78 ABSTRACTS OF CHEMICAL PAPERS. An a 1 y t i c a 1 C h em i s t ry. Kjeldahl's Method of Estimating Nitrogen. By C. ARNOLD (Arch. Pharrn. [3], 24, 785--794).-The author shows that this method (Abstr., 1884, 364; 1885, 930) is also applicable to nitrates, nitro- and cyanogen-compounds, in the presence of beneoic acid, sugar, mercury, and copper sulphate ; further, in the case of alkaloids and otber compounds that can be treated by this method, oxidation by potassium permanganate is unnecessary if the boiling be continued until the warm liquid acquires a bluish-green tint, becoming colour- less on cooling ; this takes place tolerably quickly in the presence of mercury and copper. That the oxidation has been completed can readily be ascertained by adding a crystal of potassium permanganate, which should give a persistent violet or green coloration. A descrip- tion of tlie apparatus used by the author is given.J. 1'. Separation of Arsenic and Antimony. By ZAMBELLI and LUZZATO ( A ~ c h . Pharm. [ 3 ] , 24, 772).-These elements can be sepa- mted, even in foyensic cases, by treating the still moist sulphides with hydrogen peroxide at 40' for some hours, then heating to near 100"AKALYTICAL CHEMISTRY. 79 and filtering. The arsenic acid formed goes into the filtrate, whilst the insoluble antimony oxide remains on the filter. The method is said to be very delicate. Detection of Thiosulphate in Hydrogen Sodium Carbo- nate. By RRENSTEIN and T. SALZER (Arch. Yharm. [3], 24, 761).- According to Brenstein, the reaction given for thiosulphate by Mylius is not characteristic, as other oxygen-compounds of sulphur, such as sulphites, give rise to the formation of hydrogen sulphide.A bettar test for tJiiosu1phate is to add to a 5 per cent. solution of hydrogen sodium carbonate a few drops of silver nitrate solution, then excess of nitric acid, and heat to boiling; even minute traces of thiosulphate give an immediate dark precipitate of silver sulphide. According to Salzer, the absence of thiosulphate is easily ascer- tained by adding a few drops of iodine solution to about 20 C.C. of a saturated solution of hydrogen sodium carbonate ; the solution must have a yellowish tint. Decolorisation of the iodine solution does not necessarily imply the presence of thiosulphate, since normal carbo- nate, the most commonly occurring impurity, produces this effect.Both authors found thiosulphate to be a constant impurity of ordinary qualities of hydrogen sodium carbonate, whilst the better qualities were mostly free from it. Estimation of Small Quantities of Silver in Burnt Pyrites. By E. Trrrr,o (Cherrz. Zeit., 10, 822; 1065--1067).-The amount of silver in burnt pyrites is ordinarily 0*003--0.008 per cent., and to obtain trustworthy results material containing 0.01 gram of silver should be taken for analysis, or about 300 to 500 prams of pyrites. Many difficulties have been encountered in endeavouring to deal with these large quantities of material so as to get all the silver. Experiments show that the decomposition of the whole mass is out of the question, therefore methods for extracting the silver, &c., were tried ; treat- ment with nitric acid or with gaseous chlorine, and subsequent extraction of the silver chloride proved unsuitable. Digestion with chlorine-water was somewhat better ; but bromine or bromine-water gives the best results.The powdered burnt pyrites is passed through a 0.25-mm. mesh sieve, and about 300-500 grams of it is placed i n an acid-proof iron basin; it is well covered with water, and while vigorously stirred, bromine is added until present in excess. After 24 hours it is heated on a water-bath for one hour with frequent stirring, treated with excess of ammonia, then with 500 grams of ammonium chloride, and 1 litre of water. The mass is boiled for an hour, and filtered. The residue is dried, ignited at a low red heat, and again treated with arnmoniacal ammonium chloride, by which means all the silver is extracted.Inasmuch as copper is quite as difficult as silver to extract by means of bromine, and as theamount of copper i n the pyrites is reduced from 5 per cent. to a trace (0.01 per cent.) by the above treatment, this may be used as an indicator of the progress of the desilvering, and in practice when the copper is reduced to this amount it may be safely inferred that all the silver is removed. The solu- tion, about 4 litres, is made acid with hjdrochloric acid; it is not J. T. J. T.80 ABSTRACTS Ol!’ CHEMICAL PAPERS. affected by daylight either before or after acidifying. The silver, copper, and lead are then precipitated by means of chemically piire zinc.When the precipitation is complete, the solution is colourless (or rose-red if cobalt is present), and is not turned blue bv ammonia. As the precipitate contains, most liksly, silver chloride and bromide, and some iron oxide, it is fused with potassium cpnide, and sub- sequently with anhydrous borax, the temperature being raised to melt the copper, The regulus is dissolved in nitric acid, and the silver either precipitated by means of hydrochloric acid or determined electrolytically. Silver determinations may also be made in ordinary pyrites, which is first roasted carefnllp, and then treated in the above manner. Good results have been obtained by this method, with roasted copper schist, with spathic iron ore containing argentiferous tetrahedrite and pyrites, and with other ores.A determination can be complded iu three days. Comparative Gasometric Assaying with Especial Reference to the Valuation of Zinc Powder and the Testing of Car- bonates. By J. BARNES ( J . SOC. Chem. Ind., 5, 145--147).-1n a, previous communication (Trans., 1881, 462), the author described a method for the valuation of zinc powder, consisting in measuring the amount of hydrogen liberated on treatment with an acid. The author has since devised a method and constructed an apparatus for gasome tric assaying, contrived originally for the valuation of zinc powder, but applicable also to the estimation of other substances, which may be made to cause or control tlie evolution of gases. The method, which is described in detail in the paper, is a comparative one, and requires a t least two graduated instruments, each provided with a suitable gas evolving arrangement. A substance of known value, taken as the standard, is placed in one, and the sample to be examined in the other instrument.The value of the sample is then calculated by the following formula: - . W . R = x ; - whereV equals the volume of gas from standard, v the volume of gas from mmple, W the weight of standard, and ?u the weight of sample. R is the amount of pure substance or its equivalent in unit of standard, and x the amount of pure substance in unit of sample. Determination and Valuation of Copper in Ores and Products for Commercial Purposes, with some Remarks on the Assay of Gold in Bar Copper. By J. TV. WESTMORKLAND ( J . Soc.Chem. Ind., 5, 48--64).-Esfin?ation of Moisture.-This, although apparently a simple process, is conducted very differently by rarious assr,yers, and the author proposes that the temperature and niaiiner in which the moistnre is to be taken should be more clearly defined than it is at the present time. Wet Assay of Coppr.-Having examined a iarge number of samples of all kinds of copper ores aud products by several wet processes, the author has given the iodide process as modified by Brown the preference-being more trustworthy arid accurate than the electrolytio process. For cupreous pyrites, burnt ores, &c., the followiiig mctliod D. A. L. v 20 D. B.ANALYTICAL CEEMISTRY. 81 mny be used. From 50 to 150 grains of the sample is dissolved in acids, evaporated to dryness with excess of sulphuric acid, diluted with water arid filtered, From this solution, the copper is separated with sodium t'hiosulphate, the precipitate dried, ignited, dissolved in nitric acid, evaporated with sulphuric acid to separate Graces of lead, diluted with water and filtered, sodium carbonate added in excess, and then acetic acid to acid reaction. The solution is then titrated with potassium iodide.As 8n alternative method, the ore may be calciued, dissolved in hydrochloric acid, the ferric salt reduced by boiling with a solution of sodium sulphite, and hydrogen sulphide passed through the cold Rolution; the precipitated sulphides are then washed, and the process conducted as before. Richer copper ores, mattes, preci- pitates, &c., are dissolved in acids, and the solution precipitated either with sodium thiosulphate or hydrogen sulphide, the sulpbides being dissolved in acids, and treated 8s before.The accuracy of this process was tested by numerous experiments, in which known weights of pure electrotype copper, silver, arsenic, antimony, lead, cadmium, bismuth, tin, manganese, zinc, and iron salts were used. The solutions were made to represent cnpreons burnt ores, mattes, cupriferous lead regul us, second quality and rich precipitates, &c. Experiments were also made i n many cases with the metals separately. The electro- deposition method may be employed in cases where it is desir,thle to have resultti of two separate and distinct processes, it being noted that with pure solutions of copper, accurate results are obtained, whilst when silver or bismuth are present they are precipitated, and in one experiment tin was also deposited with the copper.In the presenoe of ferric salts, the deposition is retarded. It was found that whilst the results with refined or bar coppers and rich precipitates agree with those obtained by the iodide process, those obtained by electrolgsing sulphuric acid solutions of mattes, copper ores, $c., are generally slightly below the triie percentage. With burntl ores, cupreous pyrites, and ores containing but little COPPOI', the author first precipitates with hydrogen sulphide, dissolving the s u lphide in nitric acid, and evaporating this soliifion with excess of sulphuric acid, the sulphates dissolved in water give a solution from which copper (and bismuth) is readily deposited.The results by the battery process are generally from 0.01 to 0.04 per cent. higher than by the iodide process, this being cniised by a slight deposit of bismuth. The Cornish Yroce.9~ of Dry Assay.-This process is considered to be inaccurate and misleading, and is also liable to aerious variation even in the hands of experienced operators. The author is of opinion that in cases where the sulphur contents of pyrites are sold to alkali works, tbe burnt ores being returned to the vendors, an accurate wet assay forms the best# check on the copper contents of the ore delivered to and received from the alkali works. In the remaining part of the paper, the methods in vogue for fixing the prices payable for copper in pyrites, copper ores.&c., are discussed. Reftmnce is also made to the avsay of goltl in bar coppry, and the opinion is expressed that t>hese assays are not conducted as carefully aa they should be, taking the value of the metal into oonsideration. I). B. 9 VOL LU.82 ABSTRAOTS OF CREMICAL PAPERS. Estimation of Ammoniacal Nitrogen in Soils, and the Amount of Assimilable Nitrogen in Uncultivated Land. Ry A. BAUMAXN (Lmdw. Versuchs-Stat., 1886, 2$7--303).--The methods employed to estimate the ammonia in soils are nnsatisfactor-y. Schlosing’s method, whereby the ammonitt is freed by means of milk of lime or sodium hydroxide, gives results too high, probably owing to the formation of ammonia from other compounds during the period (2-3 days) required for the process. In Bonssingault’s process as modified by Sclilosing, a hydrochloric acid extract of the soil is treated with magnesia usta, but the ammonia collected in acid mu$ not be estimated by titIration, but must be azotometrically determined ; nor must cork nor caoutchouc be employed in the formation of the apparatus, The liquid in the evolution flask must be boiled for at least an hour, and the magnesia usta must be fresh.With these two processes, it is found that humous soils, when treated with sodium hydroxide, continuously yield ammonia ; if the soil, after treatment with magnesia, be further treated with sodium hydroxide, another supply of ammonia is obtained, b u t only in the case of soils rich in numus. Knop’s process, in which the ammonia is decomposed by brominated sodium hydroxide solution, is vitiated by the fact, noticed by Rnop and others, that in the presence of the soil itself a con- traction of the volume of gas enclosed in the apparatus occurs ; this contraction was stated by some to be due to absorption by the organic matter, whilst others considered the oxides of iron in the clay to be the real cause; also it had been stated that the presence of borax prevented this contraction.To ascertain the cause of the con- tract.ion, a sample of stiff soil poor in humus, was taken from uncul- tivated land a t a depth of 15-20 cm.; this sample seemed free from humus, although root-fibres were visible ; at a distance of 40 m. from this spot another sample was drawn from a depth of 2-10 cm., this was close to a pine SO years old.The sample showed humus, and when dry was grey. At another spot, close to a 100 year old pine, a third sample was taken, and this when dry wa8s dark-brown, These last two samples overlaid a soil similar to the first sample, Employing these samples, the author argued that if the contraction was due to the clay and the iron, and not to the humus, then identical contractions would be observed when equal quantities of all samples were subjected to like conditions. On the other hand, if the phe- nomenon is caused by the humus, then that sample richest in humus would produce the greatest effect on the volume of the gas, and when borax was employed no contraction should occur. It appears that the ferruginous soil poor i n humus produced no contraction, but rather an increase of volume after the first 2 hour ; this is explained by the probable presence of nitrogenous matter decomposable by the brominated alkali.The presence of borax assisted the evolution of gas. Experimenting with humous samples, it was found that the volume of gas evolved by Knop’s method stands in no ratio to the percent- age of ammonia present ; this remarkable result was most manifest in the case of the third sample, which contained most humus, for then the contraction was evident when only 5 grams were used,AXALP TlCXL CIIEMISTRY. 83 and in spite of borax the contraction amounted to 71 per cent. when 1UO grams was employed, This contraction is clearly due to oxida- tion of the organic matter of the soil, and consequent removal of part of the gaseous contents of the azotometer.Working with sandy and chalky Hoils, the author found that the results obtained by Knop’s method are quite as untrustworthy as in the case of humous samples. The following method is recommended for the estimation of am- monia in soils:--200 grams of the soil is mixed with 100 C.C. of hydrochloric acid (1 : 4), then diluted with 300 C.C. of distilled water and allowed to remain for two hours, with frequent shaking. Should much chalk be present, then more than 100 C.C. acid must be added, and the whole amount of liquid made up to 400 C.C. ; all heating must be avoided. Of the filtrate, 200 C.C. (= 100 grams soil) is placed in the evolution vessel of the azotometer together with 5 grams of freshly ignited magnesia usta, and the vessel closed by a doubly bored india-rubber stopper ; through one of the borings a glass tube passes to the bottom of the flask, whilst the other opens immediately below the stopper, and is connected with an aspirator, whilst the other tube supplies air ozonised by passing through potassium per- manganate and concentrated sulphuric acid.The passage of this gas for 10 minutes oxidises organic matter, so that no subsequent con- traction occurs, The rest of the process is conducted in the ordinary way. Percentage of Ammonia i n Soils.-Soils of various characters were examined, and the percent’ages as found by Schlosing’s and the other methods are given : there appears to be a, wide difference between the first and the other two methods. The percentages at different periods in summer and at different depths are recorded.The conclusions drawn are that the percentage of ammonia in uncultivated soils varies with the character of the soil, loams containing most, the quantity increasing with the increase of clay; chalks and sands are poor in ammonia, but in sands rich in hurnus there is a large supply of organic compounds which readily decompose and yield ammonia, although generally speaking, the percentage of organic matter is no indication of t$he amount of ammonia present, and the percentage of ammonia in a soil does not seem to vary with the weather, but does decrease with depth from the surface. Percenta>ge of Nitrates i ? ~ Urnnunured Soils.-Baumann employs Schlosing’s method of estimating nitrates if the solution of 1000 grams soil in water made up to 3000 C.C.shows the brucine reaction, even if the reaction is only obtained after evaporating the solution to one-half of its original volume. If this test should fail, then diphenylnniine is to be employed; this should indicate 1 in 1,500,000 or 0.6- 1 mgrm. per litre ; should this fail, so also will Schlosing’s method, and it will be necessary to evaporate 1 litre of extract with some alkali to dryness, add alcohol, boil, filter, and then evaporate the alcohol and dissolve in 40 C.C. water. The solution must be again tested by brucine and diphenylamine, and if indications of the presence of nitrates are visible, the Marx-Trommsdorf’s method (Zeit. anal. C‘lzem., 1868, 412, and 1870,171) is to be employed for the quantitative8 -1 ABSTRACTS OF CHElllICAL PAPERS.estimation ; this will indicate 0*00001 per 300-400 grams soil. Examination of many soils shows that when the soil is uncultivated the percentage of nitrates is very small, especially in forests, where i t only appears as traces. The author then reviews the work of Warington, Scliliisinp, and Miint>z on nitrification, and considers that the absence of nitrates in such soils as he refers to, is due to the normal temperature ( 5 > ) being so close to that a t which nitrification first occurs; also water is necessary for nitrification, and in forest soils, therefore little nitri- fication takes place, because of the great dryness of that soil in summer, due especially to the enormous transpiration continually occurring, which renders the soil almost "air-dry ; " a further cause for the absence of nitrification may be faund in the want of animal Analysis of Gas Coal.By L. T. WRIGHT (J. SOC. Chem. Ind., 4, 656-667) .-Proximate amZysis.-T he first determination is that of moisture. T'arious temperatures have been recommended at which the sample of coal shall be dried. Since ZOO' is a temperature easily Recured and maintained constant by means of a water-bath, the author has adopted it as the standard. The estimation of moisture by loss of weight in drjing a t 100" until the weight of the substance become9 constant is not, however, free from error ; as it has been noticed by different observers, that after a time the coal not only ceases to lose weight but actually gains. Hinrichs attributes this increase to the slow oxidation of pyrites and other substances in the coal ; the author.however, considers it to be due to the absorption of gases into the pores of t h e coals left vacant by the expelled moisture. He has also found that the increase of weight, which only exhibits itself when t h e con1 has been nearly dried, has been going on during the whole period of the drying process, SO that where accuracy is required it is preferable to weigh the water as such. According to Hinrichs the total volatile matter of coal is determined with accuracy by taking 1 to 2 grams of undried pulverised coal, heating for three and a half minutes over a Bunsen burner, and then immediately igniting without cooling, for the same period over a blast gas lamp (white heat).The greatest difference which Hinrichs obtained amounted to 0.29 per cent. The author has reFeated this method and obtained very fair results, although not quite so accurate as the above. The author adopts the following method:-Take about 2 grams of finely pulverised coal and let i t form an even layer on the bottom of a thin platinum crucible. Weigh without cover. place the crucible (with cover on) in an upright position, then apply a powerful gas flame. Note when the gases cease issuing from under the lid; allow one minute further heating, remove the gas flame, place the crncible and cover in a desiccator tor about five minutes to cool, and then weigh without cover as soon as possible. For the determination of ash in refractory cokes and such substances as gas carbon, heat to redness 2 grams of the coal or coke placcd on a piece of platinurnfoil, in R combustion tube through which ft gentle current of air is drawn.The ash should be saved for the determination of sulphuric acid. nutriment for the growth of the ferment. E. w. P.ASALFTIC hL C HENISTRT. 85 k'or the determination of the total sulphur, the best and eimplest method is that suggested by Nakamura, which consists in heating the coal below a red heat in contact with alkaline carbonates, when the coal whether bituminous or not rapidly undergoes complete oxidation. For the purpose of distinction hetween the sulphur which goes over into the volatile matter, the sulphur left in the coke and that which is finally left in the ash combined as sulphate, three determinations are required :-(1) Total sulphur by Nakamura's method.(2) Sulphur which is converted into sulphurous anhydride by combustion of the coke in air, obtained by roasting a quantity of coke representing a known quantity of the coal, and aspirating the gaseous products of combustion through a solution of iodine or bromine. (3) Sulphur in ash. This can be done either by boiling the ash with hydrochloric acid, filtering and determining the sulphuric acid in the filtrate, or by fusion with alkaline carbonates. Proximate Analysis of Australian Xhale. Sp. gr. = 1.0401. Water lost at 100" - .......... 0.44 Volatile matter ............ 77 69 Corrected for sulphur. Fixed carbon .............. 5.56 9, 7 9 Ash ...................... 15.83 Sulphur in volatile matter.. . . ? * . . . . 9 9 fixed carbon.. 9 , ash. ............. 100.00 The practical method of examining coal for gas-making purposes partakes of two forms : (1) a partial imitation of the process of gns manufacture on a small scale ; (2) analysis of coal by conducting a gas manufacture in a setting of clay retorts with large plant for exhausting, condensing, scrubbing, purifying, measuring the gas, and so on, as is in actual use. The laboratory practical analysis is undoubtedly of great value ; .it will, however, be necessary in interpreting the results, to recollect that the method of heating the coal is different to that used in practice with clay retorts. As far as quality and volume of gas are concerned, the best results are obtained with the small iron retort.The difference varies with the kind of coal employed. With the very finest cokinq coals the difference is very small, and as the coking quality of the coal increases so the difference between the two methods of testing increases. With coals (not cnnnels) which scarcely intumesce a t all, the difference becomes very high. Cannels also vary in tlie same manner, the difference in the results being always connected with differences in the qualities of the cokes. Since the gas from the iron retort is not scrubbed, a deduction of about 3 per cent. should be made from the results of the small apparatus to compensate for the small loss of illuminating power suffered by the qav 9 286 ABSTRACTS OF CHEMICAL PAPERS. of the large experimental works in the washing process.When this allowance is made there is but little difference between the two methodstof testing in the case of coking coals. Estimtion of Hydrogen XulphiJe mtd Carbonic A d ydTide in Crude Coal-gas.-The author prefers the use of a method admitting of the employment of a tolerably large quantity of gas collected regularly during an interval of )time sufficiently long to afford an idea of the average composition of the gas supply to be tested. The reagent used for absorbing the hydrogen sulphide is cupric phosphate. Absorption tubes charged with cupricLphosphate gain in weight under the action of pure coal-gas ; the increme of weight, however, soon reaches a limit, and the phosphatelmsby be saturated by passing 3 cubic feet of pure dry-coal gas slowly through tbe tubes.The carbonic anhydride is absorbed in soda-lime tubes, one half f u l l of soda-lime and one half of calcium chloride. To increase its absorptive power for carbonic anhydride the soda-lime is used in a moist condition. In cases where ammonia exists in the gas its removal is best effected by passing the crude gas before it is dried through a 12-inch U-tube filled with broken pumice saturated with syrupy phosphoric acid. Cyanogen.-This substance is estimated by paming a measured quantity of crude gas freed from ammonia through a (j-tube filled with soda’lirne, and then making a combustion of the residue as in an ordinary nitrogen determination. D. B. Analysis of Explosives. By G. LUNGE ( O h m . Ifid., 9, 273-274). -To render the author’s nitrometer suitable for the determination of nitrogen by Cmm’s method in substances like dynamite and gun-cotton, which cannot be introduced into the decomposition tube in the liquid form, and at the same time to avoid the error due to the evolution of carbonic.anhydride to which Hempel’s modification (this Journal, 40, 472) is liable, a small funnel-tube bent into a swan-neck is fitted by a rubber stopper to the cup in which the weighed substance has been placed. Through this the sulphuric acid required to dissolve the substance is poured. Any carbonic anhydride evolved can escape, but loss of nitrous fumes is prevented by the acid which remains in the bend of the swan-neck. When the substance has dissolved (which in the case of gun-cotton may take three-qnarters to one hour) the solution is drawn into the graduated tube.The acid in the funnel tube follows and rinses the cup. The stopper can now be removed and further rinsings given. The siliceous earth suspended in the acid has not been found to cause any inconvenience. Three analyses of gun-cotton reported agree very closely. M. J. S. Estimation of Glycerol in Wine. By SAMUELSON (Chem. Zeit., 10, 933-934) .-Great discrepancies are observed in the estimation of glycerol by different chemists; this is probably due to want of uniformity in working, therefore the following mode of procedure is recommended. After adding milk of lime, evaporate only as far as to leave the mass just moist, then add somewhat more than 50 C.C. of 96 per cent. alcohol, and evaporate the alcoholic extract to 5 c.c ;ASALYTICAL CHE3lISTRY.87 extract with absolute alcohol and ether, evaporate the extract until free from alcohol, then dry the residue for a t least an hour. Estimation of Solid Matter in Wines. By E. ROEILHON (Cornpt. rend., 103,498) .-When the total residue in wine is estimated by evaporation in a vacuum, the weight of the residue obtained is lower the greater the surface of the evaporating dish, owing to the loss of part of the glycerol. Three dishes each containing 10 C.C. o€ wine were placed under the same receiver and kept in a dry vacuum for 8-24 hours. The results axe given in grams of solid matter per litre of wine. Dilute alcohol D. A. L. Rousil- with 10 p. c. Bordeaux. Girs. lon. Coupage. glycerol. 28 sq. C.C. surface. 22.4 30.8 34.2 25% 34.8 70 7 9 ,, .. 22.0 30-3 33.0 25.1 33.2 with sand 5 mm. 21.2 .29.1 30.4 23.8 31.7 70 9 9 I n order to obtain comparable results, flat dishes of the same diameter should be uscd, arid these should contain equal quantities of wine and be placed ,in the same position in the receiver of the air- Pump. C. H. B. deep . . . . . ” . . . * * . . } Estimation of Acidity of Malt. By E. PRIOR ( B i d Ceiitr., 1886, 647).-The usual way to estimate the acidity of malt is to digest a weighed quantity of ground malt in water for two hours, with frequent agitation, to filter quickly, estimate acidity in an aliquotl part, and calculate as lactic acid ; the author found that half an hour’s digestion with water sufficed to show acidity; he recom- mends a 20 per cent. dilution by Folume of neutral alcohol as the fluid for extraction, the percentage of acidity when this is employed remaining constant €or 24 hours.His method is to dilute ordinary commercial absolute alcohcl with four volumes of water, 500 C.C. of this is used to 100 grams of ground malt, digested in the cold fcr four hours with frequent agitation, filtered, and 100 C.C. titrated with baryta-water. Freaence of Nitrites and Nitrates in Milk an Evidence of Adulteration. By M. SCHRODT (Bied. Centr., 1886, 629).-Nitrous or nitric acids are not norrnally found in milk, and when found in a, suspected sample should be taken as evidence of dilution, spring water which is often added to dilute it, generally containing either nitrites or nitrates. The objection may possibly be made that the cow’s fodder contained nitrates or nitrites ; to put this to the proof, the author fed two cows for five days on beets, to which he added 10 grams daily per head of potassium nitrate, notwithstanding which rio trace of nitrates, &c., was found in the milk ; he therefore thinks tlie evidence afforded by their presence is conclusive.The method used was that introduced by Soxhlet, the reagent being diphenylamine. J. F. J. F.88 ABSTRACTS OF CHEMICAL PAPERS. Further Notes on the Methods of Examining and Chemistry of Fixed Oils. By A. H. ALLEN (J. SOC. Chsm. Ind., 5, 65-72, and 288--283).-Speci$c Gravihy of Oils.-A convenient instrument for ascertaining the density of fixed oils is Westphal's hydrostatic balance. A counterpoised thermometer suspended from a piece of thin platinum wire is attached to one end of a graduated lever. On immersing the thermometer in a liquid, it loses a certain weight.The equilibrium is restored by hanging on the lever a series of riders, whlch are adjusted in weight so as to make the reading very simple. As t'he employment of a thermometer as a plummet renders the instrument unsuited €or determinations of density a t loo", or other high temperature, the author substitutes in such cases a plummet of thick gIass rod. For the determination of the density of fats the author some time ago recommended the use of a Sprengel tabe, and urged that the density should be taken at the boiling point of water. I n all cases, however, where there is snfficient substance a t disposal, the Sprengel tube has been abandoned in favour of the plummet.Coeflccients of Expansion of Oils.-A series of tables illustrating the rates of expansion of fats and oils are given, showing (1) that the rates of expansion of the fluid fixed oils are not sufficiently different to be of any value for their recognition; (2) that of the fluid fixed oils examined (sperm oil, bottle-nose oil, whale oil, porpoise oil, seal oil, menhaden oil, neats-foot oil, lard oil, olive oil, arachis oil, rape oil, sesame oil, cotton-seed oil, niger-seed oil, linseed oil, and castor oil) all with the exception of whale oil expand sensibly equally for the same increase of temperature; and (3) that with the exception of whale oil the correction in density for the fluid fixed oils examined may safely be taken at 0.64 for 1" C.(water a t 15.5" = lU00). Viscosity of Oils.-The author is of opinion that Redwood's new form of viscosimeter bids fair to become the recognised standard instrument of the future. For many purposes, however, and especially as a convenient test by oil merchants, the following instrument is likely to grow in favour. It consisks of a simple arrangement by which a small paddle-wheel (actuated by a falling weight) is caused to revolve in the sampla of oil maintained at a definite temperature by an outer ressel of water. The manipulation is very simple, Snd the results expressed by the number of seconds required by the weight to fall through a given space are very constant. Bromine and Iodine Absorptions of Oils.-In order to facilitate the comparison between the results of Mills (ibid., 2,435, and 3, 366) and Hub1 (Abstr., 1884, 1435), the author has multiplied the bromine absorptions obtained by Mills by -, so as to obtain the equivalent iodine absorptions, and has compared the results with * the experi- mental numbers for iodine absorptions obtained by Hiibl.The figures which are tabulated in the original paper indicate that the drying oils (containing linoleic acid) assimilate the largest proportions of the haloyds, alzd their capacity in this respect might probably be employed as a measure of their drying properties. The fish liver oils, however, fully eqiial the vegetable oils in their assimilating power for haloids. Hiibl's results in the main confirm those of Mills.127 80ANALYTICAL CHEJlIST RY. 89 Va7enfa’s Acetic Acid Test.-The author has tried this method (Abstr., 1884, 1018) on a number of oils and finds that a slight variation in the strength or proportion of the acid employed is not of importance, and that the temperature at which turbidity occurs with any particular specimen is readily observed and fairly constant. Concordant results have also been obtained from several samples of butter, and it appears probable that further experience may prove the method to afford a simple means of distinguishing butter from but terin e. Deferininnfion of Glycerol.-The difficulties attending the deter- mitiation of the glycerol produced by the saponification of fixed oils have recently been orercome by a method originally suggested by W a n k l p and Pox (Abstr, 1886, 395), and perfected by Benedikt and Zsigmondy.It depends on the saponification of the oil, and oxidation of the glycerol thus formed by potassium permanganate in alkaline solution, with formation of oxalic acid, carbonic anhydride, and water. The excess of permanganate is then destroyed by a sdphite, the solution filtered, the filtrate acidified with acetic acid, and pmcipitated with a calcium salt. As the precipitate contains calcium sulphate and silicic acid in addition to calciuni oxalate, the amourit of oxalic acid is determined either by titration of the pre- cipitate with permnnganate in acid solution, or by estimating the alkalinity of the ignited precipitate. For the determination of glycerol in fats, the author recomniends modifying the method in the jollowing manner :-5 grams of the sample of fixed oil is placed in a six-ounce bottle, together with a solution of 2 grams of caustic potrsh in 12 C.C.of water. The bottle is secnrely closed and heated in a water-oven or in boiling water for 8 or 10 hours, the contents being frequently agitated. When the product is perfectly homo- geneous and all oily globules have disappeared, tbe bottle is opened, and the soap diluted with hot water, when a perfectly clear solution should be obtained, except in cases of sperm oil, wax, and other sub- stances yielding insoluble alcohols on saponification. The soap solution is then treated with a moderate excess of acid in the usual way, and the liberated fatty acids are separated from the aqueous liquid con- taining the glycerol, ahich latter is then ready for oxidation with a1 kaline permanganate as above described.By C. J. ELLIS (J. SOC. Chenz. Id., 5, 150-152 and 361-362).-The author has made some experiments with the view of extending the application of Maumenk’s test to drying oils and fish oils, to which it caniiot be directly applied without some slight modification, owing to the violent action which ensues when these oils are mixed with concentrated sulphuric acid. To overcome this difficulty i t was found necessary to mix with a drying or fish oil some liquid which will moderate the action of the acid on the oil. The author employed a mineral lubricatiiig oil of 0.915 sp. gr. for this purpose, and as on mixing sulphuric acid with such an oil a certain increase of temperature takes place, it is necessary to deter- mine the rise due to each gram of the mineral oil.To accelerate the action of sulphuric acid on the mineral oil, a certain proportion of I). B. Maumen6’s Test for Oils.90 ABSTRACTS OF CHEMICAL PAPERS. colza oil was added, for which the standard number, when not mixed, was accurately determined and found to be 35.8”. Contrary to ex- pectation, it was found that the smaller the quantity of mineral oil iii the mixture the greater is the value representing the rise due to each gram of the mineral oil, providing the rise due t o each gram of the vegetable oil remains constant whatever the mixture. To calculate the rise in temperature due to each gram of the mineral oil the following formula is employed :--y = a + bz, iii which -y represents the rise in tempeiature due to each gram of mineral oil, x is the fraction of the mixture coiisisting of mineral oil, and cc and b are constants depending on the conditions of the experiment and the particular mineral oil employed.In order to obtain the most concordant and trustworthy results, the maximum temperature attained should not exceed 60”, and it is for this reason that the author prefers the use of a, mineral oil as the re- tarding reagent. D. B. Employment of Congo-red in Titrating Aniline. By P. JULIUS (C‘hein. I d , 9, 109-110).-CCongo-red, the compound of tetrazodiphenyl with naphtholsnlphonic acids, is turned blue by acids, and recovers its red colour with an excess of alkali. When used as an indicator i n titrating aniline or its homologues with a mineral acid, the point is taken at which a bluish-violet, not changed by small further additions OE acid, is produced.A much larger excess is required to produce a pure blue. The results do not vary more than 0.2 per cent. from the theoretical numbers. Hufner’s Method of Estimating Urea. Ry E. PFL~GER and K. BOHLAND (P’cger’s Archiv, 39, l-l7).-Pfliiger and Schenk pre- viously proved (ibid., 38, 385) that Hiifner’s method of estimating the nitrogen in urine gave results which were too low and variable to found a calculation of the total nitrogen on. Owing to the improve- ment made in Bunsen’s method, the authors have been able to ascertain whether Hufner’s method was sufficiently accurate for the determination of urea only.They find that the results are always too high and also very variable. The variatipn ranges from 1 per cent. to 10 per cent., and does not therefore admit of compensation. J. P. L. M. J. S. Estimation of Urea in Human Urine with Sodium Hypo- bromite. By E. PFL~~GER and K. BOHLAND (Pfliiger’s Archiv, 39, 143-158) .-Pfluger’s new method of estimating urea by hypobromite (PJliiger’s rlrchiv, 38, 503) can be applied to the estimation of urea in liunian urine provided the nitrogenous extractives are first removed. J?or this purpose, a given volume of urine acidified with hydrochloric acid (1 of acid to 10 of urine) is precipitated with sufficient phospho- tungstic acid to ensure the separation of all the extractives, the mixture made up to a known volume, and allowed to remain 24 hours a t least previous to filtration.The acid filtrate is carefully neutralised with powdered lime, and again filtered through a dry filter. Great stress is laid on the use of pure soda and bromine for theASALPTICAL CHEJIISTRT. 91 preparation of the hypobromite. of urine by this process was + 1.3 per cent. The mean error of several analyses J. P. L. Qualitative Tests for the Dyes Found in Commerce. By 9. N. WITT (Chem. Ind., 9, 1--7).-A table of reactions for the identi- fication of about 80 artificial colouring matters taken singly. Many commercial dyes are mixtures : in this case, the powder strewn upon wet filter-paper or colourless sulphuric acid will generally give streaks of more than on6 colour. Where the mixture is more intimate a solution must be made, and the colouring matters withdrawn frac- tionally by dyeing small pellets of wool or silk in it.The principal adulterant for azo-dyes is sodium sulphate. It is best detected after precipitating tlie colour by pure sodium chloride. M. J. S. Detection of Artificially Coloured Red Wine (Clarat). By J. HERZ (Chem. Zeit., 10, 968-969 ; 998).-To 30-50 C.C. of the wine, or if the quantity of colouring matter in the wine is small, 100 C.C. concentrated to 30 c.c., 20-30 C.C. of a saturated solution of magne- sium sulphate, and 10-20 C.C. of soda solution are added, stirring well; if necessary the treatment is repeated until the liquid is colourless, or nearly so. The filtrate is made acid with dilut,e sul- phuric acid (1 : 3), and if sulphonic acid colours are present the red colour reappears.The most commonly used member of this group, acid-magenta (rosanilinesulphonic acid), yields a, violet-red solution, and can be estimated by comparing the tint with magenta solutions of known strength. One mgram. of magenta per litre can be distinctly detected in 30 C.C. of wine without previous concentration. When archil (orseille) colours are present, the filtrate is bluish, and when made acid turns a litmus-red colour. To test for magenta under such circumstances, Blarez’ method of shaking with lead dioxide is used ; $his destroys the orseille and natural colour. Cazeneuve’s method is not recommended. To test for other colours in the magnesium hydroxide precipitate, the gelatinous mass is stirred up with hot water, allowed to settle, and the liquid decanted off.I f only the natural colour of the wine is present, or bilberry has been used, this liquid is yellow-brown ; if archil has been used, dark-violet,; if yonceau, onion or ponceau red ; if casskshe, pale-red or dark-yellow ; if cinicoline bcrdelaise, a yellow-red to yellow-brown liquid, which when poured on sulphuric acid gives a violet ring. By shaking the coloured liquid with ainyl alcohol, Tonceau yields an onion-red residue ; vinicoline, a dark-brown one ; ccmwsine, a dirt-y-green, violet at the edge, turned yellow by strong hydrochloric acid. The precipitate is a dark-grey or brownish-grey colour when the natural or vegetable colours only are present; with archil, it is violet; with magenta (acid o r ordinary), dirty white ; with cassissine, dirty yellow-brown ; with vinicoline, crimson-red.The precipitate is mixed with sand, dried, and extracted with ether ; the extract contains any ordinary magenta which can be identified in the usual manner by dyeing wool, or cassissine which dyes wool red-brown and leaves a yellow-brown residue in the dish. The dyed wool becomes yellow when treated with strong hydrochloric acid and colourless with ammonia. When92 ABSIL’RACTS OF CHEMICAL PAPERS. wine is shaken with amyl alcohol, and the coloured extract evaporated, the residue, if it contains tho gubstances named, behaves in the manner described below :- Wilh concentmted c-A 7 H:S0,. HCI. NaHO. Archil . . . . . . . . violet-red bl ue red bl 11 e Bordeaux, B... . oarmine carmine carmine carmine Ponceau, RRR.. dark-red crimson crimson brown Cassissirie.. . . . . violet-purple yellow yellow- red Vinicoline Bor- brown delaise . . . . . . cherry-red brown red brown whilst the wine after extraction i~ cherry-red with ordinary maqefcta, violet-red with acihmngenfu, dark-cherry with Bordeam, yellow-red with p o i z c m ~ . Wine coloured with magenta produces a violet froth. The detection of vegetable colouring matters in presence of the natural colour of wine or otherwise is a matter of great difficultg, and most of the known metliods are ineffectual ; it is, however, effected by the authorwith comparative facility in the following manner :-I0 15 C.C. of wine is shaken with 5 C.C. of a saturated solutiop of tartar emetic, and then examined by reflected and transmitted light either a t once or, if no immediate chanqe has taken place, after some time.This treatmelit produces with genuine red wine always a cherry. red colour, and with other substances as follows :-Red-poppy (Papaver r?iwus), dark cherry-red ; cherry, violet : commercial elder colonring matter, Ed-violet ; hilberry (Vaccin. niyrtill), blue-violet ; priret-berry, piire violet. White wines artificially colonred, and red wines mixed with artificial colours have been successfully examined in this manner ; in the latter case the wine some time after treatment is compared with a genuine red wine to distinguish more readily the change of colour. Old solutions of privet do not give the d o u r change. Sodium hydrogen carbonate produces with pure wine, brown-red ; with wine coloured with pure eldedwry, grey-violet ; and with hilberry, brown- green.Tartar emetic appcars to form an antimony lake with tho colouring matters. With practice, all the above-mentioned coloum cart be detected in 30-50 C.C. of wine. In the subsequent communication (loc. cit., 998), the author acknowledges the priority of Ambiihl and Elsner’s recommendation of the use of tartar emetic for the purpose in question. They, homver, recommend hot solutJions ; the author finds cold better. Fermented bilberries give the violet colour even better than unfermented berries, especially when fresh, innsrriuch as oxidation interl’eree with t,he delicacy after a time. The distinctness of this colour is increased by diluting the wine.D. A. L.78 ABSTRACTS OF CHEMICAL PAPERS.An a 1 y t i c a 1 C h em i s t ry.Kjeldahl's Method of Estimating Nitrogen. By C. ARNOLD(Arch. Pharrn. [3], 24, 785--794).-The author shows that thismethod (Abstr., 1884, 364; 1885, 930) is also applicable to nitrates,nitro- and cyanogen-compounds, in the presence of beneoic acid, sugar,mercury, and copper sulphate ; further, in the case of alkaloids andotber compounds that can be treated by this method, oxidation bypotassium permanganate is unnecessary if the boiling be continueduntil the warm liquid acquires a bluish-green tint, becoming colour-less on cooling ; this takes place tolerably quickly in the presence ofmercury and copper. That the oxidation has been completed canreadily be ascertained by adding a crystal of potassium permanganate,which should give a persistent violet or green coloration.A descrip-tion of tlie apparatus used by the author is given. J. 1'.Separation of Arsenic and Antimony. By ZAMBELLI andLUZZATO ( A ~ c h . Pharm. [ 3 ] , 24, 772).-These elements can be sepa-mted, even in foyensic cases, by treating the still moist sulphides withhydrogen peroxide at 40' for some hours, then heating to near 100AKALYTICAL CHEMISTRY. 79and filtering. The arsenic acid formed goes into the filtrate, whilst theinsoluble antimony oxide remains on the filter. The method is saidto be very delicate.Detection of Thiosulphate in Hydrogen Sodium Carbo-nate. By RRENSTEIN and T. SALZER (Arch. Yharm. [3], 24, 761).-According to Brenstein, the reaction given for thiosulphate by Myliusis not characteristic, as other oxygen-compounds of sulphur, such assulphites, give rise to the formation of hydrogen sulphide.A bettartest for tJiiosu1phate is to add to a 5 per cent. solution of hydrogensodium carbonate a few drops of silver nitrate solution, then excess ofnitric acid, and heat to boiling; even minute traces of thiosulphategive an immediate dark precipitate of silver sulphide.According to Salzer, the absence of thiosulphate is easily ascer-tained by adding a few drops of iodine solution to about 20 C.C. of asaturated solution of hydrogen sodium carbonate ; the solution musthave a yellowish tint. Decolorisation of the iodine solution does notnecessarily imply the presence of thiosulphate, since normal carbo-nate, the most commonly occurring impurity, produces this effect.Both authors found thiosulphate to be a constant impurity of ordinaryqualities of hydrogen sodium carbonate, whilst the better qualitieswere mostly free from it.Estimation of Small Quantities of Silver in Burnt Pyrites.By E.Trrrr,o (Cherrz. Zeit., 10, 822; 1065--1067).-The amount ofsilver in burnt pyrites is ordinarily 0*003--0.008 per cent., and toobtain trustworthy results material containing 0.01 gram of silvershould be taken for analysis, or about 300 to 500 prams of pyrites. Manydifficulties have been encountered in endeavouring to deal with theselarge quantities of material so as to get all the silver. Experimentsshow that the decomposition of the whole mass is out of the question,therefore methods for extracting the silver, &c., were tried ; treat-ment with nitric acid or with gaseous chlorine, and subsequentextraction of the silver chloride proved unsuitable.Digestion withchlorine-water was somewhat better ; but bromine or bromine-watergives the best results. The powdered burnt pyrites is passed througha 0.25-mm. mesh sieve, and about 300-500 grams of it is placed i nan acid-proof iron basin; it is well covered with water, and whilevigorously stirred, bromine is added until present in excess. After24 hours it is heated on a water-bath for one hour with frequentstirring, treated with excess of ammonia, then with 500 grams ofammonium chloride, and 1 litre of water.The mass is boiled for anhour, and filtered. The residue is dried, ignited at a low red heat, andagain treated with arnmoniacal ammonium chloride, by which meansall the silver is extracted. Inasmuch as copper is quite as difficult assilver to extract by means of bromine, and as theamount of copper i nthe pyrites is reduced from 5 per cent. to a trace (0.01 per cent.) by theabove treatment, this may be used as an indicator of the progress of thedesilvering, and in practice when the copper is reduced to this amountit may be safely inferred that all the silver is removed. The solu-tion, about 4 litres, is made acid with hjdrochloric acid; it is notJ. T.J. T80 ABSTRACTS Ol!’ CHEMICAL PAPERS.affected by daylight either before or after acidifying.The silver,copper, and lead are then precipitated by means of chemically piirezinc. When the precipitation is complete, the solution is colourless(or rose-red if cobalt is present), and is not turned blue bv ammonia.As the precipitate contains, most liksly, silver chloride and bromide,and some iron oxide, it is fused with potassium cpnide, and sub-sequently with anhydrous borax, the temperature being raised to meltthe copper, The regulus is dissolved in nitric acid, and the silvereither precipitated by means of hydrochloric acid or determinedelectrolytically. Silver determinations may also be made in ordinarypyrites, which is first roasted carefnllp, and then treated in the abovemanner. Good results have been obtained by this method, withroasted copper schist, with spathic iron ore containing argentiferoustetrahedrite and pyrites, and with other ores.A determination can becomplded iu three days.Comparative Gasometric Assaying with Especial Referenceto the Valuation of Zinc Powder and the Testing of Car-bonates. By J. BARNES ( J . SOC. Chem. Ind., 5, 145--147).-1n a,previous communication (Trans., 1881, 462), the author described amethod for the valuation of zinc powder, consisting in measuring theamount of hydrogen liberated on treatment with an acid. Theauthor has since devised a method and constructed an apparatus forgasome tric assaying, contrived originally for the valuation of zincpowder, but applicable also to the estimation of other substances,which may be made to cause or control tlie evolution of gases.Themethod, which is described in detail in the paper, is a comparativeone, and requires a t least two graduated instruments, each providedwith a suitable gas evolving arrangement. A substance of knownvalue, taken as the standard, is placed in one, and the sample tobe examined in the other instrument. The value of the sample isthen calculated by the following formula: - . W . R = x ; - whereVequals the volume of gas from standard, v the volume of gas frommmple, W the weight of standard, and ?u the weight of sample.R is the amount of pure substance or its equivalent in unit ofstandard, and x the amount of pure substance in unit of sample.Determination and Valuation of Copper in Ores andProducts for Commercial Purposes, with some Remarks onthe Assay of Gold in Bar Copper.By J. TV. WESTMORKLAND( J . Soc. Chem. Ind., 5, 48--64).-Esfin?ation of Moisture.-This,although apparently a simple process, is conducted very differentlyby rarious assr,yers, and the author proposes that the temperatureand niaiiner in which the moistnre is to be taken should be moreclearly defined than it is at the present time.Wet Assay of Coppr.-Having examined a iarge number of samplesof all kinds of copper ores aud products by several wet processes, theauthor has given the iodide process as modified by Brown thepreference-being more trustworthy arid accurate than the electrolytioprocess. For cupreous pyrites, burnt ores, &c., the followiiig mctliodD.A. L.v 20D. BANALYTICAL CEEMISTRY. 81mny be used. From 50 to 150 grains of the sample is dissolved in acids,evaporated to dryness with excess of sulphuric acid, diluted with waterarid filtered, From this solution, the copper is separated with sodiumt'hiosulphate, the precipitate dried, ignited, dissolved in nitric acid,evaporated with sulphuric acid to separate Graces of lead, dilutedwith water and filtered, sodium carbonate added in excess, and thenacetic acid to acid reaction. The solution is then titrated withpotassium iodide. As 8n alternative method, the ore may be calciued,dissolved in hydrochloric acid, the ferric salt reduced by boiling witha solution of sodium sulphite, and hydrogen sulphide passed throughthe cold Rolution; the precipitated sulphides are then washed, andthe process conducted as before.Richer copper ores, mattes, preci-pitates, &c., are dissolved in acids, and the solution precipitatedeither with sodium thiosulphate or hydrogen sulphide, the sulpbidesbeing dissolved in acids, and treated 8s before. The accuracy of thisprocess was tested by numerous experiments, in which known weightsof pure electrotype copper, silver, arsenic, antimony, lead, cadmium,bismuth, tin, manganese, zinc, and iron salts were used. The solutionswere made to represent cnpreons burnt ores, mattes, cupriferous leadregul us, second quality and rich precipitates, &c. Experiments werealso made i n many cases with the metals separately. The electro-deposition method may be employed in cases where it is desir,thle tohave resultti of two separate and distinct processes, it being notedthat with pure solutions of copper, accurate results are obtained,whilst when silver or bismuth are present they are precipitated, andin one experiment tin was also deposited with the copper.In thepresenoe of ferric salts, the deposition is retarded. It was found thatwhilst the results with refined or bar coppers and rich precipitatesagree with those obtained by the iodide process, those obtained byelectrolgsing sulphuric acid solutions of mattes, copper ores, $c., aregenerally slightly below the triie percentage. With burntl ores,cupreous pyrites, and ores containing but little COPPOI', the authorfirst precipitates with hydrogen sulphide, dissolving the s u lphide innitric acid, and evaporating this soliifion with excess of sulphuricacid, the sulphates dissolved in water give a solution from whichcopper (and bismuth) is readily deposited.The results by the batteryprocess are generally from 0.01 to 0.04 per cent. higher than by theiodide process, this being cniised by a slight deposit of bismuth.The Cornish Yroce.9~ of Dry Assay.-This process is considered to beinaccurate and misleading, and is also liable to aerious variation evenin the hands of experienced operators.The author is of opinion that in cases where the sulphur contentsof pyrites are sold to alkali works, tbe burnt ores being returned tothe vendors, an accurate wet assay forms the best# check on the coppercontents of the ore delivered to and received from the alkali works.In the remaining part of the paper, the methods in vogue for fixingthe prices payable for copper in pyrites, copper ores.&c., are discussed.Reftmnce is also made to the avsay of goltl in bar coppry, and theopinion is expressed that t>hese assays are not conducted as carefullyaa they should be, taking the value of the metal into oonsideration.I). B.9 VOL LU82 ABSTRAOTS OF CREMICAL PAPERS.Estimation of Ammoniacal Nitrogen in Soils, and theAmount of Assimilable Nitrogen in Uncultivated Land. RyA. BAUMAXN (Lmdw. Versuchs-Stat., 1886, 2$7--303).--The methodsemployed to estimate the ammonia in soils are nnsatisfactor-y.Schlosing’s method, whereby the ammonitt is freed by means of milkof lime or sodium hydroxide, gives results too high, probably owingto the formation of ammonia from other compounds during theperiod (2-3 days) required for the process.In Bonssingault’sprocess as modified by Sclilosing, a hydrochloric acid extract of the soilis treated with magnesia usta, but the ammonia collected in acid mu$not be estimated by titIration, but must be azotometrically determined ;nor must cork nor caoutchouc be employed in the formation of theapparatus, The liquid in the evolution flask must be boiled for atleast an hour, and the magnesia usta must be fresh. With thesetwo processes, it is found that humous soils, when treated with sodiumhydroxide, continuously yield ammonia ; if the soil, after treatmentwith magnesia, be further treated with sodium hydroxide, anothersupply of ammonia is obtained, b u t only in the case of soils rich innumus.Knop’s process, in which the ammonia is decomposed bybrominated sodium hydroxide solution, is vitiated by the fact, noticedby Rnop and others, that in the presence of the soil itself a con-traction of the volume of gas enclosed in the apparatus occurs ; thiscontraction was stated by some to be due to absorption by the organicmatter, whilst others considered the oxides of iron in the clay to bethe real cause; also it had been stated that the presence of boraxprevented this contraction. To ascertain the cause of the con-tract.ion, a sample of stiff soil poor in humus, was taken from uncul-tivated land a t a depth of 15-20 cm.; this sample seemed freefrom humus, although root-fibres were visible ; at a distance of 40 m.from this spot another sample was drawn from a depth of 2-10 cm.,this was close to a pine SO years old.The sample showed humus,and when dry was grey. At another spot, close to a 100 year oldpine, a third sample was taken, and this when dry wa8s dark-brown,These last two samples overlaid a soil similar to the first sample,Employing these samples, the author argued that if the contractionwas due to the clay and the iron, and not to the humus, then identicalcontractions would be observed when equal quantities of all sampleswere subjected to like conditions. On the other hand, if the phe-nomenon is caused by the humus, then that sample richest in humuswould produce the greatest effect on the volume of the gas, and whenborax was employed no contraction should occur.It appears thatthe ferruginous soil poor i n humus produced no contraction, butrather an increase of volume after the first 2 hour ; this is explainedby the probable presence of nitrogenous matter decomposable by thebrominated alkali. The presence of borax assisted the evolution ofgas.Experimenting with humous samples, it was found that the volumeof gas evolved by Knop’s method stands in no ratio to the percent-age of ammonia present ; this remarkable result was most manifestin the case of the third sample, which contained most humus, forthen the contraction was evident when only 5 grams were usedAXALP TlCXL CIIEMISTRY.83and in spite of borax the contraction amounted to 71 per cent. when1UO grams was employed, This contraction is clearly due to oxida-tion of the organic matter of the soil, and consequent removal of partof the gaseous contents of the azotometer.Working with sandy and chalky Hoils, the author found that theresults obtained by Knop’s method are quite as untrustworthy as inthe case of humous samples.The following method is recommended for the estimation of am-monia in soils:--200 grams of the soil is mixed with 100 C.C. ofhydrochloric acid (1 : 4), then diluted with 300 C.C. of distilled waterand allowed to remain for two hours, with frequent shaking. Shouldmuch chalk be present, then more than 100 C.C.acid must be added,and the whole amount of liquid made up to 400 C.C. ; all heatingmust be avoided. Of the filtrate, 200 C.C. (= 100 grams soil) isplaced in the evolution vessel of the azotometer together with 5 gramsof freshly ignited magnesia usta, and the vessel closed by a doublybored india-rubber stopper ; through one of the borings a glass tubepasses to the bottom of the flask, whilst the other opens immediatelybelow the stopper, and is connected with an aspirator, whilst theother tube supplies air ozonised by passing through potassium per-manganate and concentrated sulphuric acid. The passage of this gasfor 10 minutes oxidises organic matter, so that no subsequent con-traction occurs, The rest of the process is conducted in the ordinaryway.Percentage of Ammonia i n Soils.-Soils of various characters wereexamined, and the percent’ages as found by Schlosing’s and the othermethods are given : there appears to be a, wide difference between thefirst and the other two methods.The percentages at different periodsin summer and at different depths are recorded. The conclusionsdrawn are that the percentage of ammonia in uncultivated soils varieswith the character of the soil, loams containing most, the quantityincreasing with the increase of clay; chalks and sands are poor inammonia, but in sands rich in hurnus there is a large supply oforganic compounds which readily decompose and yield ammonia,although generally speaking, the percentage of organic matter is noindication of t$he amount of ammonia present, and the percentage ofammonia in a soil does not seem to vary with the weather, but doesdecrease with depth from the surface.Percenta>ge of Nitrates i ? ~ Urnnunured Soils.-Baumann employsSchlosing’s method of estimating nitrates if the solution of 1000 gramssoil in water made up to 3000 C.C.shows the brucine reaction, even ifthe reaction is only obtained after evaporating the solution to one-halfof its original volume. If this test should fail, then diphenylnniineis to be employed; this should indicate 1 in 1,500,000 or 0.6-1 mgrm. per litre ; should this fail, so also will Schlosing’s method, andit will be necessary to evaporate 1 litre of extract with some alkali todryness, add alcohol, boil, filter, and then evaporate the alcohol anddissolve in 40 C.C.water. The solution must be again tested bybrucine and diphenylamine, and if indications of the presence ofnitrates are visible, the Marx-Trommsdorf’s method (Zeit. anal.C‘lzem., 1868, 412, and 1870,171) is to be employed for the quantitativ8 -1 ABSTRACTS OF CHElllICAL PAPERS.estimation ; this will indicate 0*00001 per 300-400 grams soil.Examination of many soils shows that when the soil is uncultivatedthe percentage of nitrates is very small, especially in forests, where i tonly appears as traces.The author then reviews the work of Warington, Scliliisinp, andMiint>z on nitrification, and considers that the absence of nitrates insuch soils as he refers to, is due to the normal temperature ( 5 > ) beingso close to that a t which nitrification first occurs; also water isnecessary for nitrification, and in forest soils, therefore little nitri-fication takes place, because of the great dryness of that soil insummer, due especially to the enormous transpiration continuallyoccurring, which renders the soil almost "air-dry ; " a further causefor the absence of nitrification may be faund in the want of animalAnalysis of Gas Coal.By L. T. WRIGHT (J. SOC. Chem. Ind., 4,656-667) .-Proximate amZysis.-T he first determination is that ofmoisture. T'arious temperatures have been recommended at whichthe sample of coal shall be dried. Since ZOO' is a temperature easilyRecured and maintained constant by means of a water-bath, the authorhas adopted it as the standard.The estimation of moisture by lossof weight in drjing a t 100" until the weight of the substance become9constant is not, however, free from error ; as it has been noticed bydifferent observers, that after a time the coal not only ceases to loseweight but actually gains. Hinrichs attributes this increase to theslow oxidation of pyrites and other substances in the coal ; the author.however, considers it to be due to the absorption of gases into thepores of t h e coals left vacant by the expelled moisture. He has alsofound that the increase of weight, which only exhibits itself when t h econ1 has been nearly dried, has been going on during the wholeperiod of the drying process, SO that where accuracy is required it ispreferable to weigh the water as such.According to Hinrichs thetotal volatile matter of coal is determined with accuracy by taking1 to 2 grams of undried pulverised coal, heating for three and a halfminutes over a Bunsen burner, and then immediately igniting withoutcooling, for the same period over a blast gas lamp (white heat). Thegreatest difference which Hinrichs obtained amounted to 0.29 percent. The author has reFeated this method and obtained very fairresults, although not quite so accurate as the above. The authoradopts the following method:-Take about 2 grams of finelypulverised coal and let i t form an even layer on the bottom of a thinplatinum crucible. Weigh without cover. place the crucible (withcover on) in an upright position, then apply a powerful gas flame.Note when the gases cease issuing from under the lid; allow oneminute further heating, remove the gas flame, place the crncible andcover in a desiccator tor about five minutes to cool, and then weighwithout cover as soon as possible.For the determination of ash inrefractory cokes and such substances as gas carbon, heat to redness2 grams of the coal or coke placcd on a piece of platinurnfoil, inR combustion tube through which ft gentle current of air is drawn.The ash should be saved for the determination of sulphuric acid.nutriment for the growth of the ferment. E. w. PASALFTIC hL C HENISTRT. 85k'or the determination of the total sulphur, the best and eimplestmethod is that suggested by Nakamura, which consists in heating thecoal below a red heat in contact with alkaline carbonates, when thecoal whether bituminous or not rapidly undergoes complete oxidation.For the purpose of distinction hetween the sulphur which goes overinto the volatile matter, the sulphur left in the coke and that which isfinally left in the ash combined as sulphate, three determinations arerequired :-(1) Total sulphur by Nakamura's method. (2) Sulphurwhich is converted into sulphurous anhydride by combustion of thecoke in air, obtained by roasting a quantity of coke representing aknown quantity of the coal, and aspirating the gaseous products ofcombustion through a solution of iodine or bromine.(3) Sulphur inash. This can be done either by boiling the ash with hydrochloric acid,filtering and determining the sulphuric acid in the filtrate, or byfusion with alkaline carbonates.Proximate Analysis of Australian Xhale.Sp.gr. = 1.0401.Water lost at 100" - .......... 0.44Volatile matter ............ 77 69 Corrected for sulphur.Fixed carbon .............. 5.56 9, 7 9Ash ...................... 15.83Sulphur in volatile matter. . . . ? *. . . . 9 9 fixed carbon..9 , ash. .............100.00The practical method of examining coal for gas-making purposespartakes of two forms : (1) a partial imitation of the process of gnsmanufacture on a small scale ; (2) analysis of coal by conducting a gasmanufacture in a setting of clay retorts with large plant for exhausting,condensing, scrubbing, purifying, measuring the gas, and so on, as is inactual use.The laboratory practical analysis is undoubtedly of greatvalue ; .it will, however, be necessary in interpreting the results,to recollect that the method of heating the coal is different to thatused in practice with clay retorts.As far as quality and volume ofgas are concerned, the best results are obtained with the small ironretort. The difference varies with the kind of coal employed. Withthe very finest cokinq coals the difference is very small, and as thecoking quality of the coal increases so the difference between the twomethods of testing increases. With coals (not cnnnels) which scarcelyintumesce a t all, the difference becomes very high. Cannels also varyin tlie same manner, the difference in the results being alwaysconnected with differences in the qualities of the cokes.Since thegas from the iron retort is not scrubbed, a deduction of about 3 percent. should be made from the results of the small apparatus tocompensate for the small loss of illuminating power suffered by the qav9 86 ABSTRACTS OF CHEMICAL PAPERS.of the large experimental works in the washing process. When thisallowance is made there is but little difference between the twomethodstof testing in the case of coking coals.Estimtion of Hydrogen XulphiJe mtd Carbonic A d ydTide in CrudeCoal-gas.-The author prefers the use of a method admitting of theemployment of a tolerably large quantity of gas collected regularlyduring an interval of )time sufficiently long to afford an idea of theaverage composition of the gas supply to be tested.The reagent usedfor absorbing the hydrogen sulphide is cupric phosphate. Absorptiontubes charged with cupricLphosphate gain in weight under the actionof pure coal-gas ; the increme of weight, however, soon reaches a limit,and the phosphatelmsby be saturated by passing 3 cubic feet of puredry-coal gas slowly through tbe tubes. The carbonic anhydride isabsorbed in soda-lime tubes, one half f u l l of soda-lime and one half ofcalcium chloride. To increase its absorptive power for carbonicanhydride the soda-lime is used in a moist condition. In cases whereammonia exists in the gas its removal is best effected by passing thecrude gas before it is dried through a 12-inch U-tube filled withbroken pumice saturated with syrupy phosphoric acid.Cyanogen.-This substance is estimated by paming a measuredquantity of crude gas freed from ammonia through a (j-tube filledwith soda’lirne, and then making a combustion of the residue as in anordinary nitrogen determination.D. B.Analysis of Explosives. By G. LUNGE ( O h m . Ifid., 9, 273-274).-To render the author’s nitrometer suitable for the determination ofnitrogen by Cmm’s method in substances like dynamite and gun-cotton,which cannot be introduced into the decomposition tube in the liquidform, and at the same time to avoid the error due to the evolution ofcarbonic. anhydride to which Hempel’s modification (this Journal, 40,472) is liable, a small funnel-tube bent into a swan-neck is fitted by arubber stopper to the cup in which the weighed substance has beenplaced.Through this the sulphuric acid required to dissolve thesubstance is poured. Any carbonic anhydride evolved can escape, butloss of nitrous fumes is prevented by the acid which remains in thebend of the swan-neck. When the substance has dissolved (which inthe case of gun-cotton may take three-qnarters to one hour) thesolution is drawn into the graduated tube. The acid in the funneltube follows and rinses the cup. The stopper can now be removedand further rinsings given. The siliceous earth suspended in the acidhas not been found to cause any inconvenience. Three analyses ofgun-cotton reported agree very closely. M. J. S.Estimation of Glycerol in Wine. By SAMUELSON (Chem.Zeit.,10, 933-934) .-Great discrepancies are observed in the estimation ofglycerol by different chemists; this is probably due to want ofuniformity in working, therefore the following mode of procedure isrecommended. After adding milk of lime, evaporate only as far as toleave the mass just moist, then add somewhat more than 50 C.C. of96 per cent. alcohol, and evaporate the alcoholic extract to 5 c.c ASALYTICAL CHE3lISTRY. 87extract with absolute alcohol and ether, evaporate the extract untilfree from alcohol, then dry the residue for a t least an hour.Estimation of Solid Matter in Wines. By E. ROEILHON(Cornpt. rend., 103,498) .-When the total residue in wine is estimatedby evaporation in a vacuum, the weight of the residue obtained islower the greater the surface of the evaporating dish, owing to theloss of part of the glycerol.Three dishes each containing 10 C.C. o€wine were placed under the same receiver and kept in a dry vacuumfor 8-24 hours. The results axe given in grams of solid matter perlitre of wine.Dilute alcoholD. A. L.Rousil- with 10 p. c.Bordeaux. Girs. lon. Coupage. glycerol.28 sq. C.C. surface. 22.4 30.8 34.2 25% 34.870 7 9 ,, . . 22.0 30-3 33.0 25.1 33.2with sand 5 mm. 21.2 .29.1 30.4 23.8 31.770 9 9I n order to obtain comparable results, flat dishes of the samediameter should be uscd, arid these should contain equal quantities ofwine and be placed ,in the same position in the receiver of the air-Pump. C. H. B.deep .. . . . ” . . . * * . . }Estimation of Acidity of Malt. By E. PRIOR ( B i d Ceiitr.,1886, 647).-The usual way to estimate the acidity of malt is todigest a weighed quantity of ground malt in water for two hours,with frequent agitation, to filter quickly, estimate acidity in analiquotl part, and calculate as lactic acid ; the author found that halfan hour’s digestion with water sufficed to show acidity; he recom-mends a 20 per cent. dilution by Folume of neutral alcohol as thefluid for extraction, the percentage of acidity when this is employedremaining constant €or 24 hours.His method is to dilute ordinary commercial absolute alcohcl withfour volumes of water, 500 C.C. of this is used to 100 grams of groundmalt, digested in the cold fcr four hours with frequent agitation,filtered, and 100 C.C.titrated with baryta-water.Freaence of Nitrites and Nitrates in Milk an Evidence ofAdulteration. By M. SCHRODT (Bied. Centr., 1886, 629).-Nitrousor nitric acids are not norrnally found in milk, and when found in a,suspected sample should be taken as evidence of dilution, springwater which is often added to dilute it, generally containing eithernitrites or nitrates. The objection may possibly be made that thecow’s fodder contained nitrates or nitrites ; to put this to the proof,the author fed two cows for five days on beets, to which he added 10grams daily per head of potassium nitrate, notwithstanding which riotrace of nitrates, &c., was found in the milk ; he therefore thinks tlieevidence afforded by their presence is conclusive.The method usedwas that introduced by Soxhlet, the reagent being diphenylamine.J. F.J. F88 ABSTRACTS OF CHEMICAL PAPERS.Further Notes on the Methods of Examining and Chemistryof Fixed Oils. By A. H. ALLEN (J. SOC. Chsm. Ind., 5, 65-72, and288--283).-Speci$c Gravihy of Oils.-A convenient instrument forascertaining the density of fixed oils is Westphal's hydrostaticbalance. A counterpoised thermometer suspended from a piece ofthin platinum wire is attached to one end of a graduated lever. Onimmersing the thermometer in a liquid, it loses a certain weight. Theequilibrium is restored by hanging on the lever a series of riders,whlch are adjusted in weight so as to make the reading very simple.As t'he employment of a thermometer as a plummet renders theinstrument unsuited €or determinations of density a t loo", or otherhigh temperature, the author substitutes in such cases a plummet ofthick gIass rod.For the determination of the density of fats theauthor some time ago recommended the use of a Sprengel tabe, andurged that the density should be taken at the boiling point of water.I n all cases, however, where there is snfficient substance a t disposal,the Sprengel tube has been abandoned in favour of the plummet.Coeflccients of Expansion of Oils.-A series of tables illustrating therates of expansion of fats and oils are given, showing (1) that therates of expansion of the fluid fixed oils are not sufficiently differentto be of any value for their recognition; (2) that of the fluid fixedoils examined (sperm oil, bottle-nose oil, whale oil, porpoise oil, sealoil, menhaden oil, neats-foot oil, lard oil, olive oil, arachis oil, rapeoil, sesame oil, cotton-seed oil, niger-seed oil, linseed oil, and castoroil) all with the exception of whale oil expand sensibly equally forthe same increase of temperature; and (3) that with the exceptionof whale oil the correction in density for the fluid fixed oils examinedmay safely be taken at 0.64 for 1" C.(water a t 15.5" = lU00).Viscosity of Oils.-The author is of opinion that Redwood's newform of viscosimeter bids fair to become the recognised standardinstrument of the future. For many purposes, however, and especiallyas a convenient test by oil merchants, the following instrument islikely to grow in favour.It consisks of a simple arrangement bywhich a small paddle-wheel (actuated by a falling weight) is causedto revolve in the sampla of oil maintained at a definite temperatureby an outer ressel of water. The manipulation is very simple, Sndthe results expressed by the number of seconds required by theweight to fall through a given space are very constant.Bromine and Iodine Absorptions of Oils.-In order to facilitate thecomparison between the results of Mills (ibid., 2,435, and 3, 366) andHub1 (Abstr., 1884, 1435), the author has multiplied the bromineabsorptions obtained by Mills by -, so as to obtain the equivalentiodine absorptions, and has compared the results with * the experi-mental numbers for iodine absorptions obtained by Hiibl.Thefigures which are tabulated in the original paper indicate that thedrying oils (containing linoleic acid) assimilate the largest proportionsof the haloyds, alzd their capacity in this respect might probably beemployed as a measure of their drying properties. The fish liver oils,however, fully eqiial the vegetable oils in their assimilating power forhaloids. Hiibl's results in the main confirm those of Mills.1278ANALYTICAL CHEJlIST RY. 89Va7enfa’s Acetic Acid Test.-The author has tried this method(Abstr., 1884, 1018) on a number of oils and finds that a slightvariation in the strength or proportion of the acid employed is not ofimportance, and that the temperature at which turbidity occurs withany particular specimen is readily observed and fairly constant.Concordant results have also been obtained from several samples ofbutter, and it appears probable that further experience may prove themethod to afford a simple means of distinguishing butter frombut terin e.Deferininnfion of Glycerol.-The difficulties attending the deter-mitiation of the glycerol produced by the saponification of fixed oilshave recently been orercome by a method originally suggested byW a n k l p and Pox (Abstr, 1886, 395), and perfected by Benediktand Zsigmondy.It depends on the saponification of the oil, andoxidation of the glycerol thus formed by potassium permanganate inalkaline solution, with formation of oxalic acid, carbonic anhydride,and water.The excess of permanganate is then destroyed by asdphite, the solution filtered, the filtrate acidified with acetic acid,and pmcipitated with a calcium salt. As the precipitate containscalcium sulphate and silicic acid in addition to calciuni oxalate, theamourit of oxalic acid is determined either by titration of the pre-cipitate with permnnganate in acid solution, or by estimating thealkalinity of the ignited precipitate. For the determination ofglycerol in fats, the author recomniends modifying the method in thejollowing manner :-5 grams of the sample of fixed oil is placed ina six-ounce bottle, together with a solution of 2 grams of causticpotrsh in 12 C.C. of water. The bottle is secnrely closed and heatedin a water-oven or in boiling water for 8 or 10 hours, the contentsbeing frequently agitated.When the product is perfectly homo-geneous and all oily globules have disappeared, tbe bottle is opened,and the soap diluted with hot water, when a perfectly clear solutionshould be obtained, except in cases of sperm oil, wax, and other sub-stances yielding insoluble alcohols on saponification. The soap solutionis then treated with a moderate excess of acid in the usual way, andthe liberated fatty acids are separated from the aqueous liquid con-taining the glycerol, ahich latter is then ready for oxidation witha1 kaline permanganate as above described.By C. J. ELLIS (J. SOC. Chenz. Id., 5,150-152 and 361-362).-The author has made some experimentswith the view of extending the application of Maumenk’s test to dryingoils and fish oils, to which it caniiot be directly applied without someslight modification, owing to the violent action which ensues whenthese oils are mixed with concentrated sulphuric acid.To overcomethis difficulty i t was found necessary to mix with a drying or fish oilsome liquid which will moderate the action of the acid on the oil.The author employed a mineral lubricatiiig oil of 0.915 sp. gr. forthis purpose, and as on mixing sulphuric acid with such an oil acertain increase of temperature takes place, it is necessary to deter-mine the rise due to each gram of the mineral oil. To accelerate theaction of sulphuric acid on the mineral oil, a certain proportion ofI).B.Maumen6’s Test for Oils90 ABSTRACTS OF CHEMICAL PAPERS.colza oil was added, for which the standard number, when not mixed,was accurately determined and found to be 35.8”. Contrary to ex-pectation, it was found that the smaller the quantity of mineral oil iiithe mixture the greater is the value representing the rise due to eachgram of the mineral oil, providing the rise due t o each gram of thevegetable oil remains constant whatever the mixture. To calculatethe rise in temperature due to each gram of the mineral oil thefollowing formula is employed :--y = a + bz, iii which -y representsthe rise in tempeiature due to each gram of mineral oil, x is thefraction of the mixture coiisisting of mineral oil, and cc and b areconstants depending on the conditions of the experiment and theparticular mineral oil employed.In order to obtain the most concordant and trustworthy results, themaximum temperature attained should not exceed 60”, and it is forthis reason that the author prefers the use of a, mineral oil as the re-tarding reagent.D. B.Employment of Congo-red in Titrating Aniline. By P.JULIUS (C‘hein. I d , 9, 109-110).-CCongo-red, the compound oftetrazodiphenyl with naphtholsnlphonic acids, is turned blue by acids,and recovers its red colour with an excess of alkali. When used asan indicator i n titrating aniline or its homologues with a mineralacid, the point is taken at which a bluish-violet, not changed bysmall further additions OE acid, is produced. A much larger excess isrequired to produce a pure blue.The results do not vary more than0.2 per cent. from the theoretical numbers.Hufner’s Method of Estimating Urea. Ry E. PFL~GER andK. BOHLAND (P’cger’s Archiv, 39, l-l7).-Pfliiger and Schenk pre-viously proved (ibid., 38, 385) that Hiifner’s method of estimatingthe nitrogen in urine gave results which were too low and variable tofound a calculation of the total nitrogen on. Owing to the improve-ment made in Bunsen’s method, the authors have been able toascertain whether Hufner’s method was sufficiently accurate for thedetermination of urea only. They find that the results are alwaystoo high and also very variable. The variatipn ranges from 1 percent. to 10 per cent., and does not therefore admit of compensation.J.P. L.M. J. S.Estimation of Urea in Human Urine with Sodium Hypo-bromite. By E. PFL~~GER and K. BOHLAND (Pfliiger’s Archiv, 39,143-158) .-Pfluger’s new method of estimating urea by hypobromite(PJliiger’s rlrchiv, 38, 503) can be applied to the estimation of urea inliunian urine provided the nitrogenous extractives are first removed.J?or this purpose, a given volume of urine acidified with hydrochloricacid (1 of acid to 10 of urine) is precipitated with sufficient phospho-tungstic acid to ensure the separation of all the extractives, themixture made up to a known volume, and allowed to remain 24 hoursa t least previous to filtration. The acid filtrate is carefully neutralisedwith powdered lime, and again filtered through a dry filter.Great stress is laid on the use of pure soda and bromine for thASALPTICAL CHEJIISTRT.91preparation of the hypobromite.of urine by this process was + 1.3 per cent.The mean error of several analysesJ. P. L.Qualitative Tests for the Dyes Found in Commerce. By 9.N. WITT (Chem. Ind., 9, 1--7).-A table of reactions for the identi-fication of about 80 artificial colouring matters taken singly. Manycommercial dyes are mixtures : in this case, the powder strewn uponwet filter-paper or colourless sulphuric acid will generally givestreaks of more than on6 colour. Where the mixture is more intimatea solution must be made, and the colouring matters withdrawn frac-tionally by dyeing small pellets of wool or silk in it. The principaladulterant for azo-dyes is sodium sulphate.It is best detected afterprecipitating tlie colour by pure sodium chloride. M. J. S.Detection of Artificially Coloured Red Wine (Clarat). By J.HERZ (Chem. Zeit., 10, 968-969 ; 998).-To 30-50 C.C. of the wine,or if the quantity of colouring matter in the wine is small, 100 C.C.concentrated to 30 c.c., 20-30 C.C. of a saturated solution of magne-sium sulphate, and 10-20 C.C. of soda solution are added, stirringwell; if necessary the treatment is repeated until the liquid iscolourless, or nearly so. The filtrate is made acid with dilut,e sul-phuric acid (1 : 3), and if sulphonic acid colours are present the redcolour reappears. The most commonly used member of this group,acid-magenta (rosanilinesulphonic acid), yields a, violet-red solution,and can be estimated by comparing the tint with magenta solutionsof known strength.One mgram. of magenta per litre can be distinctlydetected in 30 C.C. of wine without previous concentration. Whenarchil (orseille) colours are present, the filtrate is bluish, and whenmade acid turns a litmus-red colour. To test for magenta under suchcircumstances, Blarez’ method of shaking with lead dioxide is used ;$his destroys the orseille and natural colour. Cazeneuve’s method isnot recommended. To test for other colours in the magnesiumhydroxide precipitate, the gelatinous mass is stirred up with hotwater, allowed to settle, and the liquid decanted off. I f only thenatural colour of the wine is present, or bilberry has been used, thisliquid is yellow-brown ; if archil has been used, dark-violet,; ifyonceau, onion or ponceau red ; if casskshe, pale-red or dark-yellow ;if cinicoline bcrdelaise, a yellow-red to yellow-brown liquid, whichwhen poured on sulphuric acid gives a violet ring. By shaking thecoloured liquid with ainyl alcohol, Tonceau yields an onion-red residue ;vinicoline, a dark-brown one ; ccmwsine, a dirt-y-green, violet at theedge, turned yellow by strong hydrochloric acid. The precipitate isa dark-grey or brownish-grey colour when the natural or vegetablecolours only are present; with archil, it is violet; with magenta(acid o r ordinary), dirty white ; with cassissine, dirty yellow-brown ;with vinicoline, crimson-red. The precipitate is mixed with sand,dried, and extracted with ether ; the extract contains any ordinarymagenta which can be identified in the usual manner by dyeing wool,or cassissine which dyes wool red-brown and leaves a yellow-brownresidue in the dish. The dyed wool becomes yellow when treatedwith strong hydrochloric acid and colourless with ammonia. Whe92 ABSIL’RACTS OF CHEMICAL PAPERS.wine is shaken with amyl alcohol, and the coloured extract evaporated,the residue, if it contains tho gubstances named, behaves in themanner described below :-Wilh concentmtedc-A 7H:S0,. HCI. NaHO.Archil . . . . . . . . violet-red bl ue red bl 11 eBordeaux, B.. . . oarmine carmine carmine carminePonceau, RRR.. dark-red crimson crimson brownCassissirie.. . . . . violet-purple yellow yellow- redVinicoline Bor-browndelaise . . . . . . cherry-red brown red brownwhilst the wine after extraction i~ cherry-red with ordinary maqefcta,violet-red with acihmngenfu, dark-cherry with Bordeam, yellow-redwith p o i z c m ~ . Wine coloured with magenta produces a violet froth.The detection of vegetable colouring matters in presence of the naturalcolour of wine or otherwise is a matter of great difficultg, and mostof the known metliods are ineffectual ; it is, however, effected by theauthorwith comparative facility in the following manner :-I0 15 C.C.of wine is shaken with 5 C.C. of a saturated solutiop of tartar emetic,and then examined by reflected and transmitted light either a t onceor, if no immediate chanqe has taken place, after some time. Thistreatmelit produces with genuine red wine always a cherry. red colour,and with other substances as follows :-Red-poppy (Papaver r?iwus),dark cherry-red ; cherry, violet : commercial elder colonring matter,Ed-violet ; hilberry (Vaccin. niyrtill), blue-violet ; priret-berry, piireviolet. White wines artificially colonred, and red wines mixed withartificial colours have been successfully examined in this manner ; inthe latter case the wine some time after treatment is compared with agenuine red wine to distinguish more readily the change of colour.Old solutions of privet do not give the d o u r change. Sodiumhydrogen carbonate produces with pure wine, brown-red ; with winecoloured with pure eldedwry, grey-violet ; and with hilberry, brown-green. Tartar emetic appcars to form an antimony lake with thocolouring matters. With practice, all the above-mentioned coloum cartbe detected in 30-50 C.C. of wine. In the subsequent communication(loc. cit., 998), the author acknowledges the priority of Ambiihl andElsner’s recommendation of the use of tartar emetic for the purposein question. They, homver, recommend hot solutJions ; the authorfinds cold better. Fermented bilberries give the violet colour evenbetter than unfermented berries, especially when fresh, innsrriuch asoxidation interl’eree with t,he delicacy after a time. The distinctnessof this colour is increased by diluting the wine. D. A. L