84 ABSTRACTS OF CHEMICAL PAPERS. A n a l y t i c a l Chemistry. Thioacetic acid as a Substitute for Hydrogen Sulphide in Qualitative Analysis. By R. SCHIFE' and N. P. TARUGI (Be?.., 1894, 27, 343'7-3439) .-Excellent results may be obtained by employing thioacetic acid instead of a current of hydrogen sulphide in qualita- tive analysis. About 2 C.C. of a 30 per cent. solution of ammonium thioacetate is simply added to the solution, made acid with hydro- chloric acid in the usual way, and the liquid boiled. The reactions given by the metals appear to be identical with those obtained with hydrogen eulphide, but the time required for precipitation is greatly lessened. The only products of decomposition of the acid in the presence of hydrochloric acid are ammonium chloride, acetic acid, and hydrogen sulphide, so that the separation of the metals remaining uiiprecipitated is not interfered with.The reagent has been used for some time, and with great success, in the laboratories of the University of Pisa. A. H. New Applications of Alkalimetry and Acidimetry. By E'. HUNUESHAGEN (Chem. Zeit., 1894, 18, 505-506, 547) .-&timation of the Hardness of Water.-The reagents wanted are a, solution of 3.786 grams of sodium carbonate in 1 litre of water and dilute nitric or hydrochloric acid of corresponding strength. 200 C.C. of the sample i s titrated directly with this acid, tincture of cochinelle serving as indicator. Each C.C. of acid shows 1" of temporary hardness (German scale). The permanent hardness is estimated as follows. 200 C.C.of the sample is mixed with a moderate excess of the sodium carbonate solution, evaporated to dryness, and the residue, after being moistened with a little water, is once more evaporated and heated to about 200". It is now dissolved in a little water, the liquidANALYTICAL CHEMISTRY. 83 filtered, and the insoluble matter washed with about 50 C.C. of water. The filtrate is then titrated as before. The Dumber of C.C. of acid deducted from the C.C. of added soda represents the permanent hardness. The total hardness may be estimated directly by using the resulting liquid from the estimation of the temporary hardness for the determination of the permanent hardness. Volumetric Estimation of Phosphoric acid by Titraliizg the Ye1 low Precip*tate.-The liquid should not contain more than 0.05 gram of phosphoric anhydride.If free hydrochloric or sulphnric acid is present, the phosphate is precipitated with a little ammonia and at once redissolved in nitric acid. After adding some ammonium nitrate and a little ammonium sulphate, the liquid is heated, prccipi- tated with a slight excess of molybdatc solution, and the precipitate collected and washed with a 1 per cent. solution of potassium nitrate, iuitil the filtrate is practically neutral. The filter is then put back into the beaker, from 10 to 50 C.C. of water is added, and, after adding a few drops of 1 per cent. solution of phenolphthaleln, the mixture is t itraked with normal alkali uutil the precipitate has all dissolved, and the liquid turned permanently red. It is advisable t o add a slight excess of the alkali, and then to tiirate back with normal acid.1 O.C. of normal alkali = 0*003077 gram of phosphoric anhydride. Acidimetric Estimation of I',urhgstic acid.-T he acid is separated in the usual manner, without troubling about any silica which may be present. The precipitate, after being washed, first by decantation, xiid then on the filter, with a 5-10 per cent. solution of potassium nitrate, is mixed with water, heated to boiling, and dissolved in excess of normal soda. The excess of alkali is titrated back with normal acid, phenolphthaleln serving as indicator; 1 C.C. of normal alkali = 0.116 gram of tungstic anhydride. L. DE K. Estimation of small Quantities of Chlorine in Fats. By R. BENEDIKT and H. ZIKES (Chem. Zeit., 1894,18, 640--641).-A tube, made of hard glass, 100-110 cm.long and 11-12 mm. wide, is drawn out funnel shaped at one end and bent at an angle of 45" upwards. The funnel is provided with a doubly perforated india-rubber cork, through which passes a narrow bent tube for admitting air or carbonic anhydride, whilst the other hole admits a tube connected with a calibrated 100 C.C. reservoir containing the oil. The tube is made ready for use by introducing in the middle two separate 15-20 cm. layers of lime held together by plugs of asbestos. The empty space between the bend and the first plug of asbestos is tilled up with broken pieces of porcelain. The tube is now laid in the combustion f timace, which must extend slightly beyond the second layer of lime. The other opening is closed with a cork through which passes a bent tube connected with a test-tube to collect any t a r ; another narrow tube carries off the gases, which iiiust be lighted, and which indicate the progress of the combustion.When the lime is red hot, the oil is very slowly admitted, n very slow current of carbonic anhydride being also passed through. When, after 3 to 5 hours, about 25-30 grams of oil has entered the tube, the supply is stopped, and, when the flame has gone out, the tube is86 ABSTRACTS OF CHEMICAL PAPERS. allowed to cool; it is then emptied, and the contents examined for chlorine in the usual manner, The test analyses are very satis- Estimation of Organic Nitrogen by Stock’s Method. By C. E. ZAY (Xtaz. Spw. Agrar., 1894, 26, 22--31).-The nitrogen of several substances was determined by Kjeldahl’s method (employing 20 C.C.of sulphnric acid containing 200 grams of phosphoric an- hydride per litre) and by Stock’s modification--using manganese dioxide (5 grams) with (1) sulphuric acid alone, and (2) with sulphuric acid containing phosphoric anhydride (20 C.C. per litre). Concordant results were obtained, and the employment of manganese dioxide effected a very great saving of time in the decomposition process. factory. L. DE K. N. H. J. 31. Volumetric Estimation of Nitric acid. By D. MONN~ER and H. AURIOL (Chem. Centr., 1894, 24, 1095-1096; from Arch. sci. phys., Genive, 31, 352--358).-The process is based on the well- known principle that nitrates are reduced by nascent hydrogen. Sodium amalgam, prepared by adding 1 part of sodium to 100 parts of hot mercury, is allowed to act on the solution of the nitrate in the presence of tartaric acid.The hydrogen is collected and measured, and a check experiment is made with the same amount of amalgam without the nitrate. The difference of hydrogen evolved is the measure for calculating the amount of nitric acid. L. DE I(. Genuineness of Basic Slag. By E’. SESTINI (Staz. Spey. 4 g ~ a r . , 1894, 26, 57--62).-Sorne samples of slag which the author had occasion to examine contained about 16 per cent. of total phosphoric anhydride and 4.4 to 5.8 per cent. soluble in dilute acetic acid (1 : 2), whilst the sp. gr., the amounts of water and of lime were about the same as are found in the case of normal slag. The results ob- tained indicate that the small amount of soluble phosphoric acid is not due to adulteration, but is connected with the amounts of ferric and ferrous oxide and alumina (about 3 per cent.) soluble in the same dilute acetic acid.After discussing Wrampelmeyer’s methods for the examination of basic slag (Abstr., 1894, ii, 119), the author points out that it is sometimes more difficult than was supposed to distinguish between an adulterated slag and one which owes its sparing solubility to its mode of production. The amount of ferrous oxide which a slag con- tains is of interest, as it does not occur in natural phosphates, and the different portions which can be separated from each other by levigation should be examined. N. H. J. 31. Estimation of Carbon in Steel.By H. KOCH (Chem. Zeit., 1894,18, 485).-Air contained in a gasometer is first passed through a washbottle filled with sulphuric acid, then through two bottles con- taining solid potash ; then cornea an empty bottle, and, finally, one filled with sulphuric acid. The purified air then passes through the lower tube of the combustion flask, where it removes t,he carbonic anhydride and any hydrocarbons formed during the process, andANALYTICAL CHEMISTRY. 87 it escapes through the upper tube. The gases are dried in a calcium chloride tube and comuletely burnt by passing them through a, combustion tube contiin- ing red hot copper oxide. The re- sulting carbonic anhydride is first dried by passing it over calcium chlo- ride and phosphoric anhydride, and then absorbed in a weighed potash apparatus to which is sealed a tube containing phosphoric anhydride.When required for the analysis of steel, the apparatus and combustion tube must first be completely freed from any carbonic acid or carbonaceous matter ; the copper oxide is therefo1.e heated to redness, a cnrrent of puri- fied air being transmitted for some time. Into the flask is introduced 8 granis of chromic acid and 120 C.C. of dilute sulphuric acid (4 : S) and the mixture boiled out. Afterwards 2 grams of the steel borings is intro- duced into the flask, the apparatus is coiinected with the combustion tube, and the operation conducted as usual. It will be noticed that the apparatus P i 4mi 4. , I V , I I I I I I i 3 0 m 1 I I I I I 16c.m 2 2 m Height of flask, 30 cm.Contents, 250 C.C. Width of ground neck, 4 cm. consists of two essential parig the flask proper and a cooling contri- vance. L. DE K. Sodium Cobalt Nitrite as a Reagent for Potassium. By J. VAN EYK (Chem. Centr., 1894, i, 1162; from Nederl. Tijdsch. Piiarrn.: 6, 1:36--139).-The sodium cobalt nitrite is prepared b y adding 100 C.C. of a 60 per cent. solution of sodium nitrite to a solution of cobalt nitrate (30 grams) iii water (60 grams). After the evolution of nitric oxide has ceased, the mixture is filtered, and the saltl pre- cipitated with alcohol. By means of this salt, potassium can be detected when diluted to the extent of 1 : 10000. Ammonium salts are also precipitated by the reagent, but not until the concentration reaches 1 in 2000. E. C. R. Assay of Nitre. By A.HELLICH (Chem. Zeit., 1894, 18, 485- 486) .-The author has noticed that commercial nitre often contains perceptible quantities of perchlorates, and consequently should be tested for this impurity. Its presence, which is difficult to account for, has until now been ignored in the various schemes for nitre analysis. L. DE K. Estimation of Mercury in the Presence of Iodine. By FRANGOIS (J. Pharm., 1894, [S], 30, 249--254).-The estimation of mercury when combined in any form with iodine is easily effected by electrolysis. The salt need not be dissolved, but merely placed in the88 ABSTRACTS OF CHEMICAL PAPERS. clectrolysing vessel with the electrolyte. The latter may be dilute sulphuric mid, but the author finds the best electroljte €or this purpose to be formed by dissolving 20 grams of pure ammonium nitrate in concentrated pure ammonia, and making up with the latter to 100 C.C.The electrolysis is carried out in a platinum crucible of which the bottom is connected with the negative pole of the battery (two Bunseii cells). The positive pole is formed by a platinum yod, of which only about 1 mm. dips into the liquid. The mercury is deposifed on the bottom of the crucible. When the action is finished, the solution is removed with a pipette, and the mercury is washed first with water and finally with alcohol, dried, and weighed with the crucible. The iodine in the solution is reduced with sulphurous acid, and estimated as silver iodide. This process may also be used for estimating mercury in the presence of bromine or chlorine. L.T. T. Qualitative Separation of Chromium from Iron and Alumi- nium. By R. B. RIGGS (Amer. J. Xci., 1894, 287, 409--411).-The mixed hydroxides or basic acetates containing about 0.1 gram of each metal are digested in 100 C.C. of water to which 10 C.C. of hydrogen peroxide and 1 gram of sodium hydroxide has been added, until effervescence ceases. After filtering off the iron, the liquid is slightly acidified with acetic acid, and the aluminium precipitated by am- monia. If the filtrate is yellow, this is a sure sign of the presence of chromium, which may, however, be confirmed, for instance, by the well-known hydrogen peroxide test. L. DE K. Apparatus for the Assay of Pyrolusite by Bunsen’s Process. By C. ULLMANN (Chew. Zeit., 1894, 18, 487).-A beaker filled with the requisite amount of solution of potassium iodide is placed on a block of wood of about the same height.In the beaker is placed a special apparatus to collect the chlorine, a kind of upright glass cou- denser, the inner tube of which is drawn out like a pipette, and the top of which is provided with a stopcock and funnel. The tube is filled by drawing out the air through the funnel and then closing the stopcock. The flask containing the sample and R sufficiency of hydro- chloric acid is now connected with a doubly-bent delivery tube pro- vided with a bulb. Heat is applied, and after the bulk of the air has coilected in the apparatus, the chlorine enters, and is, of course, rapidly absorbed by thepotassium iodide. When all the chlorine has been boiled off, the stopcock is opened, and the beaker is placed on the table.After rinsing the apparatus and delivering tube, the liberated iodine is titrated as usual with thiosulphate. L. DIE K. Standardising Potassium Permanganate. By Miss C. F. ROBERTS (Amel.. J. Sci., 1894, 286, 290 --292).-The author recom- mends standardising potassium permanganate by means of a solution of ferrous chloride, made by dissolving a known quantity of electro- 1s tically - prepared metallic iron. About 10 grams of iron ammonium sulphate is dissolved in 150 C.C.ANALYTICAL CHEMISTRY. 89 of water, 5 C.C. of a satuiaated solution of potassium oxalate is added, and the whole is then heated with a sufficiency of ammonium oxalate niitil a clear solution is obtained. This solution is then decomposed in a beaker between two platinum electrodes, the iron being depo- sited on a weighed piece of platinum foil of a size convenient to bc inserted in a rather large weighing bottle.After about 16 hours with a current of 2 amperes, a sufficient amount of metal will have precipi- tated, when it is washed and dried in t.he usual way. Estimation of Iron in the Ash of Vegetable or Animal Matter. By M. RIPPER (Chem. Zeit., 1894,18,133-134).-The ash is dissolved in strong hydrochloric acid, the solution mixed with a few C.C. of hydrogen peroxide, and evaporated to di-yness on the water bath ; the residue is then just moistened with a few drops of hydi*ochloric acid, dissolved in about 20 C.C. of water, and transferred to a beaker. About 1.5 grams of potassium iodide is added, the beaker covered with a watch glass, and heated for 10 minut'es at 60" ; the liberated iodine is then titrated in the usual manner with N/100 sodium thio- sulphate.The process, which hitherto has only been used for the estimakion of fairly large quantities of iron, only occupies one or two hours, and compares favourably with the gravimetric methods. Electrolytic Separation of Iron and Cobalt from Zinc. By G. VORTMANN (Claem. Ccntr., 1894, i, 877 ; see Abstr., 1894, ii, 34).- In separating iron from zinc by the electrolysis of an alkaline tartrate solution, only a single accumulator, giving a current of 0.07-0.1 ampBre, should be used instead of the two formerly recommended ; it then becomes unnecessary to rsdissolve the iron, since it is deposited free from zinc.The electrolysis should be conimenced in the cold, but towards the end a temperature of 50-60" promotes tahe deposition. The cathode (a disc of platinum, silver, or silvered copper, 50 mm. in diameter) should be from time to time replaced by a new one, and the operation continued as long as any gain in weight takes place. After the removal of the iron from the solution, the zinc is deposited by using two accumulators in series, with an E.M.J?. of 4 volts. Cobalt is also better separated from zinc in an alkaline tartrate solution by using 2 volts than by 4. Addition of potassium iodide diminishes the deposition of cobaltic oxide on the anode, but as this cannot be completely prevented the anode must likewise be weighed. The operation is performed as for iron, but with a warm solution.Separation of Arsenic, Tin, or Antimony from Lead, Copper, Silver, Cadmium, Cobalt, Nickel, &c. By P. JAXXASCH (Uer., 1894, 27, 3335-3336).-Elernents, such as arsenic and tin, the chlo- rides of which are comparatively volatile, may be separated from other metals by dissolving the mixture in nitric acid or aqua regia in a special glass vessel, evaporating, and heating the residue at a suit- able temperature in a current of dry hydrogen chloride, the vessel being placed in a nickel air bath, which can be heated up to 450" ; L. DE K. L. DE K. M. J. 8.90 ABSTRACTS OF CHEBIICAL PAPERS. the more volatile chlcrides distil over, and are collected. the apparatus were given in a previcus paper (Abstr., 1894, ii, 330).Action of Organic Matters on Potassium Permanganate. By A. ZEGA (Chem. Zeit., 1894, 18, 2--3).--The author has proved that t,rustworthy, compa'rable results in the titration of organic matters contained in potable waters may be obtained by operating in the following manner. 50 C.C. of the sample is put into a 100 C.C. flask, mixed with 5 C.C. of the usual permanganate solution, 5 C.C. of dilute sulphuric acid (1 : 2) added, and the mixture heated f o r 20 minutes on a water bath ; the excess of permanganate is then titrated back by oxalic acid. The process is particularly useful when the water contains volatile organic matters. The standard permanganate is checked under the Details of C. F. B. same conditions. L. DE K. Analysis of Petroleums. By A. RICHE and G.HALPHEN (J. Pha~m., 1894, [ 5 ] , 30, 289-300).-The object of this work was to end methods for distingixishing (a) between petroleums of Russian and American origin, and ( b ) between crude oils, and mixtures of refined and residual oils. As a rule, the &ussian oils contain less volatile oils and are denser than the American ; €or fractions of the same boiling point, Russian oils are generally denser than American, the mean sp. gr. of t,he fraction 140-160°, for example, being, f o r Russian oils 0.782, and for American 0.755. The refractive indices vary directly with the sp. gr., and so for fractions of the same boiling point are higher for Russian than for American oils. Russian oils are generally poorer in light oils and richer in heavy oils than American, whilst the latter are richer in solid paraffins, and are often rendered turbid by cold from separation of these substances.The American petroleums are generally almost exclusively composed of saturated hydrocarbons, whilst the heavier portions of the Russian oils generally contain hydrocarbons of the CllHsn series. None of these differences are, however, sufficiently certain to form satisfactory bases for analysis. The authors have found the best test to be the solubility of the oils in a mixture of equal volumes of chloroform and alcohol. 4 grams of the refined petroleum t o be tested (of which the density at 15" has already been determined) is introduced into a flask, and the alcohol-chloroform mixture gradually added from a burette until the liquid, which is at first rendered turbid, again becomes clear; the quantity of solution required is then noted.The fractions chosen should be those of about sp. gr. 0.800 to 0.830, as at these densities the difference is most marked. Thus, for sp. gr. 0.820 American refined oils required from 8 to 11 C.C. of the alcohol-chloroform f o r solution, mihe mean being 9.5 c.c., whilst for Russian oils the numbers were 4-3 to 4.8 c.c., giving a mean of 4 5 C.C. F o r crude oils, a much larger amount of solvent is required; Russian crude oils of sp. gr. 0.851 to 0.877 requiring about 15 c.c., whilst refined oils of the same sp. gr. or mixtures of refined oils with about 10 per cent. of residuals, only required from 4 to 5 C.C. WithANALYTICAL OHEMISTRT. 91 Amel-ican crnde oils, of which the sp.gr. varies from 0.788 to 0.822, the solvent required also amounts to about 15 c.c., whilst with the same oils refined, or mixtures of refined with residuals, from 5 to 7 C.C. only are required. A few of the very light American crude oils (of about 0.785 sp. gr.) require much less of the solvent. F~dl tables and curves of solubility are given in the paper, and a special form of burette for keeping the alcohol-chloroform out of contact with the air is described. L. T. T. Assay of Ethereal Oils. By J. KLIMOXT (Chem. Zeit., 1894, 18, 641-642 ; 672--673).-The author's process is based on the fact that ethereal oils react strongly with bromine, whilst paraffin oil gives scarcely any reaction. For the assay of oil of turpentine, for instance, the following reagents are required.Solution of bromine in chloi~oform of about 1 per cent. strength ; pure chloroform, treated with strong sulphuric acid, washed, and redistilled ; pure tnrpent'ine, made by first washing oil of turpentine with aqueous soda, and after- wards collecting the fraction distilling over at 168-170'. 0.5 C.C. of this is put into a, stoppered 20 C.C. flask and accurately weighed; chloroform is then poured in up to the mark, and the solution put into ;t little burette. 10 C.C. of the bromine solution is introduced into another little flask, and the turpentine solution slowly added until the colonr of the bromine has entirely disappeared. If now a sus- pected sample of turpentine is at once treated in the same way as the pure specimen, its lesser bromine-decolorising power will indi- cate a more or less marked adulteration.The author has tabulated the results of experiments with almost every known ethereal oil, including 9 specimens of refined turpentine and 11 of inferior brands ; also experiments with adulterants, such as resin oil and petroleum. The figures given are not bromine numbers, but represent the equivalent amount of turpentine. L. DE K. Detection of Methylated Spirit in Tinctures, &c. By A. ASHBY (Analyst, 19, 261-271) .-The author has sat,isfied himself that, of the numerous methods proposed, the test with sodium nitro- prusside in the presence of ammonia is the best. The red colour will appear within 10 or 15 minutes. The constituent the reaction is chiefly due to has yet to be ascertained.When examining alcoholic liquids free from solid matter, the test may be applied directly by mixing the sample with an equal bulk of it 1 per cent. solution of sodium nitroprusside and adding a few drops of ammonia; but in the case of officinal tinctures, 25 C.C. of the sample is distilled, and the first 5 C.C. which pass over tested. Ethereal solutions are distilled to dryness, and successive portions of the distil- late are tested; if, however, the sample is very weak in spirit, it is best to add 2 or 3 C.C. of' strong, pure alcohol to the distillate before applying the test. If the not very probable presence of a sulphide be suspected, it is best to add some fixed alkali before distilling. Gravimetric Estimation of Sugar by means of Alkaline Copper Solutions.By L. GRUNHUT (Chew,. Zeit., 1894,18,447- L. DE K.92 ABSTRACTS OF CHEMICAL PAPERS. 448) .-Owing to the great difliculty of completely oxidising cuprous oxide, the results obtained by weighing the copper oxide are often much too low ; weighing as cuprous oxide on a filter has also its dis- advantages. The author strongly recommends the process wherein the cuproEs oxide is collected on a, weighed asbestos filter contained in a glass tube. After first igniting in a current of air to burn off organic matters, the oxide is reduced in a current of dry hydrogen and the residual metal finally weighed. L. DE K. Gravimetric Estimation of Glucose. By F. GAUD (Compf. rend., 119, 478-479).-50 C.C. of the freshly-prepared alkaline copper solution is mixed with an equal volume of water, boiled for :L few minutes in a porcelain dish, and then placed on a water bath, the water in which is boiling; 25 C.C.of the sugar solution, contain- ing about 1 per cent. of glucose, is then added all at once. Re- duction under these circumstances takes place below loo", a con- dition which is essential to prevent the destructive action of the alkali on the glucose. After 10 minutes, reduction being complete, the liquid, which should have a deep blue colour, is decanted off, and the precipitate is washed with boiled water until the washings are no longer alkaline to phenolphthalein. The precipitate is then transferred to a sp. gr. bottle holding 20 to 25 c.c., the capacity of which a t 0" is known, and the bottle is filled up to the mark with boiled water and weighed at a temperature t .Let P be the weight of the liquid and precipitate, the total volumc of which is equal to the capacity of the flask at the temperature f , that is t o say, Vt = V,[1 + 3P(t - to)]. The sp. gr. of dried cuprous oxide is A = 5.881, and the sp. gr., d, of water at the temperature t is known ; then the weight p of cuprous oxide is given by the expres- sion P - V t d p = - 1 - d / A . To obtain perfect results, the weight P should be reduced to a vacuum. The change in the sp. gr. of cuprous oxide caused by ordinary changes of temperature is too small to affect the result. The weight of cuprous oxide is not strictly proportional to thc weight of glucose present, and it is necessary t o prepare a table showing the relation between various values of the two numbers.' h e author obtained the following results. Cuprous oxide. Glucose. Cuprous oxide. Glucose. Milligram& Milligrams. Milligranis. Milligrams. 10 5.413 100 46.22 20 9-761 200 91.047 30 14.197 300 138.842 50 23.036 400 188.928 C. H. B. Modification of the Copper Test for Glucose. By ALLEIN and E'. GAUD (J. Phurm., 1894, [ 5 ] , 30, 305--307).-The authors findANALTTICAL OHPhlISTRY. 93 (this vol., i, 123) that the free potash 01’ soda in Fehling solution causes the decomposition of a part of the glucose to be estimated, and thus causes the lorn results known to be obtained by that method. The following solution gives results free from this error. 8.7916 grams of pure (electrically deposited) copper is dissolved in 93 grams of sulphuric acid, the solution is diluted with its own volume of water, and the whole made up to 1000 C.C.with strong ammonia. A deep blue solution is thus obtained, which is perfectly stable, and of which 10 C.C. corresponds with 0.05 gram of glucose. The estimation is con- ducted in a flask fitted with a triple-bored cork to admit the end of the burette containing the glucose solution and tubes for passing a current of hydrogen. 10 C.C. of the amraonio-copper solution and 10 C.C. of ammonia are introduced into the flask on a, water bath, and heated to about 80°, amd the liquid containing the glucose is then added drop by drop until the solution becomes colourless. If desired, the solution may be reoxidised (by substituting a stream of air for that of hydrogen as long as the reproduced blue colour deepens) and a second estimation be then performed, The cuprous oxide dis- solving in the ammonia to a clear, colourless solution, renders the end of the reaction very sharp and exact.Note.--No reference is made to the earlier processes of Pavy and others, in which ammoniacal copper solutions are employed.-EDrroRs. Estimation of Crystallisable Sugar in Raw Sugars. By M, KARCZ (Clzem. Centr., 1894, 17, 845-846 ; from Zeit. Zuck. Ind., 23, 21--24).-Thirty or fifty grams of the sample is mixed in a dish with an equal weight of absolute glycerol, and placed for some time iii a desiccator. The glycerol soon dissolves the adhering syrup, but leaves the crystals intact. After pouring the glycerol into a giass funnel filled with cotton wool, and provided with a cover containing a calciiim chloride tube, an aliquot part of the filtrate is examined in the polariscope.The polarisation deducted from that of the original raw sugar gives the amount of crystallisable cane sugar in the sample. L. DE K. Estimation of Cane Sugar in Beer Wort. By K. AMTHOR (Chern. Centr., 1894, i, 932-933 ; from Zeit. Nahi-uiiy.smittelunters. Hygiene, 8, 80--81).-A criticism of Jais’ process (Abstr., 1894, ii, 123). Whereas Jais, in estimating the nialtose and inverted sugar, boils for only two minutes, the reduction tables are constructed for an ebulIition of four minutes, which yields higher results. Moreover, the hydrochloric acid used attacks constituents of the wort other t.han cane sugar, and augments their reducing power.(Chein. Zeit., 1694, 18, 748).--Twenty to thirty grams of the sample is rubbed with 250 C.C. of water, and mixed with excess of a soluticli of iodine in potassium iodide ; the starch combines with the iodine, and forms a corn paratively heavy compound, which settles long before any appreciable quantity of yeast has gone down. The milky liquid is syphoned off, and the iodide of starch repeatedly lixiviated L. T. T. M. J. S. Estimation of Starch in Compressed Yeast. By F. FILSINC~E~;94 ABSTRACTS OF CHEMICAL PAPERS, with water until all the yeast cells have been removed. The starch is finally collected on a weighed filter, dried at 105", and weighed. The iodine is practically expelled during the heating. In calculating the percentage, it must be remembered that commercial potato starch generally contains 15 per cent.of water. The test analyses are very satisfactory. When, however, the amount of starch is below 10 per cent., the results will be somewhat too low. L. DE K. Estimation of Carbohydrates. By E. SCHULZE (Chem. Zeit., 1694,18,527-528) .-The author points out that if carbohydrates, on boiling with dilute sulphuric acid, yield other products besides dextrose, the action of the acid should not be unduly prolonged. If a mixture of carbohydrates be inverted, it is almost impossible to get a good result, as some of them may already have become largely decomposed before the inversion of the others is anything like com- plete. L. DE K. Separation of Uric acid, Adenine, and Hypoxanthine.By M. KR~GEK. (Zeit. physiol. Chenz., 1894, 20, 170--175).-1n a hot solution containing the three substances, copper sulphate and sodium thiosulphate precipitate only adenine and hypoxanthine. I n a cold solution of these two substances, the same reagents precipitate adenine only. W. D. H. Estimation of Xmthine-like Substances in Urine, By M. KR~GER and C. WULYF (Zeit. physiol. Chem., 1894, 20, 176-185).- The new name alloxuric substances is suggested for those of the uric acid group. The alloxuric bases which occur i n small quantities in urine are xanthine, guanine, hypoxanthine, carnine, paraxanthine, and heteroxanthine ; a specific reagent for their precipitation is a mixture of copper sulphate with sodium hydrogen sulphite, but it also precipi- tates uric acid.100 C.C. of urine is precipitated with 10 C.C. of these reagents, and the precipitate allowed t o collect for two hours. Uric acid is separately determined in another sample of urine. The absolute amount of the nitrogen in this precipitate varies from 2.6 to 8 milligrams per 100 C.C. of urine, the average being 4.53. The proportion of uric acid nitrogen to the nitrogen of the allox- uric bases varies from 2.1 : 1 t o 7.6 : 1. The mean of 19 analyses gives 3-82 : 1. The average uric acid nitrogen in the 24 hours is 0.2333 gram ; and of nitrogen in alloxuric bases, 0.0481 gram. Estimation of Acidity in Milk. By &I. SCHAPFER (Staz. Sper. Agrar., 1894, 26, 164-167 ; from Bent. BZGttey j". Landw.).-The apparatus used in the method described, which is a modification of the Soxhlet-Henkel method, consists of two cylindrical bulbs connected by a naryow graduated tube.The lower bulb holds just 50 c.c., and is provided at its lower end with a small bulb, holding 2 C.C. The upper of the two large bulbs is of about the same size 8s the lower, and is stoppered. I n making a determination, 3 C.C. of phenolphthale'in aolutiori is poured into the apparatus (filling the W. D. H.ANALYT tCAL CHEMISTRY. 95 lowest poi*tion of the apparatus), then the milk to be examined, until i t reaches the 50 C.C. mark, and lastly, 4 normal soda solution (2-2-5 c.c.). The apparatus is then corked, and the contents mixed. More soda is gradually added until alkalinity is produced, The amount of alkali added is read off in the narrow tube. I n mixing the alkali, the apparatus niust not be shaken (as froth would be pro- duced), but so inclined that the liquid runs into the upper bulb.The method is of use in ascertaining whethei. milk is sufficiently free from acid to keep. It will probably also be of use in test,ing milk intended for cheese-making, and will furnish evidence of milk having been more or less skimmed (since milk slways becomes more acid when left at rest), and the presence of such milk as an adult era11 t. N. H. J. &I. Soxhlet’s Areometric Estimation of Fat in Milk. By H. TIXPE (Chem. Zeit., 1894, 18, 392-394) .-This deservedly popular process has one great inconvenience, namely, that the ether sometimes refuses to properly separate from the alkaline solution, so that only a small amouiit can be drawn off.The author now recommends that the sample zf milk should first, be diluted with three volumes of water. The ethereal layer then separates with the greatest ease. L. DE K. Loss of Total Solids in Milk on Keeping. By E. J. BWAK (Analyst, 19, 241--244).-The author accidentally noticed that milk placed in the usual weighing dishes will, if not soon evaporated, yield a residue which may be as much as 1 per cent. too low. If imme- diately before evaporation the milk is cnref ully neutralised with K/lO soda, the loss will not be so great, owing to tho formation of a stable lactate. Contrary to Bell’s statement, the author fhds lactic acid to be sensibly volatile in the preseace of water. The results of several experiments are tabulated, and show the disproportion between the acidity and the loss in total solids.L. DE I(. Periodic Estimation of Volatile Fatty Acids in the Butter produced during a Year. By L. CAMTONI and L. CARCANO ( S t a ~ . Sper. Agrar., 1894, 26,131-137) .-With the view of ascertaining the causes of the variations in the amount of volatile fatty acids in butter, samples of butter from three dairies were examined weekly for a year. The results, which are given in tables, do not show any great differ- ences or regularity. This is, perhaps, due to the fact that in the Lombardy dairies calving does not take place at definite periods, and a kind of compensation may thus take place between conditions which raise and lower respectively the amount of volatile fatty acids. A table is also given showing the results of experiments with Zeiss’ butter-refractometer as well as the vol;tt.ile fatty acids.Modification of the Reichert-Meissl Butter Process. By C. BGXTE (Clzem. Zeit., 1894, 18, 204--206).-The author criticises the sulphuric acid process lately proposed by Kreiss and modified by N. H. J. M.96 ABSTRACTS OF CHEMICAL PAPER& others, and has finally adopted the following plan. 5 grains of hutter-fat is introduced into an Erlenmeyer litre flask, and heated for H time in a drying oven at 100"; 10 C.C. of sulphuric acid (sp. gr. 1-8;%5) is added, and the mixture well agitated until all the fat has dissolved. The flask is now put into water at 30-32" for 10 minutes ; 150 C.C. of water is added, and then strong solution of potassium permanganate until the liquid acquires a transitory pink colour.'l'he liquid is then subjected to the usual distillation and titration. L. DE K. Estimation of Lecithin in Plants. By I!!. SCHULZS (Zeit. physiol. Chem., 1894, 20,225-232, 252).-A critical reply to v. Bitto in reference to his method (Abstr., 1894, ii, 402). Analysis of India-rubber Wares. By R. HENRKJUES (C7Leu). Zeit., 1894, 18, 411412, 441-444; compare Abstr., 1892, ii, 399). -The author gives further instructions for the analysis of rubber wares. Adulteration with fatty matter or fatty surrogate (falctis) may be detected by treating a weighed quantity of the sample with alcoliolic soda, as previously described, and noticing the loss in weight. If the sample contains much added mineral matter, it is best to first treat i t with moderately strong acid before boiling with the alkali ; as the latter dissolves small quantities of rubber, a correction must be made by deducting from the weight of the surrogate a quantity cor- responding with 2.5 per cent. of the rubber actually found ; soluble sulphur is, of course, allowed for. Asphalt, whether true bitumen or the artificial product, is another adulterant. In the absence of surrogate, 1 grnm of the finely-divided sample is soaked for an hour in 30 C.C. of nitrobenzene. The insoluble mass is thrown upon a filter, gently pressed with a small pestle, and further washed with another 30 C.C. of the solvent; the mass is then transferred by means of a wash- bottle to a porcelain dish and boiled with water until all odour of nitrobenzene has disappeared; it is then dried and weighed. As rubber is not altogether insoluble in nitrobenzene, a correction must be made by deducting 2.5 per cent. from the asphalt for true rubber dissolved ; soluble sulphur must also be allowed for. If the sample contains also oil surrogate, this muat be first removed by treatment with alkali, in which asphalt is practically insoluble. The process becomes still more complicated if, besides asphalt, lamp-black is also present ; this withstands the action of all ordinary solvents, and re- mains in consequence with the rubber. The author has found that in pure rubber there is a fairly constant atomic relation between the hydro- gen and the carbon, which may be taken as 16 : 10. The residue containing the rubber + thelamp-black is therefore submitted t o an organic combustion, and any excess of carbon put down to lamp- black. The test analyses given by the author are remarkably satisfactory cousidering the nature of the analysis. The procass does not, as get, W. D. H. provide for i\r host of other possible adulterants. L. DE K.