44 ABSTRACTS OF CHEMICAL PAPERS. An a1 y t i c a1 Chemistry. sensitive Litmus Paper. By RONDE (Cheiuz. Cent?.., 1896. ii, 11 30 ; from Phavnz. Zeit., 41, 736).-Coarsely powdered litmus is digested during one day with 12-1 5 parts of water, and the dark-blue mixture after being treated with concentrated sulphuric acid until the colour turns to bright red, is heated on a water bath to expel carbonic anhydride; dilute sulphuric acid is then added until filter paper dipped in the liquid just appears violet-red. When cold, the liquid is filtered, and by addition of more dilute sulphuric acid, or of powdered litmus, it is adjusted so as to give either a blue or a red stain to filter paper. By FRANK A. GOOCH and C. F. WALKER (Amer. J. Xci., 1897, [iv], 3, 293-300, and Zeit. unoyg. Chem,, 14, 423--431).-The authors have worked out a new process for the estimation of iodides.An excess of solution of potassium iodate is first added, and then dilute sulphuric acid, which causes iodine to be set free ; solution of potassium hydro- gen carbonate is then added in slight excess,followed by a very slight ex- cess of standard arsenious acid ; finally, the undecomposed arsenious acid is titrated with standard iodine, using starch solution as indicator. The amount of iodine to be estimated is five-sixths of the iodine thus found. To ensure success, the amount of iodide should be something like 0.08 gram dissolved in 150 C.C. of liquid. Small quantities of bromides or By JULIUS A. REICH (Chem. &it., lS96,20,985).-The substance is gently heated with a little strong sulphuric acid in a platinum crucible covered with a watch glass, which is placed with its convex side downwards, and is moistened with a drop of water.I f fluorine is present, a coating of silicic or boric acid will soon be visible ; the latter readily dissolves in M. J. S. Application of Iodic Acid to the Analysis of Iodides. chlorides do not interfere with the process. L. DE K. Detection of Fluorine in Silicates and Borates. a drop or two of water. L. DE K.ANALYTICAL CHEMISTRY. 45 Estimation of Sulphur in Iron. By WILHELM SCHULTE (Chew. Cents.., 1896, ii, 1132 ; from Stald und Eisen, 16, 865).-The iron (10 grams) is dissolved in 200 C.C. of hydrochloric acid (1 vol. of sp. gr. 1.19 diluted with 2 vols. of water), and the gases passed through 45-50 c,c.of a solution containing 5 grams of cadmium acetate, 20 grams of zinc acetate, and 200 C.C. of glacial acetic acid per litre ; when the reaction i s complete, and the absorbent solution is warmed to 30-40" by the escaping steam, 5-7 C.C. of copper sulphate solution (80 grams of crystallised copper sulphate and 320 grams of concentrated sulphuric acid per litre) is added, whereby the precipitated sulphides are immediately converted into copper sulphide. This is collected, washed with hot water, roasted, and then strongly ignited until con- verted into cupric oxide, from the weight of which the sulphur is calculated. Cupric acetate cannot be used for the original absorption, in consequence of the presence of hydrogen phosphide, by which some copper would be precipitated. M.J. S. By EDUARD DONATH and K. POLLAK (Zeit. ungw. Chem., 1897, 555--557).-The authors have proved that the only trustworthy process for estimating ammonia in gas-liquors is the distillation method. The gas-volumetric process, consisting in liberating the nitrogen by means of bromine dis- solved in aqueous soda, gives results which are decidedly too high. This is caused by a partial decomposition of the thiocyanates always By KARL ULSCH (Zeit. Elektrochem., 1897, 3, 546-547).-The nitric acid is electro- lytically reduced to ammonia, employing as the cathode a spiral of about forty turns of soft copper wire, about 1.4 mm. thick, wound on a glass tube of about 15 mm. diameter, the coil being afterwards stretched out t o a length of about 70 mm.; before use, the copper spiral is heated to dull redness and quenched in water ; a platinum wire, 1 mm. thick and 20 em. long, is employed as the anode. These electrodes reach almost to the bottom of a test-tube, and are held in place by a rubber stopper, through which a glass tube passes to permit OF the escape of gases ; the current of 1-25 amperes (or 1.5 amp. per sq. dm.) employed is furnished by two accumulators in series. The nitrate is dissolved in semi-normal sulphuric acid. Until some 90 per cent. of the nitrate is reduced, no hydrogen is evolved, and 10 minutes after the first bubbles of hydrogen come off the reduction is completed. The results obtained with potassium nitrate are very accurate so long as at least two-fifths of the sulphuric acid originally present remains unneutralised, but un- fortanately, many substances, especially chlorides, which accompany naturally occurring nitrates, are prejudicial to the accuracy of the method.T. E. Estimation of Nitrates in Soil. By L. KUNTZE (Chem. Centr., 1896, ii, 1133 ; from Zeit., Ver. Rubenxuck. Ind., 1896, 761).-An attempt to utilise the reaction with brucine and sulphuric acid as a quantita- tive one, by comparing the depth of colour produced by a single drop of the soil extract with that yielded by a drop of a standard Estimation of Total Ammonia in Gas-liquors. present in the samples. L. DE K. Estimation of Nitric Acid by Electrolysis.46 ABSTRACTS OF CHEMICAL PAPERS. nitrate solution, failed t o give constant results, the depth of colour varying both with the size of the drops and with the quantity of sulphuric acid added.31. J. s. Estimation of Nitrogen in Nitrated Guanos. By V. SCHENKE (Chem. Zeit., 1896, 20, 1031--1033).-The author again states that the only method which gives satisfactory results with ‘6 nitrated guanos” is the one proposed by him in 1893, which he termed the Ulsch-Kjeldahl method. In this process, the nitric nitrogen is first reduced by means of reduced iron and dilute sulphuric acid, and the whole is then boiled with sulphuric acid and a drop of mercury until the organic matter is destroyed. The ammonia is then determined as usual (Abstr., 1894, ii, 67). L. DE K. Detection of Nitrites in Potable Water. By A. GAWALOWSKI (Chem. Centr., 1896, ii, 1009; from Zeit. Nahrungsm. Hgg.War., 10, 315).-The reaction with potassium iodide and starch is only trust- worthy when the iodide is free from iodate, and has been preserved in the dry state and in the dark in bottles of yellow or blue glass; dilute hydrochloric acid should be used for acidifying, since dilute sulphuric acid is apt to be reduced by dust to sulphurous acid. It is necessary that the starch should be washed immediately before it is gelatinised, and that the starch-paste be freshly made. M. J. S. Analysis of Phosphor-bronze, Phosphor-copper, Phosphor- tin, &c. By MAX WICKRORST (J. Arner. Chem. Xoc., 1897, 19, 396-398).-Estinzcction, of Phosphorus alone.-One gram of the sample is heated with a mixture of 15 C.C. of nitric and 5 C.C. of hydrochloric acid. When oxidised, the acid liquid is diluted, mixed with ammonia in excess, and the whole made up to 200 C.C.; hydrogen sulphide is then passed bhrough the solution until the copper, lead, &c., are entirely precipitated. Ammonia and magnesium mixture are added t o 100 C.C. of the filtered liquid, and after a few hours the magnesium phosphate is collected, washed with dilute ammonia containing a little ammonium sulphide, and weighed, as usual, as magnesium pyro- phosphate. Complete Analysis of Phosphor-bronze.-Half a gram of borings, treated with 5 C.C. of strong nitric acid until fully oxidised, is diluted with water and the insoluble matter collected on a filter and washed with dilute nitric acid. The filter and contents are then ignited in a weighed porcelain crucible, being gradually heated to redness ; the weight is that of the stannic oxide plus phosphoric anhydride.After being fused with 0.5 gram of dry sodium carbonate and 1 gram of sulphur, with the cover on the crucible, until the excess of sulphur is volatilised, it is allowed to cool, and the fused mass dissolved in water ; excess of ammonia and 1 gram of ammonium chloride are then added, and the phosphoric acid precipitated with magnesium mixture. The precipitate should be purified by redissolving it in a little hydro- chloric acid and reprecipitating with ammonia, before finally weighing as magnesium pyrophosphate. From the latter, the phosphorus can beANALYTICAL CHEMISTRY. 47 calculated and also the corresponding amount of phosphoric anhydride ; the latter deducted from the weight of the stannic precipitate gives the stannic oxide, which is then calculated to tin.As, however, a little of the phosphoric acid passes into the original nitric acid filtrate, it is advisable to make a special duplicate experiment in which the phosphoric acid in the filtrate can be estimated by the molybdic process. To the acid filtrate, after it is neutralised with ammonia, 5 C.C. of nitric acid is added, the whole diluted to 150 c.c., introduced into a large platinum dish, and electrolysed, the lead being precipitated as dioxide on the dish and the copper on the negative electrode, which consists of a piece of platinum foil 2 inches square. The liquid contains any iron or zinc which may be present; these are precipitated as sulphides by adding ammonia and ammonium sulphide, and the precipitate, after being collected and ignit'ed in a porcelain crucible, is weighed ; the mixed iron and zinc oxides are dissolved in hydrochloric acid and the ferric oxide precipitated with ammonia, the zinc oxide being found by difference.L. DE K. Estimation of Phosphoric Acid in Potable Waters. By CHARLES LEPIERRE (Bull. Xoc. Chim., 1896, [iii], 15, 1213-1217).- The estimation of phosphoric acid in potable waters by weighing the ammonium phosphomolybdate is inexact, because the solubility of the precipitate in dilute nitric acid introduces errors of the same order of magnitude as the quantities to be determined. The author finds (1) that a t equal temperatures the intensity of the yellow colour of the liquid after addition of the molybdate is directly proportional t o the quantity of phosphoric acid present, a t any rate up to 0.03 gram per litre, and (2) the intensity of the coloration increases with the temperature, and between 10" and 30" is directly proportional to the temperature, the increase in intensity with rise of temperature being represented by straight lines.Solutions of potassium chromate form a convenient colour scale, but a scale can also be made by means of dilute solutions of a phos- phate, 50 C.C. of which is mixed with 2 C.C. of a molybdate solution prepared by dissolving 150 grams of ammonium molybdate in 1000 C.C. of water, and pouring this solution into 1000 C.C. of nitric acid of sp. gr. 1.2. The tubes are well corked, and the corks covered with paraffin ; the solutions remain unaltered for two or three months, but those containing more than 0.010 gram of phosphoric acid per litre gradually deposit a precipitate after that time.When kept a t 40° for several days, no precipitate forms even in the solutions that contain 0.025 gram of phosphoric acid per litre. A litre of the water is evaporated after addition of nitric acid, and the residue is repeatedly evaporated with nitric acid to separate all the silica; it is then dissolved in 50 C.C. of nitric acid, mixed with 2 C.C. of the molybdate solution, and the intensityof the yellow coloration compared with the scale of units. The presence of silica introduces an error, but the coloration due t o phosphoric acid reaches its maximum intensity almost instantaneously, whereas with silica the intensity increases slowly.Observations made48 ABSTRACTS OF CHEMICAL PAPERS. after sufficient intervals of time will show whether the intensity is increasing or is constant. Solubility of Phosphates in Citric Acid and Ammonium Citrate. By OTTO FOERSTER (Cli,enx. Zeit., 1Y97,20, 1020--1021).-The author’s experiments prove that there is yet a good deal to be investi- gated before the citrate-solubility controversy is finally settled. It appears that this solubility is greatly influenced by the amount of soluble silica. Another strange thing is that some phosphates were found to be more soluble in Wagner’s ammonium citrate solution than in a 1.4 per cent. solution of citric acid,whilst the majority of phosphates are more readily soluble in the latter.Arsenical Sulphuric Acid, a Source of Error in Naumann’s Process for Estimating Phosphoric Acid. By GUSTAV LOGES and KARL MUHLE (Chern. Zeit., 1896, 20,584).-1n Naumann’s process for estimating phosphates soluble in citric solution, the citric acid solution of the basic slags is boiled with sulphuric and nitric acids until the organic matter is destroyed. The phosphoric acid is then estimated by the magnesia or molybdate method. If there is m y arsenic present in the sulphuric acid employed, this will be oxidised t o arsenic acid, and will then be precipitated along with the phosphoric acid. The use of a pure acid is therefore in- dispensable. L. DE K. By LEONARD DE KONINGH (,I Amer. Chem. Xoc., 1897, 19, 385--388).-When using Thomson’s process for estimating boric acid, it is necessary to completely expel by boiling any carbonic anhydride that may be present; the author states that there is no fear of loss of boric acid, even if the boiling is continued for 15 minutes, which is far longer than is necessary. When dealing with articles of food, the presence of phosphoric acid has t o be taken into account ; a process of separating it by means of calcium chloride is given, differing greatly in detail from that pre- viously published by Thomson. A new process is also described by which the boric acid can be estimated after removal of the phosphoric acid by means of magnesium mixture ; the filtrate is mixed with excess of sodium carbonate and heated, the precipitate of magnesia is removed by filtration, the filtrate evaporated to dryness to render the rest of the magnesia insoluble, and the residue is then treated with a little water and filtered.The boric acid can be titrated according to Thomson’s directions. As a test experiment, 0.1 gram of boric acid mas dissolved in aqueous soda, and mixed with 100 grams of oatmeal and incinerated ; from the ash, 0.095 gram of boric acid was recovered. C. H. B. L. DE K. Estimation of Boric Acid in Foods. L. DE K. Estimation of Silica in Blast-furnace Slag. By G. H. MEEKER (J. Amer. Chem. Xoc., 1897, 19, 370-374).-The author recommends the following process, which is particularly suitable if the sample con- tains an admixture of spinel, 0.5 gram of the finely ground slag is placed in a small dish, moistened with 3 C.C. of water, 10 C.C.of hydro- chloric acid added, and the whole well stirred. When the slag has nearly all dissolved, 40 C.C. of dilute sulphuric acid (1 : 1) is added, andANALYTICAL CHEMISTRY. 49 the mixture boiled until fumes of sulphuric acid are given off. When cold, a little water and 10 C.C. of hydrochloric acid are added, and the whole is boiled for a minute, The insoluble matter, consisting of nearly chemically pure silica, is collected on a filter, and after being washed with hot dilute hydrochloric acid and then with water, is ignited and weighed. L. DE K. Separation of Silicic and Tungstic Acids. By JAMES S. DE BENNEVILLE (J. Arner. Clwn. Xoc., 1897, 19, 377-379).--The author confirms the statement that tungstic acid cannot be accurately separated from silica by means of ammonia, as the latter is also sensibly soluble in that liquid ; the solubility is certainly diminished if it has been strongly ignited, but in that case the tungstic acid also becomes less readily soluble.The best plan is to first weigh the mixed oxides, then to remove the silica by evaporating with hydrofluoric and sulphuric acids, and weigh the residue of tungstic acid. L. DE K. Estimation of Potash and Phosphoric Acid in Fodders. By HARVEY W. WILEY (J. Anzer. Chew,. Xoc., 1897, 19, 320-322).-This is a slight modification of the Lindo-Gladding process for estimating potash, and admits of the determination of the phosphoric acid in the same portion of the sample. Eight grams of the substance is burnt as completely as possible over a small flame, the ash is dissolved in 50 C.C.of water mixed with 5 C.C. of hydrochloric acid, and transferred to a 200 C.C. measuring flask. Some ferric chloride is added to precipitate the phosphoric acid, and then 10 C.C. of strong ammonia and 10 C.C. of a 20 per cent. solu- tion of ammonium carbonate. After heating on the water bath for an hour, the liquid is left over-night, made up to the mark, thoroughly shaken, and poured on to a large filter; when this is drained and has become somewhat dried, i t is put back into the flask. Dilute nitric acid is added to dissolve the ferric phosphate, the solution is made up to a definite bulk, and an aliquot part is used for the estimation of the phosphoric acid. Fifty C.C. of the filtrate containing the potassium is then evaporated to dryness in a platinum dish, the bulk of the ammonium chloride expelled by heating, and the residue, after being moistened with 1 C.C.of dilute sulphuric acid (1 : l), is again gradually heated to redness. The potassium is then estimated by means of platinic chloride by the Lindo-Gladding method. The washing with alcohol, solution of am- monium chloride, and the final washing with alcohol are best done in a Gooch crucible. After weighing the double chloride, the contents of the crucible are dissolved in boiling water, and the crucible is dried and re-weighed. L. DE I(. Electrolytic Estimation of Cadmium. By S. AVERY and BENTON DALES (J. Arner. Chem. Xoc., 1897, 19, 379--382).-The authors, having in view the excellent results obtained when electrolgsing zinc from a formate solution, have applied the same process t o cadmium with great success.VOL. LXXIV. ii. 450 ABSTRACTS OF CHEMICAL PAPERS. About 0.1 gram of cadmium is dissolved in dilute sulphnric acid, 6 C.C. of formic acid of sp. gr. 1.2 is added, and then solution of potas- sium carbonate until a slight turbidity is produced. This is removed by a little more formic acid ; finally 1 C.C. more of this acid is added, and the mixture is diluted to 150 C.C. and submitted to electrolysis. L. DE K. Distribution of the Precious Metals and Impurities in Copper, and Suggestions for a Rational Mode of Sampling. By EDWARD KELLER (J. Anzer. Chem. Xoc., 1597, 19, 243--25S).-In order to get a fair sample of unrefined copper, plates from 8 to 10 inches square and 1 inch thick should be cast, and from these drillings should be taken at least 1 inch from the edge through the entire plate.The author believes that when, on the solidification of a metal, the small amounts of impurities segregate or liquate, and consequently concentrate towards the centre, the degree of concentration is greatest for those whose atomic volumes shorn the greatest difference as compared with that of the metal in which these impurities are By L. WOLMAN (Zeit. Elektrochem., lS97, 3, 537-545).-The author gives the results of comparative trials of the principal electrolytic methods which have been proposed for the determination of the heavy metals. Copper.-The best results are obtained with solutions in dilute sul- phuric or nitric acid. Solutions containing ammonium oxalate require more time and attention, and from ammoniacal solutions the metal is apt to separate in the spongy condition.Silver is easily precipitated a t 50" from solutions containing 2 to 3 per cent. by volume of nitric acid (sp. gr. 1.36) or from solutions con- taining potassium cyanide. Lead is best determined at 50" in solutions containing about 20 per cent. by volume of nitric acid (sp. gr. 1.36). The peroxide precipitate must be dried a t 180-190". Mangu?zese.-Fairly good results are obtained with solutions contain- ing 1 to 1.5 per cent. by volume of strong nitric acid, using a very small current and warming to 50". Solutions in potassium oxalate, acetic acid or ammonium pyrophosphate do not give satisfactory deposits . Zinc is best deposited from strongly alkaline solution at 50' by means of a fairly strong current.Potassium or ammonium oxalate solutions are also satisfactory, but acetic o r citric acids, ammonium acetate, potassium cyanide, or ammonium pyrophosphate are unsatis- factory. Cobalt and nickel are best deposited from ammoniacal solutions ; Classen's ammonium oxalate and Brand's ammonium pyrophosphate methods give good deposits, but somewhat high numbers. Iron is deposited slowly but completely from solutions in ammonium oxalate. Copper may be separated from zinc or nickel by first depositing the copper from an acid solution. Zinc and nickel are separated by first depositing the zinc from an present. L. DE K. Quantitative Electrolysis of Heavy Metals.ANALYTICAL CHEMISTRY. 51 alkaline solution containing Rochelle salt.If the electrolysis is tOQ long continued, the zinc is apt to contain nickel. Zinc and cadmium are separated accurately by precipitating the latter a t 50’ with a feeble current from a solution containing 1 or 2 per cent. by volume of dilute sulphuric acid. Lead is separated from zinc, copper, or silver by precipitating it at the anode from a solution strongly acidified with nitric acid. I n the case of silver, however, the lead peroxide contains some silver, and the silver deposit simultaneously formed a t the cathode is spongy. Copper is not completely precipitated from the strongly acid solution. Copper and silver are separated by employing an E. M. F. lower than that required to deposit copper, namely, 1-3 to 1.5 volts.The process is slow, so that it is easier to precipitate the silver as chloride. T. E. Detection and Estimation of Traces of Lead in Waters. By J. C. BERNTROP (Chnz. Zeit., 1896, 1020).-The author has found that lead phosphate is absolutely insoluble in water containing free sodium phosphate, and utilises this fact for the detection and estimation of lead in drinking waters. A few litres of the sample is, if necessary, rendered somewhat hard by adding calcium chloride, and excess of sodium phosphate is then added. The precipitate which forms during 24 hours contains all the lead. It is collected on a filter, and after being dissolved in nitric acid, the lead is recovered from the solution and identified by the usual methods. L. DE K. Elstirnation of Lead in Lead Ores.By RICHARD K. MEADE (J. Arne?.. Chem. Soc., 1897, 19, 374--377).-The author recommends the following process. About 1 gram of the sample is treated in a platinum dish with 50 C.C. of a mixture of concentrated sulphuric and nitric acids (1 : 3). The dish is covered and heated on a sand-bath until the action of the acid has ceased. The cover is then removed, rinsed into the dish, and 15 C.C. of bydrofluoric acid added, the dish being heated until fumes of sulphuric acid begin to make their appearance. When cold, the residue is diluted to about 100 c.c., and the undissolved lead sulphate is collected and washed first with a 2 per cent. solution of sulphuric acid and then with alcohol. The precipitate is detached from the filter and ignited, the paper being burnt separately in the usual way.If a sample of galena should contain calcite, this may be removed by a preliminary treatment of the ore with dilute hydrochloric acid, but it is always advisable to wash the lead sulphate thoroughly with acidi- fied water in order to remove the last traces of calcium sulphate. L. D E K . Volumetric Estimation of Lead. By J. HOWARD WAINWRIGHT (J. Amer. Chem. SOC., 1897, 19, 389--393).-The author recommends a process differing but slightly from that given in Crookes’ ‘‘ Select Methods.” To estimate the amount of metallic lead in litharge, 1 gram of the sample is dissolved in 10 C.C. of nitric acid (sp. gr. 1-20>, the solution is neutralised with excess of ammonia, and a large excess of 4-252 ABSTRACTS OF CHEMICAL PAPERS. acetic acid is added; the liquid is then heated to boiling anda solution of potassium dichromate run in from a delicate burette until the bulk of the lead has been precipitated.The dichromate solution should be made of such a strength that 1 C.C. represents a quantity of lead not greatly differing from 0.01 gram. The solution is now again boiled until the lead chromate has become orange coloured, and the titration is then cautiously continued until the precipitate settles promptly. The dichromate is now added a drop at the time, until a few drops of the supernatant liquid put on to a white porcelain tile gives a distinct red coloration with a drop of a 2 per cent. solution of silver nitrate. Other metals which are precipitated by potassium dichromate or which exercise a reducing action on it should be first removed.Red lead should be dissolved in dilute nitric acid with the addition of oxalic acid ; white lead may be a t once dissolved in acetic acid. L. DE I(. Sodium Peroxide as a Third [Iron-zinc] Group Reagent.-By SAMUEL W. PARR (J. Amer. Chem. Xoc., 1897, 19, 341-348).-The metals of this group may be, to some extent, separated by means of aqueous potash or soda ; but in the presence of chromium, the separation is far from complete, as this metal is partly retained by the precipitate, particularly in the presence of zinc. The author recommends dissolving the ammonium sulphide pre- cipitate in nitric acid, and after neutralising with soda, heating the solution with excess of sodium peroxide. The chromium will then be converted into sodium chromate, and the other metals, except zinc and aluminium, will be left insoluble as hydroxides, or peroxides; they Estimation of Manganese in Iron Ores in the Dry Way.By BUTTGENBACH (Chern. C'entr., 1896, ii, 1134 ; from lieu. LTnivers. ; Berg.- Hiittenm-Zeit,, 55, 368).-Two fusions are made; the one with an ac'ld flux, the other with a basic flux The regulus of the former contains only the iron of the ore, the slag containing the manganese as silicate. The regulus of the latter contains the iron with the reducible manganese, and its excess of weight over the former should agree with the excess of 'weight found in the slag of the acid fusion (allowing for the oxygen required to form manganous oxide). Should they not agree, the difference is due to the non-reduction of p w t of the manganese in the ore, in consequence of a deficiency in the amount of iron present, Ores rich in manganese can be smelted without the addition of lime; the regulus then contains the iron free from manganese and containing only traces of carbon.By CARL EKGELS (Zeit. E'lektrochem., 1897, 3, 286-289, and 305--308).--Further details are given of the method previously described (Abstr., 1896, ii, 276), of determining manganese by depositing it electrolytically as peroxide from solutions containing chrome alum, or alcohol and ammonium acetate. Manganates and permanganates are first reduced by a small excess of hydrogen peroxide in acetic acid solution, the hydrogen peroxide are usually free from phosphoric acid. L. DE K. M. J. 8.Estimation of Manganese by Electrolysis.ANALYTICAL CHEMIS'l'RI'. 53 destroyed by a little chromic acid, the ammonium acetate and chrome alum added, and the free acetic acid neutralised by ammonia; the solution is then electrolysed in an etched platinum basin a t about SOo, and the adherent deposit thus obtained is heated for 3 or 4 minutes to redness to convert it into Mn,O,. With 0.4 gram of manganese in 150 C.C. of solution containing 2 grams of chrome alum and 10 grams of ammonium acetate, the precipitation is complete in about la hours with a current density of 0.S to 1 ampere per sq. dcm. The results of the test analyses given are excellent. Heavy metals must first be removed when they occur along with manganese, and when the filtrate from them is dilute or contains substances which would interfere with the electrolytic separation, the manganese is best precipitated with hydrogen peroxide and ammonia, and then redissolved and treated as above described, The determinations can be made with smaller currents, but the time required is correspondingly longer.At ordinary temperatures, 1.25 volts and a t 80" 1.1 volts are sufficient t o produce decomposition, but in the cold the precipitation is compIete only after a very prolonged action of the current, and the precipitate is brittle and does not adhere well to the dish. In order t o obtain a sufficiently adhesive precipitate, the electrolysis must be conducted at 80' and in presence of chrome alum; the latter may be replaced by alcohol, though this is less certain in its results, and requires the application of at least 2 volts.The reason of the variations in the properties of the precipitate is discussed a t length ; the author considers that MnO, and Mn,O, are the primary pro- ducts of the electrolysis, and that they are deposited as a brittle, metal- like coating. If, however, oxygen is simultaneously separated a t the anode, it is supposed t o act, in the nascent state, as a reducing agent, rendering the coating porous and not brittle. The chrome alum acts partly by reducing the manganese peroxide, partly by taking up some of the oxygen and so modifying the evolution of gas that the coating is not mechanically disturbed. The inferior results obtained with alcohol are due to the fact that it merely acts in the second way and not in the first.With a current density of 0.8 t o 1 ampere per sq. decimetre and with 0.2 to 0.3 gram of manganese, the quantity of chrome alum employed tuay be varied from 1 to 3 grams ; with larger or smaller quantities, the deposit of manganese peroxide is brittle and non-adhesive. T. E. Estimation of Iron in Limestones. By D ~ S I R ~ DE PAEPE (Chenz. Zeit., 1896, 20, 1004).-5 to 10 grams of the sample is boiled for 5 minutes with strong hydrochloric acid, some water added, and the liquid filtered ; after the addition of nitric acid, the mixture is boiled, and the iron and alumina are precipitated by ammonia, collected on a filter, and washed. The filtrate is made up to a definite volume, and may then be used for the estimation of lime, magnesia, alkalis, Bc.The ferric precipitate is dissolved in hydrochloric acid, diluted to 125 c.c., and an aliquot portion mixed with excess of tartaric acid ; the iron is then precipitated with ammonia and ammonium sulphide, the precipitate being finally converted into ferric oxide. L. DE K.54 ABSTRACTS OF CHEMICAL PAPERS. Qualitative Separations with Sodium Nitrite in Absence of Phosphates. By GILLET WYNKOOP (J. Amel.. Chem. SOC., lS97, 19, 434-436).-The author states that iron, chromium, and aluminium may be completely separated from zinc, manganese, cobalt and nickel by. adding to the solution, if necessary, a few drops of hydrochloric acid and then boiling with excess of sodium nitrite. It is advisable to first reduce the iron to the ferrous state, as it then gives a more granular precipitate. The author’s test experiments, although only qualitative, are per- Electrolytic Separation of Nickel and Cobalt from Iron.By 0. DUCRU (Compt. rend., 1897, 125, 436-439).-When a solution containing a ferric and a nickel salt is precipitated with excess of ammonia, part of the nickel is precipitated and part remains in solntion, but i f an electric current is passed through the liquid in which the precipitate is suspended, the whole of the nickel is deposited on the cathode. A very small quantity of iron is deposited a t the same time, and a correction must be made by dissolving the metal and precipitating the iron after peroxidation ; for practical purposes, however, the correction, which is always small, is rarely necessary. The metal is dissolved, the solution evaporated to dryness with a slight excess of sulphuric acid, the residue dissolved in water, mixed with 5 to 10 grams of ammonium sulphate, heated, placed in the electrolysis apparatus, mixed with excess of concentrated ammonia, and subjected to the action of a current of 1.5 to 2.5 amphres for about 4 hours.The results quoted in the paper indicate that the method is very accurate. I n the case of steels, 0.25 to 0.30 gram is dissolved in aqua regia, evaporated with sulphuric acid, and treated as above. It is not necessary to separate the silicon and carbon, and traces of manganese, chromium, and phosphorus do not affect the accuracy of the estimation. Traces of manganese are almost always deposited with the small quantity of iron, but the error due to this cause is negligable. The author notes that a small portion of the metal deposited is insoluble in hydrochloric acid, but dissolves in aqua regia, and the latter solution gives the reactions of ferric salts.Small quantities of chromic acid prevent the precipitation of the nickel by electrolysis from the ammoniacal solutions. C. H. B. Estimation of Molybdenum lodometrically. By FRANK A. GOOCH (Amer. J. Xci., 1897, [iii], 237-240, and Zed. anorg. Chem., 14, 317-322).-The author defends his process (Abstr., 1897, ii, 76) against the criticisms of Friedheim and Euler, and states that the results they obtained are vitiated by serious arithmetical errors. fectly satisfactory. L. DE K. L. DE K. Analysis of Commercial Calcium Carbide and Acetylene, and Purification of the Latter.By GEORG LUNGE and EDWARD CEDERCREUTZ (Zeit. ccngw. Chew,., lS97, 651-655).--The authors point out the great difficulty there is in obtaining a thoroughly re-ANALYTICAL CHEMISTRY. 55 presentative sample on account of the impossibility of obtaining a fine powder without moisture being absorbed and acetylene given off. About 50-100 grams of pea-sized lumps must therefore be taken for analysis ; these are placed in a generating flask and brine is slowly added through a tap funnel until effervescence ceases ; the gas is collected in R large gasometer over brine, or, better still, over water previously saturated with the gas. One hundred grams of chemically pure calcium carbide should yield 34,877 litres of gas a t normal temperature and pressure.Frequently the sample contains phosphorus, which is evolved as hydrogen pliosphide, and sulphur which is partially given off as hydrogen sulphide and volatile sulphixr compounds. To estimate these, 50-70 grams of the sample is treated as directed above, and the gases allowed to pass through a ten-bulb absorption tube filled with '76 C.C. of a 3 per cent. solution of sodium hypochlorite. This oxidises the phosphorus and sulphur to the corresponding acids ; after pre- cipitating the first with ammoniacal magnesium chloride, the second may be precipitated in the filtrate by adding hydrochloric acid and barium chloride. Commercial acetylene may be purified by passing it first through a milk of chloride of lime, or, better still, over moist lumps of that substance.This treatment fully removes phosphorus and sulphur compounds. It may then be both dried and freed from ammonia by Estimation of Cyanogen by Silver Nitrate, using Potassium Iodide and Ammonia as Indicators. By WILLIAM J. SHARWOOD (J. Amer. Chern. Soc., 1897, 19, 400--434).-The author, after criti- cising the various processes in use, recommends the following scheme. To the solution containing the cyanogen, 5 C.C. of ammonia and 2 C.C. of a 5 per cent. solution of potassium iodide are added, and then standard solution of silver nitrate until a faint, permanent cloudiness is pro- duced. If the solution contains sulphides in small amount, 5-10 C.C. of a solution made by dissolving 0.5 gram of iodine and 2 grams of potassium iodide in 100 C.C.of water is used in place of the potassium iodide, but a special check should be made in such case. If the amount of sulphide is large, it must be removed by means of a solution of sodium plumbite; an aliquot part of the filtrate is then titrated. If zinc is present, a large excess of alkali should be added ; in this case, the cyanogen found represents, not only the potassium cyanide, but also the double zinc compound. By estimating the zinc, the amount of free potassium cyanide may be readily calculated, as 1 part of zinc corresponds with 4 parts of potassium cyanide. A similar allowance must be made if small quantities of copper are present. If calcium, magnesium, or manganese are present, ammonium chloride must be added, whilst soda is used in presence of aluminium or lead.For technical purposes, it is best to prepare a silver nitrate solution containing 1,305 grams of this salt per 100 C.C. ; taking samples of 10 C.C. each, 1 C.C. of the silver represents 0.1 per cent. of potassium passing it through sulphuric acid of 1.6 sp. gr. L. DE K. cyanide. L. DE K,56 ABSTRACTS OF CHEMICAL PAPERS. Detection of Rhodinol in Ethereal Oils. By HUGO ERDMANN and P. HUTH (J. p. Chenz., [ii], 56, 27-42). See this vol., i, 36. Recalculation of Wein’s Table for Starch Estimation. By WILLIAM H. KRUG (J. Anzer. Cl~enz. Soc., 1897, 19, 452--454).-Wein has constructed a very useful table for calculating the amount of starch or dextrin from the amount of metallic copper obtained when employing Allihn’s process for the estimation of starch.This table is based on the factor 0.90. There being some slight difference of opinion regarding the true molecular formula of starch, Ost states that the factor 0.925 should have been employed. The author has, therefore, recalculated the table, using the average factor 0.92. L. DE K. Estimation of Carbohydrates in Food-stuffs. By WINTHROP E. STONE ( J . Anzer. SOC., 1897,19, 183-197; 347-349).-The author has worked out a scheme for the estimation of carbohydrates in food- stuffs, as the methods generally employed are becoming obsolete, and do not satisfy the present demands. I n the author’s process, the sample is finely ground or grated, and from 50 to 100 grams is first extracted, preferably in a Soxhlet apparatus, with 500 C.C. of strong alcohol to dissolve out any sugars ; the residue is then exhausted with 500 C.C.of cold water to remove dextrin and soluble starch, and the undissolved part is air-dried. Two grams of the latter is boiled with 100 C.C. of water for half an hour, and, after cooling to 65”, it is digested with diastase a t this temperature, in order to hydro- lyse the starch. The residue left after this treatment is then boiled in a reflux apparatus with 100 C.C. of water and 2 c,c. of hydrochloric acid, t o convert the gums, p.entosans, hemicelluloses, &c., into reducing sugars ; finally, the residue is boiled with aqueous soda (1 *25 per cent.), and the crude fibre which is left is dried and weighed. 71. DE K. Improvements on Squibb’s Volumetric Method for Estimat- ing Acetone. By LYMAN P.KEBLER (J. Arner. Chem. Xoc., 1897, 19, 316--320).-The process differs from the original method (Abstr., 1897, ii, 466) in so far t h a t the use of pure acetone and the drop end- reaction are dispensed with. The following solutions are required : 1. A 6 per cent. solution of hydrochloric acid. 2. A decinormal solution of sodium thiosulphate. 3. An alkaline solution of potassiumiodidemadeby dissolving 250 grams of the pure salt, making the liquid up to 1 litre, and adding 800 C.C. of aqueous soda containing 257 grams of soda per litre. 4. An approximately four-fifths normal solution of sodium hypochlorite, made by mixing 100 grams of 35 per cent. chloride of lime with 400 C.C. of water, and adding a solution of 120 grams of soda crystals in 400 C.C. of hot water; when cold, the clear liquid is decanted, diluted t o 1 litre, and mixed with 25 C.C.of aqueous soda of sp. gr. 1.29. 5. A starch solution made by rubbing 0.125 gram of starch with 5 C.C. of cold water and then adding 20 C.C. of boiling water and again boil- ing for a few minutes; when cold, 2 grams of sodium hydrogen car- bonate is added. The sample of acetone to be tested should be diluted with water to 50 or 100 times its weight. Twenty C.C. of the solution 3ANALYTICAL CHEMISTRY. 57 is put into a stoppered flask, 10 C.C. of the acetone solution is added, and excess of solution 4 run in from a burette. After thoroughly shaking, the mixture is rendered acid by adding solution 1, excess of solution 2 is run in, and after a few minutes a little of solution 6 is added and the excess of thiosulphate re-titrated.A blank experiment without the acetone having been made, the per- centage of the latter can be readily c:tlculated by remembering that 1 mol. of acetone requires 3 mols. of free iodine to form iodoform. L. D E E . Estimation of all the Volatile Fatty Acids in Butter. By ED UARD WRAJIPELMEYER (Lnizdw. Kemuchs.-Xtut ., 1897,49, 21 5-218). -The filtered fat (about 5 grams) is heated in a 700-SO0 C.C. flask with 20 C.C. of glycerol sodium hydroxide (prepared by dissolving 100 grams of caustic soda in 100 C.C. of water and mixing 20 C.C. of the solution with 180 C.C. of glycerol) over a flame until frothing ceases and the liquid is clear. 250 C.C. of hot, boiled distilled water is care- fully added, then a drop of indicator (litmus), and finally 50 C.C.of dilute sulphuric acid (20 C.C. of acid sp. gr. = 1.84 to 1 litre). The flask is immediately closed with a double bored cork fitted with a bulb- tube connected with a condenser at least 0.5 metre long, and a second tube for steam distillation. For the production of steam, boiled dis- tilled water is employed, and the steam is passed through a connecting tube of copper (30 cm. long and 1.4 cm. wide) which is well heated with a flat burner. The distillation of 1.5 litres (collected in two portions of 1 and 0.5 litre) takes about an hour and a half. The distillate is filtered and half in each case (500 C.C. and 250 C.C. respectively) titrated. All the water used in the process must be previously boiled to remove carbonic anhydride.I n order to avoid any substance being mechanically carried over in the distillation, the bulb of the distilling tube must have a bent tube a t the upper opening. Check experiments must be made with each apparatus, and potash must not be substituted for soda. N. H. J. M. Estimation of Lactic and Succinic Acids in Wines. By JOSEPH A. MULLER (Bull. SOC. G'him., 1896, [iii], 15, 1203-1206).- The succinic and lactic acids in wines are usually estimated together and expressed in terms of the equivalent quantity of sulphuric acid, but they can readily be separated by taking advantage of the fact that barium succinate is almost completely insoluble in alcohol of 80-83", whereas barium lactate dissolves. The wine is mixed with quartz sand, evaporated to dryness under low pressure, and the residue dried over potash.It is then extracted with ether, the ether distilled off or allowed to evaporate spontaneously, the residue dissolved in water and the hot solution titrated with carefully purified barium hydroxide solution, using phenolphthalein as indicator. The neutral liquid is evaporated to dryness, mixed with 5 C.C. of warm water and 25 C.C. of absolute alcohol, and allowed to remain for several hours, after which it is filtered, the residue washed with alcohol of SO", and the barium estimated in both the residue and the solution. Experiments made with liquids of known composition indicate that the results are too5s ABSTRACTS OF CHEMICAL PAPERS. low in the case of both acids, the error being greater in the case of the lactic acid, by reason of its volatility (compare this vol., i, 0).C. H. B. Estimation of the Acidity of Milk. By ARTHUR DEVARDA (C‘hem. Cent?*., 1896, ii, 1003 ; from Ostew. Molkerei-Zed., 1896, Sept.). -The author uses a stoppered flask holding 100 C.C. to the bottom of the neck, and with a cylindrical neck of 6 C.C. capacity graduated in half cubic centimetres. One hundred C.C. of milk, together with the necessary amount of 4 per cent. alcoholic phenolphthalein solution, reach to the zero mark, and when sufficient N/lO alkali has been added to produce the red colour, the reading on the neck gives a t once the degree of acidity; 5 C.C. of alkali corresponds with 1 Soxhlet degree, IT. J. S. Detection of Filicic Acid in Cases of Poisoning by Fern Extract. By ICARO BOCCHI (Chem.Centr., 1896, ii, 1137 ; from Boll. Chim. Farm., 1896, 20, 609).--The viscera are cut up, dried on the water bath, and extracted with a mixture of 1 part of absolute alcohol and 3 parts of ether; the extract is then evaporated, the greenish-brown residue treated with lime-water until :I colourless solu- tion is obtained, and the filtered solution acidified with acetic acid and shaken with carbon bisulphide. The residue from the bisulphide solu- tion may be further purified by dissolving it in ether, and treating the solution with lime-water or neutral cupric acetate; the usual tests for filicic acid can then be applied. Filicic acid does not pass into the urine, but is decomposed in the organism. Analysis of Linseed Oil and Linseed Oil Varnish. By WALTHER LIPPERT (Zeit.angw. CJmz., 1897, 655-657. Compare Abstr., 1897, ii, 529).-A criticism of Amsel’s so-called water test for ascertaining the presence of rosin and rosin oil in linseed oil. This test consists in saponifying the sample with alcoholic potash and then adding water, when in the presence of adulterants a more or less decided turbidity will be noticed. M. J. 8. The author proves the test to be utterly untrustworthy. Examination of Resins, L. DE K. By KARL DIETERICH (Chenz. Cent?*., 1896, ii, 1137-1139; from Ber. deut. pha~m. Ges., 6, 247).--Peru balsam.-To obtain the saponification number, 1 gram is digested in the cold for 24 hours with 50 C.C. of light petroleum and 50 C.C. of Nj2 alcoholic potash; 300 C.C. of water is then added and the liquid titrated with acid, The normal value is 260-270.For ascertaining the acid number, a dilution of 1 : 200 is necessary. Phenolphthalein should be added repeatedly and the titration continued until the super- natant liquid shows a distinct red colour after the brown, flocculent precipitate has subsided ; values of 68-80 are obtained, whilst adultera- tions raise the acid number and lower the saponification number. The ether number, obtained by subtracting the acid number from the saponification number, varied only from 188 to 196 in the specimens examined by the author. Commercial samples contain 1.5-3 per cent, insoluble in ether. The ethereal solution serves for the estimationANALYTICAL CHEMISTRY. 59 of the ciniiamein and the resin salts (poruresinotannol cinnamate), for which purpose it is shaken with 20 C.C.of a 2 per cent. solution of sodium hydroxide, and the two liquids carefully separated. Evaporation of the ethereal solution gives the cinnamein, and precipitation of the alkaline solution with hydrochloric acid the resin. The former should amount to 65-75 per cent., and the latter to 20-28 per cent. of the balsam. Gum Ammoniacunz.-The acid number is obtained by distilling 0.5 gram in a current of steam and conveying the distillate directly into 40 C.C. of N/2 potash. To test for galbanum resin, 5 grams is boiled with 15 grams of strong hydrochloric acid for 15 minutes, 15 C.C. of water is added, and the liquid filtered through a wetted, double filter; the clear filtrate is supersaturated with ammonia, when a blue fluo- rescence reveals the presence of galbanum.I n estimating the saponifi- cation number, heat must; be avoided. The following fractional method gives the resin number and gum number. Two quantities of 1 gram each are digested for 21 hours in stoppered flasks with 50 C.C. of light petroleum and 25 C.C. of normal alcoholic potash ; one quantity is then diluted with 500 C.C. of water and titrated with N/2 sulphuric acid, which gives the resin number; whilst to the other, 25 C.C. of N/2 aqueous potash and 75 C.C. of water are added, and after another 24 hours the mixture is titrated as before, giving the saponification number. The difference is the gum number. A good specimen should show high acid and resin numbers, and a low gum number. Detection and Estimation of Santonin in the Flower Buds of Artemisia maritima.By K. THAETEB (As*ch. Pharm., 1897, 235, 401-414).--To estimate the quantity of santonin, the flowers are ex- tracted with ether, the residue left on evaporating the ether is digested with milk of lime, the filtrate treated with aluminium acetate, excess of magnesia added, and the thoroughly dried mass again extracted with pure anhydrous ether. Three different kinds yielded 2.26, 2.43, and 2-78 per cent. of santonin respectively. The author finds that Kippen- berger’s method (Ber. deutsch. phawn. Ges., 4), recommended by Thomas (Zeit. anal. Chem., 34, 294), is not generally applicable owing to the extreme difficulty of extracting santonin by means of glycerol containing tannin. Tannin precipitates santonin from aqueous but not from alcoholic solutions.With sulphuric acid alone, santonin does not give any coloration, but it is easily detected by means of furfuraldehyde sul- phuric acid, with which it gives first a carmine-red coloration on warm- ing, becoming bluish-violet, and finally dark blue ; after prolonged digestion, a black precipitate is formed. By this means 0.0001 gram may easily be detected. The author has compared the effects of this test on a large number of alkaloids, bitter principles, glucosides, &c., but only the following gave characteristic colour reactions. a-Naph- tho1 gives first a carmine-red and then a violet coloration which per- sists for an hour. P-Naphthol turns orange and then cherry-red. Veratrine gives various colours-green, red, blue-becoming violet, and finally brown; and with sulphuric acid alone a yellow, then car- mine-red, and finally a brown colour.Picrotoxin with furfuraldehyde sulphuric acid produces a very persistent violet, and piperine a light- M. J. 8.60 ABSTRACTS OF CHEMICAL PAPERS. green coloration, passing through sea-green to bluish-green and indigo blue. With sulphuric acid alone, the latter gives a reddish-yellow Caffeine, By GUILFORD L. SPENCER (J. Amerr.. Chern. Xoc., 1897, 19, 279--28l).-The author prefers estimating caffeine in tea by means of Gomberg’s volumetric iodine method (Abstr., 1897, i, lag), instead of the gravimebric process. To obtain a suitable solution, 5 grams of finely ground tea is boiled for half an hour with about 400 C.C. of water; a considerable excess of recently prepared iron hydroxide is added, and after digesting for an hour the liquid is cooled and diluted to 500 C.C. An aliquot part is then filtered ofi and titrated by Gomberg’s method. Modification of the Thalleioquinine Test for Quinine. By F. S. HPDE (J. Anterr.. Chena. Xoc., 1897, 19, 331---332).-Some of t8he sus- pected alkaloid (0*003-0*005 gram) is dissolved in 5 C.C. of water with the aid of one drop of dilute sulphuric acid (1 : 4), and a clear solution of bleaching powder is added until the blue fluorescence has just disappeared ; a few drops of dilute ammonia (1 : 3) are then added, when a clear, emerald-green coloration should appear. Tbjs is more certain than the ordinary method of testing with chlorine or bromine Behaviour of Proteids with Aldehydes. By ERNST 0. BECK- MANN with H. SCHARFENBERGER GEN. SERTZ and 0. ELSNER (CIum. Centr., 1896, ii, 930-93%).-See thisvol., i, 55. Estimation of Albumin in Urine. By WASSIL~EFF (Chem. Centrr., 1896, ii, 1012 ; from St. Yetersb. Wocl~ensch~., 1896, 331).-For gravi- metric estimation, the urine is mixed with 4 volumes of 95 per cent, alcohol, and the vessel immersed for 3-5 minutes in hot water, The precipitate is then collected, dried, weighed, and its ash (which never exceeds 1 per cent.) subtracted. For clinical purposes, albumin in urine can be titrated with salicylsulphonic acid. 10-20 C.C. of urine (acidified with acetic acid if alkaline) is diluted with water, mixed with 2 drops of a 1 per cent. aqueous solution of ‘‘ Fast Yellow ” (Echtgelb), and titrated with a 25 per cent. solution of salicylsulphonic acid until a permanent brick-red colour is obtained. One C.C. precipitates 0.01006 gram of albumin, and even at a dilution of 1 : 50000 produces a distinct turbidity. coloration which becomes brown on warming. E. w. w. L. DE K. water and ammonia. L. DE K. M. J. S.