Improvements in the Manufacture of Gunpowder. C. H. Curtis, London, and G. G. André, Derking, Surrey. Eng. Pat. 19,068, November 4, 1891. GUN-COTTON (trinitrocellulose), which generally contains about 12 per cent. of dinitrocellulose soluble in ether (mixed with a little alcohol), is mixed with 6-16 per cent. of dinitrocellulose and manufactured while wet into pellets; or the usual process for making gun-cotton is so modified as to produce a nitrated cellulose containing 18-28 parts of the dinitro- to 88 parts of the trinitro-compound, which is then made into pellets. The pellets are dried and treated with a solvent capable of dissolving only the dinitrocellulose, which is thus made to thoroughly impregnate the trinitrocellulose and to bind it together into hard unfriable granules when the solvent has evaporated. This hardening process differs from those previously suggested in that they only serve to harden the surface of the granule, and leave it friable. By varying the proportion of dinitrocellulose within the above-mentioned limits any requisite degree of explosiveness may be obtained, for the larger the proportion of dissolved cotton present the slower the rate of combustion. By dissolving nitroglycerin in the solvent used for the dinitrocellulose this explosive may be combined with the rew gunpowder. Suitable proportions are 44 parts by weight of trinitrocellulose, 12 parts by weight of dinitrocellulose with or without 40 parts by weight of nitroglycerin.-A. G. B. T piece of ash-free filter paper (a quarter of a 9 cm. filter paper). On then boiling the solution and precipitating, the paper becomes disintegrated into fine fibres to which the sticky hydrocarbons become attached, while at the same time the precipitate of barium sulphate is caused to settle more rapidly. The operation of filtering is thereby much accelerated, and a clear filtrate is invariably obtained. -H. S. P. Lead in Glass- Wool. L. Blum, Zeits. Anal. Chem. 31, 292. THE author calls attention to the fact that some glass-wools, as sold in commerce, contain lead, being made from glass containing lead. Acids should therefore not be filtered through such wool, as they will thereby become contaminated with lead. It was also found that sulphuretted hydrogen causes such wool to blacken, owing to the formation of lead sulphide.-H. S. P. XXIII.-ANALYTICAL CHEMISTRY. APPARATUS. Composition of Glass Suitable for Chemical Apparatus. Absorption Apparatus for Estimating Sulphur in Iron. L. Blum. Zeits. Anal. Chem. 31, 290–292. THE apparatus consists of two parts, the bulb A with the long tuber attached, and the bulb B, affixed to the cylinder c. A and B have each a capacity of about 200 cc. The tuber is inserted into the cylinder c, and reaches nearly to the bottom, and fits into the neck of the bulb B by a ground hollow stopper, to which is also attached the exit tube m. Ammoniacal peroxide of hydrogen or hydrochloric acid saturated with bromine is poured into the apparatus at the mouth i of the bulb A, and rising in the cylinder c forms a lute. The gases from the flask in which the iron to be tested is dissolved and led into the apparatus at i, and passing over the moistened surface of the bulb A and down r bubble up through the liquid in c, by which means the oxidation of the sulphuretted hydrogen is completely effected, and the residual gases freed from sulphur then pass away by the tube m. It is advisable to stand the apparatus in cold water to prevent overheating by the steam that distils over from the generation flask. If bromine-saturated hydrochloric acid be used it is impracticable to precipitate the sulphur directly, because of the sticky, resinous bromine hydrocarbon compounds INORGANIC chemistrY.— QUALITATIVE. On the Qualitative Separation and Detection of Strontium and Calcium by the Action of Amyl Alcohol on the Nitrates. P. E. Browning. Journ. Anal. and Applied Chem. 6 (1892), 343–345. For the qualitative separation and detection of strontium and calcium, the author proceeds as follows:-The substance, not more than 0.2 grm., is heated to boiling with 5 cc. of amyl alcohol, and the filtrate and washings examined for calcium. The residue, if very small, may be simply a deposit of some calcium compound (see above), but if strontium nitrate is present, even in very small quantities, a decided residue is obtained; the residue is dried, ignited, treated with a drop of nitric acid, and again heated with amyl alcohol; if strontium is absent, the whole passes into solution, but if present, a distinct residue or deposit is obtained, even when the amount present is equivalent to 0·0005 grm. SrO.-F. S. K. containing a suitable liquid; within this is placed a moveable index, composed of magnetic metal, which may be moved up or down by a magnet from the outside. When applied to a Six's thermometer, either a single or two "vernier" tubes are arranged between the limbs of the thermometer, each tube being provided with an index, or a single tube bent so as to form two limbs may be employed. -J. C. Č. Improvements in Spatulas. C. W. Fox, Saugus, Mass., U.S.A. Eng. Pat. 16,830, May 19, 1892. THE blade of this spatula is made of india-rubber, and is provided with a broad thin steel core, which gives the blade the necessary flexibility.-J. C. C. INORGANIC CHEMISTRY.— QUANTITATIVE. Eudiometric Determination of Nitric Acid. Glaser. THE error in eudiometric estimations of nitric acid, caused by the oxidation of the nitric oxide to nitrous acid by oxygen contained in the liquid over which the gas is measured, the author finds can be entirely obviated by measuring the gas over a 1 per cent. aqueous solution of potassium iodide. Any nitrous acid that is formed is at once reduced to nitric oxide, accompanied by the separation of iodine. By adopting this plan it is possible to estimate the nitric acid, both in saltpetre and in water with great accuracy, but in the case of water analysis it is necessary to wash the gas with caustic soda solution to remove any carbonic acid.-H. S. P. The Determination of Sulphur. T. P. Treadwell. Ber. 1892, 25, 2377-2382. THE author has shown that the sulphur in insoluble sulphides can be determined by heating with iron powder in a current of carbon dioxide and decomposing the resulting sulphide by hydrochloric acid, the sulphuretted hydrogen evolved being determined either by absorption in ammoniacal hydrogen peroxide and subsequent precipitation as barium sulphate, or volumetrically by absorption in a titrated iodine solution, the excess of iodine being titrated back by sodium thiosulphate. The latter method is only applicable in absence of arsenic and antimony. (This Journal, 1891, 951.) Sulphates, when present with the sulphide, are also reduced under these conditions so that in the analysis of roofing slate containing sulphate of lime the total sulphur (both as sulphide and sulphate) is obtained. The sulphur contained as sulphide is, however, completely liberated as sulphuretted hydrogen without the sulphate being attacked, when the mineral is decomposed by granulated tin and hydrochloric acid. The decomposition is effected in a large test-tube 20 cm. long and 2.5 cm. broad, on the bottom of which a thin layer of finely granulated tin is first placed; the weighted substance wrapped in tin foil is then introduced, and on the top of this a layer of 6 cm. of the granulated tin is placed. The tube is fitted with a stopper provided with a stop-cock funnel and a bulb-exit tube, the latter leading to a Peligot's tube and then to the absorption bulbs containing ammoniacal hydrogen peroxide. A current of hydrogen is first passed through the apparatus for 3-5 minutes, the cock of the tap-funnel being open; the funnel is then charged with concentrated hydrochloric acid and the decomposition started by running in the acid until the layer of tin is about half covered. The contents of the tube are warmed and in 10-20 minutes the decomposition is complete. The heating is continued until the hydrogen peroxide solution in the first absorption vessel (two are generally used) is no longer yellow. More hydrochloric acid is then added so that the layer of tin is completely covered, and the heating continued for half an hour, a current of hydrogen being again passed through the apparatus after the contents of the Peligot's tube have been heated to boiling. The sulphur in the hydrogen peroxide solution is then determined as previously described. Analytical data confirm the accuracy of the method which is specially applicable to the determination of the sulphide in burnt pyrites and in roofing slate. Arsenic and antimony when present are liberated as hydrides, apparently quantitatively; the author is engaged in working out a method for the determination of these two metals based on this behaviour. A sketch of the apparatus employed is given. -C. A. K. Estimation of Antimony: the Gooch Crucible. T. Paul. Zeits. f. Anal. Chem. 1892, 537--543. THE author precipitates the antimony as sulphide and heats the precipitate in an atmosphere of carbon dioxide to from 200°-230° C., according to Rose's method, whereby it is freed from much admixed free sulphur and entirely converted into Sb2S3. (See Rose's Handbook d. anal. Chem. 6 Aufl (Finkener's) Bd. 11. 295.) The sulphide is filtered through a Gooch crucible of the ordinary pattern; a special air-bath for heating the crucible at 200-230° in a current of carbon dioxide is figured in the original. It consists essentially of a glass tube fashioned in the form of an ordinary clay pipe with a constriction in the bowl; in this constricted portion the crucible rests. The stem of the pipe is passed through the side of a cylindrical air-bath, and carbon dioxide is conducted through it. The temperature of the bath is kept at 230° until the weight of the crucible is constant. The paper concludes with directions for using the Gooch crucible; these are only such as are generally followed.-A. G. B. Estimation of Thallium by Titration. K. Sponholz. THE liquid must contain the thallium as a thallous salt, and must be acid. A normal or normal solution of bromine in water is run in from a burette, the end of which dips under the surface of the liquid. The termination of the reaction is indicated by the yellow colour imparted to the oxidised solution by the excess of bromine: this excess must be taken into account by subtracting from the number of cc. of bromine water used, the number of cc. necessary to similarly colour an equal volume of distilled Two atoms of bromine are equivalent to one atom of thallium. The strength of the bromine water is determined by potassium iodide and sodium thiosulphate, and this valuation must be made anew every time the burette is filled.-A. G. B. water. Estimation of Phosphoric Acid in Basic Slag. A. F. Jolles. Zeits. f. Anal. Chem. 1892, 516-519. MAERCKER'S method of dissolving the slag in sulphuric acid gives low results because the calcium sulphate encloses particles of calcium phosphate which thus escape action. H. and E. Albert (Zeits. f. angew. Chem. 1891, 280) boil 5 grms. of the finely-powdered slag with 40 cc. of hydrochloric acid (16°-20 B.) until only a few cc. of a thick magma of silicic acid remains. Water is added, and the flask shaken until the lumps of silica are broken up, the volume is then made up to 500 cc. and filtered. To 50 cc. of the filtrate 15 cc. of Joulie's solution (400 grms. of citric acid made up to a litre with ammonia of sp. gr. 0-9) and magnesia mixture are added. H. and E. Albert claim that the silica is rendered sufficiently insoluble by this method, and evaporation to dryness is thus avoided. The author, however, has tested the process against the molybdate method, and has always obtained results which are too high, because the ammonium magnesium phosphate contains silica, the separation of which is not perfectly effected by the boiling down with hydrochloric acid. Titration of the dissolved magnesium precipitate with uranium acetate also gives too high results. The author recommends that the hydrochloric acid solution be evaporated to dryness and the residue heated to 120°, and that 50 cc. of citrate solution be employed; otherwise his process is similar to that of H. and E. Albert.-A. G. B. A Method for the Determination of Barium in the Presence of Calcium and Magnesium. F. W. Mar. Journ. Anal. and Applied Chem. 6 (1892), 346–351. BARIUM chloride is soluble to an extent not exceeding 1 part in 20,000 in pure concentrated hydrochloric acid, but the solubility increases very rapidly with the diminution in the strength of the acid; in concentrated hydrochloric acid containing ether its solubility is not more than 1 part in about 120,000. For the separation of barium from calcium and magnesium, the chlorides of the metals are dissolved in the least possible amount of boiling water, and precipitated with 25 cc. of concentrated hydrochloric acid, with the addition of 5 cc. of ether after cooling. The acid should be added drop by drop at first, allowing the precipitate formed to redissolve as long as possible, as in this way the precipitate is obtained in a coarse crystalline condition; after standing for a few minutes the precipitate is collected on a Gooch filter, washed with hydrochloric acid containing about 10 per cent. of ether, and dried at 150-200. The method is accurate and rapid.—F. S. K. A Method for the Quantitative Separation of Strontium from Calcium by the Action of Amyl Alcohol on the Nitrates. P. E. Browning. Journ. Anal. and Applied Chem. 6 (1892), 333–339. THE solubility of strontium nitrate in boiling amyl alcohol is extremely small, being equivalent to about 0.0010 grm. Sr in 25 cc. of the alcohol; calcium nitrate, on the other hand, dissolves freely in boiling amyl alcohol, although traces of some solid substance, possibly the calcium salt of some organic acid, separate on the sides and bottom of the vessel in which solution takes place. The separation of the two nitrates can, therefore, be effected with the aid of amyl alcohol; the mixed nitrates are heated with the solvent to the boiling point of the alcohol, the filtered extract evaporated, the residue treated with nitric acid, dried, and again extracted as before, the solution being filtered through the original filter. The combined residues, after correcting for the solubility of strontium nitrate, give the weight of the last-named salt; the solution contains the calcium nitrate which is estimated as sulphate. The average error in the determination of the strontium is less than 0.0001 grm. SrO; in the case of the calcium it varies from 0.005 to 0.0017 grm. CaO.-F. S. K. Valuation of Tin Crystals and of Zinc Dust. A. Fränkel. Mittt. k.k. Techn. Gewerbe-Museums, 1892, 223-227. Tin-Crystals. The customary method of estimating tin in stannous chloride by means of a standard solution of iodine has the drawback that the iodine solution is unstable, and hence requires to be frequently restandardised. This is also the case with the potassium permanganate solution employed in the method of Löwenthal and Stromeyer (Ann. 117, 261) in which the stannous salt is heated with an excess of ferric chloride and the ferrous chloride which is produced, determined, and also with the alkaline potassium manganate employed by Jolles (Chem. Zeit. 12, 597). Potassium bichromate solution, which has been recommended by Goppelsröder and Trechsel (Zeit. f. Anal. Chem. 16, 364) is free from this objection. The authors have, therefore, investigated the accuracy of the method, which consists in adding the stannous chloride to excess of an acidified solution of potassium bichromate, adding potassium iodide, and titrating with sodium thiosulphate. The length of time required for the completion of the reaction between chromic acid and potassium iodide was first ascertained. This was found to depend upon the state of dilution of the solution, as Zulkowski (Zeits. f. Anal. Chem. 8, 74) has shown. With a decinormal solution of potassium bichromate (4.91C7 grms. per litre) half an hour is required. The influences of the excess of chromic acid and hydrochloric acid used, were also studied, and the following method finally adopted: 3-4 grms. of the stannous salt are dissolved in 30-40 cc. of hydrochloric acid (10 per cent.) and diluted to 500 cc. 50 cc. of this solution are taken and treated in a stoppered flask with 50-60 cc. of decinormal potassium bichromate solution and allowed to stand for 15 minutes. Then 10-15 cc. of potassium iodide solution (1:10) and 5-10 cc. of hydrochloric acid (1:10) are added, and the whole allowed to stand for 30 minutes. The solution is then diluted to 200 cc., a small quantity of starch paste added, and the liberated iodine titrated with decinormal sodium thiosulphate. A series of determinations by this method gave percentages of SnCl2, 2 H2O, ranging from 88.37 to 88 84, the iodine test giving a mean of 88.39. Streng's direct method of titration by potassium bichromate, with potassium iodide and starch as indicators, gave with the same sample 86 13 per cent. Zinc Dust.-Drewsen's method of estimating zinc in zinc dust consists in adding to the latter an excess of potassium bichromate and sulphuric acid, and titrating with ferrous sulphate. The following modification of it is proposed: 0.5 grm. of zinc dust is placed in a dry flask, 50 cc. of seminormal potassium bichromate solution (24.58 grms. per litre), and 5 of dilute (1:5) sulphuric acid added, and the mixture shaken for 10 minutes. Then further 7-10 cc. of dilute sulphuric acid are added. and when the zinc is dissolved 40-45 cc. of potassium iodide solution (1:10) and 20 of dilute sulphuric acid. After standing for five minutes 200-300 cc. of water are added, and the iodine estimated by a seminormal solution of sodium thiosulphate, the starch-paste indicator being added near the close of the titration only. Satisfactory results are obtained. The amount of zinc present is calculated by multiplying the number of cc. of potassium bichromate reduced, by 0.016275. Examination of Iodine containing Chlorine. F. Ulzer and A. Friedreich. Mitt. k.k. Techn. Gewerbe-Museums, 1892, 287-289. A RAPID and exact method of determining chlorine in commercial iodine consists in dissolving 10--20 grms. of the sample in 100 cc. of carbon disulphide in a small separating funnel. 5-10 cc. of water is added and the cylinder agitated until the aqueous layer, which contains the iodine chloride, is saturated with iodine, standard potassium iodide solution is then run in with constant shaking. So long as iodine chloride is present, a turbidity is produced in the aqueous layer on the addition of potassium iodide, owing to the liberation of iodine; the termination is reached when no further cloudiness is obtained. Trial estimations gave very good results.-W. J. P. F The following substances do not give the reaction: formaldehyde, chloral, butylchloral, glyoxal, benzaldehyde, o-hydroxybenzaldehyde, vanillin, p-homosalicylic aldehyde, benzophenone, naphthyl-phenylketone and benzil. a- and B-dinitronaphthalenes give an intense red colouration with aldehydes and ketones of the fatty series. Trinitroresorcinol gives no reaction. Picric acid gives an orange yellow coloration with aldehydes and ketones. Creatinin gives an orange yellow coloration with picric acid, but no reaction with the other nitro-compounds. Iridole gives no reaction with the above-mentioned nitro. compounds.-A. K. M. Examination of Ethereal Oils. Jahresbericht von Schimmel and Co., Leipzig, April 1892. ACCORDING to investigations carried on in the laboratory of the above-mentioned firm, pure oil of lavender contains only a small quantity of constituents boiling at a low temperature. Cineol, which is found in considerable quantity in spike oil (the oil from Lavandula spica), was not found in lavender oil, nor was camphor found to be present, although according to older authorities it is said to be an important constituent of lavender oil. The chief The components of lavender oil are an alcohol having the formula C10H18O, and the acetic ester (ether) of this alcohol. alcohol boils at 197-199°, has a specific gravity of 0.869 at 20°, and rotates the plane of polarised light in a tube 100 mm. long, -10° 35'. Its refractive index for sodium light is 1.64. When heated with hydrogen-potassium sulphate, with sulphuric acid, or similar water-abstracting reagents, it yields a mixture of hydrocarbons, among which were recognised dipentene and terpinene. On oxidation, the alcohol formed an aldehyde, C10S160, identical with citral (Semmler's, geranium aldehyde). The alcohol of lavender oil forms an additional compound with four atoms of bromine. With hydrochloric acid it gives a liquid compound having the composition C16H18Cl. The action of the acid anhydrides upon it produces the corresponding ester. The acetic, propionic, butyric, valeric, and benzoic esters were prepared. The acetic ester, which occurs naturally in lavender oil, possesses in a high degree the characteristic smell of bergamot. Bergamot oil contains about 40 per cent. of this ester. The alcohol of lavender oil is identical with linalool, which was found by Semmler in Linaloë oil, and it is the source of the perfume of this oil. Schimmel and Co. place pure linalool in the market. It is an almost colourless liquid, easily soluble in alcohol, and with a peculiarly fine odour. Its specific gravity is 0.878 at 15°, its boiling point 197-198°, and its optical rotating power + 2°. Besides linalool, geraniol is also found in Linaloë oil. (See also this Journal, 1891, 157, 269, 362, and 1892, 706.) Schimmel and Co. find Thom's method for testing oil of cloves accurate to about 1 per cent., but they do not find Panjatow's method for detecting geranium oil in oil of roses practicable, owing to the fact that when pure rose oil has been exposed to the air and to light, an aldehyde is formed in it which gives the same characteristic colour reaction as citral the aldehyde of geranium oil does with the reagent used, viz., a solution of magenta decolorised by sulphurous acid.-H. S. P. Distinguishing Gum Arabic from Gum Senegal. L. Liebermann. Pharm. Post through the Deutsch-Amerik. Apotheker Zeit. 11, 71. THE woody particles insoluble in water are usually of a red colour in gum arabic, and of a black colour in gum senegal. Apart from these woody particles, both gums are perfectly soluble, which is not the case with some other kinds of gum, e.g, cherry gum (Kirschgummi). An aqueous solution of gum arabic gives a blue precipitate with caustic potash solution and a few drops of cupreous sulphate, which collects together and rises to the surface, whilst gum senegal gives a blacker coloured precipitate, which is equally diffused through the liquid in small flocculent particles. On heating, these precipitates are only very slightly soluble, and there is no reduction of copper on boiling. These two points distinguish these gums from dextrin. Gum arabic and dextrin, when heated for some time with caustic soda solution, yield amber-yellow liquids, but gum senegal gives scarcely any colour, or only a faint yellow colour. If it be desired to distinguish the gums in presence of dextrin, the solution is precipitated with caustic potash and cupreous sulphate, heated to dissolve the dextrin compound, filtered, the precipitate washed with distilled water, and then dissolved in hydrochloric acid. The gums are then precipitated with a large excess of alcohol and the precipitate allowed to settle for from 12 to 24 hours, when the liquid is poured away, and the precipitate washed with alcohol. The alcohol retained by the gums is then allowed to evaporate, and the gums are dissolved in water and tested as above described.-H. S. P. The Detection of Veratrine. E. Laves. Pharm Zeit. 1892, 338. THREE to four drops of a 1 per cent. aqueous solution of furfuraldehyde mixed with 1 cc. of sulphuric acid (conc.) form the reagent, 3 to 5 drops of which are used for the test, which is made in a small porcelain dish. In presence of veratrine, dark green and blue striae are formed as the reagent mixes with the substance to be tested, and on agitating, the whole turns to a dark green colour, which becomes violet on prolonged standing-more rapidly on gentle warming. The violet coloration is fairly permanent. -C. A. K. Baudouin's Reaction for Sesame Oil. V. Villavecchia and G. Fabris. Zeits. Ang. Chem. 1892, 509-510. MANY varieties of olive oil obtained from the provinces of Bari, Brindisi, Lecce, and other places, and of undoubted purity, give a red coloration in Baudouin's reaction which has hitherto been regarded as characteristic only of sesame oil. The reaction is generally carried out as follows:0.1 grm. of sugar are dissolved in 10 cc. of hydrochloric acid of sp. gr. 1.18 in a test tube, 20 cc. of the oil to be tested added, the whole thoroughly shaken, and then allowed to stand. After a short time the lower aqueous solution is coloured more or less carmine red, in presence of sesame oil, 5 per cent. of which can thus be readily detected. Under these same conditions the olive oils referred to above produce the same red coloration. In the opinion of the authors the coloration in both cases is due to the existence or the formation of levulose in the hydrochloric acid solution of the sugar; this point is still under investigation. Meanwhile, the following modification of the reaction serves to distinguish the above olive oils from sesame oil. The test is carried out as described, the shaking after the addition of the oil being continued for one minute. The aqueous solution separates immediately in the case of pure olive oil, and remains colourless for at least two minutes. The somewhat milky oily layer remains of a greenish or yellowish colour, but in the presence of the least quantity of sesame oil it turns red, and it is this colouring of the oil, rather than that of the aqueous solution, that is taken as the indication of the presence of sesame oil, as experiment showed that the coloration in the case of olive oil takes place by preference in the aqueous solution.-C. A. K. ORGANIC CHEMISTRY.-QUANTITATIVE. Analysis of Lard. C. Amthor and J. Zink. Zeits. f. Anal. Chem. 1892, 534-537. (This Journal, 1891, 801.) THE lard was rendered by the authors themselves from different individuals, and at different seasons. The mean values of the results obtained are given in the following table: The iodine absorption of the fatty acids decreases considerably with their age and with their exposure. Thus the mean iodine absorption of fatty acids which had been kept in closed vessels for 42 days was 461, and of those which had been kept in open vessels 44.34, the same value for the fresh acids was 51 61. With regard to the controversy as to the conditions under which the iodine absorption of fats should be determined (compare this Journal, 1891, 954; 1892, 183; Chem. Zeit. 1892, 651), the authors find that in several analyses of the same sample of lard the mean values obtained are practically the same, whether the iodine solution be old, new, or made at the moment, after Fahrion's recommendation (this Journal, 1892, 183), but the values vary least among themselves when the iodine solution is some 45 days old. In the case of lard, at all events, there is no necessity for a large excess of iodine; equal values were obtained when the excess varied from 21.2 to 84 per cent.-A. G. B. The Determination of Indigotine in Indigo. F. Ulzer. Mitt. k.k. Techn. Gewerbe-Museums, 1892, 215-217. FURTHER experience of the rapid method of indigo-testing devised by the author (this Journal, 1892, 63), has shown that the results accord with those obtained by other methods only with samples of high quality, being in other cases too low. This has been discovered to be due to the reducing action, in the presence of alkali, of the impurities contained in the indigo on the indigotin, and is obviated by the employment of hydrogen dioxide along with the alkali used. The improved process is as follows:-About 1 grm. of the finely-powdered sample is mixed with 50 cc. of a 5 per cent. aqueous solution of hydrogen dioxide and boiled for 10 minutes. The mixture is cooled, diluted to twice its volume, filtered through a weighed paper of 10-12 cc. diameter, and the residue washed successively with warm water, dilute hydrochloric acid, water, and boiling alcohol, which is passed through it until the filtrate is no longer brown or red but pure blue, the filter being thereupon dried and weighed. The dried mass is then burnt, and the weight of ash found deducted from the previous weighing, the difference representing the indigotin. To determine the indigo-red (indirubin) the alcoholic solution is evaporated and the residue weighed. The following results were obtained by this method : 6 |