ANALYTICAL CHEMISTRY. 61A n a l y t i c a 1 C h e m i s t r y .Specific Gravity of Liquids. By L. SIEBOLD (Analyst, 1879,189).-From experiments carried out by the author, i t is clearlyshown that hydrometers afford reliable indications of the specificgravity of liquids, no matter whether their gravity is due to substancedissolved or in suspension. L. T. 0's.Analyses of Organic Compounds containing Fluorine andBoron. By $4. LANDOLPH (Ber., 12, 1586--1588).-1n the determina-tion of carbon and hydrogen in such compounds, the author recom-mends the use of fused lead chromate, which is placed before thecopper oxide and only heated gently, as otherwise the boric acid isvolatilised. To determine the fluorine and boron, the compound isdecomposed by a solution of calcium chloride.The fluorine is thusseparated as calcium fluoride, and the boric acid remaining in solutionis determined as magnesium borate. P. P. B.Direct Separation of Manganese from Iron. By F. BEILSTEINand L. JAWEIN (Bey., 12, 1528--1531).-The author describes twoprocesses, both of which are preferable t o the ordinary method ofseparating the iron as basic acetate. The first depends on the factthat all the manganese is precipitated as peroxide or sesquioxide froma solution of manganocyanide of potassium, on addition of iodine,whereas no precipitate is produced in potassiuni ferrocyanide byiodine. The details are as follows : The solution of ferric and man-ganous salt is poured into excess of concentratcd solution of potassiumcyanide. A minute insoluble residue always remains, which containsonly iron ; it is removed by filtration.Iodine is then added until allthe cyanide has been decomposed, and the slight excess is removed byaddition of a few drops of soda. The precipitated oxide of manganeseis filtered off, washed, and dissolved in hydrochloric acid, and esti-mated as sulphide. The only disadvantage of the process is the,largeamount, of iodine required (about 30 grams), but as it can be nearlyall recovered by addition of crude nitric acid to the filtrate fromthe manganese precipitate, this inconvenience is removed.The second process depends on the conversion of salts of manganeseinto peroxide by boiling with strong nitric acid and potassiumchlorate. The salts are dissolved in nitric acid, sp.gr. 1.35, and afterthe solution has been heated to boiling, potassium chlorate is addeduntil all manganese is precipitated. The liquor is then diluted andfiltered. Tbe precipitate contains iron, but by dissolving it in hydro-chloric acid and repeating the process, it contains only an infinitesimaltrace of iron. Both of these processes arenpplicable to the estimationof manganese in cast-iron and steel."1878, Trans., 269) ,-W. R.W. R.* The second of these procesees has been suggested by Hanmy (this Journal62 ABSTRACTS O F CHEMICAL PAPERS.Estimation of Organic Nitrogen in Natural Waters. By IT.PELLET (Compt. swzd., 89, 523). -The ammonia is estimated byBoussingault's process; the nitric acid, i n three litres of water, bySchloesing's method ; and the total nitrogen by evaporating threelitres of water to dryness, with addition of a small quantity of magne-fiia, mixing with a small quantity of starch, and heating with soda-lime in the ordinary way. The starch converts the nitric acid intoammonia, if the nitric acid does not exceed 0.23 gram of potassiumnitrate.c. w. w.Notes on Some Analyses of Waters. By T. L. PHIPSOK (Chem.News, 40, l).-The author considers that a very long experience isnecessary for a chemist to decide whether a water is fit for drinkingpurposes or not;; other questions such a s its effect in attacking anddissolving lead, or corroding iron pipes or boiler plates have often tobe decided by the chemist.For deciding the questions as to the adaptability of water for drink-ing purposes, mnch stress has been laid upon the proportion of organicmatter, but this is a mistake, because some waters containing as muchas 6 or 8 grains per gallon may be drnnk with impunity, whilst otherscontaining much less are known to be exceedingly injurious, if notfatal.Four or five grains of crenate of ammonia per gallon is not at; allhurtful, whilst putrid organic matter, nnrnerous Bacteria and Micro-coccus and minute white fnngoid growths are sources of imminentdanger.He gives the following as examples of water which he has ana-(1.) Well near Sleaford (Lincolnshire). -Water not quite clear,slightly alkaline with decided saline taste, and well a6rated with airand carbonic acid ; contains some minute green alge ; total residue,169 grains per gallon, which is composed of :-lysed :--Organicmatter. NaCJ.Na,CO, Na2S04. X2S04. MgC1, Si02. Fe203. CaC03.2-0 76.0 44.0 35.0 2.0 1.5 1.0 0.5 7.0Total, 169 grains. There were traces of phosphoric acid and ofbromine.(2.) St. Anme's Well, Bztxton (Derbyshire), contains mineralmatter 18 grains, organic matter 2.0; total, 20 grains per gallon.The mineral matter is composed chiefly of calcium carbonate andsodium chloride, with a little caleium sulphate and traces of iron,silica, czesium, and strontium, but no lithium o r rubidium. Thewater is beautifully dear and tasteless, and is said to have a constanttemperature of 80" to 82" F., sp. gr. at GOo 1.003. The fact thatthis water cures g o d is owing probably to its great purity, and to itsbeing drunk warm and in large quantities.( 3 .) Well o n Wirnbledon Common (Surrey), contains mineralmatter 26 grains, organic matter and nitric acid 6 grains; total, 3ANALYTICAL CHEMISTRY. 63grains per gallon. The mineral matter is composed principally ofcalcium carbonate and sulphate, with a small proportion of alkalinesalts. It iswell aarated, and contains no phosphoric acid. A single dropof a very dilute solution of potassium permanganate gave a rose tintto 200 C.C. of the water, which persisted for several hours. This is anexample of a good well water.(4.) Well in the Lotuer Bagshot Sand, near Esher (Surrey).-Thewell is 40 feet deep, and is sitxated about 40 feet from a small ceme-tery.The water is beautifully bright,, clear, and odourless. Itattacks and dissolves lead easily, and shows decided indications ofnitrates and ranch chlorides. It contains nitric acid and organicmatter 7.0 grains, sodium chloride 14.0, sulphates, carbonates, &c., 37.8 ;total, 58.8 grains per gallon. A very deceitful water; certainly im-pregnated and likely to get worse, sp. gr. 1.0032. A spring muchfarther from the little cemetery gave nitric acid and organic matter3 grains, mineral matter 21 grains ; total, 24 grains per gallon. Thiswater dissolves lead easily.(5.) A Yellow Water (South of EngZand), supposed to be ferrugi-nous, remains clear even on boiling, but gives off a strong marshyodour. Total residue, 2.1 grains per gallon, consists principally ofthe ulmates of lime and ammonia, a little carbonate of lime, and tracesof chlorides, &c.(6.) Well at Midland Bank, Birmingham, contains mineral matter(after calcination) 58.71 grains ; total residue, 81.62 grains per gallon.This water contains a very large amoiint of nitrates and ammonia.Itis a bad water for household use, and it is said to corrode metals.(7.) Welt in, an Brt9ciul Mcc.liure Mawufactory near 8oufhanyton.-The water contains free sulphuric acid 1500 grains, phosphates, cal-cium sulphate, alkaline salts, &c., 18.20 grains.(8.) Welt at Albuny B~l,rracl~s, London.-Organic matter and nitricacid 8 grains, mineral matter ‘72 grains per gallon. Supposed to havecaused an outbreak of typhoid.(9.) Wdl near Huntingdon, contained calcium sulphate 35-89 grains,calcium carbonate 15.37, sodium chloride 16.00, organic matter andnitric acid 5.00, silica, magnesia, oxide of iron, &c., 8.74.(10.) Water f~orn n Scul1er.y Pwnp in Bolton Street, Piccadi1ly.-Total residue, 1024 grains per gallon.It contained abundance ofphosphates, resembIFd dilute urine, and was said to have caused sicknessand diarrhcea.(11.) We& at Putney, X.W.-The total residue varies from 38 to120 grains, and some containing from 38 to 48 grains, of which 7 or 8grains are composed of organic matter and nitric acid, have been usedfor many years for drinking purposes without having produced anybad effects. Others that yield 60 gmins of total residue, of which 10grains are composed of organic matter and nitric acid, have been pro-scribed by the medical authorities.Evidence is quoted from ananalysis made by the late Dr. D. Thomson to show that althoughthese waters are highly contaminated, they have not changed in com-position for 25 years.The author generalises as follows :64 ABSTRACTS OF CHEMICAL PAPERS.(1.) The depth of a well has no influence on the quantity of solid(2.) The purer a water, the more easily does it dissolve lead.(3.) Boiler deposits from all parts of the world with a few ex-ceptions, consist almost entirely (over 90 per cent.) of calcium car-bonate.residue which a water contains.(4.) The presence of phosphoric acid is always a bad indication.W. T.Rapid Estimation of Pure Sugar in Raw and Refined Com-mercial Sugars.By P. CASAMAJOK (ClLem. News, 40, 74-76 ; 97-98 ; 107 and 131).-h Payen's process, two alcoholic solutions satu-rated m7ith sugar are used, and filially absolute alcohol, t o wash out thelast traces of the sugar-saturated solutions. The first solution isobt'ained by taking alcohol of 85 per cent. and adding to this 5 percent. of strong acetic acid ; this mixture is saturated with sugar. Thisaddition of acetic acid was made in order to decompose the sucrates,which were a great nuisance to chemists in former days, but in addi-tion, it seems to make the mixture better able to remove the impuritiesof gummy sugar. This first of Payen's solutions was the one adoptedby Dumas in a process published several years ago, which, however,lias never been studied by sugar analysts: the author's process isbased on this.Dumas proposed to agitate 100 C.C. of the first, Payeneolution with 50 grams of sugar, filter, and observe the alcohometricdegree corresponding to 15'. For every per cent. of sugar less than100 the solution is said to indicate I per cent. lesd than 74. For sugarshaving 87 per cent. or more pure sugar, the results agree very closelywith those of the saccharometer, even within 0.1 per cent., but forsugars of lower grade the results obtained are not satisfactory.As nearly one-half of the raw sugars which occur in commercestand below 87 per cent., there seemed to be lit8tle use in a processwhich was declared to he inapplicable to sugars of low grade.Theauthor found, however, after tryiug the process several times, that,although the results obtained were mostly nnfavourable, it was im-possible to dismiss it entirely ; for, upon reflecting upon khe results,it was found that many questions arose which required to be solved,and on their solution the author based the hope of modifying thisprocess so as to apply i t to the analysis of cane-sugars of all grades.By employing methyl instead of ethyl alcohol, the author suc-ceeded in obtaining, with an alcohometer, results that agree veryclosely with those of the optical saccharometer, and that with cane-sugars of all classes from the highest to the lowest. After making agreat number of trials, it was found that methyl alcohol of 83.5" ofthe alcohometer (or 87 per cent..), when saturated with sugar, standsat) 77.1".This solution is the one that has given the most accurateresults. It is easily obtained by taking methyl alcohol, standingat 834" br the alcohometer, and saturating it with sugar by the pro-cess which Numa Grav suggested t o Payen. Since the solution isliable to alteration from loss of alcohol, it is best to test it beforeusing it. When the degree is lower than required, it, may he raisedby adding more alcohol. If a certain volume V of alcohol and water,whose alcohometric degree is d is to be raised to D, with stronANALYTICAL CHEMTSTRY. 65alcohol of degree A, if the volume of the latter to be added is called x,we shall hare V d + xh = (V + x)D, whence x = v(' - 'I.Thusto raise 1000 C.C. of alcohol a t 81 to 83-5 with alcohol of 92 per cent.where d = 81, D = 83.5, V = 1000, and A = 92, the volume of alcohol'Oo0 2*5 = 294.1 C.C. If the addition of 92 to be added is, x =of alcohol has been too great, the degree may be diminished by add-ing water very gradually and stirring up the mixture with an excessof sugar. To ascertain the quantity of water the above formula maybe used, but it must be notred that A = 0, and as both numerator anddenominator have become negative quantities, the signs may bechanged when x =Next in importance is the weight of commercial sugar to be takenfor 100 C.C. methyl alcohol solution saturated with sugar. At first anarbitrary quantity may be taken and the result noted, which may becorrected by the following consideration.The lowering of the alco-hometric degree depeiids on the water and the soluble impurities con-tained in the sugar. If a cert,ain weight of sugar is taken, say45 grams, the result by the alcohol process may be 91.5 per cent. ofsugar. If the same sugar is tested by the optical saccharorneter andyields 93 per cent. of sugar, it shows that the alcohol process hasgiven too low a result, and this because the solution was too dense.The first result shows in the sugar 100 - 91.5 = 8.5 of impuritiesand water, whilst, it ought to be 100 - 93 = 7. To obtain 93 thereforea weight must be taken equal t o - 4'5':7 = 37.05 grams.8.5After trjing many experiments with solutions of different strengths,it was found that each solution required a different weight.For thesaturated solution of 77.1" of the alcohometer, which is the standardsolution employed by the author, the weight is 39.6 grams for 100 C.C.of the solution. Instead of using 100 C.C. the author for a long timeused only 50 C.C. To be able to use a cylinder in which this volumewould give indications. aicohometers had to be employed of smalldiameter. For 50 C.C. of standard solution the proper weight is19.8 grams, i.e., half of the one for 100 C.C. This weight was obtainedby calculation. Using this weight with 50 C.C. of standard solu-tion, 15 conseciitive tests of raw and refined sugars were made, theresults obtained showing that the difference between the percentageof pure sugar by the saccharometer a n d that by methyl alcohol wasvery slight, the greatest deviation being 0.7.If the operations aremade a t temperatures different from 15" C. or 60" F. the correctionscan be made by using either of the tables of Gay-Lussac or those forthe instrument of Tralles. Another correction for the variation oftemperature relative to the volume of standard solution to be takenfor a, weight of sugar equal to 19.8 grams is given in t h e table-At 15" C. 20". 25". 30". 35". 40".19.8 grams 19.7 19.6 19.5 19.4 19.3A - D8.5V ( d - D):DVOL. XXXFIIT. 66 ABSTRACTS OF CHEMICAL PAPERS.The following table contains corresponding corrections for methylalcohbl of various strengths saturated with sugar :-Degrees of thealcohometer Degrees of Degree of thebefore saturation saccharometer Grams of sugarsaturation.with sugar. (Ventzke). in 100 C.C.92.5 91.8 1.7 0.4483.5 77.1 13.2 3.4382.7 76.5 - -81.5 75.0 - -Method of procedure im testi.ng.-The sugar to be tested should notbe weighed until everything is ready. The cylinder is filled with thestandard solution to a line indicating 50 c.c., and 19.8 grams of sugarare weighed out. This is transferred to a mortar and the standardsolution poured i n ; the whole is then ground until all lumps andlarge crystals are broken up. The contents of the mortar are nowfiltered into the cylinder and washed out with the filtered solution.The filtered solution is then tested with an alcohometer and a ther-mometer in succession. To the alcohometric degree, corrected fortemperature, is added the difference between 100 and the alcohomctricdegree of the standard solution. This sum represents the percentageof sugar.D. 13.Behaviour of Various Sugars with Fehling’s Solution. ByF. SOXHLET and others ( R i d Centr., 1879, 370).-Soxhlet questionsthe accuracy of the prevailing opinion, that under all circumstances5 mols. of copper are reduced in alkaline solution by one of sugar, andstates that the qnantity of copper reduced varies with the dilution of theFehling7s reagent and the amount of the latter present in excess. Inthe early part of the titration a large excess is present, as is also thecase when the oxide of copper formed is weighed, the liquid stillremaining blue. Soxhlet, in common with the rest, finds it the bestplan to keep two solutions, one of Rochelle salt and soda, and the otherof copper sulphate, a sufficient quantity of each being measured outand mixed before each experiment. When a + per cent.solution ofdextrose was used it was found that from undiluted Fehling’s solution5.05 mols. of cuprous oxide, and from diluted only 4.85 mols., are pre-cipitated by 1 mol. of sugar in titration. Similar differences are seenwhen the gravimetric method is used, 5.5 mols. and 4.85 mols. beingreduced according as the Fehling’s solution was in large excess or onlyjust, so. As the amount of sugar is an unknown quantity, the sameconditions cannot be exactly preserved during each experiment, andSoxhlet is therefore of the opinion that an accurate analysis by thegravimetric method is impossible. On the other hand, Marcker,Behrend, and Morgen hold that if certain conditions are maintainedthroughout, the analysis gives accurate results. They recommendusing the same quantity of Fehling’s solution and the same volume ofliquid in every experiment and calculating the result by means of a nempirical table.Their method is as follows :-25 C.C. of each part ofthe Fehling’s solution is mixed with a certain quantity of sugar sohANALYTICAL CHEMISTRY. 67tion coniaining not more than 0-12 gram dextrose, and the whole madeup with water to 100 C.C. and heated on a water-bath for 20 minutes.The cuprous oxide is then filtered off, washed with 300 C.C. of hotwater, and reduced in hydrogen and weighed.From the various numbers obtained, the authors have compiled thefollowing table, by means of which the amount of sugar may be calcu-lated :-Reduced Cu.mgrms.196194.7188.5182.0175.1167.9160.4Dextrose.mgrms.111.1110105100959085Reduced Cu.mgrms.152.5144.4135.8127.0'117.8108-298.3Dextrose.mgrms.80a57065605550or the amount may be calculated by the formula-u = -19.26 + 2.689 b -0.006764 P,where a is the copper and b the dextrose.The authors consider that by the use of the above table the processgives very satisfactory results.Soxhlet has also found that the quantities of cuprous oxide obtainedby reduction with milk-sugar vary in the same manner as with dex-trose, according to the strength of Fehling's solution employed, from7.4 to 7.67 mols.of copper to 1 of milk-sugar. Rodewald and Tollensmaintain, however, that accurate results are obtainable when certainprecautions are taken, the experiments being all carried out; under thesame conditions of volume. strendh, &c. : under the conditions which 0 , 2 they employ, 1 mol, of milk-sugar reduces 7.47 mols. of copper sulphate.J. K. C.Estimation of AcetyI by Means of Magnesia. By H. SCRIFF(Ber., l2,1531---1533),--This process has an advantage over the use ofsoda, inasmuch as magnesia seldom has a decomposing influence on theproducts of the reaction. Thc magnesia is prepared by precipitating thesulphate or chloride with caustic soda, excess being avoided. 5 gramsof the paste are boiled with 1 to 1.5 grams of the acetyl-deri-oativeand 80-100 C.C.of water for four to six hours in a flask with invertedcondenser. After the reaction is over, the liquid is evaporated to one-third of its volume and filtered. the magnesia is then estimated in thefiltrate by the usual process, and from its amount that of t,he acetyl canbe deduced. W. R.Test for Phenylglyoxylic Acid. By L CLAISEN (Bey., 12,1505).-Concentrated sulphuric acid, added to a solution of phenylglyoxylicacid in benzene, gives a deep red coloration, changing to intense blue-.violet. On addition of water, the colouring matter remains dissolvedin the benzene and may be obtained by evaporation. The amides andethers of this acid, as well as benzoyl cyanide, give the same reaction.f 68 ABSTRACTS OF CHEMICAL PAPERS.Metanitrophenylglyoxylic acid produces a carmine, and orthoni troben-xoyl cyanide a bluish-green colour, analogous to that produced by treat-ing isatin with benzene and mlphuric acid.W. R.Citrate of Iron and Quinine. By F. W. FLETCHER (Anahyst,1879, 191--193).--The author has applied the following modificationof Paul's method for testing quinine to the determination of thequantity and purity of the alkaloi'd in citrate of iron and quinine.20 grams of citrate of iron and quinine are dissolved in 50 C.C. ofwater, and shaken with excess of strong ammonia. The mixture istreated with 25 C.C. of ether, and shaken until the alknlo'id is dissolved ;the two liquids are separated, and the aqueous sclution shaken withether a second and third time.The ethereal washings are mixedtogether and evaporated to the consistency of a paste at the ordinarytemperature, and finally dried at 120". It is then weighed; theweight multiplied by 5 gives the percentage of alkalo'id present. Thealkaloi'd is converted into basic sulphate by adding the requihitequantity of acid. The weight of alkalo'id multiplied by 30.86 gives thenumber of C.C. of decinormal H2S04 required. The liquid is heateduntil all the substaace is dissolved, the solution allowed to cool spon-taneously, and the crystalline mass filtered through calico. Thevolume of the filtrate is taken, and to it 20 C.C. of ether and excess ofammonia are added, and the mixkure well shaken.Itl is then allowedto stand for six hours, when, at the junction of the two liquids, crys-tals of cinchonine and quinidine will be found. These are collected ona weighed filter, dried a t 120", and weighed.The crystalline residue is dried a t 100" and weighed, and the weightmultiplied by 1.18 gives its value as erystallised sulphate of quinine.Iodic Acid as a Test for Morphine. By J. C. BELL (artalyst,1879, lsl).--Todic acid is shown by the author to be most unsatisfac-tory as a distinguishing test'for morphine. Other organic bodies, suchas ipecacuanha and guaiacum, reduce iodic acid with separation ofiodine. And, moreover, the statement that, the colour is not destroyedby ammonia in the case of morphine is incorrect.L.T. 0's.L. T. 0's.Nitric Nitrogen in Guano. By R. R. TATLOCK (Ohem. Nezos, 39,268-270).-The autlior was led by experiments made some years agot o believe that a large proportion, and in some cases nearly the wholeof the nitrogen present in guano as nitrates was converted by thesoda-lime combustion psoctesfi into ammonia, and estimated as such,and the extent of this change he has since found to depend on therelative proportion of the organic matter to the nitrates present.He was surprised to find that it was the practice of chemists oflarge experience in such analyses to determine the ammonia as if thenitrates present were not decomposed ; thus a much larger percentageof that substance would be represented than what really existed.When citrates are heated with soda-lime, no ammonia is produced,but when heated with soda-lime in presence of organic matter am-monia is produced, and its quantity depends on the nature and propor-tion of the organic matter employed.The author experimented witANALYTICAL CHEMISTRY. 69potassium nitrate in presence of different quantities of starch, sugar,camphor, albumin, and wood charcoal, .and the following are some ofthe results obtained :-20 of starch- to 1 of nitrate gave 50.74 per cent; or” the nikric3 of camphor to 1 of nitrate gave . . . . . . 26.38 per cent.1+ of wood charcoal to 1 of nitrate gave.. 11.56nitrogen as ammonia.?,6 of albumin 7, 7 , . . 49-94 ,,6 of sugar 9 7 9 , . . 63.35 ,,30 of sugar ,, ,1 . . 97.40 ,,They vary somewhat, however, even with the same proportions ofthe same organic materials.The author critieises the various processes for estimating the nitricnitrogen in guanos, and concludes that Crum’a (Proc. GlasgowYhiZ.SOC., 1848, 162) is the best, the nitric acid being determined inthe nitrometer as nitric oxide. It sometimes happens, however, thata little free nitrogen is evolved a t the same time, by the action ofthe strong sulphuric acid on nitrogenous organic matters. This canbe determined by introducing a warm solution of ferrous sulphate intothe nitrometer, which absorbs only the nitric oxide present. Theauthor has not yet arrived at a satisfactory solation of the queshion.Tatlock’s results (Chenz. NPUIS, 39, 2gl) are criticised by B.J.Grosjean. He says that he published (ibid., 25, 2Q-5) some results onthis subject, in which he drew attention to the conversion of nitricnitrogen into ammonia by the soda-lime process, but this fact is statedboth in Fresenius’s “ Quantitative Analysis ” and in Church’s“ Laboratory Guide,” The author described encouraging results forthe conversion of all the nitric nitrogen into ammonia by the combus-tion of nitre with sugar and iron filings. His best results wereobtained by mixing the nitre with a caustic alkaline solution in a re-tort, adding iron filings, and distilling the mixture to IL pasty mass,which was allowed to cool, powdered, mixed with soda-lime, and acombustion made to determine the residue of the organic nitrogen.W. T.Perchloric Acid as a Test far Alkaloids.Ey G. FRAUDE(Beg-., 12, 1558-1560) .-Perc*hloric acid of sp. gr. 1.13-1.14 has noaction on quinine, quinidine, cinchonine, cinchoiiidine, morphine,code‘ine, papaverine, vcratrine, caff e‘ine, atropine, nicotine, nor conine.When boiled with brucine, it gives a dark sherry colour, with strych-nine a reddish-yellow, and with aspidospermine an intense red. Iodicanhydride and sulphuric acid give with brncine an intense orange-yellow ; morphine, deep violet, then orange brown ; and curarine,pink. These reactions are suitable as lecture experiments.W. R.Koettstorfer’s Process for Butter Analysis. By G. W. WIGNER(Ancrlyst, 1879, 183).-The author points out that for the analysis oEsamples of genuine butter this process may be used, but i n cases ofdoubt, a complete analysis should be made.L. T. 0’s70 ABSTRACTS OF CHEMICAL PAPERS.Coefficients of Expansion of Butter, Lard, Fats, &c. ByG. W. WIGNER (Analyst, 1879, 183--185).-By comparing the sp. gr.of butter and lard fat, &c., a t different temperatures, the coefficients ofexpansion have been determined.Butter fat between 100" and 212" F., has the coefiicient 0,0434 perdegree F. Between 150" and 193" the coefficient is slightly greaterthan this number, but remains the same for all other temperatures.L a d Fat and Butteri.ne.-Thc coefhients of expansion of these iwobodies are almost identical, that of lard fat being 0.0420 per degree F.Specific Gravities of Fats, Resins, &c. By H. HAGER ( P h a m .J.Trans. [3], 10, 287).-The fat is melted, dropped into a flat vesselcontaining alcohol, in such 9~ manner that the point from which thedrops are allowed to fall is not more than three centimeters distant fromthe surface of the alcohol, and that each drop is allowed to fa,ll on adifferent spot. The fat globules thus deposited are then removed to aliquid, consisting of either alcohol, water, or glycerol, o r mixtures ofthese, until after careful stirring and reduction or increase of thedensitly, by the addition of one or another of the above liquids, the fatglobules are held in equilibrium in any part of the liquid. The sp.gr. of the latter is then determined, and this of course at the sametime represents the sp. gr. of the fat.Many of the following sp. gr.'smay be used as criteria for distinguishing the various bodies investi-gated :-L. T. 0's.Sp. gr. at 15-16' C.0.938-0.940,, several months old . . . . . . . . 0.936-0.9370.924-0.9300-931-0.9320.925-0.9290.937-0*9400*936-0*9380*950-0*9520.945-0.9460.938-0.9391*016-1*0181-014-1.0150.965-9.9660.9640.967-0-9690.959-0 9520.973-0.976Butter fat, clarified by settling . . . . . . . .Artificial butter . . . . . - . . . . . . . . . . . . . .Hog's l a d , fresh . . . . . . . . . . . . . . . . . . . .Beef tallow . . . . . . . . . . . . . . . . . . . . . . . .Sheep's tallow . . . . . . . . . . . . . . . . . . . . .Beef and sheep's tallow, mixed 1 : 1 . . . .Butter of cacao, fresh . . . . . .. . . . . . . . . .Butter and beef tallow, 1 : 1 . . . . . . . . . .Expressed oil of nutmegs . . . . . . . . . . . .Ditto, extracted with CS, . . . . . . . . . . . .Ditto, crystalline . . . . . . . . . . . . . . . . . . . .Stearic acid, melted, and in drops.. . . i.Wax, yellow . . . . . . . . . . . . . . . . . . . . . . . .yellow and resin, 1 : 1 . . . . . , . . . .,, old.. . . . . . . . . . . . . . . .. . . . . 0.940--0.9427 7 very old. . . . . . . . . . . . . .7 7 crystalline . . . . . . . . . . . . . .,, African .. .. .... .. .. .. .. . . . . .. 0.960,,7 9 ,, and paraffin, 1 : 1 . . . . . . . . 0.916-0*9197 9 ,, and yellow ceresin, 2 : 1 . . 0*942-0.943Ceresin,yellow ....................... 0.985-0.928Wax, Japan . , . . . . . .. . . . . . . . . . . . . . . . 0.977-0.978,, ,, very old . . . . . . . . .. . . . , . . 0.968-0.970,, white, very old and true . . . . . . . . 0.963-0.964J J 0.916-0.925 ,, new . . . . . . . . . . . . . . . . . . . ANALYTICAL CHEMISTRY. 71Sp. gr. at 15-16' C.Wax, Japan, new, and stearicacid, 1 : 1 . .Wax, sp. g:. 0.963, and stearic acid,Ceresin, very white, pure ............ 0*905-0*9080.945sp. gr. 0.9b3, mixed, 1: 1 . . . . . . . . . . 0.975 .. whit,e ...................... 0.923-0-924Araucaria wax ......................Resin (fir. pine), yellow, transparent . .,, whitish, opaque.. ..............Shellac, lightl-coloured. ...............,, darker.. . . . . . . . . . . . . . . :. ....Dammar, old.. ......................Benzo'in, Siam ......................Guaiac resin, pure ..................Copal, East and West Indian..........,, Pennng ....................,, Borneo ....................Amber ............................Sand arac ..........................Mastic ............................Balsam of tolu, old brittle ............0.9901*083-1.0841*044-1*0471.11 5-1*1141.12'31.0751.2351*063-1*8001 -14 5-1 $1 551*165-1.1701.236-1 23 71*074-1*0941 '0.38 -1.0441.056-1.0601.231-1 *232D. B.Testing Drugs. By L. SIEBOLD (AnaZyst, 1879, 190--191).-Themethod for the detection of mineral adulteration in flour by means ofchloroform (C. Himly, Year Book of Pharmacy, 1877) may be ap-plied for the same purposc to drugs. The powdered drug is shakenwith chloroform when the mineral matter sinks to the bottom, and inthe cases of acacia, tragacanth, starches, myrrh, Rarbadoes aloes,jalap, saffron, cinchonas, nux vomica, mustard, white pepper, capsi-cum, and guarana, the drugs float on the top. By pouring the chloro-form off, the lower stratum of mineral matter may be collected andweighed.I n some cases, however, such as gamboge, scammony, opium, Socotrinealoes, liquorice root, ginger, colocynth, coussa, ipecacuanha, cinnamon,and cardamoms, a portion of the drug sinks with the mineral matter.The test may, however, be applied qualitatively, since adulterationmay be detected by a careful inspection of the sediment.L. T. 07s.Testing Malt. By W. SCHULTZE (Ried. Centr., 1879, 375-377).--Malt is usnally mashed a t from 70" to 75" C. : the author finds,however, that the yield obtained at this temperature is always smallerthan when the mashing takes place a t 60", 65", or 70". The extract is,however, much inore quickly prodiiced at the former temperature,only 20 minutes being required at 70" as against 18.5 hours at 60".No more extract is obtained after the starch has been converted intomaltose and dextrin, and it is therefore unnecessary to continue themashing longer. J. I(. C