Apatite (containing 39.55 per cent of phosphoric acid), superphos phate (6.91 per cent of total and 5.50 per cent water-soluble phosphoric acid), and a basic phosphate, obtained by supersaturating a sodium phosphate solution with calcium chlorid, were treated with 100 cc. of tri-ammonium citrate solution. The basic phosphate was found to contain 37.41 per cent of phosphoric acid and 48.11 per cent of calcium oxid, corresponding to the formula 130aO-4P2O5. The following results were obtained:

If the water-soluble phosphoric acid in the superphosphate be subtracted from the percentages given, the figures for 5 and 1 gm. will be changed to 38.3 and 64 per cent, respectively.

A citrate solution made up according to Wagner's directions1 was used in treating one sample each of the bone meal and Thomas slag, with the following results:

The temperature of the room was about 4° higher when the last two tests were made, which possibly explains the higher results obtained. The solubility found for Thomas slag is seen to vary but little with different amounts of substance weighed out, compared with the data obtained for bone meal.

The author claims that the solubility of the different kinds of phos phates in citrate solution can not be considered indicative of their comparative agricultural value. He recommends the general adoption of Wagner's citrate solution, the digestion to take place always at the same temperature.-F. W. WOLL.

'Chem. Ztg., 18, p. 1935.

2 Wagner, Chem. Ztg., 19, p. 1420, recommends 17.5° C.

The variation in the composition of superphosphates and the evolution of hydrofluoric-acid gas in superphosphates made from phosphorites (Rpt. Agl. Chem. Soc. Bologna, 23 (1894–95), pp. 23-25).-Determinations at different dates from September 22, 1894, to June 15, 1895, of soluble and reverted phosphoric acid in 3 samples of superphosphates made from mineral phosphates, showed a regular and marked increase of soluble phosphoric acid, while with superphosphates made from bone no such increase was observed. Further investigation showed that this increase was due to the gradual evolution of hydrofluoric-acid gas in case of the superphosphate made from the mineral phosphates.

The determination of the milk sugar content of milk, as well as the specific gravity of the milk serum: A contribution to milk analysis, E. VON RAUMER and E. SPÄTH (Ztschr. angew. Chem., 1896, Nos. 2, pp. 46-19; 3, pp. 70-73).-Seventy-four analyses of fresh milk are given. After long experience the authors give the following as the simplest method found for the preparation of the milk serum: A beaker weighing with glass rod about 60 gm. is partly filled with from 200 to 260 cc. of milk, and weighed on a balance sensible to 5 mg.; 2 cc. of 20 per cent acetic acid is added and the whole heated on a water bath for a half hour. After cooling, water is added to the original weight and the whole stirred and filtered. The first drops are apt to be turbid and should be refiltered. If the filtrate is still turbid, alumina cream is added to a weighed portion and the whole heated and then brought to weight again. This gives a clear serum. The specific gravity is then determined with a 50 cc. picnometer, which the authors deem the only way in which results absolutely correct to the fourth decimal can be obtained.

For the determination of milk sugar 25 cc. of milk is well shaken in a half liter flask with 400 cc. of water, 10 cc. of Fehling's solution (without the Rochelle salt), and 3.9 to 4.2 cc. normal potassium hydroxid, the liquid filled to the mark with water and filtered. The filtrate should contain an excess of copper sulphate, and accordingly should not react alkaline. The sugar is determined gravimetrically in 100 cc. portions of the filtrate by the Soxhlet-Allihn method. If the milk serum is used the same procedure is followed, but the results are uniformly 0.1 to 2 per cent higher than those obtained with the milk. The milk sugar in curdled milk is determined in the serum.

Analyses are also given which show that the specific gravity of the serum decreases on standing, most noticeably after the first 24 hours. The content of milk sugar also falls off rapidly and especially at a slightly raised temperature. Both the specific gravity and the sugar content can, however, be determined with fair accuracy during the first 24 hours. The authors believe these determinations to be of especial value in detecting watered milk.

Experiments on polarization go to show that if the milk is polarized

without previous heating too low results are obtained, while if basic lead acetate is present during the heating much of the sugar is actually destroyed. The following method gave excellent results where it could be used: Fifty cubic centimeters of milk was heated to boiling in a 100 cc. measuring flask, cooled to 17.50, and 10 cc. of basic lead acetate added. The whole was then made to 100 cc., filtered, and polarized in a 200 mm. tube. If the serum, obtained as above, was used 100 cc. was taken and directly cleared with basic lead acetate, filtered, and polarized in a 220 mm. tube. In many cases the results compared closely with those obtained by the gravimetric method. However, the polarimetric method, while it results in a great saving of time, in many cases gives too high results. The authors give results to show that this can not be due to any reagents used, but attribute it to the occurrence in some milk of a dextrin-like body which does not reduce Fehling's solution but which gives greater rotation to the right. Ritthausen, Schmoeger, and Landwehr have also proved the presence of such carbohydrates in milk. These carbohydrates may vary very much in quantity, and appear to be present especially in the colostrum.

The following conclusions are drawn: The specific gravity of a normal milk serum varies between 1.0260 and 1.0330. The content of milk sugar varies between 4.25 and 5.20 per cent if determined in the milk direct, but is from 0.1 to 0.2 per cent higher in the serum. An addition of water can only be proved in a curdled milk when results on the original fresh milk are at hand for comparison, and then only within 24 hours. The determination of milk sugar by polarization is not allowable on account of the occurrence in certain milk of dextrin-like bodies which affect polarized light but do not reduce copper solution.-C. L. PARSONS.

Cold saponification: Saponification and Reichert-Meissl numbers, R. HENRIQUES (Ztschr. angew. Chem., 1895, No. 24, pp. 721-721).After comparing cold and warm saponification on several oils, such as linseed, cotton-seed, olive, cocoanut, etc., to the advantage of the former, the author states that he considers cold saponification especially to be recommended in the determination of the Reichert-Meissl number of fats and oils. By the old methods ethers of the volatile fatty acids are formed, but this can be entirely avoided by cold saponification. The results obtained are also several tenths higher.

Five grams of the fat is dissolved in a porcelain dish in 25 cc. of petroleum ether, 25 cc. of a 4 per cent solution of sodium hydroxid added, and the whole allowed to stand overnight. Saponification is complete in the morning. The whole is then evaporated to complete dryness on a water bath. The residue is transferred to a flask and distilled as usual. The results are some 0.5 cc. higher than by the usual method of Wollny.-C. L. PARSONS.

The determination of pentoses and pentosans by distillation with hydrochloric acid, MANN, KRÜGER, and TOLLENS (Ztschr. angew. Chem., 1896, No. 2, p. 33).—The authors have carefully studied the

methods of determining the pentoses by distillation with hydrochloric acid, and find the former factors not absolutely correct. They give the following factors:

It was found that the results obtained by the phloroglucin method proposed by Conucler1 are as reliable as those obtained by the phenylhydrazin method. The phloroglucin method is carried out as follows: Two to five grams of the substance is distilled with 100 cc. hydrochloric acid of 1.06 sp. gr. (12 per cent HC1) exactly as prescribed by Flint and Tollens. To the distillate twice as much phloroglucin, previously dissolved in a little hydrochloric acid, 1.06 sp. gr., is added as there is furfurol assumed to be present. The volume is made up to 400 cc. with acid of the same strength, the solution well stirred and allowed to stand overnight. The precipitate is then collected on a tared filter, washed with 150 cc. of water, dried, and weighed. The following factors are used:

The formation of furfurol during the distillation of certain substances which are known not to contain pentosans is supposed to be due to the presence of hitherto unknown bodies which are probably formed by the oxidation of starch and which are easily decomposed.—W. H. KRUG. The determination of pentoses and pentosans by distillation with hydrochloric acid, B. TOLLENS (Ztschr. angew. Chem., 1896, No. 7, p. 194). The author has decided that the factors given in his last

Chem. Ztg., 1894, No. 51.

2 Landw. Vers. Stat., 42, p. 381.

1

paper on this subject are too complicated, and proposes to return to the simple factor 1.84, previously published by him and Mann. The factors now given are:

On the detection of pentoses by the phloroglucin-hydrochloric acid precipitate method, B. TOLLENS (Ber. deut. chem. Ges., 29 (1896), No. 8, pp. 1202-1209).-As is well known, pentoses and pentosans may be detected by warming the solution with an equal volume of hydrochloric acid and a little phloroglucin; the mixture will assume a cherry-red color, giving a characteristic absorption band in the spectrum between the lines D and E. The solution quickly becomes turbid from the separation of a dark deposit which soon renders the recognition of the absorption band impossible, and the difficulties in applying the test to impure solutions, such as plant juices, urine, etc., are very considerable. If, however, the brown deposit produced in the reaction be dissolved in alcohol the characteristic color and spectrum band are reproduced. The author's method consists in filtering and washing the dark deposit and dissolving it on the filter in alcohol. A colored solution is produced that is sufficiently permanent to be examined satisfactorily with the spectroscope. The author has tried this method with various sugars and upon several natural products and finds it much more certain and considerably more delicate than the direct application of the test.-A. M. PETER.

The composition of wood gum, S. W. JOHNSON (Jour. Amer, Chem. Soc., 18 (1896), No. 3, p. 214).—The gum of corncobs was found to consist almost wholly of xylan, but birchwood gum on hydrolysis yields a sirup which gives only a very small amount of crystals on long standing or fractionating with alcohol. "Seeding" with crystals of xylose does not increase the yield. Vegetable ivory, when extracted with sodium hydroxid solution, yields a large amount of mannan, which is easily obtained pure. It is probably accompanied by an alkali-soluble substance of lower carbon content. All these substances are difficult to obtain on account of their hygroscopic nature. They are most easily dried in vacuo at 110 to 112°.-w. II. KRUG.

On the ammonia derivatives of some sugars, C. A. LOBRY DE BRUYN and F. H. VAN LEENT (Neue Ztschr. Rübenz. Ind., 36 (1896), No. 7, p. 78; Rec. trav. Chim. Pays-Bas, 12, p. 286; 11, p. 98).-Dry ammonia gas has no action on lactose or its anhydrids at normal temperatures. In a concentrated aqueous solution of ammonia the normal rotation of [a] = +53° falls to [a] = + 30° in 5 days. Lactose hydrate dissolves quite readily in methyl alcohol containing about 20 per cent of ammonia. This solution within 11 to 18 days deposits

D

1Ztschr. angew. Chem., 1896, No. 2, p. 33.