. The ammoniacalsolution is free from lead, zinc, and iron, but contains all the cadmium,which can then be determined by means of sodium sulphide solution,either volumetrically by tit ration, nsing ferric hydroxide as indicator,or gravimetrically by precipitating, and weighing the precipitateafter drying for some hours at 140-145O.D. A. L.Volumetric Estimation of Zinc and Copper. By E. DONATRand G. HATTENSAUR (Chenz. Zeit., 14, 323--325).--Various experi-ments have been made by the authors. They find that for titratingzinc by Schaffner’s method, it is better to use sodium hydrosulphidANALYTICAL CHEMISTRY. 113(prepared by adding a known volume of dilute sodium hydroxide toan equal volume of the same solution previously saturated withhydrogen sulphide) than a solution of the crystalline sulphide ofcommerce ; howerer, in using this reagent in solutions containingtartaric acid and ammonia, the iron commences to precipitate beforeall the zinc is converted into sulphide. The estimation of zinc by usingexcess of ferrocyanide, after the removal of the iron, and titrating backwith permanganate does not answer, since in the cold a clear solutioncannot be obtained, whilst if warm, decompositions occur whichcause irregularities.It is noticed that ferrocyanide precipitates zincbu h not iron in the presence of tartaric acid and ammonia, and that theexcess of either of these substances does not seriously disturb therelative quantity of 1 mol. ferrocyanide to 2 atoms of zinc ; therefore1 C.C. of a solution containing 33.5 grams of potassium ferrocyanideper litre corresponds with 0.010 gram of zinc. As small an excess ofammonia as possible, and a hot solution, are favourable to the preci-pitation. The zinc precipitate is not decomposed by acetic acid;therefore, by placing drops of this acid and the solution under exami-nation in contact, in the presence of iron, a coloration indicatescomplete precipitation of' the zinc.The following method it; based onthese considerations :--3 t o 4 grams of material is dissolved in hydro-chloric acid with some nitric acid, diluted to a definite volume withwater, an aliquot part filtered, treated with 20-25 C.C. of concen-trated tartaric acid solution, a slight excess of ammonia added, andthe liquid warmed to about 80". The ferrocyanide is now run in untilthe precipitation of the zinc is complete, as indicated in the mannerdescribed above. The proportion of iron to zinc ili the solution underexamination should be the same as that present in the solution usedfor standardising the ferrocyanide.Under similar circumstances, copper is precipitated in a likemanner, but the precipitation is greatly influenced by ammonia ;fherefore the solutio~r for titration should be neutral or nearly so.The ferrocyanide is standardised from a solution of copper of knownstrength, and cannot, be approximated to by the weight of fcrrocyauideemployed, inasmuch as the composition of the copper precipitate isuncertain.Copper and zinc may be estimated in the same solutionby this method ; first both are titrated, then the copper is precipitatedout of another portion of solution, and the zinc alone titrated, &c.D. A. L.Estimation of Lead by Phosphomolybdic Acid. By H. BEUF(Bull.. Soc. Chinz. [3], 3, 852--855).-To the boiling neutral solutionof the metal, an aqueous solution of phosphomolybdic acid is addeduntil the supernatant liquid is coloured yellow by the excess ofreagent used.After washing, the precipitate is dried at 90-100°and weighed. It forms a dense, white powder which is insoluble inwater (1 in 500,000) and aqueous ammonia, but dissolves in nitricand in acetic acids ; it contaiiis 54.8 per cent. of lead, and correspondswith the formula Mo,Pb2,P2H,,0,,,; at a high temperature it loses7 mols. H20.By decomposition of the precipitate with dilute sulphuric acid andeinc at a gentle heat, a brown liquid is obtained, wbich may beYOL, LX. 114 ABSTRACTS OF CHEMIUAL PAPERS.titrated for lead by a solution of permanganate which has beenpreviously standardised against a solution resulting from tho similartreatment of a known weight of a lead salt.The phosphomolybdic acid is made by evaporating to dryness asolution of ammonium phosphomoly bdate in nitric acid.Iron iseliminated by a previous treatment with sodium hydroxide, copper,potassium, and ammonium by washing the mixed pbosphomolybdateswith ammonia-water, but the presence of zinc or arsenic vitiates theestimation. T. G. N.Separation of Copper from Arsenic by the Electric Current.By L. W. MCCAY (C‘laem. Zcib., 14, 509).-Under the influence ofthe current from four to six Meidinger elements, alkaline arsenatesremain in solution, whereas copper is completely and quantitativelyprecipitated, and has been estimated wikh good results.Moreover,t h e copper is quite free from arsenic, and the solution may be safelyemployed for the determination of the original amount of the lattermetal. D. A. L.Estimation of Aluminium in Commercial Aluminium. By0. KLEMP (Zeit. anal. Chsm., 29, 388--390).-The prccess employedfor zinc (Abst , 2890, 1190) cannot be applied to aluminium since,always evolved, but bv dissolving the aluminium in potash, and burn-ing the hydrogen in Fi=eserlins’ apparatus, a very accurate estimationcam be made. About 1 gram of the metal in filings is placed in a150 c c. flask with a little vaselin to prevent frothing, and the potashsolution (35 grams of KOH in ‘LOO c.c.) is added gradually, withwarming towards the close.Estimation of Alumina in Bread, &c.; Solubility of Alu-minium Phosphate in Acetic Acid.By W. C. YOUNG (dnalyst,15, 61-63 ; 83-S4).-1n Duprk’s process, the aluminium is precipi-tated as phosphate from an acid solution containing ammoniumchloride and acetate, and is collected after remaining all night in thecold. Test analyses, with weighed quantities of alum, show thatunder these conditions the resalts are much below the truth. Thebest result (from a mixture of alum, sodium phosphate, and aceticacid) was obtained by boiling the mixture both before and after theaddition of ammonium acetate, and filtering immediately. Theamount of ammonium acetate must not be too small, nor that ofacetic acid too large. For 0.1 gram of potash alum, there mas used1 gram of ammonium acetate and 5 C.C.of ordinary acetic acid. Thepresence of ammoriiuni chloride has little effect when the liquid isfiltered immediately after boiling, but lowers the result i E the pre-cipitation is performed iu the cold, or the mixture is allowed to coolbefore filtering. M. J. S.Estimation of Iron Oxide and Alumina in Phosphates. ByR. JONES ( C h e w Zeit., 14, 269--271).-The author criticises themethods employed until recently, and recommends as the bestthe following Combination of Glaser’s method, slightly modified, witheven with big ;i; ly dilute alkali and a large excess of iodate, hydrogen isThe operation takes about 45 minutes.M. J. SANALYTICAL CHEMISTRY. 115part of Stutzer's method. The phosphate is dissolved in hydrochloricand nitric acids, made up to a definite volume, an nliquot parttaken and treated with ft quarter of its volume of sulpluricacid of sp.gr. 1.84, and its om11 volume of 95 per cent. alcohol,making up to definite volume with the alcohol ; after 12 honrs,the gypsum is collected, and when weighed, gives good results forcalcium i n absence of much magnesium. The alcohol is driven offfrom the filtrate, which is rendered alkaline with ammonia andboiled to completely drive off the ammonia. The precipitate may beweighed, and half the weight taken as iron oxide and alumina, whichgives good results, or it may be treated with molybdic solution, thephosphoric acid 'separated in the usual way, the iron and aluminaprecipitated wlth dilute ammonia, redissolved in hydrochloric acid,reprecipitated, &c., and weighed.Titration of Chromates, Barium Salts, and Sulphates.ByP. SOLTPIEN (Clteni. Centr., 1890, ii, 217-218 ; from Pharm. Zeit.,35, 372) .-The titration of solutions of barium salts with potassium&chromate, and inversely the titration of chromates with bariumsalts, may be readily performed with either hzematoxylin or logwoodextract as indicator. A solution of barium chloride is prepared equalto one of potassium dichromate, and for the determination of bariumsalts, potassium dichromate is run in from the burette until a dropplaced on a warmed porcelain plate with a drop of haematoxylin justshows the formation of a blue-black coloration. The solution to betitrated must be neutral, and may not contain more than the meresttrace of either acetic acid or ammonia.Chlorides and nitrates do not in-terfere with the reaction, nor does rosolic acid, which latter may be useda s an indicatlor for the titration of soli~tions of salts of barium. If thesolution of a chromate contains sulphat,es, the titration with barhimchloride gives the total qnaritity of the two salts, from which mustbe deducted the amount of the latter as determined gravimetrically.For the determination of the combined snlphuric acid, an excess ofbarium chloride solution is added, and the excess determined bytitration with pot.assium dichromate.Salts of aluminium, copper, and iron must be removed from thesolutions. J. W. L.Estimation of Antimony by Marsh's Method.By A. VANBYLERT (Ber., 23, 2968-2971 ; compare Kuhn and Saeger, Abstr.,ISYO, 1187).-For the estimation of antimony in alloys of tin,silver, and antimony, the author recommends the following process :-A three-necked Woulffe's bottle is connected on the one hand with annpparatus for evolving carbonic anhydride, and on the other with thuusual calcium chloride tube and hard glass tube. The central neckof the Woulffe's bottle is fitted with a wide tube reaching to thebottom of the flask. About 0.5 gram of the alloy is dissolved in20 C.C. of mercury at 60", and poured, after cooling, into the fiask.100 C.C. of 10 per cent. sulphuric: acid is then added through one oLcthe side tubes, and the air expelled from the apparatus by carbonicanhydride.A freshly prepared sodium amalgam, obtained by dis-solving 5.5 grams of sodium in 25 c . ~ . of mcrcury, is then aclded dro]~D. A. IJ.i;116 ABSTRACTS OF CHEMICAL PAPERS.by drop through the wide tube. The sublimate of antimony quicklyappears in the heated tube ; the apparatus is then periodically shaken,care being taken that no liquid is allowed to pass into the centraltube. After the evolution of gas has ceased, carbonic anhydride isagain passed tbrough the apparatus t o expel all the hydrogen andhydrogen antimonide. The mercury solution is then poured off,dried with filter-paper, and divided into two equal portions, one ofwhich is returned t,o the cleaned and dried apparatus, and coveredwith 75 C.C. of a 10 per cent. sulphuric acid, whilst the other is mixedwith 3 grams of sodium, and added drop by drop through the centraltube as before.The results obtained are fairly accurate, but might possibly beimproved by employing lirdrogen in place of carbonic anhydride fordriving out the air. Another soiirce of error is the oxidation whichtakes place in the manufacture of the alloy, and during its solution inthe mercury.H. G. C.Estimation of Hardness of Natural Waters. By E. JJ. NEUGE-JMJER. (Zeit. anal. Chem., 29, 399-401 ).-The author proposes thefollowing modifications of Clark's test. The standard water is amixture of 8 vols. of calcium sulphate solution of 12" of hardness with2 vols. of a 12" magnesium sulphate solution. Of this mixture,100 C.C. is used. The soap solution is of such strength that 12 C.C. isrequired for the 100 C.C.of standard water. The following new tablehas been drawn up from titrations of the standard water diluted tothe required degrees :-Hardness.. .......... 0" lo 2" 3" 4" 5" 6"C.c of soap solution . . 0.6 1.7 2.8 3.9 4.9 5.9 6.9Hardness.. .......... I 8" 9" 10' 11" 12"C.c of soap solution ,. 7.8 8.7 9.6 10.4 11.2 1200F Oand a, special burette (titanometer) constructed, the readings of whichgive a t once the degrees of hardness.The Analysis of Sulphurous Waters. By D. VITALI (Chem.Cen.fr., 1890, ii, 166 ; from L'Orosi, 13, 73-778).-Thiosulphates maybe detected by the addition OE potassium nit,rite and a mineral acid oracetic acid. Nitric oxide is liberated, and the solution is therebycoloared yellow ; later, sulphur is precipitated, and the liquid becomesmilky.'Ibis reaction is extremely delicate, 0.0001 per cent. of thio-sulphate being detectable. I n testing for nitrites in presence of thio-sulphates with potassium iodide and starch, this reaction of the tbio-sulphate may prevent the formation of the blue iodide of starch, ayellow coloration being produced instead. In testing for iodine, inpresence of thiosulphates, with potassium nitrite and an acid, it isbetter to evaporate the water to dryness after neutralising withsodium carbonate, and then to extract the residne with absolutealcohol, which dissolves the iodide, leaving the thiosulphate undis-solved. In the case of waters containing large quantities of calciumsnlphnte, it is recommended to first precipitate the calcium as calciumcarbonate before determining the silicic anhydride. Xu order toprevent tho precipitation of sulphur during the evaporatior of waterM.J. SANALYTICAL UHEMISTRY. 117for determination of the total solid residue. the author recommendsthat a current of hydrogen should be passed through the water.J. W. L.Examination of Water for Contamination by Gas Works,By F. DICKMAEN (Zeit. miat. Cirem., 29, 398--399).-1n a specimen ofwater from a brook which had suffered contamination from a neigh-bouring gas works, and by which poultry had been poisoned, thoauthor detected traces of a substance giving the reactions of di-pbenylamine. Owing to the stability of this compound and thesensitiveness of its reaction with nitric acid, its presence might beused as a proof of contamination by tar-water, if i t should be foundto be a constant constituent of that liquid.Detection and Estimation of Organic and Inorganic Poisonsin Corpses.By A. SEYDA (Chem. Zeit., 14,31--32, 51-53,128 -129,181-184, and 198-200) .-The author describes his system ofexamining corpses. The chemical examination proper is preceded bya preliminary examination of the blood, urine, and contents of thestomach and accessories.The blood, when not too decomposed to show the absorption bands,is examined spectroscopically ; if dry, it is dissolved in water, and madevery slightly alkaline with sodium hydroxide ; the two oxyhaemoglobinbands amd the intermediate haemoglobin band merge into one con-tinuous band in partially decomposed blood; the faint band in thered is attributable to methaemoglobin in alkaline material or tohmmatin in acid blood.In the former case, fiwther examination isonly made under exceptional circumstances, such 8s the presence ofhydrogen sulphide not emanating from ordinary putrefactive pro-cesses ; but, in the latter case, the red band being due to hsematin, isregarded as indicating the presence of other reducing agents, or ofacids, potassium chlorate, ferricyanides, nitroglycerol, or nitro-benzene, which are tested for in the urine and in parts of the body.Other isolated lines in the red are carefully noted, and carbonic oxideis sought for ; the presence of hsmatin, already reduced or otherwise,indicates the absence of carbonic oxide poisoning ; carbonic oxide haemo-globia is recognised in presence of oxyhEmoglobin and hEmoglobinby the mere displacement of the absorption band towards the red partof the spectrum when ammonium sulphide is added ; chemical testswith sodium hydroxide with or without calcium chloride are con-sidered of little value.In the urine :-Notes are taken of the quantity, colour, odour, of re-actions as to the presence of blood, albumin, and sugar, of the action inalkaline copper solutions, of the behaviour toyvards barium chloridebefore and after treatment with hydrochloric acid, and of the occur-rence of balsams, alkaloids, soluble metallic poisons, and of such saltsas potassium chlorate, iodide, or bromide.The contents of tlie stomach are examined in the dark for phosphor-escence, which is only due to phosphorus in acid mixtures, otherwisefungoid growth may cause it.The, udour of the contents of thestomach may be acid, alkaline, like decayed cheese, or they may havea specific odour, or be putrid ; sometimes they are well preserved, andM. J. S118 ABSTRACTS OF OHEMICAL PAPERS.have a sweetish, repulsive odour, observed by the author in cases ofar.ienical poisoning. Food iizagmas and vomits are carefully sampled,treated with alcohol and then with ether, and examined macro- andmicro-scopically. The alcoholic extract is examined for oxalases andforeign bases and acids ; matters found in the folds are also examined.The arsenic test is applied, and an aqueous extract of the magma isexamined f o r soluble poisons.For the chemical examination proper, parts of organs, finely divided,arc heated with water in a boiling water-bath for several hours,acidified q-ith tartaric acid, and distilled with steam.Two fractionsare made: the first will contain the more volatile matters, such asalcohol, aldehyde, acetone, chloroform, nitrobenzene, ethereal oils,turpentine, camphor, amines, and their volatile corn binations ; theother, the remains of these more volatile products, and any lessvolatile substances, such as fatty acids, phenol, hydrocyanic acid, &c.A plain distillation of alkalilie material is required at times ; if, how-ever, the presence of chloral bydratc or hydrocyanic acid and ferro-cyanides is suspected, the finely divided orgaris are first moistenedwith potash or with hydrogen sodium carbonate respectively.Asteam distillation from alkaline solution, when necessary, is bestejfected, not directly, but with the liquid obtained by extraction withwater and tartaric acid. The residue from the acid distillat'ion isemployed in testing for alkaloids.The odour, colour, opdesceiice, quantity, &c., of the first fraction ofthe acid distillate are noted. I t is tested with silver nitrate, both innitric acid and in ammoniacal solution, with sodium nitroprusside,potash, and acetic acid ; with alkaline permanganate ; with iodine andpotash (iodoform reaction) ; with zinc-dust and hydrochloric acid fornitrobenzene ; with hydrochloric acid and alcoholic phloroglucinol forotheyeal oils (a reaction frequently takes place, but too much depend-ance is not to be placed on it; it is better to examine the urine for someof these substances, also for altered camphor).The reagents for tur-pentine oil, in tangible quantities, are a mixture of fresh guaiacumalcohol and fresh citronella oil. Resorcinol and potash are betterreagents for detecting chloroform than the isonitrile test. A quantita-tive examination for these substances is seldom possible, with theexception of alcohol, but this cannot be estimated by distillation in thepresence of amines ; therefore it is oxidised to acetic acid as follows.A portion of the distillate is redistilled, the first portions collectedare dried with potassium carbonate, distilled again, treated withsodium dichromate and sulphuric acid, rendered alkaline with potash,boiled t o eliminate amine bases, then acidified, steam distilled, andthe distillate, containing the acetic acid, titrated ; any sulphuric acid,if accidentally present, being estimabed and allowed for.The authorat,taches great importance to the estimation of alcohol, especially inthe case of children. Various parts of corpses of persons addicted t calcohol yield distillates containing alcohol, but not pure ethyl alcohol ;therefore such distillates reduce alkaline permanganate, &c., give theiodoform reaction, and yield an inflammable distillate on redistillationfrom potassium carbonate. Alcohols readily evaporate from corpses,the more volatile disappearing firstANALYTICAL CHEMISTRY.119Passing on to the second fraction from the acid distillation, phenol, asn normal product of the decomposition of nlbuminoids, is frequentlydetected by Millon's reagent, less readily by bromine-water. Largequantities of phenol are estimated by filtering the fraction containingit from the fatty acids, extracting with ether, drying the etherealresidue over sulphuric acid, and weighing, taking precautions againstthe phenol creeping over the edge of the evaporating dish. Asregards hydrocyanic acid, the distillate is tested with copper sulphateand guaiacum, and if the reaction is noticed, i t is confirmed by someother test, and the acid determined as silver cyanide.Phosphorus isgenerally recognised by the phosphorescence, but if this does notoccur phosphorous acid must be tested f o r and estimated in the residue.I n examining the distillate for phosphorus, it is redistilled, using anupright bulbed tube, the distillation being continued for an hourafter phosphorescence has ceased to appear; the new distillate istreated with silver nitrate, &c., and also is tested for phosphorus byoxidising with chlorine-water and adding ammonium molybdate.I n examining for alkaloids, special care must be taken not to mistakeptomaines for other alkalo'ids, and to allow for the impure form of thelatter. The material is treated with alcohol containing tartaric acid,the extract filtered, evaporated, dissolved in water, filtered, neutra-lised with potash, concentrated, treated with alcohol, separated fromthe potassium tartrate, and the neutral aqueous fiolution is testedwith alkaloid reagents (a preliminary test with tmtaric acid and iodicacid being made for morphine) and examined systematically ifrequired.A portion is rendered alkaline, and steam-distilled fornicotine, coniine, miline, &c. I n the absence of these volatile bases,another portion is extracted with ether successively when neutral,acid, and alkaline, ehen with chloroform while still alkaline, andfinally is made ammoniacal and extracted with aniyl alcohol. Thet8hird portion is reserved for the direct confirmatory examination ofany alkaloid icdicated in the other portions.The aqueous residue istested for narceine and curarine, whilst the various extracts areexamined separately. The residue from the dkaline-ether extract istested with phosphoric acid for aconitine. Vitali's atropine reactionis liable to be hidden by xaiithoprotcin colour reactions. For strych-nine, a double test is made: first., a drop of vanadic acid solution ismixed with the Gesidue, dissolved in sulphuric acid, which is subse-quently diluted with concentrated sulphuric acid, and then solidammonium vanadate is dusted over another portion of the samesolution. Attention is called to the fact that colocynthine givesreactions with both vanadic and chromic acids, which resemble thoseof strychnine with the same reagents. The atuyl alcohol extract,after purifying, serves for confirming the presence of morphine.Another portion of the original extract is examined for metallicpoisons soluble in alcohol.The examination for metallic poisonsgenerally is made with the residue from the first distillation or that,from the alcoholic extraction ; in the latter case the alcohol is expelledby warming. The residue is heated with hot water, potassium chlorate,and hydrochloric acid until the organic tissue is destroyed a r d allchlorine driven off, the magma is treated with tartaric .acid, the120 ABSTRAOTS OF OHEMIOAL PAPERS.largely diluted with water and, after 24 hours, filtered. Tho insolubleportion is treated (if required with more chlorate and hydrochloricacid, then) with alcohol, and with ether to extract fat, and is ignited,The ash is treated with very dilute hydrochloric acid, and the residuedried, ignited, weighed, and proved conclusively to be silica byfusion with sodium carbonate ; or is examined for silver, lead, barium,and strontium.The soluble portion is marde up to a definite volume,and must be free from chlorine and chloric acid.For mercury, a portion is nearly nentralised by means of potash,and is digested with brass wool €or 15 minutes a t 70" ; if the brass isvisibly amnlgnmatecl, the presence of mercury is confirmed by heatingin a test-tube, &c. ; but when the presence of niercury is not so evident,the brass wool is burnt with copper oxide in a current of air in a tabe25 cm. long, drawn out to a doubly bent capillary, in which anymercury is accumulated, and identified with iodine. To estimatemercury, the hot hydrochloric solution is treated with hydrogensulphide, the precipitate collected on an asbestos filter (or if arsenic ispresent, i t is first digested with yellow ammonium sulphide) washedwith hydrochloric acid, dissolved in nitric acid, filtered through someasbestos, and the washings and filtrate diluted and treated with phos-phorous acid.After 24 hours, the calomel is filtered off, washed with,water, alcohol, and ether, and weighed on a tared filter. Tbi*oughoutthe estimation, high temperatures and contact with organic matterare to be avoided.To detect antimony, some of t,he liquid partly neutralised withammonia is placed in a bright platinum dish with a piece of zinc forsix hours ; the brownish-black antimony flakes obtained in this mannerbeing more trustworthy than Marsh'R test.This test does not answei-in the presence of tin ; but by fusing the mixed oxides wit11 sodiumhydroxide, most of the tin can be separated. To estimate antimony,hydrogen sulphide is passed through the slightly acid, and at firstboiling, solution until it has cooled down ; after three days, most ofthe hydrogen sulphide is driven off by carbonic anhydride, the preci-pitate washed with an acetic acid solution of ammonium acetate,treated and washed with a solution of sodium sulphide, containinghydrogen sulphide, the solution treated carefully with hydrochloricacid, warmed, and then boiled. After 24 hours, the precipitate is col-lected, treated with an acetic acid solution of ammonium acetate, thencarefully with nitric acid, evaporated, the residue moistened withsodium hydroxide, intimately mixed with dry sodium carbonate, dried,introduced by small quantities at a time into fused sodium nitrate in a,silver crucible, and the mass, when cold, treated with water.After24 hours, the precipitate is washed with 45 per cent. alcohol, contain-i n g soda, digested with il hot solution of tartaric and hydrochloricacids for half an hour, filtered, and washed with a dilute solution oftartaric a d hydrochloric acids. The filtrate and wasbings combinedare concentrated in a water-bath, the excess of acid reduced withammonia, and the antimony precipitated by hydrogen sulphide as a pureorange-coloured sulphide which is converted into oxide by Bunsen'smethod.The purity of the antimony is ultimately confirmed by itsvolatilityANALYTICAL CHEMISTRY. 121To detect arsenic, mercury and antimony being absent, variousobvious precautions are observed in applying the Marsh test to someof the liquid, and when a mirror is obtained, the tube containing i tis divided by a diamond into four parts, of which one is used for theodour test, another for solubility in freshly-prepared sodium hypo-chlorite, a third for dissolving in nitric acid arid testing with silvernitrate, whilst the fourth is dissolved in nitric acid, and convertedinto arsenic sulphide by colourless ammoniuin sulphide. To estimatearsenic, hydrogen sulphide is passed through the warm hydrochloricsolution for 12 hours, and after remaining three to five days in aclosed flask, most of the hydrogen sulphide is driven off by a currentof carbonic anhydride ; the procedure then resembles that described inthe antimony estimation, but various points are to be observed :-1.Ammonia or ammonium carbonate are the only solvents used forarsenic sulphide on the filter. 2. The fusion is conducted in a porcelaincrucible with fusion mixture and potassium nitrate. 3. The arsenic isalways weighed as magnesium pyroarsenate. 4. The alkaline solutionis not precipitated directly with magnesia mixture, but is first Rub-mitted to the following treatment :-Neutralisation with nitric acid,expulsion of carbonic anhydride and nitrous acid, precipitation withhydrogen sulphide, and conversion into arsenic acid. 5.The ammo-nium magnesium arsenate is redissolved in hydrochloric acid and re-precipitated by ammonia. 6. Small quantities of magnesium arsenateare converted into pyrosrsenate by dissolving in very dilute nitric acid,evaporating in a porcelain crucible over a water-bath, and carefullyand gradually igniting the residue ; the method is susceptible of greataccuracy ; as little as 0.0093 to 0.0077 gram of arsenic in a portion ofa dead body has been estimated. 7. Large quantities of pyroarsenateare preserved as such for reference ; small quantities aro convertedinto metallic arsenic in a Marsh’s apparatus, and are preserved in ELsealed tube.In the absence of mercury, antimony, and arsenic, the hydro-chloric acid solution is made alkaline with soda, acidified with aceticacid, hydrogen sulphide passed in at the boiling point of the liquid anduntil cold, sodium carbonate added to distinct alkalinity, and the wholeallowed to remain corked up until clear.The solution serves for de-tecting and estimating tin. The precipitate is washed with sodiumsulphide containing hydrogen sulphide, oxidised with nitric acid, evapo-rated, moistened with sodium hydroxide, mixed with fusion mixture,dropped into molten nitre in a silver crucible, extracted with water,supersaturated with hydrochloric acid, filtered, and the hydrochloricacid solutions submitted to the ordinary methods of analysis, weighingany metal isolated in a definite form. The separation of iron, alumina,and zinc in the presence of calcium and magnesium phosphates iseffected in the following manner :-The hydrochloric acid filtrate fromthe hydrogen sulphide precipitate is concentrated on the water-bath,treated with chlorine-water, evaporated, the residue dissolved in verydilute hydrochloric acid, filtered, the solution supersaturated with am-monia, excess of the latter nearly expelled on the water-bath, the pre-cipitate removed, the solution acidified with acetic acid, and hydrogensulphide passed into the boiling hot liquid until it gets cold.The zin122 ABSTRACTS OF CHEMICAL PAPERS.sulphide is weighed. The ammonia precipitate is dissolved in nitricacid, treated in a platinum dish with tin, which by repeated and carefultreatment with nitric acid is converted into stannic oxide, and with i tthe phosphoric acid into insoluble stannic phosphate, from which thealumina is washed out by yery dilute nitric acid, and estimated inthe solution by precipitating with ammonia, igniting the precipitate,fusing the ignition residue with sodium carbonate, extracting withwater, filtering, &c.The alkaline filtrate from the hydrogensulphide precipitate is examined for tin :-It is acidified with hydro-chloric acid, boiled, hydrogen sulphide passed through until it is cold ;after remaining for 24 hours in a warm place the precipitate is collected,washed with an acetic acid solution of ammcinium acetate, and ignitedwith the filter.The residue is moistened with nitric acid, evapo-raked, ignited, and to get rid of the iron present it is washed into asilver crucible, dried, and treated for I.~df-nn.hour with moltensodium hydroxide, extracted with water, filtered, the filtrate acidifiedwith hydrochloric acid, and the tin obtained in the usual manner.I n concliision, it is pointed out that not only are potassium, sodium,calcium, magnesinm, iron, and niaiiganese noimally present in thehuman body, but that aluminium, copper, and zinc are always met with,and less frequently tin and lead. These extraneous metals are derivedfrom food, cooking utensils, medicine, &c. ; aluminium comes fromvarious sources, and even after death may be introduced iii the dust,when the post-mortem takes place i n the countr-y.The authoi* in advo-cating his employment of a solution prepayed directly froin the corpseinaterial for the detection of arsenic, points out that consideriiig thesensitiveness of the arsenic reaction a concentrrttion of the solutionis not necessary, that the “brown speck” on the porcelain lidreferred to by Otto cannot interfepe in his method, which, more-over, obviates any chance of vitiation through arsenical hydrogensulphide. It is also shown that the presenceof chlorides and nitratesdoes not stop the formation of gaseous hydrogen arsenide, providedthat the zinc and hydrochloric acid are in excess and the evolutionof hydrogen is allowed to proceed snfficiently long ; but the presenceof free nitric acid stops the evolution temporarily.It is still doubtfulwhether solid hydrogen areenide is converted into the gaseous modi-fication by zinc and hydrochloric acid.Ordinarily only one poison is found in a corpse, but nevertheless i tshould he borne in mind that there is the possibility of more thanone being present.Detection of Paraffin in Beeswax. By H. HAGEI~ (Zeit. anat.Cliem., 29, 480481 ; from Pharrn. Centi-allmlle, 30, 565).-A fewgrams of the substancs in fine, air-dried shavings is gradually heatedin a small, porcelain capsule, until fumes begin to rise. A half-litrewide-mouthed bottle is then inverted upon the capsule, and whenfilled with white vapours is closed and set aside until the fumes bavecondensed upon its walls. The sublimate is then dissolved in 3 C.C.of chloroform, the chloroform evaporated in a test-tube, and theresidue boiled with 4 C.C.of soda solution. If paraffin was present, it,will after cooling be found floating on the clear solution. A drop ofD. A. LANALYTICAL CHEMISTRY. 123the chlorofoi*m solution niay also be evaporated on a slip of glass andexamined microscopically.The fumes from pure beeswax are not so whiteas from paraffin, andare only obtained at a higher teinpe~ature (300-320"). The sub-limate gives a coloured solution with chloroform, and a coloured andturbid solution with soda. The residue from the chloroEorm solutionis a dnll film ; paraffin on the contrary gives separate grains in aclear field. M. J. S.Condition of the Sulphuric Acid in Plastered Wines, and aMethod of Distinguishing between Plastered Wines and Winesmixed with Sulphuric Acid.By L. Roos and E. THOMAS (Compt.reitd., 111, 573--577).-Wines which have been mixed with calciumsulphate do not contain potassium hydrogen sulphate. The liberatedtartaric acid interacts with the organic potassium compounds in thewine, and forms a new quantity of potassium hydrogen tartrate.Direct experiment shows that when calcium sulphate is added to asolution of potassium hydrogen tartrate, and an acetate, malate,citrate, or succiaate, the liquid contains no free sulphuric or tartaricacid, but acetic, citric, malic, or succinic acid is liberated. No potas-sium hydrogen sulphate could be detected i n plastered wines by thefollowing method, which will detect the addition of 0-25 gram ofsulphuric acid per litre.'l'he proportion of chlorine and the total sulphuric acid in the wineare estimated. 50 C.C.of the wine is mixed with a small quantity ofammonium acetate, and exactly procipitated with a standaiad solutionof barium chloride. The filtrate is evaporated to dryness, heatedgently, and the chlorine in the residue is estimated. If only normalpotassiuni sulphate is present, the reaction is &SO, + BaC'I, =BaSOl + 2KC1, and the chlorine in the residue should be equal tothe chlorine of the barium chloride, plus the chlorine originallypresent in the wine ; i f the acid sulphate i s present, the reaction iaKHSO, + BaCl, = BaSOp + HC1 t KC1, and the free hydrogenchloride is cxpelled i n the process of evaporation, the loss increasingwith the quantity of hydrogen-sulphate present.Estimation of Dissolved Solids in Wine.By E. L ~ S Z L ~ (Chem.Zeit., 14, 438, 455).--Results are quoted, showing t,he unsatisf'nctsrycharacter of estimations of " extractives " made by drying residuesfor 2$ hours. The author suggests determining the alcohol both bydistillation and by an alcoholometerat 15" ; the difference between t.hetwo observations being due to the "extractives" present may beutilised as a measure for them, and he finds that miiltiplying thisdifference by 0.32 gives numbers for the quantity of dissolved solidmatter in 100 C.C. of wine concordant with actual determinations.'l'he alcoholometer should not have a greater range than lo", or at theoutside 12", and f o r wines of sp.gr. greater than 1.000, a saccharo-ueter showing volume percentages is emplcyed.Detection of Methylated Nitrous Ether. By J. MUTER(Analyst, 15, 48).-Much of the sweet spirits of nitre in commerce isprepared from- methylated spirit instead of from pure ethyl alcoholC . H. B.D. A. L124 AliSTRAOTS OF CHEMICAL PAPERS.as prescribed in the pharmacopceia. The two may be discriminatedby dissolving a fragment of solid potash in a sample. Themethylated ether darkens, the colour varying from deep-yellow toorange-red, while the odour of methylated spirit becomes verydistinct. The ether from pure spirit loses its odour of ethyl nitrite,and retains only that of ethyl alcohol, and it does not dnrkeu beyondthe faintest straw colour.On treating with Hubl’s reagent the dis-tillate obtained after digestion with potash, tho methylated sample willabsorb 0.4 to 0.7 per cent. of iodine, but that from pure spirit none.Analysis of Carbolic and Sulphurous Disinfecting Powders,By J. MUTER (AnaZyst, 15, 63--68).-The author calls attention totbe ambiguities in the usual forms of specification for disinfectingpowders. Whilst the contracts are nominally for “ carbolic acid,” i tis commonly understood that the powder may contain chiefly cresoland other high-boiling tar phenols. There have, however, been caseswhere the supply of the more costlyabsolute phenol has been insisted on.The omiesion of the word “ available ” before “ sulphurous acid” some-times renders a literal compliance with a specification impossible.For the estimation of the phenols, the author still employs his ownprocess (Abstr., 1888, 92), with the single modification that 150 C.C.of a 10 per cent.solution of sodium hydroxide is now used instead of200 C.C. of a 5 per cent. solution. The cresol, measured in contactwith brine, retains about 5 per cent. of water. Since anhydi-ouscresol increases in volume by about 5 per cent. when shaken with3 volumes of brine, whilst that containing water does not increase o rmay even diminish, this furnishes a rough but ready test for t h epresence of water. For more accurate work the water must be dis-tilled out. Naphthalene, which is usually present in commercialcresol, may be estimated as follows :-50 C.C. is shaken with 200 C.C.of a 10 per cent.solution of sodium hydroxide. The phenols dissolve,leaving the naphthalene floating. The solution is removed, thenaphthalene washed with a 5 per cent. soda, solution, then rapidlyfiltered off. It is rinsed from the filter with water and again collectedon a pair of filters. After drying as far as possible by pressingbetween blotting paper, the filters are separated and the inner onewith its contents is weighed, using the outer one as a tare.For estimating the available sulphurous acid, 2 grams of the powderis washed 011 a filter with dry ether until the phenols and tarrymatters are removed. As soon as the ether has evaporated, thecontents of the filter are thrown into a bottle containing 50 C.C.ofN/10 iodine solution, and after half an hour the residual iodirie istitrated by thiosulphate. This method is unsatisfactory when thebasis of the powder is lime.I n sulphurous powders which have undergone osidation, the amountof original sulphurous acid cannot be ascertained if the mixture hadconsisted of gypsum and calcium sulphite, but where the basis issilica, the sulphates present may be regarded as oxidised sulphites,and where sodium hydrogen sulphite has been mixed with gypsum, theestimatiou of calcium, sulphuric, and sulphurous acids in an aqueousextract will give the necessary data.M. J. S.M. J. SANALYTIOAL OHEMISTRY. 125Arabinose.Milligrams.17 -018 *620 '321 -923.525 -126 *728.329.931 *533.1---Detection of Diresorcinol as an Impurity in SyntheticallyPrepared Phloroglucinol.By J. HERZIG and S. ZEISEL ( J h a t s h . , 11,421-423) .-The presence of diresorcinol, as an impurity in phloro-glucinol, scarcely affects its melting point, or the nnmbere obtainedon estimating carbon and hydrogen. It may be best detected bydissolving a few milligrams of the sample in about 1 C.C. of concen-trated snlphuric acid, adding 1-2 C.C. of acetic anhydride, andwarming the mixture for a few minutes i n a water-bath. If diresor-cinol-or its tetrethyl ether or tetracetyl derivative-be present, abluish-violet colour, which disappears on the addition of much wateror of an excess of alkali, will be produced. By this means, the presenceof 0.4 per cent.of diresorcinol may be clearly shown, and the delicacyof the test is probably much greater. G. T. M.Estimation of Sugars by means of Copper PotassiumCarbonate Solution. By H. OST (Bey., 23, 3003-3011 ; compareAbstr., 1890, 103l>.-A solution cmtaining 23.5 grams of crystallisedcopper sulphate, 250 grams of potassium carbonate, and 100 grams ofhydrogen potassium carbonate per litre has the following advantagesover Fehling's solution for the gravimetric determination of sugars :-(1) It is unchanged by keeping. (2) Its action on cane-sugar isrelativcly slight. (3) After 10 minutes boiling, the precipitation ofcuprous oxide is practically complete, and thus more concordantresults are obtained. (4) The monosaccharoses precipitate almosttwice asmuch cuprous oxide from this solution as from Pehling'ssolution.( 5 ) The quantity of precipitate obtained from differentkinds of sugars varies considerably, thus rendering it possible todetermine the composition of mixtures. The solution may also beemployed for Folumetric estimations, as the end reaction is sharp ; thetime required for boiling, is, however, longer than with Fehling'Hsolution. For gravimetric determinations, 30 C.C. of the copper solu-tion is mixed with 25 C.C. of the sugar solution, water is added andthe liquid boiled for 10 minutes, filtered through an asbestos filter,and the cuprous oxide reduced in a stream of hydrogen. The follow-ing table shows the quantity of copper precipitated by differentsugars :-Copper.Xilligmins.50556065707580859095100----Invert-sugar.Milligrams.15 *216 *618'019 *420.822-323-785 *226 *628 *129 ' 5--Dextrose.Milligrams.15 *617 *O18 '519 '921 '422 *924 -425.827 -328 6830 -3---Levulose.Milligrams.14 *716 '117 * 518 *920 '321 -723.024 -325 -727 -128 -5&lac tose .Milligrams.17 '419 *120 *822'524 *225'927 -629 -381 '132 -834 -5-ABSTRACTS OF CHEMICAL PAPERS.Copper.Milligrams.1051101151201251301351401151501551601651701751801861901952002c5210215220225230235240245250255260265270275280285290295298 -7CI-vInvert-sugar.Milligrams.31 '032 *433 -935.336 -838 *239 *741 '142 *644-045 -547.048 -550 '051 -553 -054.556 -057 -559-160 *762 '4184 *165 -867 *569 *371 -17s -974 -876.778 *680 -582 -584.787 *I89.792 -395 -198 '0103 '0--Dextrose.Milligrains.31'833 -334 *836 '337'839 '340 *s42.343-s45 -346.848.34g9 * 851 -452 *954'556.057 *c;a9 -260 *862 *464 '165.867.869 *270-972'774 *576 *478 -480 -582.885.187'589.992.494 -997.6100 '4102 *8--Levulose.Milligrams.29 -831 -232.634.035.436 *838 *23'3.641 -042.543 .945-346.748'149.561 -052 -554 -055 *557 -058 fi60.261 .S63.565.26G 968 -770 *G72 -574 -476 -578 *881 '183 -585 -988 691 '394 '297 -299 '0~Galactose.I'Jilligrams.36 '238 -039 -741 '443 -144 *846 548 -350.051 *853.655 *457.259 *o60.862 *764 -566 '468 -370 *372 '374 *376 *378 -380 *382 -484 -586 -688.991.293.595 -998.31GO * 7103 -3106 *l109 -0112 -0115'1117 *O-- Arabinose.Milligrains.34 -736.337 *939 *541 142 -844.346 '047'649 '350 '952 *G54 '355 .957 *559 -260 962 -761 '466 *268.069.871.673 -575 477 -379 *381 -383.485 -587 -689 -892 -294 -697 *I99 '6102 -3105 *1107 -9109 -5--In the case of lactose, the factor copper/l~ctose = 1.31 to 1.57 forsolutions containing from 125 to 198 milligrams of sugar.Forvolumetric work, a n indicator must be employed; after 20 minutesboiling, 198 milligrams of lactose precipitate 190 milligrams of copper,Ra5ose, C,H,Ola + 5H20, does not affect the copper solutionANALYTICAL CHEMISTRY. 127but after hydrolysis, it has the highest reducing power, 50 milligramsprecipitating 150 milligrams of copper.Estimation of Sugar in Milk. By M. KUHN (Bied. Centr., 19,628 ; from 3IiZc*hzeit., 18, 926) .-Results obtained by Tollens’method agree better with those obtained by Soxhlet’s method, whenonly so much serum solution is employed that the colour is bluishafter boiling. If so much sugar solution is used that the liquid isgreenish after reduction, results will be obtained which are 0.1 to0-15 per cent.too low. The phosphotungstic acid method is notrccornmended. If Soxhlet’s method is not used, the lead acetatemethod should be employed.Estimation of Ash in Raw Sugar. By W. MINOR (Chenz. Zeit.,14,51O).-Stammer objects to the use of oxygen, and recommends airfor the incineration of raw sugay in estiniations of ash. The authorhas investigated the point, and sees no reason for disqualifyingoxygen nor any special virtue in the atmospheric nitrogen, and asburning the charred sugar in oxygen takes 25 minutes, whilst corn-biistion in air, with the aid of mechanical agitation, requires from6 to 15 hours, he considers the oxygen method is distinctly to berecommended.11. A. L.J. B. T.N. H. J. M.Estimation of Starch. By 0. REINKE (Zeit. anal. Cl2etn., 29.472-475 ; froin Zed. XpiritiiLdusf.) .-The author divides the processeshitherto proposed into those with and without high pressuw, andrecommends the following as the best of the respective methods :-With high pressure : 3 gi-arns of the finely ground substance is stirredwith 25 C.C. of a 1 per cent. solution of lactic acid and 30 C.C. ofwater in a metallic beaker, then covered and heated for 24 hours ina digester (Soxhlet’s or Lintiier’s) a t 3-& atmospheres pressure, thenmixed with 50 C.C. of hot water and, after cooling, made up to 250 C.C.and filtered. 200 C.C. is then inverted by cohobating with 15 C.C. ofhydrochloric acid (1.125 ~ p . gr.) for Z& hours, then neutralised withsoda, made np to 500 c.c., and 25 C.C.of it titrated with Febling’ssolution. Without high pressure : 3 grams of the substance is boiledwith 50 C.C. of water, and then digested for au hour a t 62.3” with0.05 gram of Lintnep’s diastase. It is then cooled, mads up to 250 c.c.,and 200 C.C. inverted with acid 8s above. For the preparation ofLintner’s wade diastase, I part of green malt is extracted for24 hours with 2 to 4 parts of 20 per cent. alcohol. The extract,filtered by suction, is precipitated with twice, or at most 2$ times, itsvolume of absolute alcohol. The upper liquor is poured off and theprecipitate thrown upon a pressure filter, then rubbed down withabsolute alcohol in a mortar, again filtered and washed with absolutealcohol and then with ether, and finally dried in a vacuum oversulphuric acid.For the purification of this raw product, the precipi-tation and digestion with alcohol, washing with ether, and dryingare repeated. By this means, fclbuminoiid impurities are renderedinsoluble and dextrino’id ex tractive matters removed. The driedproduct is a loose, yellowish-white powder, which has no action o128 ABSTRACTS OF CmMICAL PAPERS.Fehling’s solution either before or after boiling with hydrochloricacid, and which does not turn brown when its solution is evaporatedon the water-bath.M. J. S. .A New Application of Molisch’s Reactions. By G, COLASANTI(Gazzetta, 20, 2C39--%05) .-Molisch (Abstr., 1886, 923) found that,the merest traces of sugar or glucosides (O*OOOOl per cent.) couldbe detected by the addition of one or two drops of an alkaline solu-tion of a-naphthol or thymol (15 to 20 per cent.), together with anexcess of coiicentrated sulphuric acid.Molisch further derived fromthis reaction a confirmation of the alleged presence of sugar in normalurine.The author finds that extremely dilute solutions of potassium orsodium thiocyanate, treated ih the same manner, show first a greenband and, on agitation, an iiiiense violet coloration resembling in allrespects that obtained from solutions of sugar. On cooling the liquid,a, compound containing the naphthalene nucleus and the sulphonicgroup separates in a mass of long, slender needles. The solution ofthiocyanate or thiocynnic acid must be very dilute, or on addition ofsulphuric acid a brown coloration is produced, and hydrogen sulphideis evolved.Urine must similarly be diluted before treatment witha-naphthol, and altogether fails to give the thymol reaction.As urine has been found to contain thiocyanic acid, Molisch’sreaction affords no confirmation of the presence of sugar in khat fluid.Reaction of Thiocyanic Acid. By G. COLASANTI (Gazzetta, 20,306--308).-1f a few drops of a solution of auric chloride (&th percent.) made alkaline with a saturated solution of sodium carbonateor a 5 per cent.. solution of potash are added to a few C.C. of a dilutesolution of a thiocyanate (0.01 per cent.), a deep violet coloration isobtained, and a precipitate of metallic gold gradually separates.Thethiocyanic acid in urine does not give the reaction, the liquid merelyacquiring a reddish coloration.Schneider’s Method €or the Estimation of Malic Acid inWine. Uy IS. NIEDERHAIJSER (C’hem. Centr., 1890, ii, 172; fromPharm. CentraZhaZZe, 31, 378-379) .-lo0 C.C. of t3he wine is neutral-ised with decinormal alkali, evaporated, incinerated, and the carbonicanhydride in the ash determined. From t,his amount, the quantity ofcarbonic anhydride equivalent to the total tartaric acid present isdeducted, the difference being then calculated into malic acid.Since, however, wines usually coiitain other substances (tannic,succinic, acetic acids), all of which rieutralise alkalis, and wouldwhen incinerated produce carbonates, the author considers the methodvalueless.J. W. L.It exhibits great hydrolytic activity.S. B. A. A .S. B. A. A.comparison between Methods for Estimating Tartaric Acid.By J. T ~ T H (Chem. Beit., 14, 63-64).---To compare the three rivalmethods for the estimation of tartaric acid, the “ original Goldenberg ’ 9method, the “ Ilorenz-Goldenberg” method, and the ‘( modifiedGoldenberg ” method, the author made simultaneous and duplicatANALYTICAL CEEMISTttY. 129estimations in crystalline calcium tartrate, in wine lees, in argol, andin tartaric acid, following rigidly the directions laid down in eachmethod ; the numbers obtained are tabulated, and from the results i tis concluded thstt the Lorenz modification of the Goldenberg methodis the best method, and is applicable in all cases, a specially valuablefactor about it being the introduction of one-third normal soda forthe titration.With regwd to Roessneck's suggested method, theauthor shows that the amount oE antimonious oxide taken up by thecalcium tartrate is not + a mol. for 1 mol., but is a variable quantity,which seems to depend on the amount of free tartrate in solution.D. A. L.Estimation of Tartaric Acid. By J. WOLFMANN (Ohem. Zeit.,14, 320 ; compare T6th, preceding abstract).-The author considerathe use of litmus tincture unsatisfactory in deeply colonred tartaricsolutions ; he has noticed neutralisation of alkali by humus in suchsolutions, and does not regard the question of the estimation oEtartaric acid as solved bay the Lorenz method, in fact, looks withgreater favour on the Goldenberg-Geromont results.He himselfendeavoured unsuccessfully to determine tartaric acid by titrationwith permangnnate. 1). A. L.Estimation of Tartaric Acid in the Crude Products ofTartaric Acid Factories. By J. TELRISZ (Crhem. Zeit., 14, :347).-In consideration of results recently published by T6th (see above),the author has made several estimations of tartaric acid in varioussamples of calcium tartrate and dried wine lees, applying, with muchprecision, both the " original '' and " modi6ed " Goldenberg-Geromont,and also the Lorenz method; the results are tabulated, and in hishands the latter method yielded undoubtedly higher results than thefirst two methods, and he agrees with Wolfmann (preceding abstract)in considering tbe modified Goldenberg-Geromont method the mosttrustworthy, up to the present time.Variations as great as 7-10per cent., noted by Tdth, in different estimations of the same sampleby this method, have not been obaerved in the preseut experiments.D. A. L.Estimation of Citric Acid in Parts of Plants. By E. Cr,AAssEN( Z e i t . anal. Chern., 29, 468--469).-The plant is extracted with verydilute ammonia and ammonium csrbonate, the liquid somewhat con-centrated, precipitated with lead acetate, and filtered. The driedprecipitate is boiled out with strong alcohol, then suspended in water,and decomposed by hydrogen sulphide. The filtrate is evaporated tozt thin ~yrup, mixed with ammonium chloride, excess of ammonia,and calcium chloride, and 3 volumes of alcohol added.The precipi-tate is filtered, washed with 75 per cent. alcohol, dried, and dissolvedin hot dilute hydrochloric acid. After cooling, it is filtered, treatedwith excess of ammonia, and again filtered, and evaporated on thewater-bath to dryness. The residue is %taken up with boilingammoniacal water, and the insolkble calcium citrate collected on aweighed filter. Traces of citrate in the filtrate may be recovered byrepeating the evaporation. M. J. S.VOL. r,x. 130 ABSTRACTS OF CHEMICAL PAPERS.Amount of Volatile Fatty Acids in Rancid Butter. By P.C o m E w A (Chern. Zed., 14, 406).--8amplen of fresh butter were takenand examiried on the 16th of Febriiary for volatile fatty acids ; theywere then exposed in vessels covered with psper, and again examinedon April 3rd, when, in all cases, a reduction in tbe quantity ofvolatile fatty acids was observed ; in a subsequent examination onApril 30th, no further change was noted, but a final test, on AuguHtgth, indicated a still further falling off in these acids.The disappear-ance of T-olatilc fatty acids in the rancid buttJer, although progressivein these experiments, was in no instance very considerable, and in nocase could volatile fatty acids be washed from the rancid butter eitherby water or sodium hydrogen carbonate.Butter and Margarine. By C. VIOLETTE (Compt. rend., 111,345-347).--'l'he acids resulting from the saponification of 50 gramsof pure, dry butter by aqueons potash are distilled in a current ofateam, and the successive portions of the aqueous distiliste (the totalvolume of which should not be less than 10 litzes) are titrated withnormal sodium hydroxide, using phenolphthalein as an indicator.The volatile acids, which solidify, and tho non-volatile acids also, areweighed, after being dried in a vacuum and melted.A table i s givenshowing the results obtained with various butters and with margarine.It is assumed, on t,he evidence of Duclaux's results, that the ratiobetween butyric and caproic acids in genuine butter remains constant,and equal to 1.645.In ordinary butters the mean proportion of volatile acids is 7.6 percent., wit,h a minimum of 7.0, and the proportion of non-volatileacids is 84.0 per cent., with a, maximum of 84-6.I n the case of abutter of high quality, the addition of about 20 per cent. of margarinewould lower the proportion of volatile acids from 8.5 to the minimumof 7 per cent., and would raise the non-volatile acids from 82.63 to84-76. In the case of ordinary butter, the addition of 9 per cent. ofmargarine would reduce the volatile acids to 7 per cent.Optical Analyses of Butters. By C. VIOLRTTE (Compt. rend., 111,348).-From his observations, the author concludes that butter andmargarine have different indices of refraction, the deviations in theoleorefractometer being -35" to -27" for bucters, and - 15" to -sofor margarines. The indications of the oleoref ractometer are suffi -ciently exact when the instrument is applied to mixtures of con-stituents giving known deviations.It is necessary to ascertain, bymeans of a large number of observations, the minimum deviationbelow which a butter may be regarded as adulterat4ed with margarine.The olcorefractomet,er may be used for the analysis of commerci81butters, but its iudicidions will not be very exact, because thesebutters will give deviations below the minimum f o r good butters, andthe proportion of margarine deduced from the results will be too low.Analysis of Lard, Cotton Oil, and Tallow. By J. MUTER andL. DE KONINGH (Analyst, 15, 48--50).--Employing the methoddescribed by themselves (Arkalyst, April, 1889) for estimating the1). A. L.C. H. B.C. H. BANALYTICAL CHEMISTRY.131liquid fatty acids in fats, and for treating them with iodine without.exposure to air, the ttiithors have obtained the following results.They regard tallow as the best material for the preparation of olei'cacid, and for this acid they find the iodine absorption to be 90 percent., and to vary at most 0.2 per cent. from theory. The olei'c acidfrom lard never gives so low a number, the average being about 93per cent., whilst that from cotton oil is found to be 135, with verylittle variation. In consequence of this wide difference, the per-centzge of cotton oil in a sample of adulterated lard can be indirectlyestimated with considerable accuracy. M. J. S.Beeswax. By A. BUISINE and P. BUISINE (Bull. SOC. Chim. [3],3, 867-8731 --The authors confirm the results previously obtainedby Hub1 and Hebner with respect to the free, total, and combinedacids of beeswax ; further, they have det,ermined the iodine numbersfor this substance, m d describe a process for estimating the alcoholspresent.This consists in the fusion of the wax with potaasiumhydroxide and pot,a,sh lime a t 250°, which causes the evolution ofIijdrogen proportionally to the amount of alcohols acted on, andfrom the residue of this experiment the hydrocarbons existing in thewax are determined by extraction with a suitable solvent. Theirresults for pure, dry, washed beeswax are summarised :-M. p. 63-64'. Entirely soluble in hot chloroform.Wax Acids.Free acids correRponding with 19--21 m i l l i p m s KHO per gram.7, 7, 97 13.5-15.5 per cent.cerotic acid.Total acids 9 ) 91-97 milligrams KHO per gram.Combined a c i h ,, 72-76 ,, 7, 7 99 , 9 , 9 7 32-85-34.67 per cent. palmitic acidRatio of free io combined acid 3.5 to 3.8.Iodine Numbers.100 parts of wa.x absorb 8-3-11 parts iodine ; which correspondsto 9-12 per ccnt. olejic wid.Wax Alcohols.Hydrogen liberated by fusion with KHO, 53.5-57.5 C.C. per gram.Wax Hydrocadbons.M. p. 49.5. Iodine fixed by 100 parts ofEstimation of Resin in Soap. By R. WIT,T,TAMS (AnaZyst, 15,81--82).-Gladding's method (Abstr., 1883, 603) yields vcry goodremits. The author prefers to work on the soap itself rHther thanon che acids separated from it.Estimation of Csmpkor. By F. FOERSTER ( B e y . , 23, 2981--2989).-A number of substances now occur in commerce, consistingPercentage 12-5-14.11 y drocar-bon 22-03.T. G. N.M. J. S232 ABSTRACTS OF OREMIOAL PAPERSof nitrocellulose and camphor, and up to the present no method isknown for estimating the amount of camphor which they contain.The author proposes to carry out the estimation by distilling thesubstances with soda solution, when the camphor readily passes over.This may be then extracted with benzene, and the specific rotatorypower of the benzene sol ution ascertained. Detailed instructionsfor carrying out the reaction, and tables of the rotation of camphorin benzene solution at different concentlrations and temperatures aregiven in the original. The results obtained are about 0.7-1.0 percent. too low, probably owing to the difficulty of driving out the lastportions of camphor.The autbor finds that sublimed camphor contains a small quantityof impurity, and that for the determination of its rotatory power, hightemperatures must be avoided in its preparation, and the camphorfinally twice recrystallised from 50 per cent.alcohol. It then meltedat 174+3-173.3", and after six crystallisations a t 1 'i6*3-176-5", andafter 10 crystallisations the solidifying point was found by Landolt'smethod (Abstr., 1890,l) to be 15'8.7" (con-.). The boiling point of thepurified camphor was 209.1" under 739 mm.Estimation of Tannin in Tea. By P. MALTSCHEFFSKY (.I. P~CWWL.[ 5 ] , 22,270-271 ; from Pharm. Zed. f. RUSS., 29,12i).-The tanuiriis precipitated by meitus of normal copper acetate, and the excess ofcopper is titrated by the aid of potassium ferrocpnide solution, Thecopper solution contains 7.657 grams of copper oxide per litre (1 C.C.=0.01 tannin), and its strength is controlled by evaporating a measuredvolume t o dryness, moistening with nitric acid, heating to redness,and weighing tbe oxide. The ferrocyanide solution is prepared bymaking up to 1 litre, 100 C.C. of a saiurated solution. To standardisethis solution, it is added, 1 C.C. at R time, to 5 C.C. of the copper solu-tion diluted to 100 c.c., until a drop of the mixed liquids gives a bluecolonr with a solution (1 : 100) of ferric chloride. A second assay,in which the additions of ferrocyanide solution are made by tenthsof a C.C.towards the end, gives the exact strength of the solution.2 grams of tea, dried at 100-107" is extracted four times with 100 C.C.of boiling water each t h e ; the filtrates are united, made up to400 C.C. ; 100 C.C. of this solution is boiled and treated with 10 C.C.of copper solution. The precipitate is filtered oft', washed with hotwater, and the filtrate and washiip are made up to 200 C.C. ; half ofthis is taken, and the excess of copper is determined approximately bymeans of the ferrocyenide solution ; the second half of tbe solutionthen serves for the exact determination of the copper. In 14 samples,the amount of tannin yaried from 6.10 to 11.08 per cent. The watervaried from 5.59 to 12 48 per cent. ; ash, 3.14 to 9.25 ; aqueous extract,17.3 to 39.4; caffeine, 1.09 to 2.88 per cent.J. T.Estimation of Urea. By P. M~QUEL (Compt. rend., 111, 501-502).-Many of the nrophagic microbes, and especially micrococciand earcinae, can develop in a neutral and even in tt slightly acidcnltivation fluid. Several grow solely a t the bottom of the vesselsand produce more or leks granular deposits without rendering theH. G. C.ANALYTICAL CHEYISTRP. 133liquid turbid, whilst at the same time tbey produce a large quantityof the soluble ferment (this vol., p. 100 ). These clear liquidsshould be used for the estimation of urea.An aqueous solution of urea is simply mixed with the cultivationfluid containing the ferment; the alkalinity i s at once estimated bytitration, and the liquid is heated at 50" for two hours in a well-closed vessel, which it nearly fills.The alkalinity is again deter-mined? and from the quantity of ammonium carbonate formed theamount of urea present is calculated.Urine and ohher organic liquids are previously heated with a sligbtexcess of ammonium carbonate, filtered if necessary, and then mixedwith the feiment, the object of this treatment being to prevent loss ofammonia from formation of double salts, neutralisation of any acidpresent, &c.A quantity of urea exceeding 10 per cent. interferes with theactivity of the ferment, and in solutions o€ 30 per cent. the ferment isit mctive. Concentrated solutions must, therefore, be diluted. Am-monium carbonate, sodium chloride in small proportion, uric acid,arnmoniacal and alkaline salts, extractive matters, albumin, and sugarin large quantity do not interfere with the results. C.H. B,Simple Mode of Estimating Urea. By C. W. HEATON andS. A. VASEY (Analyst, 15, 106--107).-The method, which does notaim at great accuracy, is suggested for the use of medical menin cases where none of the special forms of apparatus is available.An &-ounce bottle is fitted with a thistle funnel and gas delivery tubewhich dips under water in a basin. In the bottle is placed 1 fluiddrachm of bromine and 10 fluid drachms of a 40 per cent. solution ofcaustic soda. A bottle full of water is inverted over the deliverytube to receive the gas ; 2 fluid drachms of urine is then poured intothe generator and rinsed in by 1 fluid dirtchm of water, and +,bebottle is shaken until gas ceases to be evolved.The receiver isthen closed by the thumb, removed from the basin, placed in anupright position, and filled up with water, the volume requiredbeing noted. Deducting 200 minims for the volume of air dis-placed by tho urine and water introduced into the generator, theremainder is equal in volume to the nitrogen evolved, and each100 minims corresponds with 0.25 per cent of urea. M. J. S.Rapid Method of Estimating Urea in Urine. By C. J. H.WARDEN (Analyst, 15, 201--203).-l'he apparatus iR a modifiedCrum's nitrometer, 630 mm. long and of about 75 C.C. capacity. Intoits lower end is ground a stopper, on which 10 narrow grooveshave been filed. The cup above the stopcock i s of 5 C.C.capacity,and is accurately marked a t 2.5 C.C. The tube is graduated toshow percentages of urea a t once, asmmirig that 1 per cent. ofurea in 2.5 C.C. of urine will yield 9.27 C.C. of gas. The hypo-brornite solution is stated to be made by dissolving 100 '' grains"(? grams) of caustic soda in 750 C.C. of water and adding 25 C.C.of bromine. The inverted tube is filled with this solution and thestopper inserted. Ifs cxterior and the cup are then rinsed an134 AUSTHACTS OF CHEhUCAL PAPERS.dried. It is stood in a vessel of brine and the stopper is removed.2-5 C.C. of urine is then placed in the cup and there mixed with itsown volume of saturated brine to increase its densify, and thismixture is allowed to enter the tube in small portions.The lasttraces are rinsed in by brine. The tube is then grasped by theright hand, the thumb being tightly pressed against tbe open end,aiid the contents thoroughly agitated. It is then transferred to avessel of water, where the heavy sollitions flow away, and thevolume of the nitrogen is read in the usufil manner. M-. J. S.Estimation of The'ine in Tea. By G. L. SPENCER (CZcex2,Centr., 1890, ii, 172 ; from J. Arner. Chern. Soc., 4, 158).-2 to 3 gramsof the finely ground tea is extracted i n a small beaker seven timeswith boiling water, the extract being each time decanted off, andthe residue finally transferred to a filter, and washed with a fewC.C. of boiling water. Ha.sic lea*d acetate is added to the extract,about 8 C.C. usually being snfficient; the precipitate is filtervd,washed with hot water, and the lead sepayated as sulphide, afterwbjch the filtrate is concentrated to about 50 c.c., with addition ofabout 5 grams of calcium hSdroxide or mngnesium oxide. The liquidis again filtered, the insoluble portion extracted with hot water, andthe tiltrate is extracted with chloroform seyen times. 'I'he chloro-form extract is distilled from a tared flask, and the weight of theresidual theine recorded after drying at 75". The method has beenin use in the Department of Agriculture. J. W, L.Estimation of Quinine. By SEATON and H. D. RICHMOND(AnaZyst, 15, 42- 43) .-In soh tions containing quinine bisulphatedissolved in an acid, and free from salts whose base is pyecipitable bybaryta, the quinine may be estimated by titration. Quinine bi-sulpbate is neutrai to methyl-o~ange, whilst the base itself has noaction on phenolphthalein. To 25 C.C. of the solution there areadded 2 drops of methyl-orange solution (0.25 gram iii a litre ofwater), and 2 drops of phenolphthalein solution (0.5 gram in a litreof 50 per cent. alcohol). Baryta solution (N/10) is then run in untilthe red colour changes to a brown, at which point all the free acid isneutralised. The addition of baryta is then continued until the pinkcolour of the phenolphthalein appears. As the pink colour develop8slowly, care must be taken not to overstep this point. The numberof cubic centimetres required for this second stage, multiplied by0.0218, gives the weight of the hepta-hydrated quinine sulphatepresent. M. J. S.Reaction for Cocai'ne. By F. DA SILVA (Compt. rend., 111,54$-349).-A small quantity of cocaine, or one of its salts, or of thcresidue obtained by evaporating a solution, is mixed with a few dropsof fuming nitric acid of sp. gr. 1.4, evaporated to dryness on thewater-bath, and the residue mixed with 2 or 3 drops of concentratedalcoholic potasli. A distinct, and peculiar (;dour, recalling that ofpeppermint, iLi developed. In lhageudorfl's systematic scheme oASALTTICAL CHEMISTRY. 135analysis, cocaine is found among the alkaloidu extracted by benzenefrom an aqueous ammoniscal solution. Of the other alkaloids of thesame group, atropine, hyoscyamine, strychnine, codeine, and eserinegive colortttions wlien treated in the same way,and eseriiie also developsa disagreeable odour resembling that of phcn.ylcarbylamine. Del-phinine, brucine, and veratrine give only indistinct odours, whichcannot be confonnded with that from cocaine. Sabadilline andnarcotine can be recognised in the same way, but the other alkaloidsgive no sensitive reactions of this order.The reaction will detect 0.5 milligram of cocaine hydrochloride.C. H. B.Detection of Colchicine in Corpses. By N. OBOLONSEI (&it.a n d Chem., 29, 493).-The finely divided viscera are rubbed upwith glass powder treated with oxalic acid, and digested for 2‘2 hourswith alcohol. The liquid is squeezed out, and the dry residue twicemashed wit.h alcohol. The extract is concentrated at a temperaturenot exceeding SO”, and the cooled residue made up to the originalvolume with alcohol. The filtered liqriid is evaporated as before, nadthis operation repeated until no clots separate on the addition ofalcohol. The residue is then dissolved in water, the solution purifiedby shaking with light petroleum, and the colchicine finally extractedwith chloroform as usual.Tbe alkaloid is best identified by means o€ the violet colour pro-duced by nitric acid ; by Erdmann’s reagent (nitrosnlphuric acid),which gives in succession green, dark-blue, violet, and yellow colourti,turning to raspberry-red on adding alkali ; also by Mandelin’sreagent (1 gram of ammonium vanadate in ‘200 grams of siilphuricacid) which gives a green colour. Colchicine is with difficultydestroyed by putrefaction of animal matter. The kidneys, bladder,and urine are best suited for forensic examination. M. J. S.Detection of Bile Constituents in Urine. By A. JOLLUS(Zeit. a n d Chenz., 29, 402 -406).--Of the various tests proposedfor detecting bile pigments in mine (Grnelin’s, Huppert’s, Vitali’s,Rosenbach’s, Ultzmann’s, Hoppe-Seyler’s, Dragendorff ’s), those ofRosenbach and Huppert, with the following modifications, give thebest results :-Rosenbach’s Test.-A large quantity of the urine is filtered throughclean, white filter-paper, the interior of the filter is touched with adrop of strong nitric acid containing nitrous acid, and the funnel isgently warmed over a flame. After a few minutes a green ring isformed round the spot moistened by the nitric acid.Huppert’s Test.-About 10 C.C. of the urine is shaken with an equalvolume of milk of lime containing 10 grams of calcium oxide in thelitre. The success of the test depends on the proper concentratiotiof the milk of lime. The precipitate is filtered off and washed into atest tube with alcohol and dilute hydrochloric acid, then filtered, andthe filtrate boiled, With only traces of bile pigments, the liquidbecomes green to blue. An estimate of the amount of bile con-stituents can be obtained from the iodine number oE the urine. If 136 ABSTRACTS OF CHEMICAL PAPERS.is the number of grams of iodine absorbed by 10 C.C. of the urine,and s the specific gravity, the iodine number is - The numbers-1for normal urine, filtered after cooling, is 6.3 to 8.1, though even inspecimens rich in uric acid it rarely exceeds 7.8. The presence ofeven traces of bile pigments raises the uumber to 9.6, and values ashigh as 17.4 have been observed.New Test for Albumin. By A. JOLLES (Zeit. ai2aE. Chem., 29,406--407).-About 8 or 10 C.C. of albuminous urine is mixed with anequal volume of concentrated hydrochloric acid, and then 2 or3 drops of a saturated solution of bleaching powder deposited quietlyon the surface. If a8 little as 0.01 gram of albumin per 100 C.C. ispresent, a white turbidity appears at the surface of contact. Thistest, being less sensitive than that with nitric acid, which latter willdetect 0*0015 gram per 100 c.c., may be used to find approximatelythe proportion of albumin present, since by diluting the urine untilthe one test gives an iiidication but the other none, the percentagemay be known to lie between the above minimum limits.M. J. S.M. J. S.Detection of Albumin in Bacterial Urines. By A. JOLLES(Zeit. anal. Chem., 29, 407--408).-The most sensitive test foralbumin in urine is that with acetic acid and potassium ferrocyanide,the lower limit of which is 0.0008 gram in 100 C.C. It is, however,necessary to filter the urine to obtain a standard with which to com-pare the turbidity produced by the test. When bacteria are present,ti clear filtrate is best obtained by shaking with infnsorial earthbefore filtering. In the case of purulent., slimy urines, rich in leuco-cytes, traces of albumin may adhere to the precipitate ; but by wash-ing this with warm potash, and testing the filtrate, the smallesttraces of albumin may be detected. M. J. S