either by ammonium acetate or by neutralisation with ammonia, soda, or potash. In the former case, a large excess of acetic acid should be used to ensure the complete removal of the lime. The precipitate is then collected on a filter and washed repeatedly with cold 1 per cent. solution of ammonium nitrate, containing di-ammonium hydrogen phosphate equivalent to 0.05 gram phosphoric anhydride per litre. It is then thoroughly washed with the same solution (hot), and finally once with water, then dried, ignited, and weighed. It may then be dissolved in hydrochloric acid, reduced with stannous chloride, and the iron determined by standard potassium dichromate. The result is calculated to iron phosphate, which is subtracted from the weight of the precipitate, and the remaining aluminium phosphate calculated to alumina. Test analyses gave good results.

If a high result is apprehended from the precipitation of calcium phosphate, the precipitate may be dissolved in hydrochloric acid, excess of caustic soda added, boiled, and filtered. The iron in the residue is then determined volumetrically; and the alumina in the filtrate, after acidifying with hydrochloric acid, by either of the above methods. It is advisable to add some phosphoric acid or ammonium phosphate to ensure the presence of an excess of the former sufficient to produce normal aluminium phosphate. When the mineral phosphates contain fluorides, it is best to heat the weighed portion of the sample with concentrated sulphuric acid until these compounds are decomposed and all the hydrofluoric acid is expelled. The residue is then treated with hydrochloric acid and the determination proceeded with. D. B.

Estimation of Nickel in Ores, Mattes, Slags, &c. By T. MOORE (Chem. News, 54, 300).-After the separation of copper, arsenic, &c., in the usual manner, the strongly acid solution is diluted, and an excess of sodium phosphate added; any precipitate formed is redissolved by the aid of acid. The whole is boiled, and while boiling, sodium acetate is added until all the iron phosphate is precipitated, the boiling is continued for a minute, and the precipitate collected and washed with hot water containing a little sodium acetate and acetic acid. The filtrate is warmed and treated with potash and bromine. The precipitate is dissolved in warm, dilute sulphuric acid, made strongly alkaline with ammonia, and electrolysed. Metallic nickel is deposited, the small quantity of iron present which has escaped precipitation floats about as an insoluble oxide, and does not influence the results, whilst chromium and manganese remain in solution; part of the latter is deposited on the anode as peroxide (compare Abstr., 1886, 921). The aluminium remains with the iron. When greater accuracy is required, the precipitation of the iron may be repeated. D. A. L.

Volumetric Determination of Chromium. By W. J. SELL (Chem. News, 54, 299-300).-The author has devised a method of applying to analytical purposes the reaction between hydrogen peroxide and chromium oxides (see Martinon, Abstr., 1886, 984). The aqueous solution of the substance to be analysed is treated with sufficient potash to dissolve the precipitate first formed, hydrogen

peroxide is then added, and the whole boiled briskly for at least 15minutes, the chromic acid is then easily determined by any method in the resulting solution of the chromate formed. Aluminium and zinc do not interfere with the reaction, but with iron the results are low. The method was devised before the publication of the paper referred to. Another use for hydrogen peroxide is suggested for qualitative purposes. The ammonium precipitate in group III is dissolved in dilute nitric acid, poured into excess of potash containing hydrogen peroxide, boiled and filtered. The solution contains the aluminium and the chromium as chromate; the iron, &c., being precipitated.

D. A. L.

Analysis of Chrome Paints. By W. L. BROWN (Chem. News, 54, 329-331).-Mixtures of lead chromate and sulphate, or chromate, sulphate and carbonate, are treated with hydrochloric acid for insoluble adulterants, whilst the solution is used for the chromium and sulphuric acid determinations. For lead, the paint is decomposed with concentrated sulphuric acid, and the lead, as in all succeeding cases, weighed as sulphate. Red chromate of lead is decomposed with nitric acid and a few drops of alcohol added to facilitate the solution of the chromium. Chrome-green is treated with hydrochloric acid; lead, chromium, sulphuric acid, and sometimes iron, pass into solution, the Prussian blue, &c., remains as residue (it is sometimes slightly attacked). Any lead found in excess of that required by the chromium and sulphuric acid, is regarded as carbonate. Full details of separation, &c., are given in the paper. D. A. L.

Estimate of Chromate in the Presence of Dichromate. By N. MCCULLOCH (Chem. News, 55, 2-3). As is well known, the blue coloration soluble in ether which results from the action of hydrogen peroxide on chromic acid is only produced with dichromates in acidified solutions, or with chromates to which acid has been added in excess of that required to convert them into dichromates. The following quick and ready method for estimating chromate in the presence of dichromate is founded on this basis. The substance, dissolved in a little water, is mixed with a few c.c. of hydrogen peroxide solution and covered with a layer of ether, standard sulphuric acid is run in gradually until after agitation the ether assumes a blue colour. From the quantity of acid used, the amount of chromate present is easily calculated. D. A. L.

Estimation of Tin and Lead in Alloys. By O. WACHSMUTH (Chem. Centr., 1886, 491).-For the rapid estimation of the composition of alloys of these two metals, the author recommends the determination of the specific gravities and melting points.

The specific gravities are not sensibly affected by the small admixture of antimony (2 to 5 per cent.), used in preparing many alloys. The specific gravities and melting points of mixtures of tin and lead, containing from 5 to 95 per cent. of each metal, are given.

C. F. C. Colour Reactions of Titanic, Niobic, Tantalic, and Stannic Anhydrides. By L. LEVY (Compt. rend., 103, 1074-1076).

various colour reactions given by these anhydrides with certain alkaloids, phenols, and with hydroxybenzoic acids in presence of concentrated sulphuric acid are described. The colours disappear when water is added, except in the case of stannic anhydride, and the substances must be free from nitrates and nitrites. Silica, alumina, and zinc and uranium oxides give no similar colour reactions.

In order to detect the four anhydrides when mixed together, the substance is strongly heated with ammonium carbonate to remove nitrites and nitrates. Different portions are then mixed with strong sulphuric acid and one of the following reagents: morphine, a crimson colour indicates titanic anhydride; codeïne, a mauve colour indicates niobic anhydride; resorcinol, an amethyst colour, tantalic anhydride (green if much titanic and niobic anhydrides are present); a-naphthol and a few drops of water, amethyst colour, stannic anhydride.

These reactions may be used in the reverse way to detect certain alkaloids and phenols.

C. H. B.

Colour Reactions of Arsenic, Arsenious, Vanadic, and Molybdic Anhydrides, and of Antimony and Bismuth Oxides. By L. LEVY (Compt. rend., 103, 1195–1196).-A continuation of the author's previous experiments (preceding Abstract). In most cases the colours are destroyed or modified by the addition of water.

Arsenic anhydride mixed with catechol gives a grey-green colour, which changes to amethyst. In this way it can be distinguished from phosphoric anhydride, which gives no coloration, and from vanadic anhydride, which gives a dark green colour, becoming paler on addition of water. With arsenic anhydride, resorcinol and quinol give respectively a sepia and a yellowish coloration, and it is evident therefore that this reaction may be used to distinguish between these three isomeric dihydric phenols.

Vanadic anhydride with resorcinol gives a dark green coloration, which becomes violet on addition of water, and it can be distinguished by this reaction from molybdic and phosphoric anhydrides.

Bismuth oxide and catechol give a greenish coloration, which becomes deeper after the addition of water. Antimony oxide and arsenic anhydride give a flesh coloration under the same conditions. Arsenites give a flesh tint with catechol, which distinguishes them from the arsenates, whilst a-naphthol gives no coloration with any of the oxides similar to that which it gives with stannic oxide.

C. H. B. Post-mortem Detection of Chloroform. By C. LUEDEKING (Amer. Chem. J., 8, 358—361).—The lungs or other viscera are made slightly alkaline by sodium carbonate, and heated on a water-bath in a flask, through which a current of air is passing. The escaping gases are passed through a short length of red-hot tube, and then tested with potassium iodide and starch paper. As the result of six experiments on dogs it is shown that chloroform may be detected at least four weeks after death, even when the carcases have been exposed to the air at full summer's heat; during the decomposition no substances are formed which interfere with the above tests. H. B.

Estimation of Glycerol in Wine and Beer. By SKALWEIT (Chem. Centr., 1886, 541).-The author isolates the glycerol in the usual way, obtaining it in a concentrated form, in which, the aggregate weight having been determined, the percentage of glycerol is estimated by determining the coefficient of refraction in the Abbé instrument. Full details are given. C. F. C.

Separation and Quantitative Estimation of Melitose (Raffinose) in Cane-sugar. By C. SCHEIBLER (Ber., 19, 28682874. Compare Abstr., 1885, 962.)-Further evidence has confirmed the formula C18H32O16 + 5H2O for melitose. The best method of drying melitose is to place it in a Liebig's drying tube, heated at 70-80° in a water-bath, and so arranged that a small quantity of air, dried over sulphuric acid, enters it by a capillary tube, whilst a partial vacuum is maintained by a Bunsen pump. After one hour's drying under these conditions, so much water is removed that on raising the temperature of the bath to 100° to complete the dehydration, fusion does not occur. Melitose dried in this way melts at 118-119°; it is very hygroscopic, and in a moist atmosphere gradually absorbs the whole of the water removed by drying.

The author finds that melitose is far more soluble than cane-sugar in commercial absolute methyl alcohol, 100 c.c. of the alcohol dissolving 9.5 and 04 grams respectively at the ordinary temperature; on this property is based a method for its extraction from the molasses of beetroot sugar. The crystalline product, dried as described, is extracted with absolute methyl alcohol, and the syrupy residue obtained on distilling off the alcohol, is taken up with a little water, and purified by precipitation with ethyl alcohol.

A method for the quantitative estimation of melitose in cane-sugar has been devised, and is under investigation. According to the details given in the paper, commercial absolute methyl alcohol is saturated with cane-sugar at a known temperature, and the rotatory power of the solution is determined. A known weight of the impure canesugar, dried by the above method, is then extracted by 100 c.c. of the solution, the temperature being kept constant, and the rotatory power is again determined. From the increase in the rotation, the amount of melitose present can be calculated. W. P. W.

Quantitative Estimation of Raffinose or Melitose. By R. CREYDT (Ber., 19, 3115—3119).-Raffinose can be determined either optically or gravimetrically after conversion into mucic acid. In the optical examination of cane-sugar and of raffinose, the amount is calculated by the subjoined formula

in which A the amount of direct polarisation, B = the polarisation at 20° after inversion, C = the difference between B and A, Z = the percentage of cane-sugar, and R = the percentage of raffinose. This method is, however, not available in the presence of other optically active substances such as dextran or invert sugar. When these are

present, raffinose is best estimated by converting it into mucic acid by the following process :

:

About 5 grams of dry substance is treated with 60 c.c. of nitric acid, sp. gr. 1.15, and the mixture evaporated to about one-third of its volume on the water-bath. After cooling, a known quantity of water is added, and the separation of the mucic acid assisted by the addition of a known weight of that acid, the whole being well stirred. The precipitate is then collected on a tared filter, dried, and weighed. The weight of the raffinose must be estimated from that of the mucic acid by aid of an empirical table which has been worked out by the author, but is not given in the paper. A. J. G.

Detection and Determination of Lactic Acid. By R. PALM (Zeit. anal. Chem., 26, 33-35).-Lactic acid, when mixed with lead acetate and alcoholic ammonia, gives a heavy, granular precipitate of the constant formula 3PbO,2C,H6O3. To examine an animal or vegetable organ for free lactic acid, it is extracted with ether (if for a lactate, acidification with sulphuric acid precedes the extraction with ether), the ethereal solution is evaporated to a syrup, and this is treated with water. The filtered aqueous solution is mixed with lead acetate, and any precipitate produced is filtered off. On adding more lead acetate to the filtrate and then alcoholic ammonia, the lactate is thrown down free from foreign substances. It may be washed with alcohol, in which it is absolutely insoluble, and the amount of lactic acid in it may be ascertained from the loss on ignition. Minute traces of lactic acid may be thrown down with greater certainty by shaking the solution (prepared as above) with an excess of freshly precipitated lead hydroxide. The precipitate produced in either way yields pure lactic acid when decomposed by hydrogen sulphide and extracted with ether. M. J. S.

Use of Turmeric as an Indicator for Citric Acid. By F. WATTS (J. Soc. Chem. Ind., 5, 214—215).—When litmus is employed as an indicator in the determination of citric acid by alkalimetry, the results obtained are frequently too low, on account of the want of sharpness in the reaction. Warington (this Journal, 1875, 925) recommends the preparation of a special litmus-paper, but even with this precaution the difficulty is not entirely removed. The author has obtained good results by using tincture of turmeric in place of litmus. Drops of this tincture are placed on a white tile, the liquid spreading out in bright yellow films. The slightest excess of alkali causes the development of the well-known red-brown colour. In using this indicator, much time is saved by employing litmus-paper to determine the near approach to neutralisation, proceeding with the addition of alkali until the litmus-paper is turned slightly blue, and determining the exact point of neutrality by means of the turmeric. D. B.

Determination of the Fatty Acids in Soap. By B. SCHULZE (Zeit. anal. Chem., 26, 27-28).-The soap is decomposed in a conical flask by dilute sulphuric acid. Ether is added to dissolve the fatty acids. The aqueous layer is then removed by a pipette. Water is