1.168 ABSTRACTS OF CHEMICAL PAPERS. A n a l y t i c a l Chemistry. A Solution of Density 3-28 suitable for the Analysis of Minerals. By M. D. KLI:IN (Compt. rend., 9 3 , 3 1 8 4 2 0 ) .-Inasmuch as the densitly of most minerals ranges from 2.2 to 3.3, they can be easily separated from one another by immersion in solutions without chemical action on them. Solutions of mercuric iodide in potassium iodide have partly fulfilled the necessary conditions, but the minute precautions required have prevented the general introduction of this method. Nearly saturated solutions of the cobalt, nickel, and cad- mium borotungstates prepared by the author have sp. gr. about 3.28-3.31. The dark tint of the two former salts renders their solution comparatively useless for the purpose, but the cadmium salt solution being of a very pale yellow colour is very suitable. The author illustrates how, for instance, by this solution a diamond can be distinguished immediately from white sapphire or colourless corundum.In order to prepare this solution, tetrasodic horotunstate (this vol., 224) is converted into barium borodecitungstate (Abstr., 1880, 612), and then decomposed by cadmium sulphate. The solution is eva- pornt,ed on a salt-water bath until a fragment of peridote swims on the surface ; then on cooling orthorhombic crystals of the cadmium salt are deposited. The crystals melt at 75' in their water of crystallisation and retain their mother-liquor, and there is formed a mobile liquid of the required density, 3.6. Estimation of Phosphoric Acid as Magnesium Pyrophos- phate.By F. A. GOOCH (An?. Chent. J., 1, 391-413).-This paper gives the results of a large number of analyses of alkaline phosphates, phosphotungstates, and phosphomolybdates, from which the following conclusions are drawn :- V. H. V.ANALYTTCSL CHEMISTRY. 1169 In determining the phosphoric oxide of aZkaZiine phosphates free from sulphates or other substances likely to contaminate a precipitate, accuracy is most conveniently and surely attained by adding to the somewhat dilute solution of the phospha.te, first ammonia in slight but distinct excess, then an excess of magnesium chloride mixture con- taining free ammonia (made by dissolving 3 parts crystallised magne- sium chloride and 8 parts ammonium chloride in water, adding water containing ammonia till the volume of the solution amounts to 48 parts, then filtering and boiling off the free ammonia,), and after the precipitate has settled, adding a further excess of ammonia till the liquid smells strongly of it.In determining the phosphoric oxide of phosphotungstates, the best results are obtained by adding to the solution containing free ammo- nia, an excess of either of the magnesia mixtures, washing the pre- cipitate with the precipitant, dissolving in hydrochloric acid, diluting if necessary, and reprecipitating with a little dilute ammonia, adding strong ammonia after the precipitate has settled ; or, when working with small amounts, by proceeding as in the case of the phosphates, taking however special care that the solution is distinctly ammoniacal before precipitating, and that no great excess of the precipitant is used.For the determination of phosphoric oxide in phosphomolyb- dates, the method recommended in the case of the phosphates serves very well when the amount, of phosphoric oxide does not exceed about 0.05 gram, but for larger amounts, the mebhod of double precipitation recommended for the phosphotungstates is more accurate. With regard to the time which should be allowed t o pass between precipitating and filtering, the author’s experiments support those of Abesser, Jnni, and Maercker (this Journal, 1874, 387) in pointing to the condition that a precipitate map safely be filtered as soon as it has completely subsided, generally after ten or fifteen minutes. By D. LTNDO (Chem. News, 44, 77, 86, 97, 129).-A moderate excess of platinum chloride solution, containing about 1 gram of platinum in ‘LO c.c., is added to the somewhat dilute solution of alkaline chlorides, the liquid evaporated to dryness on the water-bath, and t,hen heated for two or three minutes.The residue is washed three or four times with alcohol of 98 per cent., then carefully crushed with a glass rod, and again washed repeatedly with alcohol, Gooch’s method of filtra- tion (Chem. News, 37, 181) being employed. The platinochloride mas dried a t 130”. The results obtained were very accurate, even in presence of 50 per cent. sodium chloride. The double salt is not hygrcscopic, and there is no need to place the crucible in a desiccator whilst cooling. When using Tatlock’s method, the author finds that the platinum tetrachloride solution employed for the first washing dissolves a small quantity of the double salt, and consequently the results obtained are somewhat too low.Even in presence of excess of platinum chloride, it would appear that sodium chloride is not entirely converted into the double salt or, more probably, that rever- sion takes place either during evaporation or on adding the alcohol. Pi*olonged heating of the residue on the water-bath promotes rever- H. W. Estimation of Potassium as Platinochloride.1170 ABSTRACTS OF CHEMICAL PAPERS. sion, which may be prevented by employing a moderate excess of platinum chloride, and taking care not to heat t,he platinochlorides too long after they have become dry. Direct experiments showed that potassium platinochloride is practically insoluble in absolute or very strong alcohol; when the platinochloride is in the form of scales, alcohol of 80 per cent.dissolves a small quantity ; when in the form of powder, alcohol of 83 per cent. dissolves a quantity sufficient to sensibly affect the results. The author finds that potassium platino- chloride is very slightly soluble in a half saturated solution of ammo- nium chloride, and proposes to use a solution of the latter to wash out any magnesium and sodium sulphates, &c., with which the platinochloride may be mixed. The double salt is first thoroughly washed with alcohol, then with an aqueous solution of ammonium chloride, and finally with alcohol. If accurate results are desired, the ammoniui u chloride solution should be saturated wit'h platinochloride before being used.Taking the atomic weight, of platinum as 197.18, the mean of the determinations of potassium chloride was 100 per cent.; taking Seubert's number 194.46, the mean was 100.55 per cent. C. H. B. Electrolytic Estimation of Zinc. By H. REINHARDT and R. IHLE (J. pr. Chem. [2], 24, 193--201).-1t is found that the usual processes for the electrolytic estimation of zinc are all to a certain extent unsatisfactory. The presence of fvee sulphuric acid and am- monium acetate are especially prejudicial t'o the results. The follow- ing modification is therefore proposed : the anode consists of a spiral of strong platinum wire, and the cathode of a cone of platinum sus- pended above the anode, the distance between the two being 5 mm., and the surface of the latter 160 sq.cm. The object of the peculiar form of cathode is to ascertain, by lowering it into the liquid, if the last trace of metal has been removed. The negative electrode is to be previously coppered, and nitric acid may be employed for removal of the deposited zinc, as the copper is not attacked in the cold. The copper coating must be even and clear, otherwise the zinc will be deposited in numerous small loosely attached grains. The zinc is deposited from a solution of the oxalate, and for this purpose the neutral chloride or sulphate solutionis mixed with an excess of potas- sium oxalate until the precipitate produced id redissolved ; the metal is precipitated from this solution a t the negative pole, and may be easily xashed with water.This process works well with alloys. E. W. P. Reduction of Iron Ores by Powdered Zinc. By T. M. BROWN (Chem. Centr., 1881, 75).-An intimate mixtme of 0.3 gram powdered ore with 10 times as much finely pulverised zinc is placed in a porce- lain crucible and covered with an equal weight of powdered zinc. The open crucible is heated for ten minutes over a Bunsen burner to dark redness. On cooling, the whole contents of the crucible are rapidly dissolved in hydrochloric acid out of contact with air, and the iron is determined by potassium permanganate in the usual manner. Bituminous ores must, be roasted before reduction. To prevent the introduction of organic matter from the lamp, a layer of borax isANALYTICAL CHEMXSTRY.1171 placed above the layer of zinc. Powdered zinc is also found to have advantages over granulated zinc and sheet zinc in the reduction of ferric to ferrous solutions. P. L. T. Separation of Cobalt and Nickel from Iron. By T. MOO-~E (Chem. News, 44, 76).-The solution, which should not contain much free acid, is mixed with ammonium sulphate in quantity sufficient to form a double sulphate with the cobalt andnickel present. The solu- tion is then diluted to about 150 c.c., mixed with a large excess of oxalic acid and well stirred. If a precipitate is formed, it must be redissolved by further addition of ammonium sulphate. The clear liquid is mixed with ammonia in slight excess, heated gently for a few minutes, filtered, and washed with water containing ammonia, or the liquid is diluted to a known volume and, after the precipitate has settled, a definite quantity of the clear liquid is drawn off and the nickel determined by electrolysis or otherwise.Separation of Tungsten from Antimony, Arsenic, and Iron, together with the Analysis of a so-called Pseudometeorite. By A. COBENZL (Monatsh. Chem., 1881, 259-265) .-The metallic mass which is the subject of this examination, was found in 1879 at Cista in Bohemia. Externally it exhibits some of the characters of meteoric iron, but as its fall was not actually observed, and as its physical and chemical characters were found on closer examination to differ from those of well authenticated iron meteorites, it has been designated a pseudometeorite. It is brittle, easily pulverisable, and has a fine-grained confusedly crystalline texture, revealed by the shim- mering aspect of its newly fractured surface, whereas meteoric irons are tough and have a hackly fracture.Moreover, the pseudometeorite is porous and has a blistered surface, the character of which is more like that of certain furnace-products than that of meteoric iron. The colour of this metal is violet-grey, differing therefore from the neutral grey of meteoric iron. Qualitative analysis showed that the chief constituents of the pseudometeorite are iron, tungsten, antimony, and arsenic, associated with small quantities of water and insoluble silicates (ferric oxide, alumina, lime, magnesia), carbon, and sulphur ; traces of bismuth and tin were also detected. The quantitative analysis presented pecnliar difficulties in the separation of the first four constituents above mentioned, respecting which no exact statements are to be found in chemical literature.The following method was found by the author to yield satisfactory results. The mineral, very finely pulverised and sifted through linen, was drenched in a flask with strong nitric acid, and heated on a water-bath €or four or five days, with occasional addition of hydro- chloric acid, till nothing was left undissolved but pure yellow tungstic acid. The solution, together with the tungstic acid, was evaporated t o dryness over a water-bath ; the dry dusty residue redissolved in very dilute nitric acid ; the solution again evaporated to dryness, and this treatment three times repeated.The residue was then dissolved in very dilute nitric acid, with addition of a small quantity of tartaric C. H. B. VOL. XL. 4 k1172 ABSTRACTS OF CHEMICAL PAPERS. acid, the solution evaporated a t 100" and filtered, and the separated tungstic acid several times washed by decantation with boiling acidu- lated water, and finally collected on a filter. The solid matter thus separated contained the whole of the tungs- ten, together with silica, and silicates undecomposible by acids, whilst all the other metals, together with lime and alumina, were found in the filtrate. The tungstic acid was treated on the filter with very dilute ammonia, which dissolved it, leaving behind the silica and sili- cates. The ammoniacal solution was evaporated t o dryness in a weighed porcelain crucible, the residue ignited, and the pure straw-yellow turigstic acid thus obtained was weighed.The undissolved silicates were dried, ignited, weighed, and analysed in the usual way by decomposition with a mixture of potassium and sodium carbonate. The filtrate from the tungstic acid was saturated with yellow ammo- nium sulphide, whereby iron sulphide and alumina were precipitated, while arsenic, antimony, and lime remained dissolved. All these con- stituents were separated and estimated by the usual methods. The sulphur, after separation of the tungsten, was weighed as barium sulphate. The carbon was estimated, together with the water, by igniting the powder in a stream of oxygen, and collecting the water and carbonic acid in the usual way. The amount of oxygen was determined by reducing the oxides present in a stream of hydrogen.The results of the analysis were as follows :- Fe. w. Sb. As. Silicates. H20. Al,O,. 56.07 25-39 9.85 5.08 1-55 0.78 0.60 CaO. 0. C. S. 0-37 0.28 0.18 0.053 The percentage composition of the silicates was found to be : SiO,. Al,O,. Fe20,. CaO. 41.2 26.8 18.0 14.0 From the physical and chemical characters above described, it may be inferred almost with certainty that this substance is not an actual iron meteorite, but a furnace-product, such as were often obtained in the early days of the metallurgic preparation of tungsten. The fol- lowing table exhibits a comparison of its chief constituents, and its specific gravity, with those of iron meteorites :- Mean of iron meteorites.Pseudometeorite. Iron.. Iron. ....... 5 6.4 Nickel ...... 3 1 7 Tungsten . . 25.3 Phosphorus.. 0 1 Antimony .. 9.9 Cobalt.. .... 0 2.6 Arsenic .... 0.1 ...... 81 to 98 p. c. Sp. gr.. ..... 6.6 7.9 I n the lump. 8.854 In powder .. 8.8993 H. w. Action of Potassium Permanganate on Potable Waters at Different Temperatures. By G. W. WIGNER and R. H. HARLAND (AnaZyst, 1881, 39).-The experiments of the authors show that theA N AL Y TICAL CHEMISTRY. 1173 amount of oxygen absorbed by potable waters when mixed with potassium permanganate varies considerably with the nature of the water and with the time, and in some cases with the temperature. The nature of the pollution also has a marked effect on the oxygen absorbed, but no general conclusions can be drawn from their results.L. T. 0’s. Estimation of Nitrates in Potable Waters. By J. WEST- KNIGHTS (AmZyst, 1881, 56--58).-The author applies the brucine test for nitric acid to the estimation of nitrates in potable waters, the blood-red colour being permanent if oxalic acid is used instead of sulphuric acid, in applying the test to nitrates. For the purpose are required a potassium nitrate solution containing 0.721 gram per litre (1 C.C. = 0.0001 gram N as NO,), a brucine solution (1 gram in 100 alcohol), a cold saturated solution of oxalic acid, and a standard xed SOZZL- tioiz prepared by evaporating 10 C.C. of the potassium nitrate solution to dryness, and adding 3 C.C. of the brucine solution and 6 drops of oxalic acid solution, evaporating to dryness, dissolving the residue in water, and again evaporating; the residue is dissolved in water and made up to 100 C.C.This solution should have a bright-red colour, and 1 C.C. of it is equal to 0.00001 nitrogen as nitrate. 10 e.c. are evaporated to dryness, and from 0.5 to 2 C.C. of brucine solution added to the residue. The brucine should be just in excess, and the colour a bright red; if it be brown, a fresh quantity of water mud be taken, and a larger quantity of brucine added ; but if the colour is pink, a smaller quantity of brucine must be used. Three or four drops of oxalic acid are added to the residue, which is treated as in the case of the standard solution, except t h t fhe residue after the final evaporation is dissolved in a small quantihy of water, filtered into a glass cylinder, and the volume made up to 50 c.c., and the colour compared with from 1 to 10 C.C.of the standard colour. If the colour produced by the water is deeper than that of 10 C.C. of the standard, i t must be diluted with two or three volumes of water, and 50 C.C. treated as before ; but if it he lighter than that of 1 c.c., a larger quan- tity of water must be used to begin with. By F. P. PERKLNS (Analyst, 1881,58-59) .-The author applies a magnesium-platinum couple in presence of sodium chloride to the reduction of nitrates in river water. L. T. 0’s. The water to be examined is prepared in a similar manner. L. T. 0’s. Estimation of Nitrates in River Water. Detection of Lead in Potable Waters. By S. HARVEY (Ana7yst, 1881, 146-148) .-Potassium dichromate is a more delicate test than sulphuric acid or potassium iodide for the presence of lead in water.It is also preferable to sulphuretted hydrogen, since in the event of copper or tin being present in the water, there is a difficulty in distinguishing between the precipitates. It is possible to detect one-fiftieth of a grain of lead per gallon by this means, even in pre- sence of sulphates. The potassium dichromate is added in the solid state. L. T. 0’s. 4 J; 21174 ABSTRACTS OF CHEMICAL PAPERS. Assay of Wood Spirit for the Preparation of Methylated Spirit. By C. BARDY ( J . Pharna. [5], 4, 129--135).-The author describes a method for the estimation of methyl alcohol in wood spirit, which consists in the comparison of the amount of iodoform obtained from the sample with that from a standard wood spirit.For the purpose, a ,standard solution of caustic soda is prepared, and a solution of 254 grams iodine and 385 potassium iodide in 1 litre of water. 5 C.C. of the wood spirit t o be assayed is diluted to 50 C.C. with water, and 10 C.C. of this solution is made up to 1 litre with water, and 5 C.C. of this is placed in a tube with 10 C.C. soda solution and -5 C.C. of the iodine solution. After well shaking, 30 C.C. of water are added, and the milkiness due to the formation of iodo- form is compared with that in a similar solution prepared from the standard wood spirit. The comparison is effected by placing behind the tubes a white card, marked with a number of black lines of various thicknesses, and noting the number of lines visible in each case.The comparison should he made immediately, since the iodo- form settles after the lapse of a short time. Should the milkiness be less than that produced by the test solution, the wood spirit is too rich in methyl alcohol, and cannot be used. If the milkiness is alike in both cases, or greater in the case of the sample than in that of the test, it is necessary to estimate the methyl alcohol by Bardet and Bordet's method. This method cannot be used if the wood spirit contains a large pro- portion of aldehyde, which is detected by diluting the wood spirit with 10 of water, and adding to 5 C.C. of the diluted spirit 2 C.C. of aqueous solution of magenta (0.20 grain per litre), decolorised by sodium thiosulphate, when a violet colour is produced, varying in intensity with the quantity of aldehyde present.The aldehyde is estimated by distilling 5 C.C. of the wood spirit, with 7 C.C. of sulphuric acid, diluted with 10 c,c. of water. The value of the spirit may then be estimated as above. In assaying wood spirit, i t is necessary that it should contain no ethyl alcohol, accidentally or fraudulently added, and no methyl ethers. The latter are estimated by saponification. The estimation of alcohol, mixed with wood spirit, is effected by diliiting 10 C.C. to 1 litre with water, and proceeding as above. I f essences are present, 11 C.C. are diluted, and the solution filtered through a moist filter-paper. To estimate varnishes prepared with met'hylated spirit, 11 C.C. are distilled to dryness, the distillate diluted to a litre, and the above method applied.By J. MUTER (Analyst, 1881, 41-46) .-The power of glycerol to prevent the precipitation of copper by potash from its solutions is made use of for the esti- mation of glycerol. 1 gram glycerol is washed into a stopper- graduated tube with a stopcock 50 C.C. from the bottom (Muter olein tube), and 50 C.C. of potassium hydroxide (1-2) added. A dilute copper sulphate solution is then carefully run in, with constant shaking, until a fair amount of copper hydroxide remains undissolved ; L. T. 0's. Volumetric Estimation of Glycerol.ANALYTICAL CHEMISTRY. 1175 the whole made up to a given volume and allowed to settle. When clear, a given volume is made slightly acid with nitric acid, a definite excess of ammonia added, and finally titrated with potassium cyanide solution, of which 1 C.C.= 1 gram glycerol. By using pure potassium cyanide made from the acid, good results may be obtained ; there is a slight deficiency, however, owing to the difficulty of working the cyanide estimation of copper to a constant point. I n cases where the amount of glycerol to be estimated is very small, the copper is deter- mined by electrolysis on platinum. L. T. 0’s. Detection of Hydrocyanic Acid. By E. LUDWIG and 5. MAUSHNER (Chem. Centr., 1881, 43-44) .-In a poisoning case sub- mitted to the authors, potassium ferrocyanide was shown to be pre- sent, and was removed by slightly acidulating and carefully preci- pitating by ferric chloride. On submitting the filtrate to distillation with tartaric acid, a distillate was obtained both from the contents of the stomach and from the residue in the bottle containing much hydrocyanic acid.The sample of potassium cyanide which is supposed was employed was subsequently found to contain a very considerable quantity of ferrocyanide. 3’. L. T. New Demonstration of Carbonic Anhydride in the Breath. By C. F. CROSS (Chem. News, 44, 141).-When respired air is blown into a solution containing a mixture of potassium iodide and iodate and starch, the carbonic acid causes the liberation of iodine, and con- sequent production of it blue coloration. Direct experiments proved that the liberation of iodine is not due to the action of the oxygen of the air, or to the accidental presence of acid vapours. Determination of Salicylic Acid in Food-stuffs by a Colori- metric Reaction.By H. PELLET and J. DE GROBERT (Compt. rend., 93, 278--280).-The method of determining salicylic acid by the titration of the residue obtained by an ethereal extract gives too high results, from the solubility of other acids in the ether. On the other hand, on substituting benzene for ether too low results are obtained from the volatilisation of salicylic acid during evaporation. The author has modified the colorimetric method proposed by R6mont : a set of eight assay-tubes are prepared containing quantities varying from 1 C.C. to 0.05 C.C. of solution of salicylic acid (1 : lOOO), and the volume filled up with water to 10 C.C. ; to these are ridded 3 to 1 drops of a dilute solution of ferric chloride (sp. gr. 1.005--1.010).In order to estimate the quantity of salicylic acid in wine,for example, 100 C.C. of it are shaken up with 100 C.C. ether and a few drops of sulphuric acid ; the supernatant ether is drawn off and rapidly distilled. A quantity of soda solution is theradded, more than capable of saturating the salicylic acid likely to be present, and the whole is evaporated to dryness on a water-bath to drive off the excess of acetic acid extracted by the ether. The residue obtained is rendered acid by sulphuric acid, 20 C.C. of benzene added, and the solution filtered ; 10 C.C. of the solution is then drop2ed into an assay-tube, mixed with 10 C.C. water, C. H. B.1176 ABSTRACTS OF CHEMICAL PAPERS. and 1 to 2 drops of ferric chloride added ; the mixture is shaken lip, and the tint compared with one of the standard assay-tubes.The author adduces experiments in support of the accuracy of this method. V. H. V. Estimation of Tannin in Tea. By A. HILL (Analyst, 1881, 95-99),-The author uses Lowenthal’s method (Zeits. And. Chew.., 1877, 3, and 201) for the estimation of tannin in tea, and from the analysis of thirty-two samples of black and green tea, it would seem that, as a rule, green tea is richer in tannin than black. The total average amount is 14.8 per cent. The auhhor’s results, although obtained from the undried leaves, agree on the whole with those of Mulder, who worked with dried leaves. L. T. 0’s. Estimation of Quinine. By A. W. BLYTH (Analyst, 1881, 162- 164).-The volumetric estimation of quinine is effected by a standard solution of Mayer’s reagent containing 13.456 grams mercuric chloride and 49.8 grams potassium iodide per litre.The complete precipitation is determined by testing a drop of the clear solution. The drop is extracted by means of a “filter-tube” consisting of a glass tube, widened a t one end into a funnel shape ; the widened end is packed with glass wool, and the t’ube used as a pipette. To estimate quinine in wine, it is first precipitated with Scheibler’s reagent, and the preci- pitate shaken with strong soda solution and ether in a tube of special construction, which cannot be described without the aid of a diagram. L. T. 0’s. Isolation of Strychnine. By A. 11. ALLEN (Analyst, 1881, 141- 142).--The author advocates the use of a mixture of ether and chloro- form in equal volumes for the separation of recently precipitated strychnine from its aqueous solution.This solvent is preferable to ether or chloroform alone, since the former has not sufficient solvent power, and the latter does not separate readily from the aqueous soh- tion. L. T. 0’s. Analyses of Milk. By B. DYER (Analyst, 1881, 59-62).-The analyses of the milk of cows, stall-fed and at grass, show that that of the former is the richer, and that an individual cow well fed can fre- quently give milk yielding an average of 8.7 per cent. solids, not fat. It is therefore necessary t o take into account the percentage of fat as well as that of “ solids not fat” before pronouncing on cases of adul- teration. L. T. 0’s. Analyses of Milk. By C. A. CAMERON (Analyst, 1881, 75-78).- The analysis of the milk of forty-two cows kept at the Government Agricultural Institution, Glasnevin, shows- (a,) That the quantity and quality of the milk improves as the age of fhe cow advances.(b.) That towards the end of the period of lactation, although the quantity of milk becomes less it improves, on the whole, in quality. ( c . ) That the milk given in the evening is as it rule richer than that given in the morning.ANALYTICAL CHEMISTRY. 1177 The results of the analysis of tbe mixed milk from eight of the forty-two cows when it is poorest, i.e., in the morning, are as follow :- Tot’al solids .......... 13.90 per cent. Solids not fats ........ 9-75 ,, Fats .................. 4.15 ,, Ash.. ................ 0.72 ,, The author bears out the statement of Dyer (preceding Abstract), that the stttndard of solids not fat is too high, and proposes that 8.5 should be taken as a minimum ; but in no case should the total solids be less than 11.5.The minimum for fats he considers too low, and would rzise it to 2.75. The author also advocates the use of a second general standard : solids not fats, 9 per cent. ; fats, 3 per cent. L. T. 0’s. Action of Organic Matter, not Sugar, in Cane and Beet Products on Alkaline Copper Oxide Solution. By J. H. TUCKEE (Chew. News, 44,29).-It has been stated that the organic substances in commercial sugars and syrups have a reducing effect on Fehling’s solution, and therefore it is recommended to remove them by means of lead acetate before determining the amount of sugar. The author finds that such organic matters exert so little influence on the result of the analysis, that their removal is not necessary, unless they are present in considerable quantity.c. H. B. Estimation of Gluten in Flour. By B~NARD and J. GIRARDIN (J. Pharni. [5], 4,127-128).-1n estimating the gluten in flour by the ordinary method, it is necessary to allow the paste to stand for a t lea,st three hours before washing out the starch, otherwise a considerable loss of gluten (5-6 per cent.) occurs ; and in all cases where two deter- minations in one sample are made, should the same conditions of time be observed. It is preferable to estimate the gluten after drying it a t 110-120“, L. T. 0’s. Method for Examination of Coffee. (By F. M. RIMNIKGTON (AnaZyst, 1881, 2).-The presence of chicory, dandelion, &c., in coffee, may be ascertained by boiling with sodium carbonate, and then heating with dilute bleaching-powder solution, when the chicory and dandelion are bleached, the coffee being unacted on.L. T. 0’s. Separation of Wool and Silk in Textile Fabrics. By A. R~MOST ( J . Phawn. [5), 4, 135--13S).-To determine the nature of the fabric, a portion is boiled with hydrochloric acid (5 per cent.), washed and dried, the warp and weft are separated and burnt sepa- rately. If an odour of burnt horn is evolved, and ammonia is evolved when boiled with sodium hydroxide, it is plunged into boil- ing basic zinc chloride. A. Solution is complete--.silk.1178 ABSTRACTS OF CHEMICAL PAPERS. B. On addition of hydrochloric acid, copious flocculent precipitate- C.Insoluble in zinc chloride, but on boiling with caustic soda. silk mixed with wool or vegetable fibres. Completely soluble-wool. Partially soluble-wool and cotton. It does not evolve an odour of burnt horn-vegetable fibres. Quantitative estimation, four pieces of the cloth of 2 grams are taken. Estimation of Size and Dye.-Three of t8he pieces are boiled with 200 C.C. hydrochloric acid (3 per cent.) for a quarter of an hour, the solution decanted, the operation repeated, the fragments washed with water and dried. The size is thus removed, and also the colouring matter, if the material is cotton, but in the case of wool and silk the latter is only imperfectly removed. Silk, however, dyed with aniline colours contains such a small quantity of colouring matter that it may be neglected ; but silk dyed black with logwood and iron can be charged indefinitely with colour- ing matter, and then the colouring matter may be roughly estimated by ignition and weighing the ferric oxide.Should the proportion of ferric oxide amount to 5, 8, or 10 per cent,, it is necessary to take account of it. Separation of Silk;.-Two of the fragments are plunged for two minutes in a boiling solution of basic zinc chloride (1.69 sp. gr.), then thrown into water, and washed in acid water, and finally in water until the wash-water gives no precipitate with ammonium sulphide. Separation of WooZ.-One of the pieces, after treatment with zinc chloride, is dried, and boiled gently with 60-80 C.C. soda solution (1.02 sp.gr.) for a quarter of an hour, care being taken not to destroy the vegetable fibres. The four pieces of cloth are then dried for an hour a t 109" and then left until the next day in the same atmosphere as the original specimen is kept, when they are weighed. The piece of cloth which has not been treated should weigh 2 grams ; but all differences above 5 mgrms. must be taken account of in estimating the proportions of size and colouring matter, silk, wool, and vegetable matter, which are obtained by diaerence from the various woights. L. T. 0's.1.168 ABSTRACTS OF CHEMICAL PAPERS.A n a l y t i c a l Chemistry.A Solution of Density 3-28 suitable for the Analysis ofMinerals. By M. D. KLI:IN (Compt. rend., 9 3 , 3 1 8 4 2 0 ) .-Inasmuchas the densitly of most minerals ranges from 2.2 to 3.3, they can beeasily separated from one another by immersion in solutions withoutchemical action on them.Solutions of mercuric iodide in potassiumiodide have partly fulfilled the necessary conditions, but the minuteprecautions required have prevented the general introduction of thismethod. Nearly saturated solutions of the cobalt, nickel, and cad-mium borotungstates prepared by the author have sp. gr. about3.28-3.31. The dark tint of the two former salts renders theirsolution comparatively useless for the purpose, but the cadmium saltsolution being of a very pale yellow colour is very suitable. Theauthor illustrates how, for instance, by this solution a diamond can bedistinguished immediately from white sapphire or colourless corundum.In order to prepare this solution, tetrasodic horotunstate (this vol.,224) is converted into barium borodecitungstate (Abstr., 1880, 612),and then decomposed by cadmium sulphate.The solution is eva-pornt,ed on a salt-water bath until a fragment of peridote swims on thesurface ; then on cooling orthorhombic crystals of the cadmium saltare deposited. The crystals melt at 75' in their water of crystallisationand retain their mother-liquor, and there is formed a mobile liquid ofthe required density, 3.6.Estimation of Phosphoric Acid as Magnesium Pyrophos-phate. By F. A. GOOCH (An?. Chent. J., 1, 391-413).-This papergives the results of a large number of analyses of alkaline phosphates,phosphotungstates, and phosphomolybdates, from which the followingconclusions are drawn :-V.H. VANALYTTCSL CHEMISTRY. 1169In determining the phosphoric oxide of aZkaZiine phosphates freefrom sulphates or other substances likely to contaminate a precipitate,accuracy is most conveniently and surely attained by adding to thesomewhat dilute solution of the phospha.te, first ammonia in slight butdistinct excess, then an excess of magnesium chloride mixture con-taining free ammonia (made by dissolving 3 parts crystallised magne-sium chloride and 8 parts ammonium chloride in water, adding watercontaining ammonia till the volume of the solution amounts to 48parts, then filtering and boiling off the free ammonia,), and after theprecipitate has settled, adding a further excess of ammonia till theliquid smells strongly of it.In determining the phosphoric oxide of phosphotungstates, the bestresults are obtained by adding to the solution containing free ammo-nia, an excess of either of the magnesia mixtures, washing the pre-cipitate with the precipitant, dissolving in hydrochloric acid, dilutingif necessary, and reprecipitating with a little dilute ammonia, addingstrong ammonia after the precipitate has settled ; or, when workingwith small amounts, by proceeding as in the case of the phosphates,taking however special care that the solution is distinctly ammoniacalbefore precipitating, and that no great excess of the precipitant isused.For the determination of phosphoric oxide in phosphomolyb-dates, the method recommended in the case of the phosphates servesvery well when the amount, of phosphoric oxide does not exceed about0.05 gram, but for larger amounts, the mebhod of double precipitationrecommended for the phosphotungstates is more accurate.With regard to the time which should be allowed t o pass betweenprecipitating and filtering, the author’s experiments support those ofAbesser, Jnni, and Maercker (this Journal, 1874, 387) in pointing tothe condition that a precipitate map safely be filtered as soon as it hascompletely subsided, generally after ten or fifteen minutes.By D.LTNDO(Chem. News, 44, 77, 86, 97, 129).-A moderate excess of platinumchloride solution, containing about 1 gram of platinum in ‘LO c.c., isadded to the somewhat dilute solution of alkaline chlorides, theliquid evaporated to dryness on the water-bath, and t,hen heated fortwo or three minutes.The residue is washed three or four timeswith alcohol of 98 per cent., then carefully crushed with a glass rod,and again washed repeatedly with alcohol, Gooch’s method of filtra-tion (Chem. News, 37, 181) being employed. The platinochloridemas dried a t 130”. The results obtained were very accurate, even inpresence of 50 per cent. sodium chloride. The double salt is nothygrcscopic, and there is no need to place the crucible in a desiccatorwhilst cooling. When using Tatlock’s method, the author finds thatthe platinum tetrachloride solution employed for the first washingdissolves a small quantity of the double salt, and consequently theresults obtained are somewhat too low.Even in presence of excessof platinum chloride, it would appear that sodium chloride is notentirely converted into the double salt or, more probably, that rever-sion takes place either during evaporation or on adding the alcohol.Pi*olonged heating of the residue on the water-bath promotes rever-H. W.Estimation of Potassium as Platinochloride1170 ABSTRACTS OF CHEMICAL PAPERS.sion, which may be prevented by employing a moderate excess ofplatinum chloride, and taking care not to heat t,he platinochlorides toolong after they have become dry. Direct experiments showed thatpotassium platinochloride is practically insoluble in absolute or verystrong alcohol; when the platinochloride is in the form of scales,alcohol of 80 per cent.dissolves a small quantity ; when in the formof powder, alcohol of 83 per cent. dissolves a quantity sufficient tosensibly affect the results. The author finds that potassium platino-chloride is very slightly soluble in a half saturated solution of ammo-nium chloride, and proposes to use a solution of the latter to washout any magnesium and sodium sulphates, &c., with which theplatinochloride may be mixed. The double salt is first thoroughlywashed with alcohol, then with an aqueous solution of ammoniumchloride, and finally with alcohol. If accurate results are desired, theammoniui u chloride solution should be saturated wit'h platinochloridebefore being used. Taking the atomic weight, of platinum as 197.18,the mean of the determinations of potassium chloride was 100 percent.; taking Seubert's number 194.46, the mean was 100.55 percent.C. H. B.Electrolytic Estimation of Zinc. By H. REINHARDT and R.IHLE (J. pr. Chem. [2], 24, 193--201).-1t is found that the usualprocesses for the electrolytic estimation of zinc are all to a certainextent unsatisfactory. The presence of fvee sulphuric acid and am-monium acetate are especially prejudicial t'o the results. The follow-ing modification is therefore proposed : the anode consists of a spiralof strong platinum wire, and the cathode of a cone of platinum sus-pended above the anode, the distance between the two being 5 mm.,and the surface of the latter 160 sq. cm. The object of the peculiarform of cathode is to ascertain, by lowering it into the liquid, if thelast trace of metal has been removed.The negative electrode is to bepreviously coppered, and nitric acid may be employed for removalof the deposited zinc, as the copper is not attacked in the cold. Thecopper coating must be even and clear, otherwise the zinc will bedeposited in numerous small loosely attached grains. The zinc isdeposited from a solution of the oxalate, and for this purpose theneutral chloride or sulphate solutionis mixed with an excess of potas-sium oxalate until the precipitate produced id redissolved ; the metalis precipitated from this solution a t the negative pole, and may beeasily xashed with water. This process works well with alloys.E. W. P.Reduction of Iron Ores by Powdered Zinc.By T. M. BROWN(Chem. Centr., 1881, 75).-An intimate mixtme of 0.3 gram powderedore with 10 times as much finely pulverised zinc is placed in a porce-lain crucible and covered with an equal weight of powdered zinc.The open crucible is heated for ten minutes over a Bunsen burner todark redness. On cooling, the whole contents of the crucible arerapidly dissolved in hydrochloric acid out of contact with air, and theiron is determined by potassium permanganate in the usual manner.Bituminous ores must, be roasted before reduction. To prevent theintroduction of organic matter from the lamp, a layer of borax iANALYTICAL CHEMXSTRY. 1171placed above the layer of zinc. Powdered zinc is also found to haveadvantages over granulated zinc and sheet zinc in the reduction offerric to ferrous solutions.P. L. T.Separation of Cobalt and Nickel from Iron. By T. MOO-~E(Chem. News, 44, 76).-The solution, which should not contain muchfree acid, is mixed with ammonium sulphate in quantity sufficient toform a double sulphate with the cobalt andnickel present. The solu-tion is then diluted to about 150 c.c., mixed with a large excess ofoxalic acid and well stirred. If a precipitate is formed, it must beredissolved by further addition of ammonium sulphate. The clearliquid is mixed with ammonia in slight excess, heated gently for a fewminutes, filtered, and washed with water containing ammonia, or theliquid is diluted to a known volume and, after the precipitate hassettled, a definite quantity of the clear liquid is drawn off and thenickel determined by electrolysis or otherwise.Separation of Tungsten from Antimony, Arsenic, and Iron,together with the Analysis of a so-called Pseudometeorite.ByA. COBENZL (Monatsh. Chem., 1881, 259-265) .-The metallic masswhich is the subject of this examination, was found in 1879 at Cistain Bohemia. Externally it exhibits some of the characters ofmeteoric iron, but as its fall was not actually observed, and as itsphysical and chemical characters were found on closer examination todiffer from those of well authenticated iron meteorites, it has beendesignated a pseudometeorite. It is brittle, easily pulverisable, andhas a fine-grained confusedly crystalline texture, revealed by the shim-mering aspect of its newly fractured surface, whereas meteoric ironsare tough and have a hackly fracture. Moreover, the pseudometeoriteis porous and has a blistered surface, the character of which is morelike that of certain furnace-products than that of meteoric iron.Thecolour of this metal is violet-grey, differing therefore from the neutralgrey of meteoric iron.Qualitative analysis showed that the chief constituents of thepseudometeorite are iron, tungsten, antimony, and arsenic, associatedwith small quantities of water and insoluble silicates (ferric oxide,alumina, lime, magnesia), carbon, and sulphur ; traces of bismuth andtin were also detected.The quantitative analysis presented pecnliar difficulties in theseparation of the first four constituents above mentioned, respectingwhich no exact statements are to be found in chemical literature.The following method was found by the author to yield satisfactoryresults. The mineral, very finely pulverised and sifted through linen,was drenched in a flask with strong nitric acid, and heated on awater-bath €or four or five days, with occasional addition of hydro-chloric acid, till nothing was left undissolved but pure yellow tungsticacid.The solution, together with the tungstic acid, was evaporatedt o dryness over a water-bath ; the dry dusty residue redissolved in verydilute nitric acid ; the solution again evaporated to dryness, and thistreatment three times repeated. The residue was then dissolved invery dilute nitric acid, with addition of a small quantity of tartaricC. H.B.VOL. XL. 4 1172 ABSTRACTS OF CHEMICAL PAPERS.acid, the solution evaporated a t 100" and filtered, and the separatedtungstic acid several times washed by decantation with boiling acidu-lated water, and finally collected on a filter.The solid matter thus separated contained the whole of the tungs-ten, together with silica, and silicates undecomposible by acids, whilstall the other metals, together with lime and alumina, were found inthe filtrate. The tungstic acid was treated on the filter with verydilute ammonia, which dissolved it, leaving behind the silica and sili-cates. The ammoniacal solution was evaporated t o dryness in a weighedporcelain crucible, the residue ignited, and the pure straw-yellowturigstic acid thus obtained was weighed.The undissolved silicateswere dried, ignited, weighed, and analysed in the usual way bydecomposition with a mixture of potassium and sodium carbonate.The filtrate from the tungstic acid was saturated with yellow ammo-nium sulphide, whereby iron sulphide and alumina were precipitated,while arsenic, antimony, and lime remained dissolved. All these con-stituents were separated and estimated by the usual methods.The sulphur, after separation of the tungsten, was weighed asbarium sulphate. The carbon was estimated, together with the water,by igniting the powder in a stream of oxygen, and collecting thewater and carbonic acid in the usual way. The amount of oxygenwas determined by reducing the oxides present in a stream ofhydrogen.The results of the analysis were as follows :-Fe. w. Sb. As. Silicates. H20. Al,O,.56.07 25-39 9.85 5.08 1-55 0.78 0.60CaO. 0. C. S.0-37 0.28 0.18 0.053The percentage composition of the silicates was found to be :SiO,. Al,O,. Fe20,. CaO.41.2 26.8 18.0 14.0From the physical and chemical characters above described, it maybe inferred almost with certainty that this substance is not an actualiron meteorite, but a furnace-product, such as were often obtained inthe early days of the metallurgic preparation of tungsten. The fol-lowing table exhibits a comparison of its chief constituents, and itsspecific gravity, with those of iron meteorites :-Mean of iron meteorites. Pseudometeorite.Iron.. Iron........ 5 6.4Nickel ...... 3 1 7 Tungsten . . 25.3Phosphorus.. 0 1 Antimony .. 9.9Cobalt.. .... 0 2.6 Arsenic .... 0.1...... 81 to 98 p. c.Sp. gr.. ..... 6.6 7.9 I n the lump. 8.854In powder .. 8.8993H. w.Action of Potassium Permanganate on Potable Waters atDifferent Temperatures. By G. W. WIGNER and R. H. HARLAND(AnaZyst, 1881, 39).-The experiments of the authors show that thA N AL Y TICAL CHEMISTRY. 1173amount of oxygen absorbed by potable waters when mixed withpotassium permanganate varies considerably with the nature of thewater and with the time, and in some cases with the temperature.The nature of the pollution also has a marked effect on the oxygenabsorbed, but no general conclusions can be drawn from their results.L. T.0’s.Estimation of Nitrates in Potable Waters. By J. WEST-KNIGHTS (AmZyst, 1881, 56--58).-The author applies the brucinetest for nitric acid to the estimation of nitrates in potable waters,the blood-red colour being permanent if oxalic acid is used instead ofsulphuric acid, in applying the test to nitrates. For the purpose arerequired a potassium nitrate solution containing 0.721 gram per litre(1 C.C. = 0.0001 gram N as NO,), a brucine solution (1 gram in 100alcohol), a cold saturated solution of oxalic acid, and a standard xed SOZZL-tioiz prepared by evaporating 10 C.C. of the potassium nitrate solutionto dryness, and adding 3 C.C. of the brucine solution and 6 drops ofoxalic acid solution, evaporating to dryness, dissolving the residue inwater, and again evaporating; the residue is dissolved in water andmade up to 100 C.C.This solution should have a bright-red colour,and 1 C.C. of it is equal to 0.00001 nitrogen as nitrate.10 e.c.are evaporated to dryness, and from 0.5 to 2 C.C. of brucine solutionadded to the residue. The brucine should be just in excess, and thecolour a bright red; if it be brown, a fresh quantity of water mudbe taken, and a larger quantity of brucine added ; but if the colour ispink, a smaller quantity of brucine must be used. Three or fourdrops of oxalic acid are added to the residue, which is treated as inthe case of the standard solution, except t h t fhe residue after thefinal evaporation is dissolved in a small quantihy of water, filtered intoa glass cylinder, and the volume made up to 50 c.c., and the colourcompared with from 1 to 10 C.C.of the standard colour. If the colourproduced by the water is deeper than that of 10 C.C. of the standard,i t must be diluted with two or three volumes of water, and 50 C.C.treated as before ; but if it he lighter than that of 1 c.c., a larger quan-tity of water must be used to begin with.By F. P. PERKLNS(Analyst, 1881,58-59) .-The author applies a magnesium-platinumcouple in presence of sodium chloride to the reduction of nitrates inriver water. L. T. 0’s.The water to be examined is prepared in a similar manner.L. T. 0’s.Estimation of Nitrates in River Water.Detection of Lead in Potable Waters. By S. HARVEY (Ana7yst,1881, 146-148) .-Potassium dichromate is a more delicate test thansulphuric acid or potassium iodide for the presence of lead inwater.It is also preferable to sulphuretted hydrogen, since in theevent of copper or tin being present in the water, there is a difficultyin distinguishing between the precipitates. It is possible to detectone-fiftieth of a grain of lead per gallon by this means, even in pre-sence of sulphates. The potassium dichromate is added in the solidstate. L. T. 0’s.4 J; 1174 ABSTRACTS OF CHEMICAL PAPERS.Assay of Wood Spirit for the Preparation of MethylatedSpirit. By C. BARDY ( J . Pharna. [5], 4, 129--135).-The authordescribes a method for the estimation of methyl alcohol in woodspirit, which consists in the comparison of the amount of iodoformobtained from the sample with that from a standard wood spirit.For the purpose, a ,standard solution of caustic soda is prepared,and a solution of 254 grams iodine and 385 potassium iodide in 1 litreof water.5 C.C.of the wood spirit t o be assayed is diluted to 50 C.C. withwater, and 10 C.C. of this solution is made up to 1 litre with water,and 5 C.C. of this is placed in a tube with 10 C.C. soda solutionand -5 C.C. of the iodine solution. After well shaking, 30 C.C.of water are added, and the milkiness due to the formation of iodo-form is compared with that in a similar solution prepared from thestandard wood spirit. The comparison is effected by placing behindthe tubes a white card, marked with a number of black lines ofvarious thicknesses, and noting the number of lines visible in eachcase.The comparison should he made immediately, since the iodo-form settles after the lapse of a short time. Should the milkiness beless than that produced by the test solution, the wood spirit is toorich in methyl alcohol, and cannot be used.If the milkiness is alike in both cases, or greater in the case of thesample than in that of the test, it is necessary to estimate the methylalcohol by Bardet and Bordet's method.This method cannot be used if the wood spirit contains a large pro-portion of aldehyde, which is detected by diluting the wood spirit with10 of water, and adding to 5 C.C. of the diluted spirit 2 C.C. of aqueoussolution of magenta (0.20 grain per litre), decolorised by sodiumthiosulphate, when a violet colour is produced, varying in intensitywith the quantity of aldehyde present. The aldehyde is estimated bydistilling 5 C.C.of the wood spirit, with 7 C.C. of sulphuric acid,diluted with 10 c,c. of water. The value of the spirit may then beestimated as above.In assaying wood spirit, i t is necessary that it should contain noethyl alcohol, accidentally or fraudulently added, and no methylethers. The latter are estimated by saponification. The estimationof alcohol, mixed with wood spirit, is effected by diliiting 10 C.C. to1 litre with water, and proceeding as above.I f essences are present, 11 C.C. are diluted, and the solution filteredthrough a moist filter-paper. To estimate varnishes prepared withmet'hylated spirit, 11 C.C.are distilled to dryness, the distillate dilutedto a litre, and the above method applied.By J. MUTER (Analyst,1881, 41-46) .-The power of glycerol to prevent the precipitationof copper by potash from its solutions is made use of for the esti-mation of glycerol. 1 gram glycerol is washed into a stopper-graduated tube with a stopcock 50 C.C. from the bottom (Muter oleintube), and 50 C.C. of potassium hydroxide (1-2) added. A dilutecopper sulphate solution is then carefully run in, with constantshaking, until a fair amount of copper hydroxide remains undissolved ;L. T. 0's.Volumetric Estimation of GlycerolANALYTICAL CHEMISTRY. 1175the whole made up to a given volume and allowed to settle. Whenclear, a given volume is made slightly acid with nitric acid, a definiteexcess of ammonia added, and finally titrated with potassium cyanidesolution, of which 1 C.C.= 1 gram glycerol. By using pure potassiumcyanide made from the acid, good results may be obtained ; there is aslight deficiency, however, owing to the difficulty of working thecyanide estimation of copper to a constant point. I n cases where theamount of glycerol to be estimated is very small, the copper is deter-mined by electrolysis on platinum. L. T. 0’s.Detection of Hydrocyanic Acid. By E. LUDWIG and 5.MAUSHNER (Chem. Centr., 1881, 43-44) .-In a poisoning case sub-mitted to the authors, potassium ferrocyanide was shown to be pre-sent, and was removed by slightly acidulating and carefully preci-pitating by ferric chloride.On submitting the filtrate to distillationwith tartaric acid, a distillate was obtained both from the contents ofthe stomach and from the residue in the bottle containing muchhydrocyanic acid.The sample of potassium cyanide which is supposed was employedwas subsequently found to contain a very considerable quantity offerrocyanide. 3’. L. T.New Demonstration of Carbonic Anhydride in the Breath.By C. F. CROSS (Chem. News, 44, 141).-When respired air is blowninto a solution containing a mixture of potassium iodide and iodateand starch, the carbonic acid causes the liberation of iodine, and con-sequent production of it blue coloration. Direct experiments provedthat the liberation of iodine is not due to the action of the oxygen ofthe air, or to the accidental presence of acid vapours.Determination of Salicylic Acid in Food-stuffs by a Colori-metric Reaction.By H. PELLET and J. DE GROBERT (Compt. rend.,93, 278--280).-The method of determining salicylic acid by thetitration of the residue obtained by an ethereal extract gives too highresults, from the solubility of other acids in the ether. On the otherhand, on substituting benzene for ether too low results are obtainedfrom the volatilisation of salicylic acid during evaporation. Theauthor has modified the colorimetric method proposed by R6mont : aset of eight assay-tubes are prepared containing quantities varyingfrom 1 C.C. to 0.05 C.C. of solution of salicylic acid (1 : lOOO), and thevolume filled up with water to 10 C.C.; to these are ridded 3 to 1 dropsof a dilute solution of ferric chloride (sp. gr. 1.005--1.010). In orderto estimate the quantity of salicylic acid in wine,for example, 100 C.C. ofit are shaken up with 100 C.C. ether and a few drops of sulphuric acid ;the supernatant ether is drawn off and rapidly distilled. A quantityof soda solution is theradded, more than capable of saturating thesalicylic acid likely to be present, and the whole is evaporated todryness on a water-bath to drive off the excess of acetic acid extractedby the ether. The residue obtained is rendered acid by sulphuricacid, 20 C.C. of benzene added, and the solution filtered ; 10 C.C. of thesolution is then drop2ed into an assay-tube, mixed with 10 C.C.water,C. H. B1176 ABSTRACTS OF CHEMICAL PAPERS.and 1 to 2 drops of ferric chloride added ; the mixture is shaken lip,and the tint compared with one of the standard assay-tubes. Theauthor adduces experiments in support of the accuracy of thismethod. V. H. V.Estimation of Tannin in Tea. By A. HILL (Analyst, 1881,95-99),-The author uses Lowenthal’s method (Zeits. And. Chew..,1877, 3, and 201) for the estimation of tannin in tea, and from theanalysis of thirty-two samples of black and green tea, it would seemthat, as a rule, green tea is richer in tannin than black. The totalaverage amount is 14.8 per cent. The auhhor’s results, althoughobtained from the undried leaves, agree on the whole with those ofMulder, who worked with dried leaves.L. T. 0’s.Estimation of Quinine. By A. W. BLYTH (Analyst, 1881, 162-164).-The volumetric estimation of quinine is effected by a standardsolution of Mayer’s reagent containing 13.456 grams mercuric chlorideand 49.8 grams potassium iodide per litre. The complete precipitationis determined by testing a drop of the clear solution. The drop isextracted by means of a “filter-tube” consisting of a glass tube,widened a t one end into a funnel shape ; the widened end is packedwith glass wool, and the t’ube used as a pipette. To estimate quininein wine, it is first precipitated with Scheibler’s reagent, and the preci-pitate shaken with strong soda solution and ether in a tube ofspecial construction, which cannot be described without the aid of adiagram.L. T. 0’s.Isolation of Strychnine. By A. 11. ALLEN (Analyst, 1881, 141-142).--The author advocates the use of a mixture of ether and chloro-form in equal volumes for the separation of recently precipitatedstrychnine from its aqueous solution. This solvent is preferable toether or chloroform alone, since the former has not sufficient solventpower, and the latter does not separate readily from the aqueous soh-tion. L. T. 0’s.Analyses of Milk. By B. DYER (Analyst, 1881, 59-62).-Theanalyses of the milk of cows, stall-fed and at grass, show that that ofthe former is the richer, and that an individual cow well fed can fre-quently give milk yielding an average of 8.7 per cent. solids, not fat.It is therefore necessary t o take into account the percentage of fat aswell as that of “ solids not fat” before pronouncing on cases of adul-teration. L.T. 0’s.Analyses of Milk. By C. A. CAMERON (Analyst, 1881, 75-78).-The analysis of the milk of forty-two cows kept at the GovernmentAgricultural Institution, Glasnevin, shows-(a,) That the quantity and quality of the milk improves as the ageof fhe cow advances.(b.) That towards the end of the period of lactation, although thequantity of milk becomes less it improves, on the whole, in quality.( c . ) That the milk given in the evening is as it rule richer than thatgiven in the morningANALYTICAL CHEMISTRY. 1177The results of the analysis of tbe mixed milk from eight ofthe forty-two cows when it is poorest, i.e., in the morning, are asfollow :-Tot’al solids ..........13.90 per cent.Solids not fats ........ 9-75 ,,Fats .................. 4.15 ,,Ash.. ................ 0.72 ,,The author bears out the statement of Dyer (preceding Abstract),that the stttndard of solids not fat is too high, and proposes that 8.5should be taken as a minimum ; but in no case should the total solidsbe less than 11.5. The minimum for fats he considers too low, andwould rzise it to 2.75.The author also advocates the use of a second general standard :solids not fats, 9 per cent. ; fats, 3 per cent. L. T. 0’s.Action of Organic Matter, not Sugar, in Cane and BeetProducts on Alkaline Copper Oxide Solution. By J. H. TUCKEE(Chew. News, 44,29).-It has been stated that the organic substancesin commercial sugars and syrups have a reducing effect on Fehling’ssolution, and therefore it is recommended to remove them by means oflead acetate before determining the amount of sugar. The author findsthat such organic matters exert so little influence on the result of theanalysis, that their removal is not necessary, unless they are present inconsiderable quantity.c. H. B.Estimation of Gluten in Flour. By B~NARD and J. GIRARDIN(J. Pharni. [5], 4,127-128).-1n estimating the gluten in flour by theordinary method, it is necessary to allow the paste to stand for a t lea,stthree hours before washing out the starch, otherwise a considerableloss of gluten (5-6 per cent.) occurs ; and in all cases where two deter-minations in one sample are made, should the same conditions of timebe observed.It is preferable to estimate the gluten after drying it a t 110-120“,L. T.0’s.Method for Examination of Coffee. (By F. M. RIMNIKGTON(AnaZyst, 1881, 2).-The presence of chicory, dandelion, &c., in coffee,may be ascertained by boiling with sodium carbonate, and then heatingwith dilute bleaching-powder solution, when the chicory and dandelionare bleached, the coffee being unacted on. L. T. 0’s.Separation of Wool and Silk in Textile Fabrics. By A.R~MOST ( J . Phawn. [5), 4, 135--13S).-To determine the nature ofthe fabric, a portion is boiled with hydrochloric acid (5 per cent.),washed and dried, the warp and weft are separated and burnt sepa-rately. If an odour of burnt horn is evolved, and ammonia isevolved when boiled with sodium hydroxide, it is plunged into boil-ing basic zinc chloride.A. Solution is complete--.silk1178 ABSTRACTS OF CHEMICAL PAPERS.B. On addition of hydrochloric acid, copious flocculent precipitate-C. Insoluble in zinc chloride, but on boiling with caustic soda.silk mixed with wool or vegetable fibres.Completely soluble-wool.Partially soluble-wool and cotton.It does not evolve an odour of burnt horn-vegetable fibres.Quantitative estimation, four pieces of the cloth of 2 grams aretaken.Estimation of Size and Dye.-Three of t8he pieces are boiled with200 C.C. hydrochloric acid (3 per cent.) for a quarter of an hour, thesolution decanted, the operation repeated, the fragments washed withwater and dried. The size is thus removed, and also the colouringmatter, if the material is cotton, but in the case of wool and silk thelatter is only imperfectly removed.Silk, however, dyed with aniline colours contains such a smallquantity of colouring matter that it may be neglected ; but silk dyedblack with logwood and iron can be charged indefinitely with colour-ing matter, and then the colouring matter may be roughly estimatedby ignition and weighing the ferric oxide. Should the proportion offerric oxide amount to 5, 8, or 10 per cent,, it is necessary to takeaccount of it.Separation of Silk;.-Two of the fragments are plunged for twominutes in a boiling solution of basic zinc chloride (1.69 sp. gr.), thenthrown into water, and washed in acid water, and finally in wateruntil the wash-water gives no precipitate with ammonium sulphide.Separation of WooZ.-One of the pieces, after treatment with zincchloride, is dried, and boiled gently with 60-80 C.C. soda solution(1.02 sp. gr.) for a quarter of an hour, care being taken not to destroythe vegetable fibres.The four pieces of cloth are then dried for an hour a t 109" and thenleft until the next day in the same atmosphere as the original specimenis kept, when they are weighed. The piece of cloth which has notbeen treated should weigh 2 grams ; but all differences above 5 mgrms.must be taken account of in estimating the proportions of size andcolouring matter, silk, wool, and vegetable matter, which are obtainedby diaerence from the various woights. L. T. 0's