ANALYTICAL CHEMISTRY.A n a1 y t i c a1 C h e m i s t r y .109Estimation of ChIorine, Sulphurie Acid, and Chromium inPresence of Organic Matter. By C. T. POMEROT ( C h m . News,48, 41-42).-To determine these substances in the presence oforganic matter, it is first of all necessary t o remove the latter byignition with an alkaline carbonate and nitrate ; some of the nitrateis thereby reduced to nitrite, which, during subsequent operations,reduces the chromates, and interferes with the estimation of chlorine ;moreover, if the sulphuric acid is precipitated in this solution, thebarium sulphate always retains chromic oxide. To obviat'e thesedifficulties the mass from the fusion with alkaline carbonate andnitrate is dissolved and filtered ; potassium nitrite and nitric acid inexcess are added to the filtrate, which is allowed to stand 12 hours inthe cold.'On adding ammonia and boiling, the chromium hydroxide isthrown down ; i t should be filtered off hot, washed with hot water,and estimated as usual. The sulphuric acid is precipitated from thefiltrate from the chromium hydroxide by means of barium nitrate,and the chlorine estimated in the filtrate from the barium sulphatein the usual way. In theanalysis of ordinary chrome yellow, even the lead passes entirely intosolution when snfficient sodium nitrite and nitric acid are added,and the mass boiled..The published results agree very well.D. A. L.Formation of Methylene-blue as a Reaction for HydrogenSulphide. By E. P~SCHER (Ber., 16, 2234--2236).-To test forhydrogen sulphide in aqueous solution, the latter is treated with one-fiftieth volume of concentrated hydrochloric acid, a few grains ofparamidodimethylaniline sulphate are added, and when this is dis-solved, 1-2 drops of a dilute solution of ferric chloride.In the caseof a solution containing 0.00009 gr. hydrogen sulphide in a litre ofwater, coloration took place in a few minutes, and in half an hour theliquid had assumed a strong blue colour, which lasted for days. Asolution of the same strength, but without hydrochloric acid? yieldedonly a light brown coloration with lead acetate. In a solution con-taining 0.0000182 gr. hydrogen sulphide in a litre of water, themethylene-blue reaction still gave a distinct blue coloration, whilst noeffect was produced either by lead acetate or sodium nitroprusside.This reaction is recommended as the most delicate and certain testf o r neutral or acid solutions of hydrogen sulphide.Paramido-dimethylaniline is most conveniently prepared from commercialhelianthin, Me2N.C6H4.N : N.C,H,.SO,H ; this is finely powdered,mixed with 5 parts of water and an excess of ammonium sulphide ;the mixture is frequently agitated, and when (after about 24 hours)the reduction is complete the amidodimethylaniline can be extractedwith ether; the ethereal solution is then agitated with a little whitelead suspended in water, and the filtrate treated with an etherealsolution of concentrated sulphuric acid. The ether is separated fro110 ABSTRACTS OF CHEMICAL PAPERS.the crystalline sulphate, and this is heated with absolute alcohol untilit is converted into slender white needles, which after being washedwith alcohol can be pressed and dried on a water-bath.A.I(. M.Volumetric Estimation of Phosphoric Acid. By G. C.CALDW ELL (Chern. Netus, 48, 61-62).-The author finds Pemberton’smethod (Abstr., 1882, 1318) efficient ; he has introduced an improve-ment to facilitate the final test filtrations. A test-tube is fitted witha double-bored cork, through which pass two tubes, a short onebent at right angles and a longer one bent at a convenient anglefor introduction into a beaker; the bore of the tube is 1 mm., theend of the longer one has a conical enlargement or mouth with adiameter of 5 mm.To prepare the apparatus for use a perforatedplatinum cone is fixed in this mouth, and while suction is applied tothe short bent tube, the mouth of the tube with the cone is justdipped into very thin asbestos pulp and then into water, the suctionbeing continued until the water comes through quite clear. Theapparatus is now ready for use; suction is again applied, the filterdipped into the liquid to be tested, and when the desired quantityof filtered solution is obtained, the filter is withdrawn ; the suctionmust be sustained all the time to prevent the cone and asbestos fromfalling out, The liquid is treated in the test-tube, and then returnedt o the beaker. Before the next test is made the filter and tube arewashed by a small quantity of the liquid, which is drawn through thefilter and returned to the beaker.D. A. L.Quantitative Separation of Potash and Soda from FerricOxide, Alumina, Lime, and Magnesia in Silicates. By W.KNOP (Chem. News, 48, 110-111). The new method is based onthe formation of alkaline silicofluorides insoluble in acidulatedmixtures of alcohol and ether, by the combined action of hydrochloricand hydrofluoric acids on silicates. When the alkalis are in excessit is necessary to add silica before treatment with the acids.The weighed substance, mixed with a small quantity of water, and,if necessary, silica also, is treated, in a platinum capsule, with fuminghydrofluoric acid, evaporated to dryness, treated with strong hydro-chloric acid and absolute alcohol, and after some time with excess ofether.In 12 hours’ time the silicofluorides are filtered off, washedwith alcohol, dried, and removed from the filter, which is incinerated.The ash and the precipitate are mixed together in a crucible withconcentrated sulphuric acid, and as soon as the evolution of siliconfluoride ceases the crucible is ignited at a low temperature, so as t oleave the alkalis as acid sulphates. The mass is treated wit.hammonia, evaporated to a paste, rendered alkaline with ammonia, andafter an hour, to ensure complete separation of iron and alumina, istreated with ammonium carbonate solution and left for 12 hours.The iron, alumina, lime, and magnesia, are filtered off and washed withammonium carbonate, &c. The filtrate is evaporated to dryness-a,quantity of ammonium hydrogen tartrate equal to the ammonium sul-phate present being added t o prevent spurting-and ignited.ThANALYTICAL CHEMISTRY. 111residue is taken UD with wat,er, tested with ammonium carbonate foriron, &c., and the Lsolution is then evaporated to dryness aud weighed.D. A. L.Methods of Analysing Columbates by means of HydrofluoricAcid. Separation of Thoria from the other Oxides. Estima-tion of Didymium. By J. L. SMITH (Chem. News, 48, 13-15 ; 29--31).-After a review of the work done on this class of minerals byRose, Herniann, Marignac, &c., the author observes that they havespecial interest, because there are always some of the rarer earthsassociated with them, and the reasons amongst others which haveprevented their being more perfectly studied are, firstly, the scarcityof the minerals, and secondly, the difficulty attached to the process ofdecomposing them.The first obstacle was removed by the discoveryof large quantities of columbite and samarskite in the United States,and the second is set aside by the methods described in this paper.I n a previous communication (Amer. J. Xci., 1877, 360), the authorhas described these minerals, columbites and tantalites from theUnited States, both as regards their mineralogy and their chemicalcomposition, but owing to the defects in the manner of working, hecould neither give the relative proportions of columbic and tantalicacids, nor could he detect conclusively cerium or thorium.All doubton this head has been dispelled hy the use of his new method.Analysis of Xamarskite (and other Columbcites containing EarthyOxides).-5 grams of the finely powdered mineral dried a t 150” (lossby ignition is determined in a separate quantity) are placed in aplatinum capsule (50 C.C. capacity is sufficiently large) moistened withwater and treated with 4 to 5 C.C. of concentrated hydrofluoric acid ;after two or three minutes, a second and similar quantity of this acidis added, and the whole well stirred; vigorous action soon sets in,and in 5 or 10 minutes all black specks have disappeared, showingthe decomposition to be so far complete. The capsule is now heatedon a water-bath ; the contents consist of a clear solution (A) contain-i n g the metallic acids, iron and manganese, and an iusoluble portion (B)containing the earths and uranium oxide.These two parts are butvery slightly intermixed. The contents of the capsule are mixed with30 C.C. of water, warmed, filtered on a gutta-percha or silver funnel,the filtrate being received in a platinum capsule, and washed with hotwater containing a few drops of hydrofluoric acid. The solutiou isevaporated over a water-bath, and before it is dry, is treated withexcess of concentrated sulphuric acid to decompose the fluorides ; itis heated until nearly all the sulphuric acid is driven off, and whencool is washed into a glass flask and boiled with very dilute hydro-chloric acid. The whole mass is then filtered and washed with hotwater.The precipitate, consisting of columbic, tantalic, and small quan-tities of tungstic and stannic acids is dried, ignited and weighed ; sub-sequently the tungstic and stannic acids are separated in the mannerrecommended by Rose, and the columbic and tantalic acids by Marig-nac’s method. The filtrate contains iron and manganese, which areestimated in the iisual manner. The insoluble portion is washed intoa platinum capsule, treated with a sufficient quantity of concentratedsulphuric acid to decompose the fluorides, heated carefully until nearl112 ABSTRACTS OF CHEMICAL PAPKRS.all the snlphuric acid is driven off, and when cool is warmed with50 C.C. of water, in which all dissolves to a green solution, exceptperhaps a very minute quantity belonging to portion A.The greensolution is heated, and after the addition of a little nitric acid, ispoured into a beaker and made up t o about 250 C.C. with water. Itis now nearly neutralised with ammonia, and the earths precipitatedwith ammonium oxalate or oxalic acid. After six or eight hours, theearthy oxaZates (C) are filtered off and washed. The soIution contaJnsall the uranium with a trace of iron ; both are estimated ; the iron ispraecipitated as sulphide, and the urmium estimated by the ordinarymethod. The earthy oxnlates are dried, ignited, and weighed. Thepowder, which is of a pale yellow colour, is dissolved in nitric acid,evaporated to a syrup, and before it solidifies is mixed with a con-centrated solution of sodium sulphate, and subsequently small crystalsof the sulphate are added (50 C.C.o€ the solution and 4 grams of thecrystals for 1 or 2 grams of mineral), a precipitate (D) soon begins toform. After 24 hours it is very abundant and is filtered off and washedwith concentrated sodium sulphate solut,ion. The filtrate containingyttrium and erbium is treated with oxalic acid or ammonium oxalate,the precipitated oxalates ignited, redissolved, reprecipitated, ignited,and weighed. The oxides are dissolved in dilute sulphuric acid,evaporated carefully, and heated until the weight is constant: thenfrom the weight of oxide and of sulphate, the relative prnportion ofyttria and erbia is found by Bahn and Runsen’s formula. The sodiumsulphate precipitates are separated by fractional precipitation ; theyare washed with solution of t,he salt, dissolved in very diluted hydro-chloric acid, nearIy neutralised with ammonia, precipitated byammonium oxalate, ignited, redissolved, reprecipitated, and weighed.No cerium was found in the first precipitate, but instead a darkoxide soluble in very dilute nitric acid, which the author calledmosandrnm oxide.In the precipitate D are found the other ceriumoxides and tthoria. After being treated in the usual manner with hydro-chloric acid, ammonia, oxalic acid, &c., and ignited, the msultingoxides dissolved in dilute nitric acid, thus again showing absence ofcerium oxide. This fact was now confirmed by treating tjhe nitricacid solution with caustic soda and chlorine, no indications of ceriumwere obtained, all the oxides passing into solution with the exceptionof a small quantity of thoria, which forms 0% per cent.of the mineral.Some experiments were now conducted on a large scale, lead vessels,capable of working more than a kilo. of mineral a t a time, were em-ployed instead of platinum ones, The conduct of the process is thesame as described above, the earthy fluorides are decomposed withsulphuric acid, &c., and after the separation of uranium oxide, &c.,the solution is boiled up with steam, nearly neutralised, and the oxa-lates precipitated. The yttria is separated as above described ; thethoria can be separat,ed quantitative1.g by treating the solution withlarge quantities of soda and passing chlorine f o r two or three hours asabove described, or less completely by dissolving the mi-ths, separatedby the sodium sulphate, in nit& acid, concentrating the solution on awater-bath, diluting abundantly, boiling with a current of steam, andadding sufficient ammonia t o the boiling solution to precipitate about ANALYTICAL CHEMISTRY.113of the oxides present. The gelatinous precipitate containing all thethoria and a little of the other oxides, is filtered off, washed, and dis-solved in dilute sulphuric acid ; this sulphate of thoria can be easilypurified by the usual methods. The filtrate from the thoria containeda mere trace of thoria,, and no lanthanum, therefore the only oxide ofthe cerium series to be looked for was didymium, which the authorrecognised by the spectroscope.Quantitative E s t i m a t i o n of Didytn ium.-Solutions of didymum oxideof known strength are placed in tubes of uniform diameter, these arecompared before the spectroscope with similar tubes containing anitric acid solution of the oxides to be tested, until the bands pro-duced by the tubes correspond in intensity ; the author in this mannerarrived at the quantity of didymium in solution.From these resultsit is evident that the samarskite examined by the author containedthoria, and other mixed oxides in place of the cerium oxide.Columbite and tantalite are more difficult to decompose thansamarskite. They must be very finely pnlverised ; to effect this withtantalite, the powder in the mortar is covered with (95 per cent.com-mercial) absolute alcohol, and the t rituration coutinued under thesecircumstances. The trituration of 1 gram of columbite requires a quarterof an hour; 1 gram of tantalito 20 minutes. They are then decom-posed by heating with hydrofluoric acid of the usual strength used bythe author, and treated as described above. D. A. L.Dialysis of Arable Land.' By A. PETERMAN (Bied. Centr., 1883,36 1--364).-The employment of strong mineral acids in the analysisof soils has led the author to think that substances pass into solutionand are estimated, which under the ordinary conditions of cultivationare not attacked, and are useless to the plant. He has, therefore,tried a system of dialysis with distilled water, which he believesfollows better the processes of nature. The dialyser employed wassomewhat, but not materially different from the ordinary apparatus.It was found that after a dialysate of 10 days, not only the soluble' salts but a notable proportion of organic matter had passed into thedistilled water, colouring it yellow, yielding a dark brown residue aftercautious evaporation, and a carbonaceous residue on ignition.This sub-stance is different from the matiire noire of Grandeau, which does notpass through parchment,-paper. The author refers to old experinientsof Risler made in 1858, showing that plants derive a portion of theircarbon from the soil, and not entirely from the atmosphere.J. F.Titration of Copper by Means of Potassium Cyanide. ByJ.J. and C. BERINGER (Chem. News, 48, lll-l13).-The authorshave studied some of the conditions which affect the accuracy of thismethod for the determination of copper, and publish the results, &c.,in this paper.E f e c t of Manizer of Workin.q.-The result is not so much affectedby the time occupied in the entire titration as it is by the rate withwhich the last 2 or 3 C.C. are run in. The manner of finishing is,therefore, of paramount importance; and in order to obtain con-cordant results, a fictitious finishing point must be adopted.VOL. XLVI. 114C.C. Cyanide required . .Do. in presence of soda. .ABSTRACTS OF CHEMICAL PAPERS.With With With WithWithout' chloride. sulphate. nitrate. carbonate.24 *2 25 *7 25-7 25 .6 25 -625 '6 26.75 26.7 26-76 26.4------- --- -----Efect of Variations of Temperatwe.IC.C.Ammonia added.. 0c.c.Cyaniderequired.. 21.8----~ ~ ~~Temperature in C. 10". 20°. 30". 47O.-----_I--------C . C . Cyanide required . . . . . . . . . . 1 23.5 ! 23.3 I 23.05 I 22.85----------lo 1 2o 30 I 50 loo2 522.5 22.8 33.0 23.3 23.4 23.7 24.2- ~~Efect of Ammonium Xalt.Grams AmN03.C.C. Cyanide required..0-___________---24 '2 24 -6C.G. Cyanide 24.9 { required.Efect of Alkali Xalts.With Withsodium sulphatc. chloride.--.------,-24 -90 24 *9525 *05 25 *OO25 '10 25 '10C.C. Soda. 0C.C. Cyanide required . . 25.8Withnitrate.1 5 10 20 5026'0 26.4 26.9 27-25 28-__-------------24.9024 '9024 *9ANALYTICAL CHEMlSTRY. 115these may be summed up thus ; when the copper, ammonia, and volumeof liquid vary, but retain their relative proportions, the quantity ofcyanide used is proportional to the copper present.If, however, thecopper and volume are constant, whilst the ammonia increases, morecyanide is required. On the other hand, if the ammonia and copperare constant, whilst the bulk increases, less cyanide is necessary.Hence the authors inferred that by adding water and ammonia inproper proportions these errors might be got rid of. They have em-ployed successfully 15 C.C. of 0.880 ammonia to each 300 C.C. of liquid,in slightly alkaline solution, titrating with the necessary precau-tions. D. A. L.Colorimetric Estimation of Gold. By A. CARNOT (Co~zpt.rend., 97, 169-170.(See below.)New Reactions of Gold. By A. CARNor (Compt. rend, 97, 105-108).-A dilute solution of auric chloride is mixed with a smallquantity of arsenic acid, two or three drops of ferric chloride solution,and the same quantity of hydrochloric acid diluted with 100 C.C. ofwater, and a fragment of zinc are added. A purple colour is soon deve-loped in the neighbourhood of the zinc, and on agitation the liquidacquires a rose or purple tint. The same colour is also immediatelyproduced if, instead of zinc, the solution be mixed with a few dropsof an acid solution of ferrous chloride, or of the solution obtained byacting on metallic iron with arsenic and hydrochloric acids. This re-action succeeds with less than 0*0001 gram of gold in 100 C.C. ofwater.With phosphoric acid in place of arsenic acid, a violet orbluish coloration is obtained. Hydrochloric acid alone gives a lessintense rose coloration. Hydrogen gas, mixed with small quantitieso€ hydrogen sulphide, may be used instead of zinc or ferrous chloride.The rose o r purple liquid is perfectly transparent, can be filteredwithout being decolorised, and is unaltered at the end of threemonths. It is, therefore, a true solution, and is not a liquid holdingfinely-divided gold in suspension. If the solution is not distinctlyacid, or if it be mixed with certain salts, especially ammonium salts,a flocculent purple precipitate is slowly deposited. If the liquid istoo acid, the reaction does not take place, but finely-divided gold isprecipitated. The same decomposition occurs if the reducing agentis added too rapidly, or if the precipitate, when once formed, is re-dissolved in hydrochloric acid.I f the reducing agent is addedgradually, a considerable excess appears to have no action on therose or purple compound. The precipitate has a composition cor-responding with the formula Au20,19Fe20,,15As0,. It would appear,therefore, that the action of the reducing agent is only partial, aurousoxide and ferric oxide being formed. The ferric chloride, the presenceof which is essential, in all probability acts as a restrainer of thereducing action. C. H. B.Estimation of Arsenic: Pearce’s Process. By 0. J. FROST(Chem. News, 48, 85-86, Compare Abstr., 1883, 1034--1035).-1nthis process, the finely-powdered substance is fused with six to ten timesi 116 ABSTRACTS OF CHEMICAL PAPERS.its weight of the sodium carbonate and potassium nitrate mixture andthen proceed as described (Zoc.cit.), the solution, however, instead ofbeing evaporated to dryness, is exactly neutralised with ammonia andnitric acid, and filtered, if necessary. Excess of neutral silver nitratesolution is uow added ; the silver arsenate filtered off and washed withcold water, and the filtrate tested to see that precipitation is complete.The amount of silver is determined and the arsenic calculated from it.This is best effected by dissolving the arsenate on the filter in dilutenitric acid (any chloride is thus left behind), and titrating the filtratefor silver, with thiocyanate. Good results have been obtained bythis method, which the author states can be worked in half an hour.Molybdic and phosphoric acids interfere with the process ; antimonydoes not, on account of the insolubility of sodium antimonate.D. A.L.Volumetric Method for the Estimation of Arsenic. ByA. H. Low (Chem. News, 48,85).--d dcfenct.: of Pearce’s process (seelast abstract) a-gainst some criticisms made by McCay. The author con-siders that dissolving the precipitated silver arsenate in dilute nitricacid and titrating the silver is better than determining the excess ofsilver as McCay does (Abstr., 1883, 1035), for any chlorine present asimpurity in the reagents would have no effect in the first case,because any silver chloride present would not be dissolved, but wouldlower the results in the latter mehhod.He also recommends the useof a platinum instead of porcelain crucible for the fusion, as it can bemore easily cooled, and, from experience, he states that it remainsa(pparent1y uninjured, even after many times using.Test for Bismuth Subnitrate. (Dingl. p07yt. J., 248, 260.)-According to Hager, bismuth subnitrate and bismuth subamenatedissolve in eight parts of nitric acid of sp. gr. 1.185, forming clearsolutions ; the latter is, however, not completely soluble in a solutionsaturated with bismuth nitrate. Hence, by treating 0.5 gram of thesubnitrate with 4 grams nitric acid, a clear solution should be obtainedwithin half an hour, otherwise arsenate is present.D.A. L.D. B.Estimation of Manganese in Iron Ores. (DirzgZ. polyt. J.,248, 259.)-Zulkowsky recommends incineration of the manganoussulphide precipitate in a plttinum capsule, and moistening the residuewith aqueous sulphurous acid. The solution is evaporated on a water-bath, the residue taken up with water and two t o three drops of dilutenitric acid, and titrated with pot.assium permanganate. For thispurpose, the solution is transferred to a flask, diluted with water to150 to 200 c.c., and boiled. To the hot solution, potassium perman-ganafe is added from a burette, and the mixture boiled after each addi-tion until the supernatant liquid shows a pale red colour. Accordingto the formula 3Mn0 + Mn,O, = 5Mn02, 1 C.C. of a decinormal solutionof potassium permanganate corresponds with 1-65 mgrm.manganese.Volumetric Method of Estimating Manganese, especially inIron and Steel. By R. SCHOFFEL and E. DONATH (Din&poZyt. J.,248, 421--484).-To conduct this method, a solution of sodium car-D. BANALYTICAL CHEMISTRY. 117bonate is required, which does not reduce potassium permanganateeren on boiling ; also a solution of potassium permanganate of knownstrength. For this purpose sodium hydrogen carbonate is convertedinto the normal carbonate by strongly heating it. A saturated soln-tion of the carbonate is then prepared, which is heated to boilingand treated with a solution of potassium permanganate until themixture assumes a faint reddish colour, which remains on continuedboiling.The fluid is kept in well-stoppered bottles, and although itscolour soon disappears, it no longer nff ects potassium permanganate.The solution of potassium permanganate is titrated with pure ironwire or with ammonium ferrous sulphate. As this reaction is illus-trated by the equation lOFeO + Mn20, = 5Fe,O3 + 2Mn0, and theequation which explains the act'ion of potassium permanganate on themanganese salt is 3Mn0 + Mn20, = 5Mn02, 10 atoms of iron cor-respond with 3 atoms of manganese ; so that by multiplying the ironvalue by 02946, the permanganate value is obtained.For the analysis of samples rich in manganese, such as pig iron,spiegeleisen, and ferromanganese, 2 to 1 grams,-for poorer samples,4 to 3 grams,-are dissolved in boiling hydrochloric acid.The mixtureis allowed t o cool, treated with a small amount of potassium chlorate,and again boiled until all chlorine has been expelled. The solution, iftoo acid, is evaporated to a small bulk, partly neutralised with sodiumcarbonate, and made up to 100 C.C. 50 to 60 C.C. of the solution ofsodium carbonate are then transferred to a flask holding 700 to800 c.c., diluted with 400 to 500 C.C. distilled water heated to boiling,and treated with standard potassium permanganate. The test solutionis then run in from a burette, the mixture being stirred all the while.From the instant the colour becomes fainter, the solution should beadded with caution, and the mixture allowed to settle from time totime. The operation is concluded as soon as the red colour of thefluid has disappeared.The amount of permanganate solution usedshould be so regulated that one-half, or at least one-third of the testsolution is required, otherwise the results will be vitiated too much bythe limits of error. D. B.Suspended Matter in Water. By E. MARCHAND (Compt. rmd.,97, 49-50).-The water under examination is placed in a flask sur-rounded by black paper, in which are cut two opposite rectangularapertures, and a beam of light is passed through the water. Thismethod of examination reveals the existence of a large number ofsuspended particles which are invisible under ordinary conditions,but which the author has found to exist in large numbers in all formsof natural water, and even in distilled water which has been exposedfor some time to the air.Some of these particles are vacuoles con-taining water or gas, others have A shape resembling thatl of discoidaldiatoms. They have a sp. gr. higher than that of sea-water (in whichthey are very abundant), and they are not athacked by dilute acids oralkalis. Amongst these suspended corpusc!es are germs of Euglenea,a fact which explains the development of green growths in all placesexposed to light and moisture, and they also include organisms whichprobably play an important part in the oxidation of the organi118 ABSTRACTS OF CHEMICAL PAPERS.matter contained in water. Although some of them have a diameteras great as 2 mni., they are so flexible that they pass through theclosest filters, and when taken into the body they can pass throughthe kidneys, and are found in the urine.Test for Glycerol and Woody Fibre (Dingl.poZyt. J., 248,259.)-According t o Reichl, minute quantities of glycerol can bedetected by boiling the solution to be examined with a small amountof pyrogallol and a few drops of sulphuric acid diluted with an equalvolume of water ; this causes the formation of a red colour, which ischanged to purple on adding stannic chloride. Carbohydrates andcertain alcohols must be absent, as they produce similar colours. Byboiling woody fibre in a solution of stannic chloride mixed wih a fewdrops of pyrogallol, a fine purple colour is formed. This reaction canbe used as a meaiis of dyeing wood.C. H. B.3.). B.Valuation of Sugar-beets by their Density.By A. v. WACHTEL(Bied. Centr., 1883, 421422).--The custom of valning beet-rootsby their weight is a rough and ready rule ; the beet-juice on analysisfrequently shows considerable differences. The author recommendsKrocker’s process, that is, to prepare several mixtures of calciumchloride or cane-sugar solutions of different sp. gr., then with asampling tool to take pieces out of the root, placing them in solutionsof different densities until they float. Considerable differences arecaused by the manner in which the roots are cleansed f o r sampling,roots which are brushed and wiped showing a greater percentage ofsugar than those cleaned with warm water. J. F.Detection of Silver Cyanide. By C. L. BLOXAM (Chenz.News,48, 49).-The following test is of use in qualitative work. Precipi-tated silver cyanide appears amorphous under the microscope ; if,however, it is treated with ammonia and warmed, it forms needles.Silver chloride, treated in similar manner, forms octohedrons. In a,mixture of the two, both constituents can be recognised in this way.Silver thiocyanate also forms needles under like conditions, theabsence of thiocyanic acid must therefore be ascertained by the irontest, Silver cyanide also crystallises when boiled with a strong s o h =tion of sodium carbonate, or when moistened with strong nitric acidand warmed. Silver cyanide can be separated from silver chloride bytreating the mixture with hot dilute nitric acid in which the cyanideis soluble.If the solution is cooled, and the tube containing it iskept still, the cyanide separates in a semi-transparent gelatinous form ;if, however, the tube is agitated, the precipitate collects suddenly intoopaque masses (generally) of microscopic needles. D. A. L.Reactions with Silver Cyanide, Ferrocyanide, and Ferri-cyanide. By c. L. BLOXAM (Chenz. News, 48, 73--74).-Whenwhite silver ferrocyanide is shaken with ammonia it forms an opa-lescent liquid. When heated it becomes brownish-grey in colour, andmetallic silver is deposited. When treated with nitric acid, the greyprecipitate jields a solution of silver nitrate and ferric nitrate, leavinANALYTICAL CHEMISTRY. 119a residue of ferric oxide, mixed with a little silver ferricyanide.Nocyanide could be detected in this precipitate by means of ammoniumsulphide. With hydrochloric acid, a solution containing both ferrousand ferric chlorides is obtained.The ammoniacal solution neutralised with nitric acid yielded silverand ammonium cyanide. When silver ferrocyanide is treated with acold strong potash solution, a heavy brown precipitate forms. Thesupernatant liquid is colourless, stains paper brown, and on being heateddeposits silver ; it contains a trace of potassium cyanide, but no ferro-or ferri-cyanide. The brown precipitate contains metallic silver, andsilver ferro- and ferri-cyanide. Apparently, then, a part of the silverferrocyanide is converted into ferricyanide with separation of silver ; andthe ferricyanide t>hus formed forms a compound with the unchangedferrocyanide, which compound is not decomposed by cold potash or warmammonia. Boiling with potash entirely decomposes the ferrocyanide,yielding a solution of potassium and silver cyanide, and leaving a residueof ferrous and ferric oxides and silver.By boiling together precipitatedsilver oxide, silver ferrocyanide, and water, the colour changes frombrown to black, and silver and silver cyanide and ferricyanide areproduced. When silver ferricyanide is treated with potash in thecold it becomes black, yielding silver oxide and potassium ferricyanide ;on boiling, the black precipitate changes to pink, and ammonia isevolved, but on continuing to boil the precipitate again becomesblack. The pink precipitate contains a compound of silver cyanidewith silver ferricyanide insoluble in ammonia, along with silvercyanide, silver ferrocyanide, and ferric oxide.The pink precipitatecan be exactly reproduced by boiling silver oxide with potassinmferricynnide, on continuing the ebullition ammonia is evolved and theprecipitate becomes black. The filtrate from the pink precipitatecontains potassium cyanide and silver ferrocyanide in large quantities,and small quantities of silver cyanide and potassium formate. Thefinal black precipitate consists of metallic silver and ferric oxide ; thefiltrate from it contains the same constituents as the pink precipitatefiltrate, with the exception that there is no potassium formate, butinstead silver ferricyanide.When ammoniacal solutions of silvercyanide and silver ferricyanide are boiled together, a buff precipitateis produced, which behaves like the above-mentioned pink precipitate(thus synthesising the compound of silver cyanide with silver ferri-cyanide). The potassium cyanide then reduces the remaining silverferricyanide to ferrocyanide thus: 2Ag6Fe2Cy,, + 4KCN + 4H20 =SAglFeCy6 + K,FeCy, + 2HCN + 2C02 + 2NH3. When silver oxideand silver ferrocyanide are boiled together, the pink precipitate, silvercyanide, and f erricoxide are produced : Ag,Fe2Cyl, + 3Ag20 = l2AgCy +Fe,03. Several equations are given t o illustrate the reactions whichprobably take place during these changes. D. A. L.Test for Gallic Acid. By S. YOUNG (ohem. News, 48,31-32).-When aqueous gallic acid is treated with potassium cyanide solution,a red coloration is produced, which disappears in a short time if theliquid remains undisturbed.If, however, the liquid is shaken ener-getically the colour reappears, but again disappears on standing. Th120 ABSTRACTS OF CHEMICAL PAPERS.colonr can be reproduced in this manner from 15 to 20 times, thesolution finally becoming brownish-yellow. Pure tannic acid, freefrom gallic, is not coloured by potassium cyanide. D. A. L.Examination of Fatty Almond Oil. By H. HAGER (Diwql.Yolyt. J., 248, 524).-The author draws attention to the fact that theoil obtained from bitter almonds differs from that of sweet almondsby the ela'idin test, as it gives only a small amount of solid ela'idin.For the examination of the oil, 1 gram is treated in a small porcelaindish with 4 drops of concentrated sulphuric acid, and the mixturestirred together with a glass rod.A yellow colour rapidly changingt o yellow-red appears, which is Boon converted into a permanentbrown with green tinge, or green with brown tinge. By mixingequal volumes of fuming nitric acid and water with 7 volu. of almondoil, and agitating the mixture, the oil from sweet almonds gives a,white colour, whilst that from bitter almonds yields light to dark-yellow shades. D. €3.By A. v. BASTELAER ( B i d Centr., 1883, 419).-In order to determine the relations of water, fat, casein, and salt,the author takes 10 grams from the centre of a butter sample, placesit in a porcelain dish of 5 o r 6 cm.diameter ; dried at 100-120" toconstant weight, the loss is water; the residue is extracted withrectified benzene, the first portion poured on without stirring t o allowthe case'in to separate, the last two portions stirred up with a glassrod, and again dried ; the loss shows fat ; residue ignited is cssejin byloss ; the ash is salt. The limiis of a large number of determinationsare given :-Butter Analysis.Pure butter fat.. .... 75 to 85 per cent.Water ............ 9 to 15 ,,Case'in ............ 1 to 3 ,,Sodium chloride .... 5 to 10 ,,J. F.Tests for Vegetable Alhaloi'ds. By R. PALM (Chem. News,48, 65-66).-The author has shown previously that the alkaloidsare precipitated by solutions of alkaline sulphides or persulphides,and moreover that in contact with a solution of sodium thioantimo-nate, solutions of the alkaloid salts form characteristically colouredprecipitates consisting of the alkaloid hydrosulphides mixed with anti-mony sulphide. When the solutions of the alkaloid and reagent aredilute, these precipitates appear as colourless turbidities, which becomeyeliow on exposure to the air ; whilst with concentrated solutionsthey are yellow t o reddish-brown, and in saturated solutions theyform resinous masses.The precipitation is more complete in dilutesolutions, and is accelerated by gently heating, or by the addition ofstrong alcohol. In most cases the yellow precipitates are dissolvedby excess of the thioantimonate; they are, with few exceptions,amorphous, and dilute acids only partially separate the alkaloid fromthem.The chemical composition of the precipitates has not beendetermined. Sodium thioantimonate produces the following changeANALYTICAL CHEMISTRY. 121with the alkaloids referred to. With quinine sulphate in diluteneutral solutions, a white turbidity ; in stronger solutions, yellowflocks, which on shaking form resinous lumps, and become darker.When hot solutions of the quinine salt and reagent are mixed, resin-ous masses form at once, which when dry fall to a fine yellow powderlike lead iodide. With cinchonine sulphate, in dilute solutions, darkyellow (leather colour) flocks form at once; they do not coagulateeither on standing or heating. With quinidine sulphate, the eifect isalmost exactly the same as with the quinine salt, with the exception thatthe whole of the precipitate does not become resinous, and when dry isof a darker yellow colour (an intense dark chrome-yellow) : t'he pre-cipitation is also more complete.With morphine hydrochloride indilute solutions,. yellow flocks are at once deposited, which are darkerin strong solutions, and when dry resemble powdered gamboge incolour. With codeine hydrochloride, a flocculent precipitate is pro-duced, which when dry resembles the qninidine precipitate in tone,being a paler yellow than the morphine precipitate. With narcotine,in concentrated hot solutions, the precipitate coagulates in resinousmasses, which when dry have the colour of dry precipitated ferrichydroxide.With strychnine nitrate, the reaction is more sensitivethan with all the other veget'able alkalo'ids, the strychnine beingentirely precipitated, and moreover the precipitate is not soluble inexcess of the reagent. I n dilute solutions of strychnine nitrate,colourless flocks separate which become yellowish in air ; in concen-trated solutions, yellow flocks form which do not coagulate on stand-ing, and when dry are of a fine intense deep golden-yellow colour.With brucine nitrate, when the reagent is added in successive portionsto a, moderately concentrated solution of this alkaloid salt, three dis-tinct precipitates are obtained : 1. Reddish-yellow, which collects inresinous masses. 2. Light golden-yellow flocks. 3. Colourless flocks,which form a crust on the surface of the liquid.When the mixedprecipitates are boiled with water, the greater part dissolves, leavingan amorphous deep orange residue. The solution deposits yellowcrystals of the double sulphide.With atropine sulphate in strong solutions, a yellow deposit isformed, which coagulates on shaking or heating, but when dry is notso dark as the dry morphine precipitate.With bebeerine hydrochloride, a dark-coloured precipitate is formedwhich coagulates in strong, and especially in hot, solutions, and whendry is greyish-brown. The alkaloids also form double sulphides withother metallic sulphides.Lead chloride can be used as a reagent for vegetable alkalcids ; itshould be dissolved in a solution of sodium chloride, which dissolvesmore of the lead salt than cold water does. The precipitates aregenerally crystalline, and consist of a mixture of lead chloride andan alkalo'id salt. Quinine and brucine form crystalline powders ;cinchonine, morphine, and code'ine small fine needles : the strychnineprecipitate when dry forms a crystalline asbestos-like felted mass.The lead chloride is not so delicate a test as the thioantimonate.A strong solution of sodium chloride completely precipitates bebeerinefrom its solutions. D. A. L122 ABSTRACTS OF CHEMICAL PAPERS.Estimation of Urea. By L. HUGOUNENQ (Compt. rend., 97, 48-49).-Urine is filtered through animal charcoal, diluted with water,and heated in sealed tubes at a temperature above 140°, the ammo-nium carbonate which is formed being estimated by means of standardacid with “ methyl orange ” as indicator. This method is applicableto albuminous urine, if the albumin is previously removed by coagula-tion, but i t is not applicable to urine containing glucose or a notablequant,ity of magnesium. C. H. B.Estimation of Gluten in Flour. By L. REED (Chern. News, 48,63)-The proposed method of estimation is based on the fact that ayellow nitro-body is produced by the action of nitric acid on albumi-noids.Half a gram of flour is put in a test tube, which is graduated fromthe bottom to about half way up into 4 parts of equal capacity, wateris added up to the 4th mark, and the tube violently shaken. Thecontents are now temporarily transferred to a dry tube, the graduatedtube is masbed, and a quarter of the liquid poured back up to 1stmark, and colourless nitric acid is added up to 3rd mark. After fiveminutes’ standing, with occasional shaking, the liquid is filteredthrough a dry filter into a dry receptacle ; a standard flour is treatedin the eame way, aad the clear yellow solutions are examined colori-metrically, the qualities of the flours being inversely as the heights ofequal colour. D. A. L