ORGANIC ANALYSIS 359 ORGANIC ANALYSIS. Angeli’s Reaction for Aldehydes. 0. Baudisch and J. H. Coert. (Bey., 1912, 45, 1775-1779.)-Angeli’s test for aldehydes is based on the fact that an aqueous solution of the sodium salt of nitrohydroxylaminic acid forms with various aliphatic and aromatic aldehydes hydroxamic acids which are readily detected in minute quantities by their characteristic iron reaction.The sodium salt of nitro- hydroxylaminic acid first splits up into nitrosyl-sodium and sodium nitrite : Na0.N : NO.ONa= NONa+ NaNO,. The nitrosyl-sodium, or perhaps free nitrosyl, combines with the aldehyde ; the hydroxamic acid produced may be isolated in the form of its internal complex copper salt. Form-hydroxamic acid may also be produced by the action of light on an aqueous formaldehyde or methyl alcoholic solution of potassium nitrate or nitrite, owing to the formation of nitrosyl or nitrosyl-potassium.In this reaction it appears probable that the combination of nitrosyl with formaldehyde or methyl alcohol takes place through intermediate stages, forming first nitroso-methyl alcohol, which then changes to methylene-nitronic acid, and, finally, to form-hyrdoxamic acid.If this course be actually followed, the formation of the nitroso alcoholate should be accom- panied by the temporary appearance of a blue-green coloration. I n performing Angeli’s test in the ordinary way no such coloration can be observed ; the change is too rapid. But in presence of certain substances-e.g., methyl acetate, acetone, ethyl acetate, or gelatin-Angeli’s salt shows with aldehydes the blue-green colour of the intermediate nitroso compound.Angeli’s salt is dissolved in a little water with a, large excess of methyl acetate, and the mixture is shaken; aqueous form- aldehyde is added, the liquid being rotated; the methyl acetate then acquires a blue-green colour, which lasts for several seconds. J. F. B. Precipitant for Amino-Acids.C. Neuberg and J. Kerb. (Biochem. Zeitsch., 1912, 40, 498-512.)-When a dilute solution of an amino-acid (for instance, glycin) is rendered distinctly alkaline with sodium carbonate, and then treated with a 25 per cent. mercuric acetate solution, a white precipitate is obtained, and, on the addition of from 4 to 5 volumes of alcohol, the amino-acid is precipitated almost completely.The following results were obtained with various amino-acids, the figures representing the percentage of the total nitrogen precipitated : Glycin, 95.7 ; d, I-alanin, 95-8 ; d-alanin, 97-3 : d, I-valin, 70.0 ; leucin, 96-6 ; Z-asparagin, 98-6 ; d-glutaminic acid, 97.2 ; d, I-serin, 94.9 ; I-cystin, 98.1 ; d, Z-phenylaknin, 97.9 ; I-tyrosin, 96.9 ; d, I-prolin, 75.4 ; I-histidin, 99.4 ; I-tryptophan, 97.7 ; d-glucosamine, 99.5 ; a-P-diaminopropionic acid, 99.4 ; d, I-isoserio, 97.8 ; S-amino- valeric acid, 96.1.w. P. s.360 ABSTRACTS OF CHEMICAL PAPERS Detection and Estimation of Arsenic in Organic Compounds. G. Bres- sanin. (Gazx. Chinz. Ital., 1912, 42, 451-455.)-The method of estimating arsenic by means of the insolubility of its iodide in sulphuric acid may also be used with organic compounds.Treated with potassium iodide in sulphuric acid of sp. gr. 1-45, precipitates of different kinds were obtained with commercial organic compounds. Thus atoxyl gave a maroon precipitate ; sodium cacodylate, a vinous red ; and arrhenal, a yellow precipitate. The precipitations were not complete, and E hrlich's '' 606 " (di-hydroxy-diamino-arseno-benzene hydrochloride) gave no precipitate in acid of this concentration, For the estimation of arsenic in such compounds it is necessary to decompose them completely by heating 0.2 to 0.4 grm.with 10 C.C. of sulphuric acid over a very minute flame (to prevent liberation of cacodylic oxide), after which the arsenic may be quantitatively precipitated a8 described before (Zoc.cit.). Ehrlich's '( 606 " is sold as a canary-yellow powder, de- composing at 175O C . without melting. It dissolves in sulphuric acid of sp. gr. 1-84 with decomposition, and the solution, when diluted to sp. gr. 1.45, now gives a, pre- cipitate with potassium iodide. Silver nitrate gives a yellow coloration, followed by a gelatinous precipitate ; copper salts give a blue precipitate ; picric acid a yellow precipitate ; and potassium ferrocyanide a green precipitate, emitting hydrocyanic acid on heating.A positive reaction is obtained with Gutzeit's test, and Fehling's solution is reduced on heating. C. A. M. Esterification of Monamino Acids by Means of Ethyl Iodide. Separa- tion of Pyrrolidonecarboxylic Acid from Glwtamic Acid.E. Abderhalden and K. Kautzsch. (Zeitsch. physiol. Chew,., 1912, 78, 115-127.)-The silver salt of pyrrolidonecarboxylic acid is readily esterified when boiled for two hours under a, reflux condenser with an excess of ethyl iodide. After the silver iodide has been removed by filtration, the ester may be obtained i n the form of a white crystalline mass by evaporating the solution under reduced pressure over sulphuric acid; it melts at 60" to 61.5" C.Under similar conditions, glutamic acid, aspartio acid, asparagin, and prolin remain unaltered. When a mixture of the silver salts of pyrrolidonecarboxylic acid and glutamic acid is boiled with ethyl iodide, the former is esterified, and may be separated from the insoluble silver salt of glutamic acid by means of benzene. Attempts to isolate pyrrolidonecarboxylic acid in the digestion products of casein resulted in the formation of an estex which could not be identified ; this ester yielded glutamic acid hydrochloride when hydrolysed with hydrochloric acid.w. P. s. Influence of Metallic Carbonates on the Estimation of the Volatile Sub- stances in Coal. E. Prost and M. Ubaghs. (BUZZ.Xoc. Chim. Belg., 1912,26, 216-223.)-Estimation of the volatile substances in coal may give an erroneous idea of the calorific value of the coal or its suitability for gas manufacture, unless account be taken of the amount of carbon dioxide in the volatile substances. I n the case of two samples of coal more than 2-5 per cent. of carbon dioxide was present in the volatile substances, whilst other samples gave amounts ranging from 0-5 to 2 per cent.The gas was found to originate from carbonates of iron, calcium andORGANIC ANALYSIS 361 magnesium in the coal. This was found to be the explanation why a coal yielding 15 to 16 per cent. of volatile substances was suitable for the manufacture of coke, whilst another coal from the same source, yielding more than 17 per cent.of volatile substances, was quite unsuitable for the purpose. C . A. M. Combustion of Volatile Organic Liquids. L. Clarke. (J. Amer. Chem. SOC., 1912, 34, 746-747.)-A combustion tube is drawn out at one end, and a tube bearing a female ground glass joint is sealed on. That part of the tube which is within the heating zone of the furnace is filled with copper oxide, held in place by plugs as usual, and the end remote from the ground glass joint is provided with a perforated rubber stopper for attachment to the usual sulphuric acid tube and potash bulb. The end of the tube bearing the ground glass joint protrudes about 4 inches from the end of the furnace.The liquid to be analysed is weighed out in a small stoppered U-tube of the type in which the stoppers close the limbs, and, by turning, may be made to put the side tubes in or out of connection with the U.To one side tube is sealed a male ground glass joint, which has been ground to fit the ground seat on the combustion tube. The stoppers of the U-tube and the joint between U-tube and combustion-tube are very lightly greased with vaseline for the upper third or fourth of their ground surfaces.The unlubricated space below, wherever pervious, is filled with air, forming an effective cushion to prevent absorption by the vaseline of any vapour, should this be soluble in paraffin. After weighing, the U-tube is con- nected to the hot combustion-tube, the other arm being connected to the usual apparatus for hying the air on its way to the combustion-tube.The stopcocks are turned, and the stream of air allowed to pass over the surface of the liquid, carrying forward vapour to the combustion-tube. If the liquid has a boiling-point as low as that of ether, it is well to immerse the lower end of the U-tube in cold water to prevent too rapid volatilisation. For liquids boiling at 100" to 125" C., it is best to immerse the lower part of the U-tube in water at 50" to 60' C.With acetone and similar liquids, neither heating nor cooling is necessary. G. C . J. Estimation of Caoutchouc as Tetrabromide. Utz. (Gummi Zeit., 1918, 26, 968-970 ; through Chem. Zeiztralbl., 1912, l., 1797-1798.)-1n order to avoid the loss of bromine which various authors have found to take place when caoutchouc tetrabromide, prepared by Budde's method, is heated with nitrio acid in presence of silver nitrate, the author recommends the method of Baubigny and Chavanne for the estimation of the bromine.The caoutchouc tetrabromide is decompofied by means of chromic and sulphuric acid mixture; a slight modification in the apparatus, consisting in the introduction of a stoppered dropping funnel, is advised. The author, independently of Hiibener, has devised a method of bromination in aqueous solution : 0.1 grm.of caoutchouc is treated in a Lintner's pressure-flask with 60 to 70 C.C. of water and 5 to 7 C.C. of bromine, and heated in the closed flask for some hours in the water-bath at 50' to 60" C. When the bromination is complete, the liquid is poured off through a filter, and the contents of the flask are extracted with boiling water.The residue of caoutchouc bromide is washed with alcohol and ether and digested with chloroform for six hours at the ordinary temperature. The362 ABSTRACTS OF CHEMICAL PAPERS bromide is reprecipitated by an excess of petroleum spirit, and, after settling, is collected on a filter, washed with akohol and ether, dried in wacuo at the ordinary temperature, and the bromine estimated in Baubigny and Chavanne’s apparatus. The bromine found ranged from 70.12 to 70-25 per cent.in six estimations; theoreti- cal, 70.14 per cent. for CloH,,Br,. A sample of Para rubber showed 93.96 and 93-72 per cent. of caoutchouc by this method. The author is investigating the error due to the presence of nitrogenous impurities in the raw rubber.J. F. B. Latent Heats of Vaporisation of Mixed Liquids. Part 111. : Mixtures of Associated with Non-associated Liquids, and the Detection of Solvates in Mixtures of Liquids. D. Tyrer. (J. Chem. SOC., 1912, Pol, 1104-1113.)-Resnlts are given for the boiling-points, vapour compositions, latent heats at constant pressure, and latent heats at constant composition for the following mixtures : Benzene and ethyl alcohol, chloroform and methyl alcohol, chloroform and acetone, carbon tetra- chloride and ethyl alcohol.I t was found that the mixture of chloroform and acetone gives a maximum boiling-point ; the three other mixtures give minimum boiling-points. From a study of the latent heats of such mixtures a method is established by which the existence of solvates in the mixtures can be detected.I n all four cases studied solvates are formed, Molecular association in liquids is the result of the combination of hydroxyl groups, and when two hydroxyl substances are mixed solvates are always formed. When a hydroxyl substance is mixed with one containing no such group, the latent heat affords a reliable test for the formation of solvates.In normal mixtures each constituent retains its own latent heat unaffected by the presence of the other constituent ; thus the latent heat of either constituent can be calculated from the latent heat of the mixture and the latent heat of the other constituent. Applying this rule to the case of an associated liquid in admixture with a non- associated liquid, the apparent latent heat of the associated constituent can be calculated at various concentrations.If no solvates are formed, the latent heat of the associated liquid in the mixture should gradually diminish as the concentration gradually diminishes, owing to the progressive dissociation of the associated mole- cules, until at infinite dilution it becomes constant, indicating complete dissociation.If this is not the case, then undoubtedly to some extent solvates are formed in the mixture. I n the mixtures above mentioned the diminution was not steady, but the latent heat in all cases, over a portion of the range of concentrations, increased with increase of the dilution. This means that solvates are formed, but it is not possible to determine their amount or composition.The acetone apparently functions in the above case as an hydroxyl substance. J. F. B. Estimation of Ferrocyanides. H. E. Williams. (J. SOC. Chem. Ind., 1912, 31, 468-471.)-The following process is stated to be free from most of the possible sources of error inherent in the Feld method: Half a gram of the ferrocyanide is dissolved in 100 C.C. of water, and the liquid distilled after adding 0.05 to 0.1 grm.of cuprous chloride (dissolved in a few drops of hydrochloric acid, or a little saturated sodium or potassium chloride solution) and 25 to 30 C.C. of 4N sulphuric acid. The distillate is passed from the condenser into absorption-flasks containing caustic alkaliORGANIC ANALYSIS 363 solution. The ferrocyanide is completely decomposed, and the hydrocyanic acid dis- tilled, by about half an hour's gentle boiling. The cyanide in the alkaline hydroxide solution is then titrated in the usual way with standard silver nitrate solution, in the presence of a few drops of potassium iodide solution.Much larger quantities of ferrocyanide may be taken if desired, and in such cases it is only necessary to add an amount of cuprous chloride equal to 2 to 3 per cent.of the weight of ferrocyanide. Sufficient sulphuric acid must be present to give an excess of acid over that required to form ferrous sulphate and sodium bisulphate. The cuprous chloride used should first be freed from cupric salt by washing with a little dilute sulphuric acid. The addition of cuprous chloride decomposes the ferrocyanide, forming cuprous cyanide, which is decomposed by the acid present, yielding hydrocyanic acid and cuprous chloride again, which latter then attacks a further quantity of ferrocyanide.Both soluble and insoluble ferrocyanides may be determined by the method, but if the substance be dry it should be finely powdered. Ferric ferrocyanide requires more cuprous chloride than above stated for its complete decomposition.For the estimation of the ferrocyanide content of cyanogen mud or spent oxide, a weighed quantity of the dried material, from which the sulphur has been rernovcd by extraction with carbon disulphide, is boiled with sodium hydroxide, filtered, washed to a definite volume, and a portion of the liquid distilled with cuprous chloride and sulphuric acid as before.The results on spent oxides are higher, gener- ally by about 0.2 to 0.4 per cent,, than those given by the titration method, and are more accurate. Thiocyanates are stated to have no influence on the distillation ; if present in spent oxide, a larger amount of cuprous chloride should be added to insure their precipitation before distilling (cf. Colman, ANALYST, 1910, 35, 295).A. R. T. Direct Estimation of Geraniol in Citronella Oil. J. Dupont and L. Labaune. (Ann. Chim. anal., 1912, 17, 210-213.)--The method depends upon the conversion of the citronellal into a nitrile, which is not affected by the potassium hydroxide in the estimation of the geraniol as geranyl acetate. Ten grms. of citronella oil (or of a mixture of geraniol and citronellal) are shaken for two hours at the ordi- nary temperature with an aqueous solution of hydroxylamine, prepared by dissolving 10 grms.of hydroxylamine hydrochloride in 25 C.C. of water, adding a solution of 12 grms. of potassium carbonate in 25 C.C. of water, and filtering the mixture. The oil rising to the surface'is decanted, dried over anhydrous sodium sulphate, and boiled under a reflux condenser for one and a half hours with twice its volume of acetic anhydride.Under these conditions the citronellal oxime is converted into citronello nitrile. After washing, neutralisation, and drying, 2 grms. of the acetylated oil are saponified with alcoholic potassium hydroxide solution, and the amount of geraniol calculated from the result in the usual way. The difference between the total products of acetylation in a preliminary test and the amount of geraniol thus found, gives the quantity of citronellal.Thus, a Ceylon oil containing 60.2 per cent. of acetylisable products was found to contain 43 per cent. of geraniol, whilst a Java oil contained 43 per cent. of geraniol and 40 per cent. of citronellal. The results given by test mixtures were in close agreement with theory.C. A. M.364 ABSTRACTS OF CHEMICAL PAPERS Oil of the Southern Cypress. A. F. Odell. (J. Amer. Chem. Soc., 1912,34, 824-826. )-The cones of the Southern cypress (Tuxodiwu distichurn) contain con- siderable quantities of a volatile oil, consisting of d-pinene, 85 per cent. ; d-limonene, 5 per cent. ; a pseudo terpene alcohol (? sabinol), 2 per cent.; carvone, 3 per cent.; tricyclic sesquiterpene, 3 per cent. ; the remainder being composed of substances boiling above 275" C. No aldehydes were found in the oil. H. F. E. H. Estimation of Hydrazine. G. S. Jamieson. (Chenz. News, 1912, 105, 268.) -According to r;he author, Rimini's method requires much time, while the following procedure, based on the general method of Andrews (ANALYST, l903,28,306),is stated to be rapid and accurate : Hydrazine salts, or the double salts of hydrazine and the metals, are analysed by titration with a standard solution of potassium iodate (3.567 grms.KIO, per litre) in presence of hydrochloric acid and chloroform. Suitable propor- tions are 20 C.C. water, 30 C.C. hydrochloric acid, and 6 C.C. chloroform. The solution of potassium iodate is run in gradually, the reaction proceeding according to the equation : The liquid is shaken between each addition of iodate, until the chloroform, which at first is coloured, becomes decolorised.Metallic double salts (such as zinc and hydrazine sulphate) may be examined by this method, and the published results are accurate. I n the case of the nickel double salt, which .is only sparingly soluble, the reaction proceeds very slowly.N,H,*EI,SO, + KIO, + 2HCl= N, + ICI + 3H,O + KC1 + H2S04. A. R. T. Detection and Estimation of Methyl Alcohol. F. Wirthle, (Chem. Zeit., 1912, 36, 700.)-In testing spirits for methyl alcohol, the first 50 C.C. (at most) of the alcoholic distillate from 100 C.C. are distilled and twice fractionated, so as to obtain first 15 C.C.and finally 3 C.C. as the final fraction. One C.C. of this distillate is oxidised by the official potassium perinanganate method, after the addition of 4 C.C. of dilute (1 : 2) sulphuric acid, the liquid filtered, and 1 C.C. of the colourless filtrate heated with 5 C.C. of concentrated sulphuric acid. On treating the cold mixture with 5 mgrrns. of solid morphine hydrochloride, a reddish-violet coloration is pro- duced in the presence of as little as 20 mgrms.of methyl alcohol (corresponding to about 5 per cent of the denatured alcohol in the spirit). At the same time a control test is made with a mixture of 20 mgrms. of methyl alcohol and 1 C.C. of 90 per cent. ethyl alcohol. If the preliminary test has given a strong violet coloration, and the spirit contains about 40 per cent.(by volume) of alcohol, 60 C.C. (from 100 c.c.) are taken, whilst with a higher alcoholic strength a proportionately smaller quantity of the distillate is used, The initial temperature of distillation in a fractional distilla- tion-flask affords information as to the amount of methyl alcohol present. Thus a, liquid containing 10 per cent.of methyl alcohol begins to distil at 75' C., and after five minutes the temperature will have risen to 76.8' C., whilst in the presence of 20 per cent. of methyl alcohol the initial temperature will be 71.5' C., rising, after five minutes, to 75.5' C. It is advisable to treat the residue from the first distillation with alcohol and to fractionally distil it again, the process being repeated so that from solutions containing about 10 per cent.of methyl alcohol about 80 C.C. in allORGANIC ANALYSIS 365 are obtained, from 20 per cent. solutions about 100 c.c., and from 30 per cent. soh- tions about 125 C.C. The mixture of methyl and ethyl alcohols is now converted into iodides by heating it under a reflux condenser with iodine and amorphous phosphorus ; these are eeparated by fractional distillation, and the methyl iodide identified by its saponification value (394.3). For the conversion of 10 C.C.of methyl alcohol into methyl iodide 31-7 grms. of iodine and 1-79 grm. of amorphous phos- phorus are required, whilst ethyl alcohol requires only 22 grms. of iodine and 1.79 grm. of phosphorus for the formation of the corresponding compound.The necessary amounts may thus be approximately calculated from the temperatures observed in the fractional distillation, a slight excess of iodine being taken. The iodides from a quantity of distillate corresponding to 10 C.C. of alcohol are distilled into 10 C.C. of water, decolorised by shaking with 14 to 16 C.C. of 10 per cent. potas- - sium hydroxide solution, separated from the supernatant liquid, washed twice with water, allowed to stand overnight, and then run into a tared flask, which is im- mediately closed and weighed.The small residue of iodides remaining in the burette is collected in a narrow calibrated tube, and its weight calculated from the specific gravity, Since 10 C.C. of methyl alcohol yield 31-35 grms. of methyl iodide, and 10 C.C.of ethyl alcohol yield 24.0 grms. of ethyl iodide, each 0.736 grm. of iodide in excess of 24 grms. corresponds to 1 C.C. of methyl alcohol. To obtain reliable quantitative results it is advisable to dilute the distillate from the spirit under examination with sufficient 96 per cent. ethyl alcohol to give a mixture con- taining approximately 10 per cent. of methyl alcohol, whilst at the same time a blank test is made with a mixture of methyl and ethyl alcohols containing 10 per cent.of the former. C. A. M. Activity of Organie Nitrogen as measured by the Alkaline Perman- ganate Method. C. H. Jones. (J. Ind. Eng. Chem., 1912, 4, 438-441.)-1t is generally recognised that soluble organic nitrogen is readily available as plant-food, whereas nitrogen in insoluble combinations may or may not be so available.Thus, though dried blood and hoof meal contain most of their nitrogen in insoluble forms, they are known to possess high manurial value, whereas there is reason to believe that the nitrogen of peat, roasted leather, garbage tankage, and some other materials, is not immediately available as plant-food, For this reason certain States of the American Union forbid the use of these latter materials in compound fertilisers offered for sale.I n other States, agricultural chemists are called upon to report on the manurial value of mixed fertilisers which may contain nitrates, ammonia, and soluble and insoluble organic nitrogenous substances. The valuation of the insoluble nitrogen has hitherto presented a difficulty for the reasons stated, and the author now proposes to differentiate between the insoluble nitrogen which is converted into ammonia by digestion under standard conditions with alkaline permanganate and that which is not so transformed. That which is converted into ammonia under the conditions described in the paper is designated active insoluble organic nitrogen, and it is suggested that this active nitrogen should be assigned a manurial value but little inferior to the soluble nitrogen, whereas the residue of insoluble nitrogen, designated inactive, is held to have a very low value indeed.These views find support in field366 ABSTRACTS OF CHEMICAL PAPERS experiments-referred to but not described in the paper-and in the fact that in dried blood, high grade tankage, fish, hoof meal, cotton-seed meal, and many other sub- stances which give manurid results in close relation to their nitrogen contents, the proportion of inactive nitrogen, as determined by the author's method, is low, whereas the same method shows that the inactive insoluble organic form is the predominant form of nitrogen in peat, garbage tankage, charred leather, and other substances of notable nitrogen content but negligible manurial value.Active insoluble organic nitrogen is determined as follows : Total insoluble nitrogen is first determined by extracting 2 grms. of the sample on a filter-paper, with water at room temperature, until the filtrate amounts to 250 C.C. The filter- paper and its contents are subjected to the Kjeldahl process.So much of the sub- stance is now weighed out as contains 50 mgrms. of insoluble nitrogen, and, if a mixed fertiliser, is extracted with water, as in the determination of total insoluble nitrogen. Highly nitrogenous raw materials are mixed in a mortar with 2 grms. of powdered rock phosphate before extracting with water, and very oily materials should have a preliminary treatment with ether.The filter and insoluble matter are dried at a temperature not exceeding 80" C., and the insoluble matter is then transferred to a 500 to 600 C.C. round-bottomed flask, together with 20 C.C. of water, a few glass beads, and 100 C.C. of an alkaline permanganate solution containing 150 grms. of caustic soda and 25 grms. of potassium permanganate per litre.The flask is con- nected to a condenser as for an ammonia determination, the receiver being charged with a measured quantity of standard acid. The contents of the flask are digested short of boiling for at least thirty minutes, local overheating of the insoluble residue being avoided by the use of wire gauze and asbestos-paper between the flame and the flask. About 95 C.C.are now distilled, preferably at such a rate that the operation requires an hour. A tendency to froth may be overcome by prolonging the digestion until this no longer manifests itself, and then distilling at a somewhat brisker rate, so that the whole process occupies about ninety minutes as usual. The ammonia in the distillate is the measure of the active insoluble nitrogen. The method has been adopted since March, 1911, by the experiment stations of New York, New Jersey, and the New England States, as the result of laboratory and field experiments in 1910.G. C. J. Action of Permanganate on Organic Substances. J. Hetper. (Zeitsch. anal. Chem., 1912, 51, 409-429.)-111 earlier papers (ANALYST, 1911, 36, 232, 422) the author has shown that many organic substances are quantitatively oxidised to carbon dioxide and water by permanganate in acid solution, whilst the members of another large group of compounds may also be estimated by means of permanganate, the reaction being a quantitative one, with acetic acid for one of the end products.The influence of alkaline permanganate has now been studied. In the majority of cases it proves a less useful reagent than an acid solution.For example, ethyl alcohol, which can be quantitatively oxidised to acetic acid in acid solution, is oxidised partly to acetic acid and partly to carbon dioxide in alkaline solution, the former reaction predominating if the action begins in the cold, the latter if the mixture is hot from the start ; but at no temperature between 0" and 100' C.can either reaction be sup-ORGANIC ANALYSIS 367 pressed entirely. Methyl alcohol is much more quickly oxidised in alkaline than in acid solution, the products being carbon dioxide and water in either case. Caffeine, theobromine, and antipyrin, which only consume from 40 to 80 per cent. of the theoretical amount of permanganate in acid solution, are almost quantitatively oxidised to carbon dioxide, water (and ammonia) in alkaline solution.Many sub- stances which are oxidised quantitkively in acid solution behave similarly in alkaline solution, but in a much greater number of instances the results in alkaline solution correspond to no single reaction. However, the author thinks that the consumption of permanganate in alkaline solution under standard conditions may prove a useful constant, especially when compared with the corresponding number obtained in acid solution.The experimental conditions followed in the author’s laboratory are as follows : The permanganate solution is :, and contains 40 gr. of caustic soda per litre. Of this solution, 25 C.C. is placed in a 200 C.C. flask with 65 C.C. of water, and the sub- stance dissolved in 10 C.C.of water. So much of the substance is taken as may be expected to reduce about 10 C.C. of the permanganate, and, if the substance is insoluble in water, it is mixed with 1 grm. of glass powder, not omitting in such a case to add 10 C.C. of water to the contents of the flask to insure uniformity of conditions in all experiments. The fiask, with its contents, is heated on the water-bath for one and a half hours ; 25 C.C.of oxalic acid (containing 75 C.C. concentrated sulphuric acid per litre) is then added, and the solution titrated with permanganate containing 40 grms. of glacial phosphoric acid per litre. The result is calculated to C.C. of +’ permanganate per gram of substance. The amount of permanganate reduced by forty-two sub- stancesunder these and some other conditions are given in the paper.* G. C. J. Influence of Peptones on the Estimation of Reducing Sugars by Fehling’s Solution. A. Bernardi. (Biochem. Zeitsch. 1912, 41,160-164.)-Although peptones do not reduce Fehling’s solution, their presence in sugar solutions causes more cuprous oxide to be precipitated than is due to the reducing action of the sugar alone.In cases, therefore, where sugar has to be estimated in solutions containing peptones, the latter should be removed by means of phosphotungstic acid previous to the estima- tion of the sugar. w. P. s. Physico-Chemical Basis of the Seliwanoff Lamdose Reaction. A. Jolles. (Biochem. Zeitsch., 1912, 41, 331-332.)-When dextrose is placed in contact with concentrated hydrochloric acid for some months, the solution yields a reaction for laevulose with the Seliwanoff test.The author finds, however, that the polarisation of the sugar is not changed, a fact which shows that either the quantity of laevulose formed is so small that it cannot be estimated, or, which is more probable, that the substance giving the Seliwanoff reaction is not laevulose. The action of dilute alkali solutions on dextrose is much more pronounced, and the change is accompanied by the formation of appreciable quantities of an acid.For instance, 200 C.C. of a 0% per cent. dextrose solution were mixed with about 40 C.C. of -& sodium hydroxide solution ; 50 C.C. of this mixture required 9-1 C.C. of Fo- acid for neutralisation when titrated at once, but after the lapse of ninety-four hours the same quantity of the368 ABSTRACTS OF CHEMICAL PAPERS mixture used only 7.7 C.C.of the acid, phenolphthalein being used as the indicator in both cases. The author also found that the Seliwanoff test would detect the presence of as little as 0.1 per cent. of lmvulose in urine containing dextrose. w. P. s. Method for Determining the Value of Commercial Starches for Use in Cotton Mills.G. M. MacNider. (J. Ind. Eng. Chem., 1912, 4, 417-422.j-The value of starch for cotton-mill purposes depends on its property of swelling and forming a viscous solution when treated with hot water, the principal starches used being maize, potato, cassava, and, to a small extent, wheat, sago, and rice. As each of these starches has a different viscosity, to obtain the best results they must each be used in a diff'erent manner. No definite relation exists between their behaviour and the size or shape of the grains.The author determines the viscosity of each starch at a high temperature under conditions similar to those in which it is used in the mills, employing for the purpose a Scott viscosimeter in the following manner : 12 grms. of the starch are weighed intc a 600 C.C.beaker, 300 C.C. distilled water added (thus making a 4 per cent. solution), and heated over a Bunsen burner, with constant stirring, to the boiling-point and boiled for ten minutes; 200 C.C. of this solution are then poured into a viscosimeter cup, which is jacketed with boiling water, the temperature allowed to become constant, which it does with most starches, at 94' C., and 50 C.C.run out into a graduated vessel, the time being measured with a stop-watch. The number of seconds so required divided by the number of seconds required to deliver 50 C.C. of boiling water (nine seconds) gives the viscosity value. The recorded results show that not only is there a' difference between different starches, but also a considerable variation in the viscosity of different samples of the same starch, only a small part of this latter difference being due to variation in moisture content. The viscosity of maize starch increases uniformly with the length of time of boiling, and corresponds to the increase in concentration brought about by continued boiling.Potato starch reaches its maximum viscosity after being boiled five minutes ; it then decreases rapidly, the increasing concentration having no effect and after thirty minutes' boiling it has fallen below that of maize starch, which has been boiled for the same length of time.Cassava starch attains itis maximum viscosity at the boiling-point, after which it decreases uniformly with the length of time of boiling, and thus resembles potato starch, though to a less extent; the former starch, therefore, has a much wider application in sizing and finishing.Wheat starch has a much lower viscosity than any of the others examined, and shows a gradual but small increase with the time of boiling. Wheat starch foams more in boiling than any other. Rice starch, even at the end of thirty minutes' boiling, still shows a value only very slightly greater than unity.The degree to which starches go into solution on continued boiling was examined, and it was noticed that those from roots and tubers developed under ground (potato and cassava) showed a high rate of solution, while This is a valuable property of maize starch as compared with others.ORGANIC ANALYSIS 369 those from the grains developed above ground were very insoluble. An examination was made of some ‘( thin boiling starches,” which usually consist of maize starch treated in some way to reduce the viscosity.Experiments are described showing the effect of boiling maize starch with borax, alkali, and boric acid, as regards viscosity changes. A small amount of all these reagents reduces, while a slightly larger amount increases, dscosity.H. F. E. H. Estimation of Thymol, Salicylates, and Allied Compounds. A. Seidell. (Amer. Chem. J., 1912, 47, 508-526.)-The method given in the author’s previous paper on this subject (ANALYST, 1909, 34, 536) is admittedly tedious, and was only suggested as a possible improvement over the ordinary bromate titration in which tribromo-phenol bromide is formed.The control of the too vigorous action of bromine on the compounds under examination has now, in the author’s opinion, been accomplished more satisfactorily. The method is as follows: A quantity of 0.1 to 0.5 grm. of thymol or other substance to be estimated is placed in a glass- stoppered bottle, with 1 to 2 C.C. of carbon tetrachloride and 100 C.C. of water.Ehmine vapour is poured in till the mixture, after thorough shaking, shows a con- siderable excess. After half an hour, 5 C.C. of carbon disulphide, immediately followed by 5 C.C. of aqueous potassium iodide (20 per cent.), are added, and the liberated iodine titrated with Fc thiosulphate. An additional amount of potassium iodide is added, and if there is no further liberation of iodine, the burette reading is taken, 5 C.C.or an excess of aqueous potassium iodate solution (2 per cent.) added, and the titration, after thorough shaking, continued till the iodine colour is discharged. The second titration corresponds to the hydrobromic acid formed by the action of bromine on the compound used. Two molecules of hydrobromic acid are equivalent to 1 molecule of thytuol; 1 C.C.Fc thiosulphate, therefore, is equal to 0.0075 grm. thymol. The results quoted vary from 99.0 to 100.8 per cent. of the theoretical. Carbon disulphide and tetrachloride serve to remove the compound from the aqueous layer, and the reaction in the organic solvent is more under control, only dibromothymol being produced. Carbon tetrachloride, unlike carbon disulphide, is little acted on by bromine on standing, and is therefore added first ; but a satisfactory end-point in the titrations is not obtainable without the addition of the disulphide as described.Salicylic acid may be determined when in admixture with benzoic acid, the latter being unacted on. Other substances tried gave variable results. A. R. T. Tetraformal-trisazine from Formaldehyde and Hydrazine Hydrate, a New Reducing Agent for Analytical Use.K. A. Hofmann and D. Storm. (Ber., 1912, 45, 1725-1730.)-Hydrazine hydrate is employed in analytical chemistry for the precipitation of silver, gold, mercury, platinum, selenium, and tellurium, but its reducing power is too great to permit of fine differentiations among the noble metals, and it also reduces solutions of copper and bismuth. The precipitates are not easily filtered, and firm mirrors are not obtainable on glass.Hydrazine hydrate also attacks glass, and is not free from alkalis. The authors, therefore, recommend, as a reducing agent far more under control and free from alkalis and acid, the com-370 ABSTRACTS OF CHEMICAL PAPERS pound tetraformal-trisazine, C,H,,N,, formed by the combination of formaldehyde with hydrazine hydrate.This substance has the structural formula- HN.CH,.N. CH,. NH For its preparation, 50 C.C. of 35 per cent. formalin are poured slfwly into 50 C.C. of hydrazine hydrate cooled with ice, and after one hour a further 30 C.C. of formalin &re similarly added. The mixture is placed in a basin, covered with paper, and left for one or two days, when the yield of crystallised compound should amount to 27 grms.The crystals are filtered off, washed with spirit, redissolved in water, and precipitated with alcohol and ether ; they may also be recrystallised as flat, silky needles from slightly alkaline hot water. At 28O C., 100 grms. of water dissolve 14 grms. of the azine. Silver nitrate yields first a white precipitate, changing im- mediately on warming to a fine adherent metallic mirror; palladium chloride also yields a mirror.Gold chloride yields first a blue colloid, and later a dark precipitate. Copper sulphate gives a reddish-brown solution. The azine is not decomposed by alkalis, from which it crystallises unchanged. In alkaline solution, copper salts are reduced to cuprous oxide, mercury, gold, and silver salts to the metals, whilst plati- num and palladium chlorides give stable, reddish-brown solutions. I n presence of excess of potassium hydroxide, chromates, molybdates, vanadates, selenites, and tellurites remain unchanged, even at the temperature of the water-bath; but on the addition of ammonium chloride, indigo blue molybdenum oxide, red selenium, or black tellurium are precipitated.Tungstates are not reduced even in ammoniacal solution. Thus tetraformal-trisazine may be recommended as a selective reducing agent which is readily prepared in a perfectly pure form. The substance, though stable towards alkalis, is very sensitive to the action of acids, even carbonic and acetic acids, forming polymerised formalazine, (CH,N,CH,),, as an insoluble white powder.Ammonia is one of the principal products of the action of hot dilute sulphuric acid, J. F. B. HN.CH,.N.CH,. I ‘ 4 H. Examination and Valuation of Turpentine. H. Wolff. (Furbenxeitung, 1912, 17, 1492 ; through Chem. Zentralbl., 1912, I., 1930.)-The colour is more conveniently, determined by a comparison with solutions of iodine, potassium dichrornate, etc., than by means of the Lovibond colorirneter.The polymerisation with sulphuric acid, which involves oxidation, should not be omitted (cj. Marcusson, Chem. Zeit., 1910, 34, 285); otherwise, adulteration with residues from the manu- facture of turpentine may be overlooked. The distillate obtained before and after the chief fraction in the distillation of turpentine are to be found on the market and are difficult to detect when added to turpentine. The greater part of the residual oil distils earlier when mixed with turpentine than it does alone.On the other hand, with concentrated or fuming sulphuric acid, more separates from the mixture than would be anticipated from the amount of residual oil. I t is evident that sulphuric acid has less action in the presence of high-boiling constituents.A separation of more than the normal 2 to 4 per cent. is sometimes shown, when the resinification is well started, even by pure turpentines, in which there is a slight natural increase ofORGANIC ANALYSIS 371 these constituents, and, consequently (petroleum being excluded), is not a proof of adulteration with these residues.A small increase in the residue on evaporation, with a large increase in the amount separated by fuming sulphuric acid, points to an adulteration with these waste products, whereas an increase in the resinifica- tion is of no value as an indication. A residue of 1 per cent. on evaporation has a noticeably bad effect on the drying-period, whilst 2 or 3 per cent. renders a turpentine useless for many purposes in the varnish and colour industry.0. E. &I. Estimation of Petroleum in Turpentine. H. Wolff. (Farbenzeituizg, 1912, 17, 1553 ; through Chem. Zentralbl. 1912, I, 1931.)-The constituents dissolved by nitric acid are, contrary to the statement of Herzfeld (Chem. Zeit., 1910, 34, 885), cyclic hydrocarbons, since they can be reduced by the slow action of nascent hydrogen, the solution being kept warm.I t was found that xylene can be dissolved in concentrated sulphuric acid alone, but that long shaking is necessary; this explains the conflicting results obtained by Marcusson and Herzfeld with turpentines containing xylene. Consequently Herzfeld's method of determining benzene has only a qualitative value, and only then if it is positive and agrees with the other constants (boiling-point, bromine number, etc.).The separation of petroleum or paraffin hydrocarbons by concentrated fuming sulphuric acid is more valuable, especially if, by the method of Ute, the refraction of the separated residue is determined. The Marcusson method is preferable for the exact determination of petroleum and benzene hydrocarbons.In the case of a strongly resinified oil, if there is a considerable separation with concentrated sulphuric acid, even when the Marcusson test gives a negative result, it is better to examine once more the substance separated with sulphuric acid by the Marcusson method. 0. E. M Tests for Determining the Purity of Oil of Turpentine. R. Mareille. (Ann. Falsific., 1912, 5, 241-251.)-Determinations of the solubility of oil of turpentine in acetic acid, the rise in temperature when mixed with sulphuric acid, and the solubility in sulphuric acid, are stated to yield results which will afford evidence as to the genuineness of a sample of turpentine.Solubility in Acetic Acid.-The acetic acid employed should contain 98 per cent. of the acid and 2 per cent.of water; it should be capable of dissolving an equal volume of anhydrous carbon disulphide at a temperature of 39 to 40" C., at which point a mixture of equal volumes of the two liquids should just become turbid ; if the temperature at which the turbidity is observed is lower or higher than this, the strength of the acid must be adjusted accordingly. Five C.C. of the oil of turpentine are placed in a stoppered graduated tube, and shaken with successive additions of acetic acid until a clear mixture is obtained at a temperature of 15' C. The quantity of the acid required to dissolve 1 part of the oil of turpentine is then calculated. For genuine samples this quantity is usually less than 4. The oil of turpentine may also be distilled, the distillate being collected in five separate portions, which are then sub- mitted to the test. The first fractions dissolve in about 3.5 parts of acetic acid, whilst the fifth fraction may require up to 4.5 parts of acid. The presence of petroleum diminishes the solubility of the oil, the fifth fraction of a sample containing 5 per372 ABSTRACTS OF CHEMICAL PAPERS cent. of petroleum requiring from 6 to 7 volumes of acetic acid. Oil of turpentine adulterated with (( benzine ” has a greater solubility than pure oil of turpentine. Thermal Test with Xulphuric Acid.-The method of carrying out this test in the case of fatty oils has been described by Tortelli (ANALYST, 1909, 34, 168). The same procedure is adopted with oil of turpentine, but, as the latter evolves very consider- able heat when mixed with sulphuric acid, it is better to perform the test on a mixture of 5 C.C. of the sample and 15 C.C. of (‘ vaseline oil.” Genuine oil of turpen- tine gives a value of from 68 to 72, and the presence of 10 per cent. of mineral oil will diminish this value by about 7. Solubility in Xulphuric Acid.-One hundred C.C. of petroleum, 20 C.C. of oil of tur- pentine, and 50 C.C. of sulphuric acid are mixed in a Rose tube, and the increase in the volume of the sulphuric acid is ascertained after the lapse of one hour. In the case of genuine oil of turpentine, the increase will be about 19.2 C.C. (that is, 96 per cent. of the oil is soluble in the acid), whilst with an oil containing 10 per cent. of mineral oil the increase in the volume of the acid will be only 17.1 C.C. The test is useless should benzine ” be present. Care should be taken to examine the petroleum used in the test, as some specimens yield notable quantities of matter soluble in sulphuric aoid. w. P. s.