Page images

covering may be only 2 cm. provided it is retentive of moisture. Seeds should, if possible, be procured from the previous year's crop, and never used when over two years old.

There are other recommendations well known in agricultural practice. J. F.

Manurial Experiments with Sugar-beets. By A. PAGNOUL (Bied. Centr., 1887, 223-227).-The experiments were made in vessels containing about 1200 litres filled with a sand freed from lime and organic matter, but containing traces of alumina, iron, phosphoric acid, and potash.

The manures employed were Chili saltpetre (15.5 per cent. N), ammonium sulphate with 20 per cent. N, ammonium nitrate, and dried blood; potassium chloride 50 per cent. ; superphosphate; Thomas slag with 7 per cent. phosphoric acid, and finely ground natural phosphate of 20 per cent., with addition of carbonate of lime.

The growth of the beets sown without any manure, and those manured without phosphoric acid was slow and weak. The ground raw phosphate and the Thomas slag produced very good effects, and the addition of carbonate of lime increased the crop in every case. The best results were obtained from a mixture of 600 parts of Chili saltpetre, 600 of superphosphate, and 300 of potassium chloride, and with another mixture of 500 parts of dried blood, 300 Chili saltpetre, 300 calcium chloride, 660 superphosphate, and 1000 carbonate of lime.

The average weight of the roots, however, shows that sand to which is added theoretically correct quantities of manures, does not produce as favourable results as natural soils.

One of the vessels was manured with crushed oil-cake placed at a certain distance from the seeds. The plants were longer in developing, but the crop of leaves was as large as in the case of the complete manure, the roots attained an abnormal length, and they passed in a slanting direction so as to reach the oil-cake, which was not thoroughly mixed with the soil as in the case of soluble manures. The sugar contents of the juice was considerably influenced by the nature of the manure; the absence of potash sensibly diminished the yield of sugar, and in one case where both soda and potash were absent the actual weight of the whole roots was diminished one-half. Soda appears to have more affinity for organic salts than potash, and forms larger quantities of those salts which are considered impurities of beet-juice. It is also shown that when manured with substances containing both soda and potash, the latter is absorbed in preference to the former.

J. F.

Analytical Chemistry.

Weil's Method for Determining Sulphur. By C. FRIEDHEIM (Ber., 20, 1483-1485).-The author made determinations of sulphur by Weil's method, and obtained results which confirm the statement previously made by him (this vol., p. 396) that a correct result can only be obtained when the error in one direction happens to equal that in the other. This remark applies to the results since given by Weil (this vol., p. 618).

N. H. M.

Detection of Sulphites in Presence of Thiosulphates and Sulphates. By A. VILLIERS (Compt. rend., 104, 1177-1178).— When a solution of a normal alkaline sulphite is mixed with excess of barium chloride, barium sulphite is precipitated, and the liquid which was originally strongly alkaline becomes neutral to litmus. If the barium chloride is added to an alkaline hydrogen sulphite, normal barium sulphite is precipitated, and free sulphurous acid remains in solution. A solution which contains a mixture of a normal alkaline sulphite and a hydrogen sulphite has a distinctly alkaline reaction even when the proportion of normal salt is very small, but after addition of barium chloride the solution becomes acid and contains free sulphurous acid.

To detect sulphites in presence of thiosulphates, the solution is neutralised with hydrochloric acid, care being taken to avoid excess, and the liquid is mixed with barium chloride, when its reaction becomes acid to litmus. If the liquid is distilled, the sulphurous anhydride passes over with the first portion of the distillate, and can be detected in the usual way. The filtered liquid can also be tested for sulphurous acid by means of iodine, without distillation.

This method is also applicable in presence of the thionic acids.

C. H. B.

New Method for the Quantitative Determination of Hydroxyl. By C. L. JACKSON and G. W. ROLFE (Amer. Chem. J., 9, 82-87).As an improvement on the ordinary method of determining hydroxyl groups in organic compounds, parabromobenzoic chloride or anhydride is substituted for acetic or benzoic chloride, and in the ethereal salts formed with the compound in question, the bromine is estimated by Carius' method. The differences to be looked for are almost twice as great as in the analyses of the acetyl or benzoyl compounds, and the parabromobenzoyl compounds are easily prepared and purified.

The method of preparing parabromobenzoic acid is described; the yield amounts to 40-70 per cent. Parabromobenzoic chloride, C.HBr COCI, has been previously prepared, but not described. It melts at 30°, and boils at 245-247°; it is not dissolved or acted on by cold water, but dissolves in benzene and in alcohol, being then converted into its ethereal salt. Parabromobenzoic anhydride,


is very insoluble in ordinary reagents, is crystalline, and melts at 212 -213°. Its best solvent is chloroform.

Parabromobenzamide, C,H,BrCONH2, was prepared from the chloride, it is nearly insoluble in cold water, is crystalline, and melts at 186°.

As an example of the application of the method, the preparation of the two following compounds is described. Phenyl parabromobenzoate, CH,Br·COOCH, was obtained by heating at about 200° phenol, with either parabromobenzoic chloride or anhydride. It forms crystals resembling naphthalene, melting at 117°, insoluble in water, but soluble in other solvents. Pyrogallol triparabromobenzoate, (C&H,Br COO),C,H,, was made by heating pyrogallol with the chloride at 100°; the excess of chloride was removed by light petroleum and the pyrogallol by sodium carbonate. The substance crystallises easily from hot benzene, and melts at 140°. The corresponding pyrogallol tribenzoate is resinous, and hardly capable of purification.

H. B.

Quantitative Estimation of Glycerol. By R. DIEZ (Zeit. physiol. Chem., 11, 472—484).-The methods hitherto employed for the estimation of glycerol in wine and beer consist in dissolving it out usually by alcohol and ether from a mixture of the beverage with chalk, and finally weighing the glycerol. Neubauer and Borgmann (Abstr., 1879, 404) found that the glycerol so obtained contained 2 per cent. of mineral constituents and 04 per cent. of nitrogen. Champion and Pellet (this Journ., 1873, 1165) devised a method in which the glycerol was obtained as nitroglycerol, and weighed in this form, but this and other methods are also liable to error. The present method is one in which the compounds of glycerol with benzoyl are weighed. There are three benzoates of glycerol, according as to whether one, two, or three atoms of the hydrogen of the latter are replaced by the group C,H,O2. The following gives the method of procedure:-Glycerol was diluted to a known extent with water (01 gram in 10 or 20 c.c.); 5 c.c. of benzoic chloride and 35 c.c. of sodium hydroxide added; this mixture was cooled and shaken for 10 to 15 minutes. The benzoyl compound which separated was collected on a weighed filter, washed with water, dried at 100°, and weighed. A mean of eight estimations gave the amount of the compound as 0.385 gram. In a second series of four estimations, the number obtained was rather higher, the mean being 0.395 gram; in these cases, the alkaline filtrate was shaken a second time with benzoic chloride and sodium hydroxide; the second filtrate contained hardly a trace of glycerol. These numbers formed a basis for the subsequent analyses, and showed that the compound formed in this way was chiefly the tribenzoate; theoretically the amount of that compound for 0 1 gram of glycerol would be 0.439 gram. Tables of the amount of glycerol in various forms of beer and wine, estimated by this method, are given, the numbers obtained being somewhat less than those given by Borgmann. The method has the following advantages: the substance weighed is solid and not hygroscopic, and admixture with inorganic and nitrogenous substances is avoided.

W. D. H.

New Sugar Reactions. By D. LINDO (Chem. News, 55, 230 and 239). With reference to the new reactions for sugar described by Molisch (Abstr., 1886, 923), the author has found that a 1 per cent. starch or gum solution gives this reaction with thymol, as does also one part of cane-sugar in 200,000 of distilled water, and that all samples of normal human urine do so also, even when diluted with 50 vols. of water. He considers, however, that this fails to prove the presence of sugar is normal urine, which as a complex fluid must be acted on by the sulphuric acid. He finds that nitrates give the same characteristic colour reactions and precipitate on dilution, and that similar results are obtained when menthol is substituted for thymol; menthol has also the advantage of not giving with nitrates or nitrites any reaction that could be mistaken for that of sugar. Chlorides or hydrochloric acid do not impair the delicacy of either test, but sugar cannot be detected by either, in the presence of notable quantities of nitrates or nitrites.

R. R.

Analyses of Sugar-cane and Beet Juices. By C. A. CRAMPTON (Chem. News, 55, 207-209). The paper details the author's method of analysing sugar juices. By working with a smaller quantity and using the same solution, first for polarisation and then for reduction by Fehling's solution, with filtration by means of Wiley's tubes, he is enabled to expedite the process. The paper contains an extended table for calculating the results from the different factors obtained.

R. R.

Separation of Acetic Acid from Formic Acid. By D. S. MACNAIR (Chem. News, 55, 229). For the quantitative estimation of acetic acid in presence of formic acid, the author recommends that the substance be distilled with dilute sulphuric acid, and the distillate boiled for 10 minutes in a reflux apparatus, with an equal bulk of chromic acid mixture, made by dissolving 12 grams of potassium dichromate in 30 c.c. of sulphuric acid, diluted with 100 c.c. of water. Acetic acid only remains unchanged, and when the liquid is distilled, passes over and can be titrated in the distillate. R. R.

Analysis of Milk. By F. G. SHORT (Amer. Chem. J., 9, 100103). About 2 c.c. of the milk is weighed into a very thin, tared glass capsule, containing a quantity of asbestos, and dried at 110° for two hours. For the determination of the fat, the capsule is wrapped in a piece of cheese cloth, crushed and pushed tightly into the tube of a continuous ether extraction apparatus, and extracted for two hours with 50 c.c. of ether; the ether then evaporated, and the flakes heated in the drying oven for three hours before weighing. The same method can be used for butter analysis; the sample is placed in a jar or bottle, the whole melted and then vigorously shaken until solid. The drying and extraction of fat is carried out as before. In place of the glass capsules, porcelain ones may be used and the dried fat and asbestos transferred by a piece of cloth to the extraction apparatus, the little dish being washed with a few drops of ether. A large

number of analyses may be conducted at once, and very concordant results are obtained.

H. B.

Determination of Butter in Milk. By H. N. MORSE and C. PIGGOT (Amer. Chem. J., 9, 108-112).-20 grams of anhydrous copper sulphate is made to cover the bottom of a mortar, and 10 c.c. of the sample carefully run into the powder. In a few minutes, when dry, the contents are ground up with a little sand, and transfered to an extraction tube plugged with cotton-wool, where it is extracted with 10 portions of 10 to 15 c.c. of light petroleum; the fat solutions are concentrated to 10 c.c. or less, and 20 c.c. of decinomal potash solution (in 95 per cent. alcohol) added; after warming for a quarter of an hour, the excess of alkali is estimated by standard acid and phenolphthalein. To saponify one gram of butter, 0-230 gram of potassium hydroxide is required. The results agree with those obtained by other methods, and 25 to 30 analyses can be made in the day.

H. B.

Colour Tests for Strychnine and other Alkaloïds. By C. L. BLOXAM (Chem. News, 55, 155).-Strychnine is dissolved in a drop of dilute nitric acid, and gently heated; to the warm solution a very minute amount of potassium chlorate is added, when an intense scarlet coloration is obtained. Ammonia changes this to brown, and gives a brownish precipitate; the mixture is then evaporated to dryness, when it leaves a dark-green residue, dissolved by a drop of water to a green solution, changed to orange-brown by potash, and again turned green by nitric acid. A table is given showing the colourchanges given by brucine, narcotine, morphine, quinine, cinchonine, and caffeïne on like treatment. The reactions given by various alkaloïds, when boiled with a mixture of potassium chlorate and hydrochloric acid, are also described. A. J. G.

Guaiacum Resin. By H. HAGER (Zeit. anal. Chem., 26, 261).—To distinguish the purified from the natural resin, the latter alone being suitable as a test for ozone, 0·15 gram is dissolved in 5 c.c. of absolute alcohol, and to the solution, filtered in the shade, 10 drops of oil of turpentine are added. If in the course of half an hour the solution turns bluish, the resin has either been purified or has been altered by heat or sunlight. Natural resin gives a solution which remains yellow for several hours. M. J. S.

« PreviousContinue »