ANALYTICAL CHEMISTRY. ‘73 A n a1 y t i c a1 C h e m i s t r y. Preparation of Starch Solution for Use in Volumetric Analysis with Iodine. By G. GASTINE (BUZZ. SOC. Chim., 50, 172 -173).-Five grams of potato-starch is mixed with 0.01 gram of mercury iodide, stirred with a little water, and poured into boiling water (1 litre). It is allowed to settle, and the clear liquid poured off. A solution prepared in this manner has been kept for more than a year without deteriorating. Use of Salicylic Acid for Preserving Standard Solutions. By H. BORNTRAGER ( Z e i t . and. Chem., 27, 641--642).-The addition of a pinch of salicylic acid to each litre of a thiosulphate solution greatly diminishes its tendelicy t o decompose. The author’s deter- minations show large variations, but not a progressive diminution in strength.M. J. S. Applications of Spectrophotometr y to Chemical Physiology. By E. LAMBLING (Arch. de Physiol., 4th Series, 12, 1--.34).-This 1)aper gives an historical account of the spectrophotometer, and of the principles upon which the spectrophotometric method depends. The practical application of the method for quantitative purposes in- xolves :-(1) The choice of a region of the spectrum; (2) the deter- mination of the coefficient of extinction of the coloured solution for that region ; and (3) the determination of the amount of absorption of the colouriig matter for the same region. When two colouring matters are mixed in a solution, they may also be estimated quantita- tively, provided the absorptive power of each of the two pigments for two regions of the spectrum be previously known.Finally, t,he application of such methods t o animal pigments (of the blood, bile, urine, &c.), is pointed out. Determination of Chlorine in Plant-ashes. By A. JOLLES ( Chem. Ceiztr., 1888, 863-864, from Zeit. Nahrunpnittel u. Hygiene, 2, 81).-The method the author proposes as the best is as follows:- The plant (10 grams) is incinerated gently in a platinum dish, moistened with an alcoholic solution of sodium carbonate, the alcohol burned off, aiid the process repeated; after this, the whole of the carbon may be burnt off without an1 fear of chlorine being lost, and N. H. M. W. D. H.74 ABSTRACTS OF CHEMICAL PAPERS. the ash may he extracted by means of water and the determination finished in the usual manner with silver nitrate.This deviation from the ordinary direct combustion of vegetable matter in the determina- tion of chlorine obviates the chance of a loss of this element which might otherwise take place. J. W. L. Method for the Determination of Bromine in Sea-water. By F. GUTZKOW (Chern. News, 58, 190-193).-25o C.C. of sea-water is mixed with a drop or two of sulphuric acid and with 100 C.C. of a solution containing about 25 grams of copper sulphate. It is then treated with sodium sulphite solution until the precipitate redissolves with difficulty, and heated until the blue colour returm; more sodium sulphite is now added, about one-third that already used, and the whole ag"n heated until blue, and then cooled. In this way, all the bromine is precipitated in a few minutes.After washing the precipitate, first with 100 C.C. of water containing 1 gram of sulplinric acid, then with a few drops of sodium carbonate solution, it is warmed with hydrochloric acid until all sulphurous anhydride is removed, and is finally treated with zinc. The filtrate and washings from the copper, which combined should not exceed 25 c.c., are titrated with sodium hypochlorite by a method of the author's which is described in detail in the original. The flask containing the liquid is fitted with a triple- bored cork ; through one hole passes a tube closed by a piece of india- rubber tubing and a clip, through the other holes two thistle-headed funnels reaching nearly to the bottom of the flask, one is roughly graduated. By blowing into the tube, a quantity of liquid is forced u p both funnels and is titrated in the graduated one, the other serving for comparing the colour ; in this way an indication of the quantity of hypochlorite required by the whole liquid is obtained, and so the operation can be finished with greater rapidity than by the ordinary mode of procedure.Determination of Fluorine in Substances Decomposable by Sulphuric Acid and especially in Natural Phosphates. Bg H. LASNE (BUZZ. Xoc. Chirn., 50, 167- liO).-Sufficient substance to yield 0.2 gram of calcium fluoride is put into a, flask containing strong sulphuric acid (50 c.c.) and pure sand (10 grams), and con- nected with two wash-bottles, the first of which contains 2.5 grams. and the second 0-5 gram of soda dissolved respectively in 25 C.C.of water. Dry air is passed into the flask, which is heated for an hour a t 180--260"; it is then allowed to cool, dry air being passed through. The soda solutions are united and boiled for a hour; phenolphthaleyn is added, and carbonic anhydride passed through the solution until it is colourless ; it is then heated for + hour at 50", and treated with ammonium carbonate, when the silica is almost com- pletely precipitated. The liquid is cooled quickly, diluted to 125 c.c., filtered through a large folded filter, and 100 C.C. collected. A few drops of tropeoline solution is added to the solution, which is carefully neutralised with dilute hydrochloric acid, It is then treated with pure sodium carbonate (equal to 0.5 gram of anhydrous carbonate), boiled until free from carbonic anhydride, and precipitated D.A. L.ANALYTICAL CHEMISTRY. 75 with a slight excess of calcium chloride. The precipitate is ignited, treated with acetic acid, evaporated to dryness, again treated with aretic acid, and the undissolved calciuni fluoride, washed, ignited, and weighed. N. H. M. General Method for the Separation and Volumetric Estima- tion of Acids : Application to Sulphuric Acid. By G. LINOSSIEI~ (BUZZ. SOC. Chin?., 50. 46-47).-The method is applicable to all acids which yield insoluble compounds when combined with metals which are precipitated by hydrogen sulphide in acid solution. The determinahn of sulphuric acid is carried out as follows :--The s o h - tion of the sulphate (containing 0.05 to 0.1 gram) contained iri a dish is treated with alcohol (2 vols.), heated almost to boiling, and precipitated with a slight excess of lead acetate.When cold it is maslied by decantation with a mixture of alcohol and water (0.5 to 1 vol.), passing the decanted liquid through a small filter. The trace of lead sulphate on the filter is washed with aqueous hydrogen sulphide into a flask ; the rest of the lead sulphate is shaken with a saturated solution of hydrogen sulphide, the liquid poured on to n filter, and the precipitate treated with a fresh amount of hydrogen sulphide and filtered. It is then washed with aqueous hydrogen sulphide until the filtrate gives no reaction with Poirrier’s orange. The whole of the sulphuric acid is then i n the filtrate, and is titrated with decinormal soda solution.N. H. 31. Volumetric Estimation of Boric Acid, and of Ammonia in Ammonium Salts. By J. MCGLASHAN (Chew. News, 58, 175- 176).-By using Poirrier’s oranges I and 11, boric acid and ammo- nium borate can be titrated directly with normal soda, but borax must first be made neutral to methyl-orange with aulphuric acid; boracite must be heated with dilute sulphuric acid, made neutral to methyl-orange with soda, and any carbonic anhydride eliminated before titrating. Ammonia is objectionable in any form except as borate or hydroxide, the latter when dilute is neutral to both orangeh. Therefore, with these indicators ammonia may be titrated with soda, in any ammonium salt, without distillation. With ammonium car- bonate and with arsenates, the end reaction is not distinct, but with hydrogen sulphide it is sharp. D.A. L. Resorcinol as a Test for Nitrates. By D. L~NDO (Chem. News, 58. 176--177).-For testing with resorcinol, 10 grams are dis- solved in 100 c . ~ . of water, and one drop of this solution, one drop of 15 per cent. hydrochloric acid, and 2 C.C. of concentrated sulphuric acid are added to 0.5 C.C. of the nitrate solution (compare Abstr., 1888, 1337) ; one in 500,000 gives a definite permanent purple colour after some time, whilst with increasing strengths of nitrate solution the colour becomes more intense, until with one in 10,000 the vivid purple- red colour is so intense as only to be distinctly seen i n the lower portion of the band. Resorcinol is valueless without hydrochloric acid, but with it, is a more delicate test for nitrates than ordinary phenol.Copper suiphate does not aid the reaction materially. The reagents76 ABSTRACTS OF CHEMICAL PAPERS. alone gave a band which, however, cannot be mistaken for the nitrate band. With nitrites, it is more delicate than with nitrates ; with chlorates it is no good, with dichromate (0.5 gram per litre) it gives a red to buff upper with a purplish lower band, the latter changing to pink or with hydrochloric acid to reddish-brown ; the colours are slightly different if the acid is run in at once, or after some delay. Perman- gsnate, N/10, with 4 vols. water, gives a, dark-orange upper and yellowish lower band ; with hydrochloric acid, these are respectively orange, red, and greenish. Hydrogen peroxide in dilute solution yields a green and brownish compound b;lnd.D. A. L. Estimation of Phosphorus in Iron and Steel. By P. W. SHIMER (Chem. News, 58, 165--168).-Dissolve 1 gram of iron in 20 C.C. of nitric acid, sp. gr. 1-20, and add to the boiling solution 10 c,c. of a solution of 20 grams of potassium permanganate in a litre of water, 2.5 C.C. a t R time, and after a few minutes 5 C.C. of hydro- cliloric acid, sp. gr. 1.12. When the action has ceased, add a mix- ture of 5 C.C. of concentrated sulphuric acid and 5 C.C. of water, and evaporate until fumes of sulphuric acid begin to come off. When cool, add 5 C.C. of nitric acid, sp. gr. 1.20, and sufficient water; boil to dissolve iron salts, filter, and wash with water. The residue serves well for the estimation of silicon. Heat the filtrate to 80" and add 5 C.C.of ammonium molybdate (5 grams of MOO,, 20 C.C. ammonia, sp. gr. 0.96, 30 C.C. nitric acid, sp. gi.. l a % ) , then keep at GO" until the liquid is clear, which occupies less than an hour when the solution is not too dilute. The yellow precipitate is washed with acid ammonium nitrate solution, dissolved in ammonia and pre- cipitated by magnesia mixture. The method gives good results both for silicon and phosphorus. A few experiments were made using sulphuric acid and ammonium sulphnte instead of nitric acid and ammonium nitrate for the molybdate precipitation, with satisfactory results. I). A. L. Influence of Sulphur on Eggertz's Carbon Colour-test. l 3 ~ T. W. HOGG (Chern. News, 58, 175).- When ordinary steels contain- ing say 0.05 per cent.of sulphur are dissolved in the usual manner, the sulphur separates out and produces a turbidity which interferes with the colour-test for carbon ; a fact easily proved by dividing such a solution into two portions, filtering one, and then comparing the colour produced in the filtered and unfiltered solutions. Conse- quently, if a common steel is compared with a standard of pure steel, the colour intensity is sure to be over-estimated, and vice versd. D. A. L. Quantitative Analysis by Electrolysis. By A. CLASSEN and R. SCHELLE (Ber., 21,2892-2899).-The current from two of Farbaky and Schenek's accumulators, fully charged by a dynamo giving R current of 20-25 amperes, was employed in making 50 analyses in the course of six days. Four to eight analyses were made simul- taneously, and the current was employed continuously day and night, except during the short intervals required for changing the platinumANALYTICAL CHEMISTRY.7 7 dishes, During this time the strength of the current decreased from 2.05 to 1.92 volt, so that one charge would be sufficient for 60 to 70 analyses. The condition of the accumulator can be ascertained from the specific gravity of the sulphuric acid, which is 1.240 when the accumulator is charged and 1.118 when i t is not charged. The pre- cipitated metal is in a more suitable condition than when a battery or dynamo is employed. Experiments which were made with Neumann’s (Abstr., 1888, 529) and Wolffs ( Z e i t . ung. Chem., 1888, 296) voltameters, employ- ing currents of various strengths, show the necessity of working under the same conditions when repeating electrolytic methods of analpis (compare Classen, Quantitative Analyse durch Electrolyse, 2nd ed., 43).In the separation of antimonyfroin tin in sodinm sulphide solution, the solution, freed from antimony, is boiled with ammonium pulphate, and the tin precipitated by electrolysis (compare Abstr., 1884, 932). A simpler and more convenient method is to convert the stannic sulphide into hydrogen stannic oxalate and electrolyse the solution. For this purpose, the solution is acidified with dilute sulphuric acid, and the sulphide oxidised with ammoniacal hydrogen peroxide, or the hot alkaline solution is treated a t once wi+h hydrogen peroxide until it becomes colourless, acidified with sulphuric acid, neutralised with ammonia, and more hydrogen peroxide added.The solution is then boiled, filtered, and the residual stannic acid washed off the filter with and dissolved in a hnt solution of oxalic acid. If there is a residue of sulphur, i t is separated by filtration, washed with a cold saturated solution of hydiwgen animonium oxalate, and the washings added to the filtrate. The solution, which should contain at least 50 C.C. of the hydrogen ammonium oxalate solution, is electrolysed with a current giving 8-10 C.C. of explosive gas per minute. The electroljtic deposition of copper from its salts, dissolved in a saturated solution of ammonium oxalate, is hastened very consider- ably if the solutian is kept acid, especially towards the end of the operation, by adding excess of a cold, saturated solution of oxalic acid.The smaller the quantity of copper, the more oxalic acid soln- tion may be added. A solution poor in copper can be mixed with the oxalic acid solution a t the commencemtnt of sthe process, but in concentrated solutions, the precipitation must be carried out in a solu- tion as nearly neutral as possible to avoid the separation of copper oxalate. If the copper solution is kept a t 40-50°, aLout two grams of copper are deposited in from 3 to 4 hours. A current giving 3-4 C.C. of explosive gas per minute mas employed, and quantitative experiments showed that the whole of the copper is deposited. Separation of Calcium, Barium, and Strontium. By KUPFFERSCHLAEGER (Bull. SOC. Clzim., 49, 854- 856).-A quantitative analysis of a mixture of the carbcnates of barium, strontium, and calcium can be carried out as follows :--The mixture is dissolved in a slight excess of very dilute nitric acid, the solution evaporated to dryness, the residue dissolved in distilled water, and the filtered solu- tion again evaporated to complete dryness, The residue of mixed nitrates is agitated with a small quantity of a mixture of absolute I?.s. I(.78 ACSTRACTS OF CHEMICAL PAPERS. alcohol and ether, and the solvent separated by filtration as soon as the solution becomes clear ; this process is repeated three times, the propor- tion of ether being gradually increased until the mixture contains equal volumes of the constituents. The residual mixture of the nitrates of barium and strontium is dried, dissolved in water, and treated with a cold saturated solution of potassium dichromate. The precipitated barium dichroniate is washed with cold, very dilute a,lcohol, and con- verted into sulphate by heating with sulphuric acid.The strontium in the filtrate is converted into sulphate by warming the solution with dilute sulphuric acid. A solution of strontium chloride is not precipitated by potassium dichrornate. F. S. K. Volumetric Estimation of Mercuric Chloride. By G. KASSRER (Arch. Phawn. [ 3 ] , 26, 595-604).-The method promises well for pharmaceutical practice. 50 grams of the organic compound contain- ing mercuric chloride is placed in a porcelain dish without previous division, 500 C.C. of water is added, and the mass is well kneaded with a pestle.500 C.C. of a solution of 0.4 gram of potassium anti- monions tartrate and 1 gram of sodium phosphate, or in place of the latter 1 to 2 grams of sodium acetate, is now added and the kneading continued. A very intimate mixture is thus obtained, and a milky liquid results from the decomposition of the mercuric chloride. About 500 C.C. is now filtered off, and titrated with decinormal iodine solution after the addition of freshly prepared starch solution, and a suffirient quantity of sodium hydrogen carbonate. The iodine employed corresponds to the amount' of unchanged tartrate remaining in the filtrate. The reaction is as follows :-4HgCI, + Sb,O, + 2H,O = 4KC1 + Sb,05 + 4HgC1. The presence of phosphoric and organic acids is admissible, but not of free hydrochloric acid.By A. STRENG (Jahrb. f. Mi%, 1888, ii, Mem., 142-152 ; continuation of Abstr., 1886, 4877.-For detect- ing tin under the microscope, t'he author avails himself of the brown colour produced on adding hydrochloric acid and platinum chloride. When the solution is effected, a drop of the liquid is transferred to an object-glass, and a grain of potassium chloride added, and the solution slightly evaporated. In this way, rhumbic crystals of pot'assinm stannous chloride (2KC1 + SnC1, + H,O) are form_ed. They-are mostly combinations of the forms mP, mPm, 03Pq Pm, Pm. When the staunous chloride has been thus detected, a drop of nitric acid is added, and the solution heated. The stnncous chloride is converted into stannic chloride, and the imperfectly soluble salt, K,SnCl, is formed.This crystallises in the regular system, and is perfectly isotropic. This reaction may be employed for detecting stannous and stannic oxides in compounds soluble in hydrochloric acid. Caesium chloride may be used instead of the isomorphous potassium salt. The author also describes methods for detecting potassium, caesium, and rubidium, sodium, and silicon. Detection of small Quantities of Germanium. By K. HAUSHOFER (Clrern. Centr., 1888, 867, from Xitzber. Akad. JIGnche~~, 1887, 133). J. T. Microchemical Reactions. B. H. B.AX ALP TIC AL C €1 ENIST R Y. 70 -By heating in an atmosphere of hydrogen snlphide, the germanium in the mineral argyrodite is converted into the sulphide GeS, which is crystalline, and may be detected and yecognised by means of ihe microscope.With concentrated sulphuric acid, it forms a white, non- crystalline substance, with concentrated nitric acid it is converted into the white, crystalline oxide, GeOz, which is soluble in dilute nitric acid and water, and crystallises out of the solution again on evapora- tion. Heated in a tube, it sublimes similarly to antimony oxide, but differs from this by its solubility in water, and also by melting t o clear colourless drops. It, is necessary to apply the potassium iodide test also, as a means of distinguishing it from the mercuric sulphide, which has also been found present in argyrodite. J. W. L. Characteristic Reaction of Bismuth. By E. L ~ G E R (BUZZ. XOC. Chi%., 50, (31-93) .-A solution of bismuth iodide and potassium iodide is sometimes used for detecting alkalo'ids, with which it gives orange-yellow precipitates.It is suggested that the reaction should be employed for detecting bismuth. A solution is prepared by dissolving cinchonine (1 gram) and potassium iodide (2 grams) in water (100 c.c.). Other alkaloids may be nsed instead of cinchonine, but thir seems to give the most sensitive reaction. The reagent must be added in excess ; the presence of too much nitric acid, and espe- cially the presence.of hydrochloric and sulphuric acids, is to be avoided. Bismuth may be detected in solutions containing only 1 part in 500,000 parts. Solutions which contain &her metals besides bismuth are precipitated with hydrogen sulphide, the sulphides of copper, lead, cadmium, mercury, and bismuth converted into nitrates, then in to carbonates, and the caiabonates of bismuth and lead separated by means of potnssium cyanide ; these carbonates are converted into clibrides, and the lead chloride separated by means of alcohol.The alcoholic solution is evaporated to dryness, c?issolved in a drop of nihric: acid and sorne C.C. of water, and treated with the rexgent. Solu- tions of the salts of mercury, cadmium, silver, copper, and lead also give precipitates of various colours with the cinchonine reagent. N. H. M. Determination of Oxygen dissolved in Water. By L. W. WINKLER (Ber., 21, 2846-2854) .-The method consists in oxidising :in excess of manganese hydroxide in presence of alkali by the oxygen present in a weighed amount of the water; potassium iodide and hydrochloric acid are then added, and the iodine which separates (which is equivalent to the amount, of dissolved oxygen) titrated with sodium thiosulphate.A solution of maiiganous chloride (free from iron) is made of such a strength that 100 C.C. contains 40 grams of the salt (MnC1, + 4H20). The soda must be free from nitrate, and the solution prepared of eight times the normal strength ; potassium iodide (10 grams) is dissolved in 100 C.C. of the soda solution, the rest of which is kept. A strong flask of about 2 litre capacity is filled with the water ; 1 C.C. of the potassium iodide soda solution is added by means of a pipette reaching nearly to the bottom of the flask, then 1 C.C. of the manganous chloride solution. The flask is closed, care being taken t h a t no bubble of air remains, and the contents mixed.80 ABSTRACTS O F CHEMICAL PAPERY.When the precipitate settles, fuming h~drochIoric acid (3 s.c.) is added by means of a pipette similar to those previously used. The flask is again closed, the contents mixed, and the yellow liquid titrated in the usual manner with sodium thiosulphate, the most convenient strength of which is 1/100 normal, so that each C.C. corresponds witli 0.055825 C.C. of' oxygen (at 0" and 760 mm. pressure). When the water to be analysed contains mucii carbonic anhydride, more of the reagent must be added, as manganous carbonate is not oxidised by the oxygen. In the case of waters containing nitrates, the process has to be modified : a soda solution containing no potassium iodide is first added, then hFdrochloric acid (twice the amount other- wise used), and after three minutes a solution of potassium iodide.A manganic chloride solution is then prepared as follows :--500 C.C. of distilled water is treated with the pure soda solution (1 c.c.), 5 to 10 drops of the manganous chloride solution, and then sufficient hydro- chloric acid is added to dissolre the precipitate. 100 C.C. of this solution is taken out and diluted with distilled water; to the rest, 100 C.C. of the water to be examined is added. After two to three minutes, both solutions are treated with potassium iodide and the iodine which separates determined as usual. From the difference in the amount of thiosulphate used, the correction for 100 C.C.of water is calculated. Concordant results obtained b j the new method are gil en. N. H. M. Ash Determination. By F. A. FL~~CKTGER (Zeit. anal. Chem., 27, 637-638).-The substance is heated in a roomy platinum capsule so gently that carbonisation takes place without combustion. It ir then cooled, a copious amount of water is added, and the whole eva- porated completely on the water-bath. On reheating the carbonaceous rebidue very gradually, it burns a t a low temperature, and very quickly. &I. J. S. Wet Methods of Organic Analysis. By J. MESSIKGER (Bey., 21, 2910-2919) .-Organic compounds, as Cross and Bevan have show11 (Proc., 1888, 76, and Trans., 1888, 889), are completely oxidised when warmed with chromic acid and concentrated sulphuric acid. I f nitrogen is present, it is evolved as such or as ammonia, whilst sulphur, phosphorus, and arsenic are converted into the corresFond- ing acids.Halogens are evolved in the free state, and metals remain as sulphate or, with a large excess of chromic acid, as chromates. To estimate the quantity of carbon, the substance (0*15-0*35 gra,m) is weighed in a small bulb or tube, and placed in the apparatus employed by Classen for the estimation of carbonic anhydride (Quudtative Analyse, 3rd ed., 239), together with chromic acid (?5--.6 grams) or powdered potassium dichromate. A gentle stream of air, free from carbonic anhydride, is passed throngh the appa- ratus to drive out the carbonic anhydride, the weighed potash bulbs and sods-lime tube are attached, and the latter connected with a calcium chloride tube to prevent absorption of moisture from the air.Concentrated sulphuric acid (30 c.c.) is then poured throughANALYTICAL CHEMISTRY. 81 tlle funnel, and the stream of air stopped. The flask is warmed very cautiously until the evolution of carbonic anhydride commences, and then the heating is immediately discontinued until the reaction is almost at an end. Pure air is passed through the apparatus for half an hour, and the tubes weighed. Tlie results obtained are very satisfactory, except in the case of substances which sublime readily, but great care must be taken when heat is first applied. I n analysing volatile substances, the bulb is broken by means of the funnel. If the substance con- tains halogens, a Dreschler’s flask of about 100 C.C.capacity, contain- ing 40 C.C. of concentrated potassium iodide solution, and a small U -tube filled with glass-wool, half of which is moistened with a solu- tion of silver nitrate and half with concentrated sulphuric acid, are interposed between the condenser and the potash bulbs. Quantitative experiments with compounds of the most varied nature gave satis- factory results (compare Cross and Bevan, Zoc. cit.). Sulphur can be estimated, except in the case of extremely volatile substances, as follows :-The substance (0.15-0.35 gram), together with potassium permanganate (1&2 grams), and pure potash (0.5 gram) is placed in a flask of 500 C.C. capacity provided with a con- denser, water (25-30 c.c.) poured down the condenser, and the mixture heated for 2-3 hours.Concentrated hydrochloric acid is then gradually poured down the condenser into the cold inixture, which should be of a reddish colour, mid after the evolution of gas has ceased, the whole is heated until the liquid beconies clear. The sulphuric acid is then precipitated with barium chloride. Potassium dichromate (2-3 grams) and hydrochloric acid (20-25 C.C. : 2 parts concentrated acid, 1 part water) may be employed instead of potas- sium permaiiganate and potash. The operation is carried out in the manner described, but after heating for about two hours a few drops of alcohol are added to determine whether all the dichromate has heen reduced; if the odour of aldehyde is perceptible, the mixture iuust be heated again and the test repeated. Both methods can also be employed for the detection of sulphiir.Numerous quantitative cbxperiments gave satisfactory results, but in the case of a few sul- phones i t was found that sulphur cannot be estimated by this method. Phosphorus, arsenic, and antimony in organic compounds can be estimated by placing the substance (0.3-0.4 gram) with chromic acid (4.-5 grams) in a flask provided with a condenser, pouring sulphuric acid (10 c.c., 2 parts concentrated acid, 1 part water) down the con- denser, and heating gently. After an hour’s time, sulphuric acid (10 c.c.) is added, and the heating continued for about an how. The mixture must always be heated very carefully, and the cold solu- tion should be perfectly clear. Some antimony compounds require only 1 gram of chromic acid and 10 C.C.of sulphuric acid. In the estimation of phosphorus, the solution is warmed with ammonium nitrste (3-4 grams) and ammoniuni molybdate solutior! (50 c.c.) for 2-3 hours, filtered, the precipitate washed 6 to 8 times by decantatioii with ammonium nitrate solution (20 grams in 100 C.C. of water), throan on to a filter, and dissolved in not more than 40- 50 C.C. of warm dilute (2 per cent.) ammonia ; a concentrated solution The process occupies about two hours. VOL. LVI. 982 ABSTRACTS OF OHEMIOAL PAPERS. of citric acid (4-5 drops) is added to the filtrate, and the phosphoric acid precipitated with magnesia mixture. In estimating arsenic, the solution is diluted to about 100 c.c., heated to about 70", and the arsenic precipitated with hydrogen sul- phide.The precipitate is washed with water containing hydroqen sulphide until free from chromium salts, and then converted into arsenic acid by means of ammoniacal hydrogen peroxide (compare Classen, Abstr., 18b3, 934). After boiling for an hour, ammonia is added to the filtered solution, and the arsenic precipitated with mag- nesia mixture. Antimony is estimated by adding potash and excess of sodium sulphide, boiling for half an hour, and precipitating the metal electrolytically. Metals are estimated by mixing the solution with excess of ammo- nium oxalate and precipitating electrolytically. Halogens can be detected by warming the substance (1-2 mgrms.) with chromic acid and sulphnric acid, and passing the gas evolved into a dilute solution of potassium iodide ; quantitative experiments gave unsatis- factory results.F. S. K. The Safety of Commercial Kerosene Oils. Bp S. B. NEWBURY and W. P. CUTTER (Amer. Chem. J., 356-362).-Although oils are regularly testea for their flashing points and conclusions drawn as to their being sa€e for burning in lamps, it is noteworthy that many modern lamps heat their reservoir of oil to temperatures above the legal flashing point, and that there are not sufficient experimental data to allow of the correct interpretation of the flashing point deter- minations into +terms expressing the inability of such oils to form explosive mixtures of vapour with air. All hydrocarbons up to and including octane, form at ordinary temperatures mixtures that can be exploded; nonane mixtures (b. p. 148-1.50") explode sharply at 79" F., and decane at 104" F.The addition of small quantities of low-boiling oils materially lowers the flashing point of another sample. Thus a sample flashing at 212" F. was made to flash at 145" F. by the addition of 5 per cent. of heptane; at 110" F. by the addition of 5 per cent. pentane or hexane, or 10 per cent. heptane; at 96" F. by the addition of 15 per cent. heptane ; and at ordinary temperatures by the addition of 10 per cent. pentme or hexane, or 20 per cent. heptane. The tem- peratures at which am oil may be kindled in an open vessel approaches the more nearly to the Bashing point the more homogeneous the oil is. The flashing point of an ordinary oil may be materially raised, with but little diminution of weight, by passing a current of air through it for several hours.H. B. Solubility and Estimation of Paraffin. By B. PAWLEWSKI and J. FILEMONOWICZ (Ber., 21, 2973-2976) -The following table gives the solubility at 20' of ozokerit paraffin of sp. gr. 0.9170 at 20°, melting at 64-65', and solidifying at, 61-63' :-ANALYTICAL CHE3lISTHY. - Solvent. Carbon bisulphide.. ................... Light petroleum, up to 75" G., sp. gr. 0'7233 ............................ Turpentine oil, sp. gr. 0 -857 ; b. p. 158- 186". .............................. Cumene (comm.), up to 160°, sp. gr. = 0 *867. ............................. Cumene (frac.), 150-160", sp. gr. = 0.849 Xylene (comm.), 138- 143",sp. gr. = 0.866 (frac.). 136-1383, sp. gr. = 0.864. . Toluene (comm.), l(B--llO",sp.gr. = 0.866 (frac ), 108.5-109 5", sp.gr. = 0-866 .............................. Chloroform .......................... Benzene ............................. Ethyl ether .......................... Isobutyl alcohol (comm.), sp. gr. = 0.804. 1l:tliyl acetate. ........................ Amy1 alcohol, 127-129", sp. gr. = 0813 . l'ropionic acid.. ...................... )'ropy I alcohol ........................ Met hpl alcohol, 65.5 - 66 5 O , ~ p . gr. = 0.798 JIetli! 1 forniate ....................... (fluci 11 acetic acid .................... ,, ,, Acetone, 55-5-56 5", sp. gr. = 0.797.. ... Ethyl alcohol, 99.5' Tr.. ............... Ethyl alcohol, 945" Tr. ................ Acetic anhydride.. .................... l?ormic acid (crjst.) ................... Ethyl alcohol, 75" Tr.. .................Paraffin (grams), dissolved by 100 grams 12 *99 11 '73 6.06 4.28 3 -99 3.95 4.s9 3 -83 3 -92 2 '42 1 -9Y 1 -95 0.288 0 '262 0 '238 0 * 819 0 -20% 0 -166 0 -141 0 ' o n 0 -060 0 -060 0 -046 0 -028 0 -013 0 *Wd03 -.- 100 C.C. - 8 *48 5 -21 3-72 3 -39 3 -43 3.77 3 -34 3 -41 3 -61 1 -75 0 -228 0 '209 0 '164 0 -056 0 -063 0 '015 - - - - - c - - Weight of solvent required to dis- solve completely 1 part d paraffin. 7.6 8'5 16 *1 23 -4 26-0 25 -1 22 -7 26 -1 25 *5 41 '3 60 '3 50.8 352 -9 378'7 419.0 453 -6 495 -3 595 -3 709 -4 l a 7 -5 1648 -7 1668 -6 2149 * 5 3856 -2 7689 -2 330000 *o The liquid constituents present in man-j- prodncts of the petroleum or ozokerit industry are soluble in glacial acetic acid, whereas vaselin, cerisin, ozokerit, and paraffln are almost insoluble.To estimate the quantity of solid paraffin in petroleum, lubricating oils, mineral oils, vaselin, &c., 5-20 C.C. of the mixture is well shaken with 100-200 C.C. of glacial acetic acid, the residual para& thrown on to a weighed filter, washed two or three times with glacial acetic acid, and then two or three times with alcohol of 75" Tr., dried and weighed, or the residual paraffin is washed, dissolved in benzene or ether, the solution evaporated, and the residue weighed. This method is quick and accurate, and can be carried out at the ordinary temperature, F. S. K. Analysis of a Mixture of Silver Chloride, Cyanide, Thio- cyanate, Ferricyanide, and Ferrscyanids. By J. TELSS~ER (BUZZ. Soc. Chim., 50, 10&106).-The mixture occurs in the analysis of materials used in the purification of coal-gas. Sodium carbonate is heated in a crucible until aahjiirous, the weighed substance and Borne81 ABSTRACTS OF CHEMICAL PAPERS.potassium nitrate are added, and the whole is heated. The product is extracted with water, which leaves a residue of pure silver and ferric oxide. In the solution, sulphur is determined as barium sul- phnte and chlorine as silver chloride. The residue is dissolved in nitric acid, the silver determined as chloride, and the iron as sesqui- oxide. Equations are given by means of which the amount of each salt present in the mixture is calculated. Estimation of Paranitrotoluene. By F. REVERDIN and C. DE LA HARPE (Bull. Xoc. Chim., 50, 44-46).-The method is based on the fact that paranitrotoluenesulphonic acid is readily converted by boiling with soda into dinitrosulphostilbene which yields a red colour when dissolred in alkali, whilst orthonitrotolnenesulphonic acid is not attacked by soda, and the alkaline solution is yellow.2 C.C. of pure orthonitrotoluene is heated in a water-bath with 6 C.C. of sul- phuric acid (containing 25 per cent. of anhydride) for three hoiirq, the product cooled and diluted to 1 litre. I n a similar manner a mixture of orthonitrotoluene (96 parts) and paranitrotoluene (4 parts) is sulphonated and the product diluted to 200 C.C. I n determining the amount of paranitrotoluene, the substance and nitrotoluene (con- taining 4 per cent. of the para-compound) are sulphonated, the pro- ducts diluted to 200 c.c., and the colours compared which are obtained by mixing 1 C.C.of each solution with 5 C.C. of aqueous soda. If the sample to be aiialysed gives a stronger colour, a measured quantity (50 to 50 c.c.) is progressively diluted with the solution of ortho- nitrotoluenesulphonic acid until 1 C.C. when heated with 5 C.C. of aqueous soda gives the same amount of colour as the solution con- taining 4 per cent. The percentage of paranitrotoluene can then be calculated. N. H. &I. Detection of Methyl Alcohol in Wood Spirit. By J. HABER- MAXN (Zcit. anal. Chem., 27, 663, from VerAandl. nut. Ver. Byutm, 26). -Commercial methyl alcohol contains impurities which reduce per- manganate energetically. Cazeneuve and Cotton shake 10 c c. of the spirit to be tested with 1 C.C. of a 0.1 per cent.solution of per- manganate a t 20'. If wood spirit is present. decolorisation takes place immediately ; with pure alcohol, 20 minutes is required. Ethereal oils, which may be present in spirit o r brandy, and would cause a similar reduction, may be removed by shaking the spirit twice with half its volume of the purest olive oil and then filtering through a well wetted filter. N. H. M. If sugar is present, the spirit must be distilled. M. J. S. Composition of Natural Brandies and the Way of Dis- tinguishing them. By X. ROCQUES (Bull. SOC. Chim., 50, 157- 164).-500 C.C. of the brandy is distilled in a, Le Bel-Henniger ap- paratus and nine fractioiis of 50 C.C. each collected, the temperatures being noted. Each fraction is subjected t o the following tests:- (1) Rosaniliue bisulphite, (2) aniline acetate, (3) sulpharic acid, (4) permanganate, (5) ammoniacal silver nitrate solution.The results of several analyses of brandies are given in tables. N. H. 31.ANALYTICAL CHEMISTRY. 85 Estimation of Sugar in Presence of Carbohydrates. By R. W. BISHOP (Chern. Centr., 1888, 952-953, from Msc. ,Won. Sci., 32, 558).-The author has carried out a number of experiments with a view to determine the conditions under which inversion may be completed without a t the same time damaging the accompanying carbohydrates. It was found that sulphuric acid has a greater power of inversion than hydrochloric acid, but it appears that hydro- chloric acid is the best for the conversion of starch into dextrose. Prolonged heating of inverted sugar with the acid seems to act on the lmvulose rather strongly, and the solution becomes less lmvo- rotatory and eventually dextrorotatory.For the inversion of cane- mgar, 0.5 C.C. of hydrochloric acid is snfficient, and the heating is con- tinued for 10 minntes at 95". For the conversion of dextrin, the solution should not contain more than 1-3 grams of carbohydrates i n 50 C.C. The inversion is performed by adding 2 C.C. of hydrochloric acid and heating for three hours at 95". J. W. L. Detection of Sugar in Urine. By C. SCHWARZ (Arch. Pharnz. [3], 26, 796, from Pharm. Zeit., 33, 465).-1 to 2 C.C. of lead acetate is added to 10 C.C. of urine and filtered ; 5 C.C. of the filtrate is mixed with 5 C.C. of normal potash solution and one or two drops of phenylhydrazine, well shaken and vigorously boiled ; in the presence of sugar, the liquid becomes lemon- to orange-yellow, and becomes opaque on adding an excess of acetic acid owing to the immediate formation of a finely divided yellow precipitate.In the absence of sugar, this precipitate never OCCUTS with urine. J. T. Detection of Chloral or Chloroform in Liquids. By C. SCHWARZ (Zeit. anal. Chem., 27, 668-669, from Pharm. Zeit., 33, 419).-Either of these substances when boiled with resorcinol and an excess of soda gives a red colour, which disappears on acidifying and is restored by alkalis. If, on the other hand, an excess of resorcinol and only a dropor two of soda solution is used, the product is n yellowish- red solution with intense yellowish-green fluorescence.0.0001 gram of chloral hydrate in 1 C.C. gives this reaction very diwtinctly when vigorously boiled with 0.05 gram of resorcinol and five drops of soda solution. Rf. J. S. Modification of the Reichert-Meissl Metho& of Butter Analysis. By M. MANSFELD (Chem. Centr., 1888, 870-871, from Milch Zeit., 17, 281-283).-1n order to obviate the possible error arising from the use of alcohol in the saponification of the butter-fat, theauthor has tried the use of alkali alone, the latter being added to the fat in a small flask and heated on the water-bath for two hours, a t the end of which time the saponification is complete, and the process is finished as Wollny prescribes. J. W. L. Densities and Refractive Indices of Certain Oils. By J. H. LONG (Amer. Chem. J., 10, 392--405).-The following values arc obtained-the densities a t 20" being compared with water at 4' as86 ABSTRACTS OF CHEMICAL PAPERS.unity, and the substance weighed in a vacuum; the refractive indices are for sodium light a t 20" :- Olive oil .......... 0.9130 Ref. index 1.4703 Cotton-seed oil. 0.9191 2.4732 Sesame oil ........ 0,9191 1.4740 Mustard oil.. 0.9121 1.4742 Castor oil.. 0.9589 1.4791 Lard oil 0.9122 1.4686 Peanut oil ........ 0.9173 1.4717 The densities and refractive indices are also given for other tempe- ratures than 20". In nearly all cases, the variations due to temperature are t,he same, namely, about -0.G0068 i n denpity and -0-0004 in refraction for each rise of 1". The author believes such determiuations may prove of value in the identification of other oils..... ...... ........ .......... H. B. Bechi's Newest Method for the Detection of Cotton-seed Oil in Mixtures. By G. Biz10 (Chem. Centr., 1888, 873, from Atti Inst. Veneto [ 6 ] , 6 ) .--The author finds that pure olive oil gives Bechi's new cotton-seed oil reaction with slightly acid silver nitrate, whilst, on the other hand, he finds that there are some cotton-seed oils which do not give the reaction at all. J. W. L. Qualitative Test for Resin Oil in Vegetable and Mineral Oils. By HOLDE (Chem. Centr., 1888, 952, from Pharna. Zeit., 33, %8).- Whereas Storch's reagent (concentrated sulphuric acid and arihjdrous acetic acid) for the detection of resin oil is not always admissible, the author recommends sulphuric acid of sp. gr. 1.53, which produces a violet coloration with resin oil.If the oil under examination becomes sc) dark-coloured with sulphuric acid as to interfere with the reaction, the resin oil may first be extracted with alcohol, when the colour test is readily performed. J. W. L. Test for Saccharin." By D. LINDO (Chem. News, 58,155).-The author modifies his test for " saccharin " (Abstr., 1888, 1350). After evaporating to dryness with nitric acid on a water-bath, a few drops of alcoholic potash is added to the cold residue ; when this is heated, a greater variety of colours is obtained than by following the original directions. D. A. L. Recent Processes for Testing Quinine. By W. LEKZ (Zeit. anal. Chenz., 2 7, 549-631) .-The foreign alkalojids in commercial quinine consist chiefly of cinchonidine and hydro-bases.Four principal methods are in use for the determination of these. In all these methods, a product (" Nebenalkaloide ") is obtained, containing the greater part of the impurities together with a certain quantity of quinine.', In this product, the cinchonidine is determined by the " tetrasulphate process." 1 gram of the mixture is dissolved in 9 grams of absolute alcohol and 3 grams of 50 per cent. sulphuric acid. The mixture is kept at 0" for 24 hours, the acid liquid isANALYTICAL CHEMISTRY. 87 removed by suction, the crystals are washed with a little absolute alcohol, and then air-dried. They are then dissolved in water, and the base is precipitated by excess of sodium carbonate. It is dried first over sulphuric acid, and then a t 115”.A correction (the amount of which depends on the percentage of cinchonidine found) must be applied. The author gives a curve for the purpose. The cinchonidine is very nearly pure. The hydro-bases are determined approximately by oxidising the quinine and cinchonidine in acid solution by a 1 per cent, solution of potassium permanganate, rendering alkaline, and shaking with ether, followed by chloroform. The residues from these solutions, although very impure, are regarded as hydroquinine. For the “chromate procew” of De Vrij see Abstr., 1887, 404; for the “ oxalate tesh” see Schaefer (Abstr., 1887, 623). The amount of oxalate prescribed is insufficient for samples containing less than 15 per cent. of water, and the cooling at 20” should be prolonged to one hour.Schaefer’s correction of 004 gram per 100 C.C. appears to be ltoo large. In Hesse’s ‘( bisulphate process” 5 grams of quinine sulphate is dissolved in 12 C.C. of normal sulphuric acid by warming, and the solution is allowed to erystallise in a naprow-necked funnel in a cold place. The mother-liquor is withdrawn by a filter-pump into a graduated cylinder, and the crystals washed with 3 C.C. of water. This solution is shaken with 16 C.C. of ether (0.721-0*728), then 3 C.C. of ammonia (0.96) added, again shaken, and left for 24 hours. The ether is removed by a pipette, the crystals are collected on a filter, washed with water saturated with ether, dried between filter- paper, washed again with ether, and dried. In the “ crystallisation process ” of Paul and Hewe, 5 ,mms of quinine sulphate is dissolved in boiling water, and crystallised out four times, using in the first case 150 c.c., next 130 c.c., and then twice 120 C.C.The united mother-liquors are evaporated at a low temperakmre almost to dryness, the residue dissolved in the smallest possible quantity of dilute sulphuric acid, made up to 20 c.c., and shaken with ether and excess of ammonia. The crystals which form are treated as in the bisulphate test. The hydro-bases crystallise in part with the quinine, therefore the mother-liquor should not be used for their determination. The chromate process gives very varying results, but on the average gives the highest yield of cinchonidine, especially with the purer samples The oxalate test gives the lowest numbers, but they are more concordant than those of the chromate process.The com- position of the bye-product is, however, variable. The bisulphate test gives results varying considerably. The alkaloids in the ethereal solution ought t o be submitted to the process a second and even a third time, but even with this improvement the whole of the cin- chonidine is not obtained, and the results vary much, but the com- position of the bye-product is more uniform than in the other processes. The crystallisation test has the same advantages as the bisulphate test if the crystallisation is repeated often enough, and is the process which i8 least influenced by the presence of hydro-bases. I t is, however, tedious. The process of the German Pharmacopoeia depends88 ABSTRACTS OF CHEMICAL PAPERS. on the fact that the precipitate produced by ammonia in solntionc: of the alkalo'ids is soluble i n excess of ammonia, but that much more ammonia is required for quinine than for the other alkalolds.The excess of ammonia required varies, however, very considerably with the temperature. M. J. S. Method for Recognising the Adulteration of Pepper by the Addition of Ground Olive Stones. By GILLET (BUZZ. XOC. chim., 50,173-174).-When 1 gram of olive stones is treated with 1 C.C. of a 5 per cent. iodine tincture, it acquires, after a quarter of an hour, a, J ellowish colour, whilst pepper is coloured brown or maroon. A series of mixtures of pepper with 5, 10, 15, and 20 per cent. of olive stones are prepared and coloured with iodine tincture. It is then easy to determine the amount of olive stones in samples of pepper by com- paring the colour obtained with these types.A New Test for the Blood in Carbonic Oxide Poisoning. By R. KATAYAMA (Virchow's Archiv, 114, 53-64).-When ammonium sulphide holding sulphur in solution and acetic acid are added to hlmd containing carbonic oxide, the result is a beautiful clear red or mse colour ; whereas normal blood becomes greenish-grey, or reddish- preen-grey on the addition of the same reagents. On examining these liquids spectroscopically, it is found that both in the case of normal and csrbonic oxide blood the absorption-spectrum indicates that a mixture of two substances is present. In the case of normal blood, there is a band between C and D and another between D and E ; this last becomes double on shaking up the mixture with air.In other words, there is a mixture of sulphur methaemoglobin (see Hoppe-Seyler, Physiol. Chem., 386) and reduced haemoglobin. I n the case of carbonic oxide there are three bands : one between C and D, due to sulphur methEmoglobin, and two between D and E, due to carbonic oxide hsemoglobin. That is to say, in spite of the pre- sence of sulphur methaemoglobin, the liquid does not become greenish, but remains red, the colour of carbonic oxide haemoglobin overpowering the olivegreen of the sulphur methaemoglobin. This test is stated to be more delicate than Hoppe-Seyler's (Abstr., 1888, 540), and is obtained with a mixture which contains one part of carbonic oxide blood to five parts of normal blood.Estimation of Albumin in Urine. By H. SCHACJMANN (Zeit. afial. Chent., 27, 635-636J-The estimation of albumin is much accelerated by collecting it, not on a paper filter, but in a filter-tube plugged with cotton-wool and connected with a filter-pump. When the washing with hot water is complete, a calcium chloride tube is connected, and dry air is drawn over the albumin whilst gradually raising its temperature to 110". N. H. M. W. D. H. M. J. S.ANALYTICAL CHEMISTRY. ‘73A n a1 y t i c a1 C h e m i s t r y.Preparation of Starch Solution for Use in VolumetricAnalysis with Iodine. By G. GASTINE (BUZZ. SOC. Chim., 50, 172-173).-Five grams of potato-starch is mixed with 0.01 gram ofmercury iodide, stirred with a little water, and poured into boilingwater (1 litre).It is allowed to settle, and the clear liquid pouredoff. A solution prepared in this manner has been kept for more thana year without deteriorating.Use of Salicylic Acid for Preserving Standard Solutions.By H. BORNTRAGER ( Z e i t . and. Chem., 27, 641--642).-The additionof a pinch of salicylic acid to each litre of a thiosulphate solutiongreatly diminishes its tendelicy t o decompose. The author’s deter-minations show large variations, but not a progressive diminution instrength. M. J. S.Applications of Spectrophotometr y to Chemical Physiology.By E. LAMBLING (Arch. de Physiol., 4th Series, 12, 1--.34).-This1)aper gives an historical account of the spectrophotometer, and of theprinciples upon which the spectrophotometric method depends.Thepractical application of the method for quantitative purposes in-xolves :-(1) The choice of a region of the spectrum; (2) the deter-mination of the coefficient of extinction of the coloured solution forthat region ; and (3) the determination of the amount of absorptionof the colouriig matter for the same region. When two colouringmatters are mixed in a solution, they may also be estimated quantita-tively, provided the absorptive power of each of the two pigments fortwo regions of the spectrum be previously known.Finally, t,he application of such methods t o animal pigments (of theblood, bile, urine, &c.), is pointed out.Determination of Chlorine in Plant-ashes. By A. JOLLES( Chem. Ceiztr., 1888, 863-864, from Zeit. Nahrunpnittel u.Hygiene,2, 81).-The method the author proposes as the best is as follows:-The plant (10 grams) is incinerated gently in a platinum dish,moistened with an alcoholic solution of sodium carbonate, the alcoholburned off, aiid the process repeated; after this, the whole of thecarbon may be burnt off without an1 fear of chlorine being lost, andN. H. M.W. D. H74 ABSTRACTS OF CHEMICAL PAPERS.the ash may he extracted by means of water and the determinationfinished in the usual manner with silver nitrate. This deviation fromthe ordinary direct combustion of vegetable matter in the determina-tion of chlorine obviates the chance of a loss of this element whichmight otherwise take place. J. W. L.Method for the Determination of Bromine in Sea-water.By F.GUTZKOW (Chern. News, 58, 190-193).-25o C.C. of sea-wateris mixed with a drop or two of sulphuric acid and with 100 C.C. of asolution containing about 25 grams of copper sulphate. It is thentreated with sodium sulphite solution until the precipitate redissolveswith difficulty, and heated until the blue colour returm; moresodium sulphite is now added, about one-third that already used, andthe whole ag"n heated until blue, and then cooled. In this way,all the bromine is precipitated in a few minutes. After washing theprecipitate, first with 100 C.C. of water containing 1 gram of sulplinricacid, then with a few drops of sodium carbonate solution, it is warmedwith hydrochloric acid until all sulphurous anhydride is removed, andis finally treated with zinc.The filtrate and washings from the copper,which combined should not exceed 25 c.c., are titrated with sodiumhypochlorite by a method of the author's which is described in detailin the original. The flask containing the liquid is fitted with a triple-bored cork ; through one hole passes a tube closed by a piece of india-rubber tubing and a clip, through the other holes two thistle-headedfunnels reaching nearly to the bottom of the flask, one is roughlygraduated. By blowing into the tube, a quantity of liquid is forced u pboth funnels and is titrated in the graduated one, the other servingfor comparing the colour ; in this way an indication of the quantityof hypochlorite required by the whole liquid is obtained, and so theoperation can be finished with greater rapidity than by the ordinarymode of procedure.Determination of Fluorine in Substances Decomposable bySulphuric Acid and especially in Natural Phosphates.BgH. LASNE (BUZZ. Xoc. Chirn., 50, 167- liO).-Sufficient substance toyield 0.2 gram of calcium fluoride is put into a, flask containingstrong sulphuric acid (50 c.c.) and pure sand (10 grams), and con-nected with two wash-bottles, the first of which contains 2.5 grams.and the second 0-5 gram of soda dissolved respectively in 25 C.C. ofwater. Dry air is passed into the flask, which is heated for an houra t 180--260"; it is then allowed to cool, dry air being passedthrough. The soda solutions are united and boiled for a hour;phenolphthaleyn is added, and carbonic anhydride passed through thesolution until it is colourless ; it is then heated for + hour at 50", andtreated with ammonium carbonate, when the silica is almost com-pletely precipitated.The liquid is cooled quickly, diluted to 125 c.c.,filtered through a large folded filter, and 100 C.C. collected. A fewdrops of tropeoline solution is added to the solution, which iscarefully neutralised with dilute hydrochloric acid, It is thentreated with pure sodium carbonate (equal to 0.5 gram of anhydrouscarbonate), boiled until free from carbonic anhydride, and precipitatedD. A. LANALYTICAL CHEMISTRY. 75with a slight excess of calcium chloride. The precipitate is ignited,treated with acetic acid, evaporated to dryness, again treated witharetic acid, and the undissolved calciuni fluoride, washed, ignited, andweighed. N.H. M.General Method for the Separation and Volumetric Estima-tion of Acids : Application to Sulphuric Acid. By G. LINOSSIEI~(BUZZ. SOC. Chin?., 50. 46-47).-The method is applicable to allacids which yield insoluble compounds when combined with metalswhich are precipitated by hydrogen sulphide in acid solution. Thedeterminahn of sulphuric acid is carried out as follows :--The s o h -tion of the sulphate (containing 0.05 to 0.1 gram) contained iri adish is treated with alcohol (2 vols.), heated almost to boiling, andprecipitated with a slight excess of lead acetate. When cold it ismaslied by decantation with a mixture of alcohol and water (0.5 to1 vol.), passing the decanted liquid through a small filter.The traceof lead sulphate on the filter is washed with aqueous hydrogensulphide into a flask ; the rest of the lead sulphate is shaken with asaturated solution of hydrogen sulphide, the liquid poured on to nfilter, and the precipitate treated with a fresh amount of hydrogensulphide and filtered. It is then washed with aqueous hydrogensulphide until the filtrate gives no reaction with Poirrier’s orange.The whole of the sulphuric acid is then i n the filtrate, and is titratedwith decinormal soda solution. N. H. 31.Volumetric Estimation of Boric Acid, and of Ammonia inAmmonium Salts. By J. MCGLASHAN (Chew. News, 58, 175-176).-By using Poirrier’s oranges I and 11, boric acid and ammo-nium borate can be titrated directly with normal soda, but boraxmust first be made neutral to methyl-orange with aulphuric acid;boracite must be heated with dilute sulphuric acid, made neutral tomethyl-orange with soda, and any carbonic anhydride eliminated beforetitrating.Ammonia is objectionable in any form except as borateor hydroxide, the latter when dilute is neutral to both orangeh.Therefore, with these indicators ammonia may be titrated with soda,in any ammonium salt, without distillation. With ammonium car-bonate and with arsenates, the end reaction is not distinct, but withhydrogen sulphide it is sharp. D. A. L.Resorcinol as a Test for Nitrates. By D. L~NDO (Chem.News, 58. 176--177).-For testing with resorcinol, 10 grams are dis-solved in 100 c .~ . of water, and one drop of this solution, one drop of15 per cent. hydrochloric acid, and 2 C.C. of concentrated sulphuric acidare added to 0.5 C.C. of the nitrate solution (compare Abstr., 1888,1337) ; one in 500,000 gives a definite permanent purple colour aftersome time, whilst with increasing strengths of nitrate solution thecolour becomes more intense, until with one in 10,000 the vivid purple-red colour is so intense as only to be distinctly seen i n the lowerportion of the band. Resorcinol is valueless without hydrochloric acid,but with it, is a more delicate test for nitrates than ordinary phenol.Copper suiphate does not aid the reaction materially. The reagent76 ABSTRACTS OF CHEMICAL PAPERS.alone gave a band which, however, cannot be mistaken for the nitrateband.With nitrites, it is more delicate than with nitrates ; withchlorates it is no good, with dichromate (0.5 gram per litre) it gives ared to buff upper with a purplish lower band, the latter changing to pinkor with hydrochloric acid to reddish-brown ; the colours are slightlydifferent if the acid is run in at once, or after some delay. Perman-gsnate, N/10, with 4 vols. water, gives a, dark-orange upper andyellowish lower band ; with hydrochloric acid, these are respectivelyorange, red, and greenish. Hydrogen peroxide in dilute solution yieldsa green and brownish compound b;lnd. D. A. L.Estimation of Phosphorus in Iron and Steel. By P. W.SHIMER (Chem. News, 58, 165--168).-Dissolve 1 gram of iron in20 C.C.of nitric acid, sp. gr. 1-20, and add to the boiling solution10 c,c. of a solution of 20 grams of potassium permanganate in a litreof water, 2.5 C.C. a t R time, and after a few minutes 5 C.C. of hydro-cliloric acid, sp. gr. 1.12. When the action has ceased, add a mix-ture of 5 C.C. of concentrated sulphuric acid and 5 C.C. of water, andevaporate until fumes of sulphuric acid begin to come off. Whencool, add 5 C.C. of nitric acid, sp. gr. 1.20, and sufficient water; boilto dissolve iron salts, filter, and wash with water. The residueserves well for the estimation of silicon. Heat the filtrate to 80"and add 5 C.C. of ammonium molybdate (5 grams of MOO,, 20 C.C.ammonia, sp. gr. 0.96, 30 C.C. nitric acid, sp. gi.. l a % ) , then keep atGO" until the liquid is clear, which occupies less than an hour when thesolution is not too dilute.The yellow precipitate is washed withacid ammonium nitrate solution, dissolved in ammonia and pre-cipitated by magnesia mixture. The method gives good results bothfor silicon and phosphorus.A few experiments were made using sulphuric acid and ammoniumsulphnte instead of nitric acid and ammonium nitrate for themolybdate precipitation, with satisfactory results. I). A. L.Influence of Sulphur on Eggertz's Carbon Colour-test. l 3 ~T. W. HOGG (Chern. News, 58, 175).- When ordinary steels contain-ing say 0.05 per cent. of sulphur are dissolved in the usual manner,the sulphur separates out and produces a turbidity which interfereswith the colour-test for carbon ; a fact easily proved by dividing sucha solution into two portions, filtering one, and then comparing thecolour produced in the filtered and unfiltered solutions.Conse-quently, if a common steel is compared with a standard of pure steel,the colour intensity is sure to be over-estimated, and vice versd.D. A. L.Quantitative Analysis by Electrolysis. By A. CLASSEN andR. SCHELLE (Ber., 21,2892-2899).-The current from two of Farbakyand Schenek's accumulators, fully charged by a dynamo giving Rcurrent of 20-25 amperes, was employed in making 50 analyses inthe course of six days. Four to eight analyses were made simul-taneously, and the current was employed continuously day and night,except during the short intervals required for changing the platinuANALYTICAL CHEMISTRY.7 7dishes, During this time the strength of the current decreased from2.05 to 1.92 volt, so that one charge would be sufficient for 60 to 70analyses. The condition of the accumulator can be ascertained fromthe specific gravity of the sulphuric acid, which is 1.240 when theaccumulator is charged and 1.118 when i t is not charged. The pre-cipitated metal is in a more suitable condition than when a battery ordynamo is employed.Experiments which were made with Neumann’s (Abstr., 1888,529) and Wolffs ( Z e i t . ung. Chem., 1888, 296) voltameters, employ-ing currents of various strengths, show the necessity of working underthe same conditions when repeating electrolytic methods of analpis(compare Classen, Quantitative Analyse durch Electrolyse, 2nd ed., 43).In the separation of antimonyfroin tin in sodinm sulphide solution,the solution, freed from antimony, is boiled with ammonium pulphate,and the tin precipitated by electrolysis (compare Abstr., 1884, 932).A simpler and more convenient method is to convert the stannicsulphide into hydrogen stannic oxalate and electrolyse the solution.For this purpose, the solution is acidified with dilute sulphuric acid,and the sulphide oxidised with ammoniacal hydrogen peroxide, or thehot alkaline solution is treated a t once wi+h hydrogen peroxide untilit becomes colourless, acidified with sulphuric acid, neutralised withammonia, and more hydrogen peroxide added.The solution is thenboiled, filtered, and the residual stannic acid washed off the filterwith and dissolved in a hnt solution of oxalic acid.If there is aresidue of sulphur, i t is separated by filtration, washed with a coldsaturated solution of hydiwgen animonium oxalate, and the washingsadded to the filtrate. The solution, which should contain at least50 C.C. of the hydrogen ammonium oxalate solution, is electrolysedwith a current giving 8-10 C.C. of explosive gas per minute.The electroljtic deposition of copper from its salts, dissolved in asaturated solution of ammonium oxalate, is hastened very consider-ably if the solutian is kept acid, especially towards the end of theoperation, by adding excess of a cold, saturated solution of oxalicacid. The smaller the quantity of copper, the more oxalic acid soln-tion may be added.A solution poor in copper can be mixed withthe oxalic acid solution a t the commencemtnt of sthe process, but inconcentrated solutions, the precipitation must be carried out in a solu-tion as nearly neutral as possible to avoid the separation of copperoxalate. If the copper solution is kept a t 40-50°, aLout two gramsof copper are deposited in from 3 to 4 hours. A current giving3-4 C.C. of explosive gas per minute mas employed, and quantitativeexperiments showed that the whole of the copper is deposited.Separation of Calcium, Barium, and Strontium. ByKUPFFERSCHLAEGER (Bull. SOC. Clzim., 49, 854- 856).-A quantitativeanalysis of a mixture of the carbcnates of barium, strontium, andcalcium can be carried out as follows :--The mixture is dissolved ina slight excess of very dilute nitric acid, the solution evaporated todryness, the residue dissolved in distilled water, and the filtered solu-tion again evaporated to complete dryness, The residue of mixednitrates is agitated with a small quantity of a mixture of absoluteI?.s. I(78 ACSTRACTS OF CHEMICAL PAPERS.alcohol and ether, and the solvent separated by filtration as soon as thesolution becomes clear ; this process is repeated three times, the propor-tion of ether being gradually increased until the mixture contains equalvolumes of the constituents. The residual mixture of the nitrates ofbarium and strontium is dried, dissolved in water, and treated witha cold saturated solution of potassium dichromate.The precipitatedbarium dichroniate is washed with cold, very dilute a,lcohol, and con-verted into sulphate by heating with sulphuric acid. The strontiumin the filtrate is converted into sulphate by warming the solutionwith dilute sulphuric acid. A solution of strontium chloride is notprecipitated by potassium dichrornate. F. S. K.Volumetric Estimation of Mercuric Chloride. By G. KASSRER(Arch. Phawn. [ 3 ] , 26, 595-604).-The method promises well forpharmaceutical practice. 50 grams of the organic compound contain-ing mercuric chloride is placed in a porcelain dish without previousdivision, 500 C.C. of water is added, and the mass is well kneadedwith a pestle. 500 C.C. of a solution of 0.4 gram of potassium anti-monions tartrate and 1 gram of sodium phosphate, or in place of thelatter 1 to 2 grams of sodium acetate, is now added and the kneadingcontinued.A very intimate mixture is thus obtained, and a milkyliquid results from the decomposition of the mercuric chloride.About 500 C.C. is now filtered off, and titrated with decinormal iodinesolution after the addition of freshly prepared starch solution, anda suffirient quantity of sodium hydrogen carbonate. The iodineemployed corresponds to the amount' of unchanged tartrate remainingin the filtrate. The reaction is as follows :-4HgCI, + Sb,O, + 2H,O= 4KC1 + Sb,05 + 4HgC1. The presence of phosphoric and organicacids is admissible, but not of free hydrochloric acid.By A. STRENG (Jahrb. f. Mi%, 1888,ii, Mem., 142-152 ; continuation of Abstr., 1886, 4877.-For detect-ing tin under the microscope, t'he author avails himself of the browncolour produced on adding hydrochloric acid and platinum chloride.When the solution is effected, a drop of the liquid is transferred to anobject-glass, and a grain of potassium chloride added, and the solutionslightly evaporated.In this way, rhumbic crystals of pot'assinmstannous chloride (2KC1 + SnC1, + H,O) are form_ed. They-aremostly combinations of the forms mP, mPm, 03Pq Pm, Pm.When the staunous chloride has been thus detected, a drop of nitricacid is added, and the solution heated. The stnncous chloride isconverted into stannic chloride, and the imperfectly soluble salt,K,SnCl, is formed. This crystallises in the regular system, and isperfectly isotropic.This reaction may be employed for detectingstannous and stannic oxides in compounds soluble in hydrochloricacid. Caesium chloride may be used instead of the isomorphouspotassium salt. The author also describes methods for detectingpotassium, caesium, and rubidium, sodium, and silicon.Detection of small Quantities of Germanium. By K. HAUSHOFER(Clrern. Centr., 1888, 867, from Xitzber. Akad. JIGnche~~, 1887, 133).J. T.Microchemical Reactions.B. H. BAX ALP TIC AL C €1 ENIST R Y. 70-By heating in an atmosphere of hydrogen snlphide, the germaniumin the mineral argyrodite is converted into the sulphide GeS, whichis crystalline, and may be detected and yecognised by means of ihemicroscope.With concentrated sulphuric acid, it forms a white, non-crystalline substance, with concentrated nitric acid it is converted intothe white, crystalline oxide, GeOz, which is soluble in dilute nitricacid and water, and crystallises out of the solution again on evapora-tion. Heated in a tube, it sublimes similarly to antimony oxide, butdiffers from this by its solubility in water, and also by melting t o clearcolourless drops. It, is necessary to apply the potassium iodide testalso, as a means of distinguishing it from the mercuric sulphide, whichhas also been found present in argyrodite. J. W. L.Characteristic Reaction of Bismuth. By E. L ~ G E R (BUZZ. XOC.Chi%., 50, (31-93) .-A solution of bismuth iodide and potassiumiodide is sometimes used for detecting alkalo'ids, with which it givesorange-yellow precipitates.It is suggested that the reaction shouldbe employed for detecting bismuth. A solution is prepared bydissolving cinchonine (1 gram) and potassium iodide (2 grams) inwater (100 c.c.). Other alkaloids may be nsed instead of cinchonine,but thir seems to give the most sensitive reaction. The reagent mustbe added in excess ; the presence of too much nitric acid, and espe-cially the presence.of hydrochloric and sulphuric acids, is to be avoided.Bismuth may be detected in solutions containing only 1 part in 500,000parts. Solutions which contain &her metals besides bismuth areprecipitated with hydrogen sulphide, the sulphides of copper, lead,cadmium, mercury, and bismuth converted into nitrates, then in tocarbonates, and the caiabonates of bismuth and lead separated bymeans of potnssium cyanide ; these carbonates are converted intoclibrides, and the lead chloride separated by means of alcohol.Thealcoholic solution is evaporated to dryness, c?issolved in a drop ofnihric: acid and sorne C.C. of water, and treated with the rexgent. Solu-tions of the salts of mercury, cadmium, silver, copper, and lead alsogive precipitates of various colours with the cinchonine reagent.N. H. M.Determination of Oxygen dissolved in Water. By L. W.WINKLER (Ber., 21, 2846-2854) .-The method consists in oxidising:in excess of manganese hydroxide in presence of alkali by the oxygenpresent in a weighed amount of the water; potassium iodide andhydrochloric acid are then added, and the iodine which separates(which is equivalent to the amount, of dissolved oxygen) titrated withsodium thiosulphate.A solution of maiiganous chloride (free fromiron) is made of such a strength that 100 C.C. contains 40 grams ofthe salt (MnC1, + 4H20). The soda must be free from nitrate, andthe solution prepared of eight times the normal strength ; potassiumiodide (10 grams) is dissolved in 100 C.C. of the soda solution, the restof which is kept. A strong flask of about 2 litre capacity is filledwith the water ; 1 C.C. of the potassium iodide soda solution is addedby means of a pipette reaching nearly to the bottom of the flask, then1 C.C. of the manganous chloride solution. The flask is closed, carebeing taken t h a t no bubble of air remains, and the contents mixed80 ABSTRACTS O F CHEMICAL PAPERY.When the precipitate settles, fuming h~drochIoric acid (3 s.c.) isadded by means of a pipette similar to those previously used.Theflask is again closed, the contents mixed, and the yellow liquid titratedin the usual manner with sodium thiosulphate, the most convenientstrength of which is 1/100 normal, so that each C.C. corresponds witli0.055825 C.C. of' oxygen (at 0" and 760 mm. pressure).When the water to be analysed contains mucii carbonic anhydride,more of the reagent must be added, as manganous carbonate is notoxidised by the oxygen. In the case of waters containing nitrates, theprocess has to be modified : a soda solution containing no potassiumiodide is first added, then hFdrochloric acid (twice the amount other-wise used), and after three minutes a solution of potassium iodide. Amanganic chloride solution is then prepared as follows :--500 C.C.ofdistilled water is treated with the pure soda solution (1 c.c.), 5 to 10drops of the manganous chloride solution, and then sufficient hydro-chloric acid is added to dissolre the precipitate. 100 C.C. of thissolution is taken out and diluted with distilled water; to the rest,100 C.C. of the water to be examined is added. After two to threeminutes, both solutions are treated with potassium iodide and theiodine which separates determined as usual. From the differencein the amount of thiosulphate used, the correction for 100 C.C. ofwater is calculated.Concordant results obtained b j the new method are gil en.N.H. M.Ash Determination. By F. A. FL~~CKTGER (Zeit. anal. Chem., 27,637-638).-The substance is heated in a roomy platinum capsule sogently that carbonisation takes place without combustion. It irthen cooled, a copious amount of water is added, and the whole eva-porated completely on the water-bath. On reheating the carbonaceousrebidue very gradually, it burns a t a low temperature, and veryquickly. &I. J. S.Wet Methods of Organic Analysis. By J. MESSIKGER (Bey., 21,2910-2919) .-Organic compounds, as Cross and Bevan have show11(Proc., 1888, 76, and Trans., 1888, 889), are completely oxidisedwhen warmed with chromic acid and concentrated sulphuric acid.I fnitrogen is present, it is evolved as such or as ammonia, whilstsulphur, phosphorus, and arsenic are converted into the corresFond-ing acids. Halogens are evolved in the free state, and metals remain assulphate or, with a large excess of chromic acid, as chromates.To estimate the quantity of carbon, the substance (0*15-0*35 gra,m)is weighed in a small bulb or tube, and placed in the apparatusemployed by Classen for the estimation of carbonic anhydride(Quudtative Analyse, 3rd ed., 239), together with chromic acid(?5--.6 grams) or powdered potassium dichromate. A gentle streamof air, free from carbonic anhydride, is passed throngh the appa-ratus to drive out the carbonic anhydride, the weighed potash bulbsand sods-lime tube are attached, and the latter connected with acalcium chloride tube to prevent absorption of moisture from theair.Concentrated sulphuric acid (30 c.c.) is then poured througANALYTICAL CHEMISTRY. 81tlle funnel, and the stream of air stopped. The flask is warmed verycautiously until the evolution of carbonic anhydride commences, andthen the heating is immediately discontinued until the reaction isalmost at an end. Pureair is passed through the apparatus for half an hour, and the tubesweighed. Tlie results obtained are very satisfactory, except in thecase of substances which sublime readily, but great care must betaken when heat is first applied. I n analysing volatile substances,the bulb is broken by means of the funnel. If the substance con-tains halogens, a Dreschler’s flask of about 100 C.C.capacity, contain-ing 40 C.C. of concentrated potassium iodide solution, and a smallU -tube filled with glass-wool, half of which is moistened with a solu-tion of silver nitrate and half with concentrated sulphuric acid, areinterposed between the condenser and the potash bulbs. Quantitativeexperiments with compounds of the most varied nature gave satis-factory results (compare Cross and Bevan, Zoc. cit.).Sulphur can be estimated, except in the case of extremely volatilesubstances, as follows :-The substance (0.15-0.35 gram), togetherwith potassium permanganate (1&2 grams), and pure potash (0.5gram) is placed in a flask of 500 C.C. capacity provided with a con-denser, water (25-30 c.c.) poured down the condenser, and themixture heated for 2-3 hours.Concentrated hydrochloric acid isthen gradually poured down the condenser into the cold inixture,which should be of a reddish colour, mid after the evolution of gashas ceased, the whole is heated until the liquid beconies clear. Thesulphuric acid is then precipitated with barium chloride. Potassiumdichromate (2-3 grams) and hydrochloric acid (20-25 C.C. : 2 partsconcentrated acid, 1 part water) may be employed instead of potas-sium permaiiganate and potash. The operation is carried out in themanner described, but after heating for about two hours a few dropsof alcohol are added to determine whether all the dichromate hasheen reduced; if the odour of aldehyde is perceptible, the mixtureiuust be heated again and the test repeated.Both methods can alsobe employed for the detection of sulphiir. Numerous quantitativecbxperiments gave satisfactory results, but in the case of a few sul-phones i t was found that sulphur cannot be estimated by this method.Phosphorus, arsenic, and antimony in organic compounds can beestimated by placing the substance (0.3-0.4 gram) with chromic acid(4.-5 grams) in a flask provided with a condenser, pouring sulphuricacid (10 c.c., 2 parts concentrated acid, 1 part water) down the con-denser, and heating gently. After an hour’s time, sulphuric acid(10 c.c.) is added, and the heating continued for about an how.The mixture must always be heated very carefully, and the cold solu-tion should be perfectly clear.Some antimony compounds requireonly 1 gram of chromic acid and 10 C.C. of sulphuric acid.In the estimation of phosphorus, the solution is warmed withammonium nitrste (3-4 grams) and ammoniuni molybdate solutior!(50 c.c.) for 2-3 hours, filtered, the precipitate washed 6 to 8 timesby decantatioii with ammonium nitrate solution (20 grams in 100 C.C.of water), throan on to a filter, and dissolved in not more than 40-50 C.C. of warm dilute (2 per cent.) ammonia ; a concentrated solutionThe process occupies about two hours.VOL. LVI. 82 ABSTRACTS OF OHEMIOAL PAPERS.of citric acid (4-5 drops) is added to the filtrate, and the phosphoricacid precipitated with magnesia mixture.In estimating arsenic, the solution is diluted to about 100 c.c.,heated to about 70", and the arsenic precipitated with hydrogen sul-phide.The precipitate is washed with water containing hydroqensulphide until free from chromium salts, and then converted intoarsenic acid by means of ammoniacal hydrogen peroxide (compareClassen, Abstr., 18b3, 934). After boiling for an hour, ammonia isadded to the filtered solution, and the arsenic precipitated with mag-nesia mixture. Antimony is estimated by adding potash and excessof sodium sulphide, boiling for half an hour, and precipitating themetal electrolytically.Metals are estimated by mixing the solution with excess of ammo-nium oxalate and precipitating electrolytically. Halogens can bedetected by warming the substance (1-2 mgrms.) with chromicacid and sulphnric acid, and passing the gas evolved into a dilutesolution of potassium iodide ; quantitative experiments gave unsatis-factory results.F. S. K.The Safety of Commercial Kerosene Oils. Bp S. B. NEWBURYand W. P. CUTTER (Amer. Chem. J., 356-362).-Although oils areregularly testea for their flashing points and conclusions drawn as totheir being sa€e for burning in lamps, it is noteworthy that manymodern lamps heat their reservoir of oil to temperatures above thelegal flashing point, and that there are not sufficient experimentaldata to allow of the correct interpretation of the flashing point deter-minations into +terms expressing the inability of such oils to formexplosive mixtures of vapour with air. All hydrocarbons up to andincluding octane, form at ordinary temperatures mixtures that can beexploded; nonane mixtures (b.p. 148-1.50") explode sharply at79" F., and decane at 104" F.The addition of small quantities of low-boiling oils materiallylowers the flashing point of another sample. Thus a sample flashingat 212" F. was made to flash at 145" F. by the addition of 5 per cent.of heptane; at 110" F. by the addition of 5 per cent. pentaneor hexane, or 10 per cent. heptane; at 96" F. by the addition of15 per cent. heptane ; and at ordinary temperatures by the additionof 10 per cent. pentme or hexane, or 20 per cent. heptane. The tem-peratures at which am oil may be kindled in an open vessel approachesthe more nearly to the Bashing point the more homogeneous the oil is.The flashing point of an ordinary oil may be materially raised, withbut little diminution of weight, by passing a current of air throughit for several hours.H. B.Solubility and Estimation of Paraffin. By B. PAWLEWSKI andJ. FILEMONOWICZ (Ber., 21, 2973-2976) -The following table givesthe solubility at 20' of ozokerit paraffin of sp. gr. 0.9170 at 20°,melting at 64-65', and solidifying at, 61-63' :ANALYTICAL CHE3lISTHY.-Solvent.Carbon bisulphide.. ...................Light petroleum, up to 75" G., sp. gr.0'7233 ............................Turpentine oil, sp. gr. 0 -857 ; b. p. 158-186". ..............................Cumene (comm.), up to 160°, sp. gr. =0 *867. .............................Cumene (frac.), 150-160", sp. gr.= 0.849Xylene (comm.), 138- 143",sp. gr. = 0.866(frac.). 136-1383, sp. gr. = 0.864. .Toluene (comm.), l(B--llO",sp.gr. = 0.866(frac ), 108.5-109 5", sp. gr. =0-866 ..............................Chloroform ..........................Benzene .............................Ethyl ether ..........................Isobutyl alcohol (comm.), sp. gr. = 0.804.1l:tliyl acetate. ........................Amy1 alcohol, 127-129", sp. gr. = 0813 .l'ropionic acid.. ......................)'ropy I alcohol ........................Met hpl alcohol, 65.5 - 66 5 O , ~ p . gr. = 0.798JIetli! 1 forniate .......................(fluci 11 acetic acid ....................,,,,Acetone, 55-5-56 5", sp. gr. = 0.797.. ...Ethyl alcohol, 99.5' Tr.. ...............Ethyl alcohol, 945" Tr.................Acetic anhydride.. ....................l?ormic acid (crjst.) ...................Ethyl alcohol, 75" Tr.. .................Paraffin (grams),dissolved by100 grams12 *9911 '736.064.283 -993.954.s93 -833 -922 '421 -9Y1 -950.2880 '2620 '2380 * 8190 -20%0 -1660 -1410 ' o n0 -0600 -0600 -0460 -0280 -0130 *Wd03-.-100 C.C.-8 *485 -213-723 -393 -433.773 -343 -413 -611 -750 -2280 '2090 '1640 -0560 -0630 '015-----c --Weight of solventrequired to dis-solve completely1 part d paraffin.7.68'516 *123 -426-025 -122 -726 -125 *541 '360 '350.8352 -9378'7419.0453 -6495 -3595 -3709 -4l a 7 -51648 -71668 -62149 * 53856 -27689 -2330000 *oThe liquid constituents present in man-j- prodncts of the petroleumor ozokerit industry are soluble in glacial acetic acid, whereas vaselin,cerisin, ozokerit, and paraffln are almost insoluble.To estimate thequantity of solid paraffin in petroleum, lubricating oils, mineral oils,vaselin, &c., 5-20 C.C. of the mixture is well shaken with 100-200 C.C.of glacial acetic acid, the residual para& thrown on to a weighed filter,washed two or three times with glacial acetic acid, and then two or threetimes with alcohol of 75" Tr., dried and weighed, or the residual paraffinis washed, dissolved in benzene or ether, the solution evaporated, andthe residue weighed. This method is quick and accurate, and can becarried out at the ordinary temperature, F.S. K.Analysis of a Mixture of Silver Chloride, Cyanide, Thio-cyanate, Ferricyanide, and Ferrscyanids. By J. TELSS~ER (BUZZ.Soc. Chim., 50, 10&106).-The mixture occurs in the analysis ofmaterials used in the purification of coal-gas. Sodium carbonate isheated in a crucible until aahjiirous, the weighed substance and Born81 ABSTRACTS OF CHEMICAL PAPERS.potassium nitrate are added, and the whole is heated. The productis extracted with water, which leaves a residue of pure silver andferric oxide. In the solution, sulphur is determined as barium sul-phnte and chlorine as silver chloride. The residue is dissolved innitric acid, the silver determined as chloride, and the iron as sesqui-oxide.Equations are given by means of which the amount of eachsalt present in the mixture is calculated.Estimation of Paranitrotoluene. By F. REVERDIN and C.DE LA HARPE (Bull. Xoc. Chim., 50, 44-46).-The method is basedon the fact that paranitrotoluenesulphonic acid is readily convertedby boiling with soda into dinitrosulphostilbene which yields a redcolour when dissolred in alkali, whilst orthonitrotolnenesulphonic acidis not attacked by soda, and the alkaline solution is yellow. 2 C.C. ofpure orthonitrotoluene is heated in a water-bath with 6 C.C. of sul-phuric acid (containing 25 per cent. of anhydride) for three hoiirq,the product cooled and diluted to 1 litre. I n a similar manner amixture of orthonitrotoluene (96 parts) and paranitrotoluene (4 parts)is sulphonated and the product diluted to 200 C.C.I n determiningthe amount of paranitrotoluene, the substance and nitrotoluene (con-taining 4 per cent. of the para-compound) are sulphonated, the pro-ducts diluted to 200 c.c., and the colours compared which are obtainedby mixing 1 C.C. of each solution with 5 C.C. of aqueous soda. If thesample to be aiialysed gives a stronger colour, a measured quantity(50 to 50 c.c.) is progressively diluted with the solution of ortho-nitrotoluenesulphonic acid until 1 C.C. when heated with 5 C.C. ofaqueous soda gives the same amount of colour as the solution con-taining 4 per cent. The percentage of paranitrotoluene can then becalculated. N. H. &I.Detection of Methyl Alcohol in Wood Spirit.By J. HABER-MAXN (Zcit. anal. Chem., 27, 663, from VerAandl. nut. Ver. Byutm, 26).-Commercial methyl alcohol contains impurities which reduce per-manganate energetically. Cazeneuve and Cotton shake 10 c c. of thespirit to be tested with 1 C.C. of a 0.1 per cent. solution of per-manganate a t 20'. If wood spirit is present. decolorisation takes placeimmediately ; with pure alcohol, 20 minutes is required. Etherealoils, which may be present in spirit o r brandy, and would cause asimilar reduction, may be removed by shaking the spirit twice withhalf its volume of the purest olive oil and then filtering through a wellwetted filter.N. H. M.If sugar is present, the spirit must be distilled.M. J. S.Composition of Natural Brandies and the Way of Dis-tinguishing them.By X. ROCQUES (Bull. SOC. Chim., 50, 157-164).-500 C.C. of the brandy is distilled in a, Le Bel-Henniger ap-paratus and nine fractioiis of 50 C.C. each collected, the temperaturesbeing noted. Each fraction is subjected t o the following tests:-(1) Rosaniliue bisulphite, (2) aniline acetate, (3) sulpharic acid,(4) permanganate, (5) ammoniacal silver nitrate solution. Theresults of several analyses of brandies are given in tables.N. H. 31ANALYTICAL CHEMISTRY. 85Estimation of Sugar in Presence of Carbohydrates.By R. W. BISHOP (Chern. Centr., 1888, 952-953, from Msc. ,Won.Sci., 32, 558).-The author has carried out a number of experimentswith a view to determine the conditions under which inversion maybe completed without a t the same time damaging the accompanyingcarbohydrates.It was found that sulphuric acid has a greaterpower of inversion than hydrochloric acid, but it appears that hydro-chloric acid is the best for the conversion of starch into dextrose.Prolonged heating of inverted sugar with the acid seems to act onthe lmvulose rather strongly, and the solution becomes less lmvo-rotatory and eventually dextrorotatory. For the inversion of cane-mgar, 0.5 C.C. of hydrochloric acid is snfficient, and the heating is con-tinued for 10 minntes at 95". For the conversion of dextrin, thesolution should not contain more than 1-3 grams of carbohydrates i n50 C.C. The inversion is performed by adding 2 C.C. of hydrochloric acidand heating for three hours at 95".J. W. L.Detection of Sugar in Urine. By C. SCHWARZ (Arch. Pharnz.[3], 26, 796, from Pharm. Zeit., 33, 465).-1 to 2 C.C. of leadacetate is added to 10 C.C. of urine and filtered ; 5 C.C. of the filtrateis mixed with 5 C.C. of normal potash solution and one or two dropsof phenylhydrazine, well shaken and vigorously boiled ; in the presenceof sugar, the liquid becomes lemon- to orange-yellow, and becomesopaque on adding an excess of acetic acid owing to the immediateformation of a finely divided yellow precipitate. In the absence ofsugar, this precipitate never OCCUTS with urine. J. T.Detection of Chloral or Chloroform in Liquids. By C.SCHWARZ (Zeit. anal. Chem., 27, 668-669, from Pharm. Zeit., 33,419).-Either of these substances when boiled with resorcinol and anexcess of soda gives a red colour, which disappears on acidifying and isrestored by alkalis.If, on the other hand, an excess of resorcinol andonly a dropor two of soda solution is used, the product is n yellowish-red solution with intense yellowish-green fluorescence. 0.0001 gramof chloral hydrate in 1 C.C. gives this reaction very diwtinctly whenvigorously boiled with 0.05 gram of resorcinol and five drops of sodasolution. Rf. J. S.Modification of the Reichert-Meissl Metho& of ButterAnalysis. By M. MANSFELD (Chem. Centr., 1888, 870-871, fromMilch Zeit., 17, 281-283).-1n order to obviate the possible errorarising from the use of alcohol in the saponification of the butter-fat,theauthor has tried the use of alkali alone, the latter being added tothe fat in a small flask and heated on the water-bath for two hours, a tthe end of which time the saponification is complete, and the processis finished as Wollny prescribes. J.W. L.Densities and Refractive Indices of Certain Oils. By J. H.LONG (Amer. Chem. J., 10, 392--405).-The following values arcobtained-the densities a t 20" being compared with water at 4' a86 ABSTRACTS OF CHEMICAL PAPERS.unity, and the substance weighed in a vacuum; the refractive indicesare for sodium light a t 20" :-Olive oil .......... 0.9130 Ref. index 1.4703Cotton-seed oil. 0.9191 2.4732Sesame oil ........ 0,9191 1.4740Mustard oil.. 0.9121 1.4742Castor oil.. 0.9589 1.4791Lard oil 0.9122 1.4686Peanut oil ........0.9173 1.4717The densities and refractive indices are also given for other tempe-ratures than 20". In nearly all cases, the variations due to temperatureare t,he same, namely, about -0.G0068 i n denpity and -0-0004 inrefraction for each rise of 1". The author believes such determiuationsmay prove of value in the identification of other oils........... ..................H. B.Bechi's Newest Method for the Detection of Cotton-seedOil in Mixtures. By G. Biz10 (Chem. Centr., 1888, 873, from AttiInst. Veneto [ 6 ] , 6 ) .--The author finds that pure olive oil gives Bechi'snew cotton-seed oil reaction with slightly acid silver nitrate, whilst,on the other hand, he finds that there are some cotton-seed oils whichdo not give the reaction at all.J. W. L.Qualitative Test for Resin Oil in Vegetable and Mineral Oils.By HOLDE (Chem. Centr., 1888, 952, from Pharna. Zeit., 33, %8).-Whereas Storch's reagent (concentrated sulphuric acid and arihjdrousacetic acid) for the detection of resin oil is not always admissible,the author recommends sulphuric acid of sp. gr. 1.53, which producesa violet coloration with resin oil. If the oil under examinationbecomes sc) dark-coloured with sulphuric acid as to interfere with thereaction, the resin oil may first be extracted with alcohol, when thecolour test is readily performed. J. W. L.Test for Saccharin." By D. LINDO (Chem. News, 58,155).-Theauthor modifies his test for " saccharin " (Abstr., 1888, 1350). Afterevaporating to dryness with nitric acid on a water-bath, a few dropsof alcoholic potash is added to the cold residue ; when this is heated,a greater variety of colours is obtained than by following the originaldirections.D. A. L.Recent Processes for Testing Quinine. By W. LEKZ (Zeit.anal. Chenz., 2 7, 549-631) .-The foreign alkalojids in commercialquinine consist chiefly of cinchonidine and hydro-bases. Fourprincipal methods are in use for the determination of these. In allthese methods, a product (" Nebenalkaloide ") is obtained, containingthe greater part of the impurities together with a certain quantity ofquinine.', In this product, the cinchonidine is determined by the" tetrasulphate process." 1 gram of the mixture is dissolved in9 grams of absolute alcohol and 3 grams of 50 per cent.sulphuricacid. The mixture is kept at 0" for 24 hours, the acid liquid iANALYTICAL CHEMISTRY. 87removed by suction, the crystals are washed with a little absolutealcohol, and then air-dried. They are then dissolved in water, andthe base is precipitated by excess of sodium carbonate. It is driedfirst over sulphuric acid, and then a t 115”. A correction (the amountof which depends on the percentage of cinchonidine found) must beapplied. The author gives a curve for the purpose. The cinchonidineis very nearly pure. The hydro-bases are determined approximatelyby oxidising the quinine and cinchonidine in acid solution by a 1 percent, solution of potassium permanganate, rendering alkaline, andshaking with ether, followed by chloroform.The residues from thesesolutions, although very impure, are regarded as hydroquinine.For the “chromate procew” of De Vrij see Abstr., 1887, 404;for the “ oxalate tesh” see Schaefer (Abstr., 1887, 623). The amountof oxalate prescribed is insufficient for samples containing less than15 per cent. of water, and the cooling at 20” should be prolonged toone hour. Schaefer’s correction of 004 gram per 100 C.C. appears tobe ltoo large.In Hesse’s ‘( bisulphate process” 5 grams of quinine sulphate isdissolved in 12 C.C. of normal sulphuric acid by warming, and thesolution is allowed to erystallise in a naprow-necked funnel in a coldplace. The mother-liquor is withdrawn by a filter-pump into agraduated cylinder, and the crystals washed with 3 C.C.of water.This solution is shaken with 16 C.C. of ether (0.721-0*728), then3 C.C. of ammonia (0.96) added, again shaken, and left for 24 hours.The ether is removed by a pipette, the crystals are collected on afilter, washed with water saturated with ether, dried between filter-paper, washed again with ether, and dried.In the “ crystallisation process ” of Paul and Hewe, 5 ,mms ofquinine sulphate is dissolved in boiling water, and crystallised outfour times, using in the first case 150 c.c., next 130 c.c., and thentwice 120 C.C. The united mother-liquors are evaporated at a lowtemperakmre almost to dryness, the residue dissolved in the smallestpossible quantity of dilute sulphuric acid, made up to 20 c.c., andshaken with ether and excess of ammonia. The crystals which formare treated as in the bisulphate test. The hydro-bases crystallise inpart with the quinine, therefore the mother-liquor should not beused for their determination.The chromate process gives very varying results, but on theaverage gives the highest yield of cinchonidine, especially with thepurer samples The oxalate test gives the lowest numbers, but theyare more concordant than those of the chromate process. The com-position of the bye-product is, however, variable. The bisulphatetest gives results varying considerably. The alkaloids in the etherealsolution ought t o be submitted to the process a second and even athird time, but even with this improvement the whole of the cin-chonidine is not obtained, and the results vary much, but the com-position of the bye-product is more uniform than in the otherprocesses. The crystallisation test has the same advantages as thebisulphate test if the crystallisation is repeated often enough, and is theprocess which i8 least influenced by the presence of hydro-bases. I t is,however, tedious. The process of the German Pharmacopoeia depend88 ABSTRACTS OF CHEMICAL PAPERS.on the fact that the precipitate produced by ammonia in solntionc: ofthe alkalo'ids is soluble i n excess of ammonia, but that much moreammonia is required for quinine than for the other alkalolds. Theexcess of ammonia required varies, however, very considerably withthe temperature. M. J. S.Method for Recognising the Adulteration of Pepper by theAddition of Ground Olive Stones. By GILLET (BUZZ. XOC. chim.,50,173-174).-When 1 gram of olive stones is treated with 1 C.C. ofa 5 per cent. iodine tincture, it acquires, after a quarter of an hour, a,J ellowish colour, whilst pepper is coloured brown or maroon. A seriesof mixtures of pepper with 5, 10, 15, and 20 per cent. of olive stonesare prepared and coloured with iodine tincture. It is then easy todetermine the amount of olive stones in samples of pepper by com-paring the colour obtained with these types.A New Test for the Blood in Carbonic Oxide Poisoning.By R. KATAYAMA (Virchow's Archiv, 114, 53-64).-When ammoniumsulphide holding sulphur in solution and acetic acid are added tohlmd containing carbonic oxide, the result is a beautiful clear red ormse colour ; whereas normal blood becomes greenish-grey, or reddish-preen-grey on the addition of the same reagents. On examining theseliquids spectroscopically, it is found that both in the case of normal andcsrbonic oxide blood the absorption-spectrum indicates that a mixtureof two substances is present. In the case of normal blood, there is aband between C and D and another between D and E ; this lastbecomes double on shaking up the mixture with air. In other words,there is a mixture of sulphur methaemoglobin (see Hoppe-Seyler,Physiol. Chem., 386) and reduced haemoglobin.I n the case of carbonic oxide there are three bands : one between Cand D, due to sulphur methEmoglobin, and two between D and E,due to carbonic oxide hsemoglobin. That is to say, in spite of the pre-sence of sulphur methaemoglobin, the liquid does not become greenish,but remains red, the colour of carbonic oxide haemoglobin overpoweringthe olivegreen of the sulphur methaemoglobin. This test is stated tobe more delicate than Hoppe-Seyler's (Abstr., 1888, 540), and isobtained with a mixture which contains one part of carbonic oxideblood to five parts of normal blood.Estimation of Albumin in Urine. By H. SCHACJMANN (Zeit.afial. Chent., 27, 635-636J-The estimation of albumin is muchaccelerated by collecting it, not on a paper filter, but in a filter-tubeplugged with cotton-wool and connected with a filter-pump. Whenthe washing with hot water is complete, a calcium chloride tube isconnected, and dry air is drawn over the albumin whilst graduallyraising its temperature to 110".N. H. M.W. D. H.M. J. S