THE ANALYST. 405 ORGANIC ANALYSIS. Use of " Standard Benzine " (Light Petroleum) for the Determination of Asphaltum in Dark Mineral Oils. Holde. (Jfitt. kgl. Materialspriijungsamt Gross-Lichterfelde West, 1909, 27, 143-148 ; through J. SOC. Chem. I d . , 1909, 28, 831.)-Asphaltum in dark mineral lubricating oils is determined by dissolving the406 THE ANALYST. oil in forty times its volume of light petroleum and separating the insoluble residue. The light petroleum to be used for this purpose is supplied by Kahlbaum, of Berlin, t o a specification of a committee of the German Union for Testing Technical Materials. This '' standard benzine " is to be uniform as far as possible, free from unsaturated and aromatic hydrocarbons, with the sp. gr. 0-695 to 0-705 at 15OC. ; at least 95 per cent.must boil between 65" and 95" C. Some analyses of the benzine supplied during the last few years are given in the paper ; unsaturated and aromatic hydrocarbons are' determined by shaking the benzine with fuming sulphuric acid, when the loss should not exceed 2 per cent. I n testing successive batches on the same mineral lubricating oil, it was found that the asphaltum present in the latter increases slightly on keeping. Analysis of Vulcanised Caoutchouc Goods. F. W. Hinrichsen and W. Manasse. (Chem. Zeit., 1909, 33, 735-736.)-1n default of a direct method for the estimation of caoutchouc, the method of precipitation by bromine appears the most promising of the indirect methods. Budde's process, consisting in swelling the rubber with carbon tetrachloride, and brominating with a mixture of bromine and iodine, is probably the best (Gummi Zeit., 1907, 21, 215).With cold-vulcanised goods, the main source of error due to the presence of chlorine in the form of a double compound formed between the rubber and the sulphur chloride, is generally sufficiently small to be neglected. The method is applicable to hot-vulcanised rubber, but the error is greater; it is not applicable in presence of antimony sulphide and lead compounds. Axelrod's process (ibid., 1907, 21, 1779) is not so useful, and the attendant errors are greater. The rubber is dissolved in petroleum of high boiling-point, brominated, and the product is precipitated by alcohol and weighed on a tared filter. In this process the precipitate contains the whole of the combined sulphur, and the factor varies according to the degree of vulcanisation. The drying of the bromide to a constant weight is by no means easy, but the chief error is due to the fact that the loading materials are first weighed as such, together with the bromide, and the weight is then corrected from the ash.I n this way, any loss suffered by the mineral loading during incineration is calculated as caoutchouc, and the error is especially large in the case of red-lead. Hiibener's method (ANALYST, 1909, 365) marks a decided advance in the estimation of caoutchouc. Estimation of Loadings.-The error, mentioned above, involved in correcting the weight of the precipitate by the weight of the ash, induced the authors to endeavour to find a method by which the loading materials may be estimated directly without incineration.This has been effected successfully by dissolving the caoutchouc in a suitable solvent and collecting and weighing the insoluble residue. The petroleum oil (boiling-point up to 300" C.) prescribed by Axelrod (Zoc. cit.) serves this purpose : One grm. of the finely-chopped sample is boiled under a reflux con- denser with 25 C.C. oi petroleum oil distilling between 230" and 260' C. until the caoutchouc is dissolved. After cooling, the solution is diluted with benzene, the residue is allowed to settle, and the liquid is decanted off through a tared Gooch filter. The precipitate is finally washed with hot benzene, petroleum spirit, alcoholTHE ANALYST, 407 a d ether, without removing it from the flask.Flask and filter are dried at 105" C. and weighed. In order to prevent the free sulphur from reacting with the un- vulcanised caoutchouc or the solvent, it is desirable to remove this sulphur by extraction with acetone before treatment with petroleum. A centrifuge is of considerable assistance in separating and purifying the residue, Certain types of rubber goods require a solvent of higher boiling-point than the above, such as paraffin oil. In many cases solvents of lower boiling-points may be used, which is desirable in order to avoid charring. Essential oil of camphor, boiling-point 175" C., is particularly useful for this purpose. The time occupied in dissolving the caoutchouc is from a half to two hours; the results are generally within 2 per cent.of the actual loading added. Estimation of Total Szdphw.-The powerful oxidising action of the NO, ions separated at the anode in the electrolysis of nitric acid may be employed advantageously for the destruction of the organic matter. The procedure is as follows : In a tall beaker of 100 to 150 C.C. capacity, a platinum-wire basket is suspended from a, platinum wire to serve as the anode and to contain the sample to be oxidised. The basket is then covered with the strongest fuming nitric acid, and electrolysed. The oxidation is complete in less than an hour, and the analysis is concluded by ordinary methods after removing the bulk of the nitric acid. A platinum spatula serves as the cathode. J. F. B. New Method of Estimating Combined Sulphur in Vuleanised Caout- chouc.T. Budde. (Gunzmi Zeit., 1909, 23, 1143-1144.)-The author shows, by means of Hiibener's method of estimating caoutchouc in hard vulcanised rubber (ANALYST, 1909, 170), that estimations of sulphur by the acetone method give too high results. Hiibener found that h a d rubber was soluble in bromine, and that on adding water and heating the solution the excess of sulphur was oxidised, whereas that in combination with the caoutchouc was not attaclred. The following modifica- tion, devised by the author, renders this method also applicable to soft rubbers: About 1 grm. of the sample is left in contact with 30 to 40 C.C. of carbon tetrachloride for twenty-four hours, and then mixed with about 250 C.C. of a solution of 6 C.C. of bromine and 1 grm. of iodine in 1,000 C.C.of carbon tetrachloride. After standing for six houm, ths liquid is mixed with half its volume of alcohol, allowed to stand for twelve hours, and then filtered. The filtrate will contain part of the excess of sulphur, resins, etc. The residue of caoutchouc tetrabromide is now treated with 10 to 15 C.C. of bromine, in which it dissolves after twenty-four hours. The solution is mixed with water and heated gently on the water-bath until the caoutchouc tetra- bromide rises to the surface, whilst the aqueous layer will contain the remainder of the excess of sulphur. This liquid is heated and filtered, and the residue heated three times with water to the boiling-point, and finally washed with alcohol and ether, and the sulphur it contains estimated by Henrique's method.Various comparative results by the two methods are given. Thus, a sample of Mexican rubber, which by the acetone method was found to contain 3.36 per cent. of combined sulphur, yielded only 1.52 per cent. by the new method. C. A. M.408 THE ANALYST. Analysis of Coal Gas and Similar Gaseous Mixtures, with Special Reference to the Estimation of Nitrogen in the Same. G. v. Knorre. (Chem. Zeit., 1909, 33, 717-719.)-The method described is a modification of that proposed by Jager (J. Gasbelewht, 1898, 764). The carbon dioxide, carbon monoxide, heavy hydrocarbons, and oxygen, are absorbed in the usual way, and the residue of gas is then passed over copper oxide heated to a temperature of 250" C. in a quartz tube from which the air has been displaced previously by pure nitrogen.The hydrogen is oxidised to water by the copper oxide, and the gas is passed to and fro over the oxide until no further diminution in volume takes place. The diminution in volume is equal to the volume of the hydrogen present. The residue of gas is again passed through the quartz tube after the copper oxide has been heated to bright redness, and the decrease in volume when the gas is brought in contact with potassium hydroxide is equivalent to that of the methane, this having been burned to carbon dioxide and the latter absorbed by the alkali. The final residue of gas represents the nitrogen. The copper oxide must be reoxidised from time to time and renewed when necessary. The total nitrogen may also be estimated by passing the original gas through the copper oxide tube heated to bright redness and filled with nitrogen.The residue of gas left after treatment with potassium hydroxide consists of the nitrogen in the original gas. w. P. s. Estimation of Fatty Acids in Soap. M. Dominiklewicz. (Chem. Zcit., 1909, 33, 728.)-The method proposed is similar in principle to the butyrometric estimation of fat in milk. From 40 to 60 grms. of the soap are dissolved in water and diluted to a volume of 1 litre. Ten C.C. of hydrochloric acid and 10 C.C. of the soap solution are then introduced into a tube having the shape of an ordinary butyrometer tube; the lower and wider part of the tube is closed by means of a rubber stopper, whilst the narrow graduated neck is provided with a metal screw-tap. The soap solution and acid are mixed by gently shaking the tube, and the latter is then placed in boiling water for five minutes, after which it is subjected to centrifugal action for five minutes.Boiling water is then added to the tube so as to bring the fatty acids into the graduated neck, and after being again whirled in the centrifugal machine, the volume of the fatty acids is read off at a temperature of 99" C. Each division of the scale corresponds with 0-1 c.c., and the weight of the acids is then calculated, the sp. gr. of the fatty acids at a temperature of 99O C. being taken into account. The results obtained by the iiiethod are stated to agree well with those yielded by the ordinary gravimetric process. w. P. s. The Alkalimetric Estimation of Hydroxylamine. A.Stahler. (Bey., 1909, 42, 2695-2696.)-Hydroxylamine is reduced by titanium sesquisulphate to ammonia., and may be estimated by titrating back the excess of titanous salt. But the employ- ment of standardised titanium solutions is inconvenient, owing to their extreme liability to oxidation. A better method, therefore, consists in distilling over the ammonia in the usual manner, since elaborate precautions to preserve the titauous reagent are unnecessary, provided an excess be employed. The same principle mayTHE ANALYST. 409 be applied to any substance which yields on reduction a volatile amine which can be titrated with acid--e.g., certain oximes, nitro- and nitroso-compounds. J. F. B. New Colour Reactions of Hydrastine, Hydrastinine and Narcotine.A. Labat. (BzdZ. SOC. Clzinz., 1909, [iv.!, 5, 742-743.)-Hydrastine, hydrastinine, and narcotine give coloured reactions with phenols in the presence of sulphuric acid. Theee colorations may be readily obtained by mixing 2 C.C. of sulphuric acid (sp. gr. 1.84) with 0.1 C.C. of an alcoholic solution of the alkaloid (1 : 300 in the case of hydrastine, and 1 : 100 in the case cf the others), adding 0.1 C.C. of the phenolic solution, and heating the tube in boiling water. With a 5 per cent. alcoholic solution of tannin or gallic acid there is produced an intense emerald-green coloration, which gradually changes to bright blue. On diluting the liquid with sulphuric acid and examining it spectroscopically, a characteristic absorption-band will be observed in the red part of the spectrum. This band can be clearly seen in very dilute solutions --e.g., 1 : 50,000 of hydrastinine, 1 : 40,000 of hydrastine, and 1 : 20,000 of narcotine.If a 5 per cent. alcoholic or aqueous solution of guaiacol or pyrocatechol be used in the test, there is produced a gooseberry-red coloration, which becomes darker, and then changes to a violet tint; whilst the use of an alcoholic solution of morphine hydrochloride (1 : 50) gradually gives a violet coloration. These reactions, however, are not so sensitive, especially as regards narcotine, as the coloration with gallic acid. C. A. IN. New Reactions of Opianic Acid, and their Use in the Detection of Hydrastine and Narcotine. A. Labat. (BUZZ. SOC. Chim., 1909, [iv.], 5, 743- 745.)-On oxidising hydrastine or narcotine by means of perrnanganate there is formed a considerable quantity of opianic acid, which forms characteristic coloured compounds with certsin phenols in the presence of sulphuric acid.A solution of 0.1 grm. of the alkaloid in 0.5 C.C. of 10 per cent, sulphuric acid is gently heated vith 2 C.C. of a 2 per cent. solution of potassium permanganate until the pink colour has disappeared. The liquid (which will be fluorescent in the case of hydrastine) is then mixed with 2-5 C.C. of alcohol, and the solution used for the tests for opianic acid. On mixing 0.1 C.C. of a 1 per cent. alcoholic solution of that acid with 2 C.C. of sulphuric acid (sp. gr. 1.84) and 0.1 C.C. of the phenolic solution, the following reactions are obtained: With a 5 per cent. alcoholic solution of gallic acid a blue coloration rapidly appears, to give place to a leaf-brown.With a 5 per cent. alcoholic solution of guaiacol the coloration is gooseberry-red, changing to an in tense blue when heated on the water-bath. Similar colorations are obtained with pyro- catechol. With a-naphthol there is a gooseberry-red coloration, and with /3-naphthol a wine-red coloration. I n both cases the liquid should be cooled, since the intensity of the colour is diminished on heating. A 5 per cent. alcoholic solution of codeine gives a violet coloration, gradually changing to blue on the water-bath, whilst an alcoholic solution of p-methylnaphthol (1 : 40) gives a bright violet colour, which fades on heating. C. A. M,410 THE ANALYST. Estimation of Indigo in Dyed Cotton.E. Knecht. ( J . SOC. Dyers and Colour., 1909, 25, 135; through Chem. Zeit. Rep., 1909, 33, 347.)-The method proposed is based on the solubility of both cotton and indigotin in 80 per oent. sulphuric acid at a temperature of from 35" to 40" C., and to the fact that the indigotin sulphate formed, when treated with water, dissociates with the separation of the whole of the indigotin. Four grams of the fabric in the form of small pieces are heated with 25 C.C. of 80 per cent. sulphuric acid for about ten minutes at the above-mentioned temperature. The solution is then diluted with water to a volume of 120 c.c., and filtered through a Gooch crucible containing sand and asbestos as a filtering medium. After being dried at 110" to 120" C., the contents of the crucible are heated on the water-bath with a little sulphuric acid for one hour, and the indigo sulphonate is then dissolved in water and titrated with either titanous chloride or permanganate solution.Test analyses showed that the method is accurate, and that the basic colours and sulphide-blues employed with indigo do not interfere with the results obtained; the first mentioned remain in the sulphuric acid solution, and the second are destroyed. Manganese bases must, however, be removed previously by treatment with bisulphite. w. P, s. Colorimetric Method for the Estimation of the Molecular Weights of Carbohydrates - 11. Differentiation of Primary from Secondary and Tertiary Alcohols. L, Wacker. (Ber., 1909, 42, 2675-2680.)-1n a previous paper (ANALYST, 1908, 33, 131) the author described a method for the determination of the molecular complexity of carbohydrates, depending on the fact that solutions of carbohydrates containing an equal number of molecules give colorations of equal intensity with an alkaline solution of phenylhydrazine sulphonic acid.This circum- stance is probably due to the presence of the same number of chromophoric groups in the molecule, and it suggests that only the terminal aldehydic or primary alcoholic group is concerned. Secondary or tertiary alcohols react far less readily, and the reaction may serve for distinguishing the primary alcohols from these others. For the est.imation of the molecular complexity of carbohydrates a colorimetric standard is necessary, which may be either maltose or dextrose; the latter sugar gives some- what bluish shades of red, which causes the results to appear 5 or 10 per cent.stronger than with equivalent solutions of di- or polysaccharides. I n the present communication the author deals more particularly with starch. If starch be boiled with 1,000 times its weight of water, a solution of "readily soluble granulose" is obtained, which remains clear when cold; the product was isolated and tested by this method, and showed a complex of 5 hexose residues. Further €ractions were obtained jn a similar manner by repeated extractions of the residue, and there were thus found : '' Sparingly soluble granulose " composed of 6 hexose groups ; " once- boiled starch cellulose'' with 7 hexose groups; and '' repeatedly boiled starch cellulose " with 8 hexose groups in the molecule. '' Achroodextrin " and " erythrodextrin " were found both to contain 4 hexose groups; glycogen from liver showed 10 hexose groups.For the classification of alcohols centinormal solutions are prepared, and 10 or 20 C.C. are placed in a series of cylinders and diluted to 100 C.C. each. A sufficient quantity of phenylhydrazine sulphonic acid is weighed out to allow 0.4 grm. for each test, andTHE ANALYST+ 411 dissolved in sufficient dilute sodium hydroxide to contain this quantity in 1 C.C. of solution, which should be used quite fresh. Immediately after adding the phenyl- hydrazine compound, 25 C.C. of a 33 per cent. solution of sodium hydroxide are added $0 each cylinder, and these are shaken every fifteen minutes during the first two to three hours to develop the coloration; the colours are best compared after seven to eight hours.If primary, secondary, and tertiary alcohols be compared with each other and with a blank (water), it will be observed that only the primary alcohols show intense colorations, and the other alcohols differ little from the blank. Amines and amino acids also give red colorations with phenylhydrazine sulphonates under the 8ame conditions. J. F. B. Estimation of Molecular Weights by the Boiling-Point Method. R. Meyer and K. Desamari. (Bey., 1909, 42, 2809-2814.)-The authors have investigated a discrepancy which has arisen in the molecular weight of tribromresoquinone. When the boiling-point of the pure solvent (benzene) was redetermined after the estimation of the molecular weight, it was found to have changed: the benzene boiling at 79.400" C. before the experiment, boiled at 79425O C . after the experiment. This difference of 0.025" C. with such dilute solutions (1.5 to 2 per cent.) makes a difference of nearly 300 units in the molecular weight of the quinone, hence the discrepancy. The difference can only be explained by an alteration of the barometric pressure during the period elapsing between the two determinations of the boiling-point, and from existing data, it is calculated that 1 mm. difference in pressure corresponds to 0.0427" C. difference in boiling-point for benzene. Thus, in molecular weight determinations it is of the highest importance to eliminate this error by correcting for any alteration of barometric pressure, For this purpose an ordinary vertical barometer is not sufficiently delicate, and the authors have employed an extremely sensitive aneroid barometer, such as is used for the determination of altitudes, which can be read to tenths and even hundredths of a millimetre. The thermometer employed could be read to of a degree. J. F. B.