and liquid excreta of animals and placed upon a floor, undergoes the following changes::

1. Solution of the vasculose and of a large proportion of the nitrogenous matter by the alkaline carbonates, the dissolved material constituting the dark brown colouring substance of the liquid drainings of the dung-heap.

2. Destruction of a great part of the cellulose of the litter by fermentation, with disengagement of methane and carbonic anhydride: the insoluble residue which persists after the loss of part of the nitrogenous matter, vasculose, and cellulose, consists of vasculose unattacked at the ordinary temperature; it is this residue, rich in vasculose partially altered by loss of water, which is often called "black butter."

3. Transformation of ammonia (arising from the fermentation of urea) into nitrogenous organic compounds, in great part soluble in alkaline liquids; this change is produced solely by living ferments.

4. During this fermentation, part of the nitrogen escapes in the free state, for when the manure is sufficiently moist very little loss of ammonia occurs. J. M. H. M.

Waste Products as Manure (Bone-meal). By A. PETERMANN (Bied. Centr., 1888, 230-239).-These experiments, in which bone- and blood-meal mixed with potassium chloride, Chili saltpetre, or bone-ash were used, have been conducted for three years on clay and sandy soils in the neighbourhood of Gembloux. It was found that the simple addition of phosphates (bone-ash) to this sandy clay was of little or no value, but that when an addition of nitrogen was also made, as in bone-meal, then there appeared an increase in the yield of the crop; the results obtained from this mixture were far surpassed, however, by those obtained by the use of blood-meal, or phosphates with saltpetre and potash. The results obtained on the sandy soil were in all cases inferior to those on the heavy land: the best results were produced by the employment of blood-meal, bone-ash, and potash. Reviewing all his experiments made since 1880, the author comes to the conclusion that the various nitrogenous manures stand in the following order as regards manurial value: Saltpetre, dried blood, dissolved wool, bone-meal, raw wool, and leather-meal. E. W. P.

Addition of Wood-ashes to Superphosphates. By V. T. MAGERSTEIN (Bied. Centr., 1888, 225-226).—An admixture of woodashes with superphosphates has long been known as advantageous, but the addition must not exceed 25 per cent., otherwise the lime which is thus added will cause a loss by reason of the formation of reduced phosphate. E. W. P.

Manuring Experiments with Various Phosphates. By E. GATELLIER (Bied. Centr., 1888, 227-228).-As superphosphates are supposed to be reduced by the bases in the soil, and thus to become equal to insoluble phosphates, experiments were instituted to test the value of varieties of phosphates on clay soils in France; in all cases bone superphosphate (16 per cent. sol.) surpassed basic slag, ordinary superphosphate, &c. E. W. P.

VOL. LIV.

3 d

Analytical Chemistry.

Thompson's Calorimeter. By SCHEURER-KESTNER (Compt. rend., 106, 941-944).-The proportion of oxidising mixture necessary for complete combustion varies with the nature of the coal and the proportion of ash which it contains. The author employed from 10.5 to 12.5 times the weight of the coal; with only 10 parts of oxidising mixture, the combustion is usually incomplete. The correction of 10 per cent. usually added to the results is much too low. Experiments with ignited wood charcoal showed that for the author's apparatus 15 per cent. was the true correction. This was used for samples of coal, the heat of combustion of which had previously been determined by Favre and Silbermann's method; the maximum difference was 1.25 per cent. The heat of combustion of 20 coals was determined by Thompson's calorimeter, and afterwards by Favre and Silbermann's apparatus. The variations were sometimes on the one side, sometimes on the other, the maximum difference being 3.5 per cent. In 11 cases, the difference was less than 2 per cent.

Although Thompson's calorimeter gives results of practical value, when an error of 2 or 3 per cent. is negligible, it is only worthy of limited confidence. Berthelot's calorimetric bomb would be much better. The main sources of error, as Stohmann has pointed out, are the heat of decomposition of the oxidising mixture and the heat of dissolution of the salts after combustion, since the proportion of each salt varies with each coal. Some finely powdered samples of coal which had been exposed to diffused light for 20 years in bottles not very tightly closed, were found to be completely unaltered. C. H. B. 1

Improved Form of Gas Apparatus. By J. T. WILLARD (Amer. Chem. J., 10, 53-56).-The apparatus is essentially a combination of Elliott's (Abstr., 1887, 1137) and Frankland's. H. B.

Preservation of Solutions of Hydrogen Sulphide. By D. LINDO (Chem. News, 57, 173-175).-The hydrogen sulphide is readily estimated in fresh solutions of this gas containing glycerol, camphor, boric acid, or thymol, by precipitation with silver nitrate and ammonia, filtering on a Gooch filter, drying, and weighing. In old solutions containing glycerol, silver nitrate does not answer, therefore the author employs sodium arsenite and hydrochloric acid for the precipitation. This precipitate is, however, troublesome to wash, and is best kept as small as possible, or if precipitated along with prepared asbestos, by adding some of the latter to the solution before precipitation, it may be washed with comparative ease.

From numerous experiments made in Jamaica in closed and open bottles at an average temperature of 26°, the author concludes that there is no advantage derived from the use of glycerol for the preservation of solutions of hydrogen sulphide.

Moreover, the other substances mentioned above do not prevent

loss in strength by escape of gas, but camphor and thymol appear to retard oxidation. D. A. L.

Rapid Determination of Hydrogen Peroxide. By CONTAMINE (Dingl. polyt. J., 267, 238).-A few c.c. of the solution under examination is introduced into a glass tube divided into tenths of a c.c., and sealed at one end. Having previously neutralised the solution with ammonia, the volume is read off. A few crystals of potassium permanganate enclosed in tissue paper are then added, after which the aperture is closed with the finger and the tube shaken vigorously. It is then opened under water and the volume read off. The difference between the two readings gives the quantity of oxygen contained in the hydrogen peroxide solution. The method is sufficiently accurate to be used in works for determining the quantity of hydrogen peroxide in bleaching solutions. D. B.

Supposed Occurrence of Hydrogen Peroxide in Animal and Vegetable Juices. By T. BOKORNY (Ber., 21, 1100-1102).— Wurster's tetramethylparaphenylamine reaction (Abstr., 1887, 295) is useless as a test for hydrogen peroxide in animal and vegetable juices, since the paper is turned blue by very many compounds, such as nitrites, moist silver oxide, quinone, aldehyde, acetone, &c. For this reason, the reaction is no proof of the presence of hydrogen peroxide. Experiments have shown that living protoplasm does not give the ordinary reactions for hydrogen peroxide. F. S. K.

Gravimetric Estimation of Chlorine. By F. STOLBA (Chem. Centr., 1887, 1240, from Listy. Chem., 11, 224-225).-In the case of small quantities of chlorine (up to 80 miligrams) in water, &c., the author precipitates the chlorine as silver chloride, collects the precipitate on a large Schleicher-Schuell filter-paper, dries, and so folds the paper that the precipitate is enveloped at the point in many folds of paper, the paper is very slowly carbonised in a platinum capsule, the carbon burnt off at a low temperature, and the remaining spongy silver weighed and tested as to its complete solubility in nitric acid. J. P. L. Volumetric Determination of Sulphuric Acid. By A. GAWALOWSKI (Zeit. anal. Chem., 27, 152-159). The principle of the method is identical with Wilsing's (Abstr., 1887, 181). The titration with sodium carbonate is, however, made in the cold after the addition of a few drops of alcohol. Boiling is said to favour the end reaction, but not to be absolutely necessary. The mixture of the sulphate with the barium chloride should be allowed to remain for orhour before titrating the excess, and a similar interval should be allowed before finally judging of the completion of the titration with sodium carbonate. Seminormal solutions are recommended. The test analyses show errors of 1 per cent. in both directions.

M. J. S. Volumetric Estimation of Sulphuric and Phosphoric Acids. By J. T. WHITE (Chem. News, 57, 165-166; 187).-Sulphuric acid in combination with fixed alkalis may be estimated by titration with

silver nitrate and potassium chromate in the following manner :Precipitate the solution with excess of barium chloride, and remove the barium remaining in solution with ammonium carbonate, then evaporate, ignite, and titrate the resulting alkaline chlorides in the ordinary fashion. If phosphoric acid is also present, precipitate it beforehand with magnesium chloride and ammonia. One portion of this filtrate is then evaporated with ammonium chloride, and the ignited residue titrated; whilst another portion is treated as above for sulphuric acid; the difference between the two results is the chlorine due to the decomposed sulphates. To estimate phosphoric acid in alkaline phosphates, decompose chlorides if present by evaporation with sulphuric acid, neutralise with soda, precipitate with excess of silver nitrate solution, neutralise with calcium carbonate, and titrate the excess of silver nitrate in the usual way. The presence of sulphuric acid does not interfere, as the solution is kept sufficiently dilute to dissolve silver sulphate. D. A. L.

Estimation of Total Nitrogen. By HOUZEAU (Chem. Centr., 1888, 82-83, from Pharm. Centr., 28, 627-628).—The process depends on the fact that all nitrogenous substances yield the whole of their nitrogen in the form of ammonia when heated with sodium acetate, sodium thiosulphate, and soda-lime.

The mixture of sodium acetate and thiosulphate is prepared by melting equal parts by weight of both salts in their water of crystallisation on a water-bath and powdering the cooled residue.

The combustion is performed in precisely the same manner as in the Will-Varren trap process, about 2 grams of the mixture with grams of soda-lime at the end of the tube supply the necessary indifferent gases to drive off the last of the ammonia.

2

J. P. L.

Sources of Loss in the Determination of Nitrogen by Sodalime. By W. O. ATWATER and E. M. BALL (Amer. Chem. J., 10, 113 — 119). The following directions are given as being necessary. The anterior portion of the tube must be tightly packed with coarse sodalime, not with fine, and there must be no appreciable open channel along the upper portion of the tube, otherwise, losses amounting to 64 per cent. of the nitrogen may occur. The anterior layer is the first part of the tube heated. The time of combustion should not be too protracted, 2 hours for instance, as this may introduce an error of 2.8 per cent., and if the tube is not packed as above, an error of even 77 per cent. may occur. Moreover, the results are far more concordant among themselves, and agree perfectly with those given by Kjeldahl's method when the above conditions are carried out. Substances like caseïn yield all their nitrogen as ammonia under these conditions, but strychnine yields only a small fraction, and the whole cannot be obtained even by Kjeldahl's method. H. B.

Absorption of Ammonia by Acid Solutions in Nitrogen Determinations. By I. S. HAYNES (Amer. Chem. J., 10, 111–113). -The absorption of the ammonia is quite complete if double the amount of acid theoretically necessary is employed, even when large

quantities of inert gases are present, and when the gases are passed rapidly through the acid, say at the rate of 90 bubbles per minute, the combustion occupying only 12 minutes. Great attention must, however, be paid to the asbestos plug in the anterior portion of the tube being packed sufficiently tightly, or particles of soda-lime will be carried forward into the acid. H. B.

Examination of Wine for Nitric Acid. By E. BORGMANN (Zeit. anal. Chem., 27, 184-187).-The method proposed by Egger (Abstr., 1885, 842), for detecting the addition of water to wine by testing for nitrates with diphenylamine and sulphuric acid, is not in all respects trustworthy. Wines which on analysis give results incompatible with an addition of water are sometimes found to show the nitric acid reaction distinctly. Some of the waters of Rhenish Hesse, however, contain so much nitric acid (in one case 1 gram per litre), and the reaction will detect so small a trace (1 part in two millions of wine when Egger's concentration method is employed), that the mere rinsing of the casks with such water would introduce enough nitrate to give the reaction. On the other hand, wines to which nitrate has actually been added, have failed to give any indication of its presence after the lapse of a few months. M. J. S.

Determination of Phosphoric Acid. By F. BENTE (Chem. Centr., 1887, 1241, from Rep. anal. Chem., 7, 533-534).-The low and discordant results obtained in the analysis of basic slag which were assigned by the author (Abstr., 1887, 397) to the insufficient time allowed for complete precipitation, are probably brought about by the varying amount of chlorine in solution. For that reason he now employs sulphuric acid to dissolve the slag, as Brunnemann and Loges have recently done. J. P. L.

Volumetric Determination of Phosphoric Acid. By C. SCHINDLER (Zeit. anal. Chem., 27, 142-146).-The molybdenum in the yellow phosphomolybdate precipitate is determined by the author's volumetric method (p. 758), and the phosphoric acid is calculated therefrom. A precipitate of constant composition is secured by adding citric acid to the ordinary molybdate solution (15 grams to 1 litre). 50 c.c. of the nitric acid solution of the phosphate (0.5 gram of substance) is mixed with so much of a solution of ammonium nitrate (750 grams per litre), that after the addition of the molybdate the mixture shall contain 25 grams of ammonium nitrate per 100 c.c. Then for each 0.1 gram of phosphoric anhydride, 100 c.c. of molybdate is added, and the mixture is heated in a water-bath to about 58°. The precipitate is allowed to deposit for 10 minutes, the supernatant liquid is filtered, and the precipitate is washed three or four times with dilute ammonium nitrate (100 grams with 10 c.c. of nitric acid per litre). It is then dissolved in 3 per cent. ammonia, treated with 10 to 20 c.c. of Fresenius' magnesia mixture, and made up to 250 c.c. After shaking, it is filtered: 50 c.c. of the filtrate is acidified with acetic acid, diluted to 300 c.c. with hot water, and titrated with lead acetate. Comparative determinations on a variety of materials show a close agreement with the magnesia method. M. J. S.