mechanical action. (See also Sebelien's method of estimating casein and albumen, p. 311.)*

§ 146. (5.) Isolation of Peptone-like Body (Galactin).-As in the previous operation, 100 cc. of milk are greatly diluted, the casein coagulated, and the whey separated by subsidence. This whey is used for the precipitation of a second portion, and the same whey from the second portion for the precipitation of a third, and so on. By this means it is possible to obtain the whey from a litre of milk in a form not too dilute. The albumen may be separated from time to time in the different fractions, or in one concluding operation; lastly, the casein must be collected and well boiled, and the liquid separated from it by filtration and strong pressure. A solution of nitrate of mercury is now added, the precipitate separated and decomposed by hydrogen sulphide, and the well-boiled filtrate precipitated by lead acetate. The amount of the peptone may be estimated from the weight of the lead oxide left on ignition of the lead compound, or the filtrate from the mercury sulphide may be treated by Kjeldahl's process, and the amount of peptone calculated from the nitrogen.

(6.) Estimation of the Total Nitrogen in Milk. Kjeldahl's t method of determining nitrogen is applicable to milk :-5 cc. of the milk are digested with 3 cc. of strong sulphuric acid at 100° for about an hour, then crystals of powdered potassic permanganate are added cautiously, and the heating continued until the mixture assumes a green colour. On cooling, the liquid is made up with water to a litre, and a fractional part neutralised by pure sodium or potassium hydroxide, and distilled in a current of hydrogen. The distillate is neutralised by decinormal acid, each cc. of which is equal to 0014 grm. nitrogen, or 00886 grm. albuminoids. (See also p. 311.) Blank experiments to ascertain the ammonia in the reagents themselves are absolutely

necessary.

$147. (7.) Isolation of the Principles Precipitated by Tannin.The whey, now free from casein, albumen, peptones, and colouringmatter, but containing mercury nitrate, must be made alkaline, the precipitate filtered off, the liquid saturated with hydrogen sulphide, any precipitate again filtered off, the liquid concentrated to a small bulk and completely precipitated with tannin. The tannin precipitate is decomposed as before described, p. 249.

*F. Klug's method of estimating albuminoids by the biuret reaction might probably be applied to milk; 4 cc. of a solution of albuminoid are mixed with 2 cc. of concentrated soda and 4 drops of 10 per cent. copper solution, and filtered. The filtrate is examined spectroscopically (see p. 80, et seq.); the absorption factors for X542·5 to X526·7 are as follows:-casein 0.72, albumen 0-79, peptone 2·65 to 2 635. (Centralbl. f. Physiol., 1893, S. 227.) + Zeits. Anal. Chemie, xxii., 366-382.

§ 148. (8.) Estimation of Urea.-The estimation of urea is of some importance, since any disorder interfering with the action of the kidneys throws (as has been well ascertained) an excess of urea on all the secretions of the body. A known quantity of milk, which should not be less than half a litre, is evaporated with constant stirring in very large flat dishes to a granular condition. The fat is next extracted by dry ether, and from the fat-free solids the urea is extracted with other substances by boiling alcohol. The alcoholic solution is evaporated to dryness; and from this dry residue absolute alcohol will extract the urea nearly pure. A litre of milk in its normal state yields about 10 mgrms. Urea must be identified by its properties as follows:

:

It is crystalline, crystallising in quadratic prisms, and polarising with a gentle blue colour under the microscope. The crystals should be heated with a little hydrate of baryta in a closed tube to 200° for some hours, when a very definite reaction ensues, ammonia and carbon dioxide being produced. The liquid may be distilled, and ammonia identified by the Nessler test. Carbonate of baryta will appear as a precipitate, and may be readily examined, converted into sulphate, and weighed; 1 part of barium sulphate = 2574 urea. A convenient method of identifying urea is also to dissolve the crystals in the smallest possible quantity of water, and then to add a drop of dilute nitric acid; the nitrate of urea is precipitated, and can be identified by its microscopic characters. Nitrate of urea crystallises on the rhombic system. The most common appearance is that of large plates, many of which lie one upon the other.

The

§ 149. (9.) Estimation of Alcohol.-A litre of milk, which, if acid, must be neutralised, is placed in a specially constructed, non-tubulated, very capacious retort, provided with a tube a metre in length. This tube is surrounded by a water jacket, through which a continuous stream of cold water runs. retort tube is pushed through a strong india-rubber stopper, and connected with a small flask holding about 200 cc., and immersed in ice and salt. The india-rubber stopper, by the aid of a second perforation, carries a piece of angle tubing, by which it may be connected with the mercury pump (fig. 4). or, where available, with the ordinary water pump so common in laboratories. The milk is now cautiously raised to the boiling point, and 100 to 150 cc. distilled over. distillate is redistilled, in the ordinary way, about one-third. All the alcohol is now in a very small compass, and the distillate should be placed in a strong assay flask, and oxidised to acetic acid by heating in a water-bath with from 30 to 70 cc. of an oxidising solution of bichromate of potash and

This

sulphuric acid.* The excess of bichromate is reduced by zinc, some phosphoric acid added, and the acetic acid distilled off by heating the flask (previously attached to an efficient condenser) over a spermaceti bath. On distilling in this way about three times to dryness, each time adding water, the whole of the acetic acid is obtained in the distillate, and may be determined by titration with a volumetric solution of soda. The amount of alcohol to which the acetic acid found is equal may be calculated by the aid of the following table :

The amount of alcohol is of importance in the analysis of decomposed milks; each part of alcohol may be taken to represent 2:06 parts of sugar.†

Under

§ 150. (10.) Volatile Acids.-Volatile acids may be separated in exactly the same way-viz., by careful distillation in a vacuum, first acidifying the milk by tartaric acid. Acetic acid in small quantities is invariably present in fermented milks; but the distillate of normal, quite fresh milk is neutral. In milks already undergoing decomposition, it is best to dilute the milk slightly, and filter through an earthenware cell, and then distil. these circumstances, it is not necessary to distil in vacuo. § 151. (11.) Estimation of the Total Acidity of Milk-Estimation of Lactic Acid.-10 cc. of d. n. acid are added to 10 cc. of milk diluted to about 100, filtered, the precipitate well washed, and in the clear filtrate the acidity estimated by d. n. soda, using phenyl-phthalein as an indicator; 10 cc. of d. n. soda will have to be subtracted from the total amount of alkali used as representing the acid originally added.

Estimation of Lactic Acid.-It is of great importance to make an accurate estimation of lactic acid in milk. A method, accurate enough for technical purposes, consists in thoroughly exhausting

* 147 grms. bichromate of potash: 220 grms. of sulphuric acid made up to 1400 cc. with water.

+ Pasteur, Ann. Chim. Physique, lviii., 323.

the milk of fat and lactic acid by ether, then taking another portion and exhausting it by carbon disulphide, when the difference between the two determinations represents the lactic acid. A more accurate method is as follows:-The milk is dried, exhausted of fat by carbon disulphide, then treated with an alcoholic solution of oxalic acid, filtered, and an excess of hydrated oxide of lead added. Any lactic acid now contained in the fluid will be present in solution as a lactate of lead. The liquid is filtered, saturated with SH, and again filtered, concentrated by evaporation, and boiled with oxide of zinc; on filtration, evaporation, and standing, crystals of lactate of zinc are produced. There are four isomeric lactic acids; that which is obtained from milk is fermentation lactic acid, also termed "ethylene lactic acid." The zinc salt has the composition 2(CHO)Zn + 3H20. It crystallises in four-sided prisms; it is soluble in 6 parts of boiling, 58 of cold water. It is nearly insoluble in hot or cold alcohol. 100 parts of the salt contain 25.8 of zinc oxide. Lactic acid itself may be obtained in a very pure state by decomposing the zinc-salt with hydrogen sulphide, when the acid presents itself as a colourless, strongly acid liquid. A drop of this acid, placed in the author's subliming cell (described in the author's work on "Poisons"), and heated very gradually above 200°, gives a white sublimate of lactide, CH4O2, a very characteristic reaction. If the heat is not gradual, this sublimate is not obtained, for it then decomposes into carbon dioxide and aldehyde.

§ 152. (12.) Detection of Metals in Milk.--The detection of the minute quantities of heavy metals which may occur in milk in cases, where, for the purpose of experiment, salts of the metals have been administered to animals or women, is best conducted by electrolysis, supplemented by the spectroscope. One of the best ways to do this is the method proposed by Dr. Reynolds.* Four to six rather stout platinum wires, half an inch in length, are made into a bundle by binding with thin platinum wire, and secured by cotton wool into the throat of the stem of a funnel so tightly, that water placed in the funnel filters through in single drops. The milk, previously acidified with nitric acid, and filtered, is placed in the funnel, and a platinum wire, connected with one pole of a battery, inserted in the funnel, so as to be about half an inch distant from the bundle of wire. This bundle is connected with the zinc terminal of the battery-a single Grove's cell is sufficient-and the current is allowed to pass until the whole has filtered through. A very decided metallic deposit

* Irish Hospital Gazette, 1873. See also "The Spectroscope in Medicine," by C. A. Macmunn, B. A., M.D. London, 1880.

may be recognised, and examined by ordinary analysis; but should the deposit be scanty or indistinct, two of the short wires are connected with a Ruhmkoff's coil attached to a sufficiently powerful battery. The wires are so adjusted opposite each other as to leave a very short interval between their points, and the succession of sparks allowed to stream between them is observed by the spectroscope. In this way the spark spectra of the metals arsenic, antimony, mercury, copper, and lead, are easily obtained, and may be identified.*

(13.) Detection of Nitrates in Milk.-Pure milk contains no nitrate, hence the detection of nitrates in milk would be fair presumptive evidence of watering. A test recommended by J. Szilasi† is as follows:-1 cc. of a solution of diphenylamine sulphate is put into a small porcelain dish, and a few drops of milk are dropped in; if nitrates are present, a blue colour will gradually appear. Soxhlet tests milk for nitrates as follows:-The milk is coagulated by a solution of calcic chloride free from nitrate, the serum is now treated with a solution of diphenylamine in concentrated sulphuric acid in the same way as with the ferrous sulphate test. The nitrate test will be found of great utility in those cases in which an impure strongly nitrated well-water has been added to milk, but with ordinary waters containing 01 grm. and under of nitric anhydride per litre, the test is not sufficiently delicate to be of much utility.

THE MILK SECRETED BY THE UNHEALTHY.

§ 153. The result of the analyses and cases shortly to be quoted shows,-(1.) That in the case of the cow, in certain diseases only, the milk constantly deviates from the normal standard; (2.) that the most marked changes are found in local diseases of the udder or mammary glands; (3.) that the animal may be labouring under a most mortal and virulent malady, and yet secrete milk which, although differing from the same milk secreted by the same animal when in health, yet, considered

Reference may be made to Boisbaudran's "Spectres Lumineux." Paris, 1874. A simple method of obtaining the wave length value of any spectroscope scale has been already given in the article on the "Spectroscope " in the present volume, p. 74, et seq.

+ Repert. Anal. Chemie., 33.