gen peroxide is shown by the appearance of a red color. Utz* states that the reaction is obtained whether or not the milk has been heated previous to adding the peroxide.

Peroxidase Methods.-Several of the methods for detecting peroxidase (pp. 173 and 174), notably the paraphenylenediamine (Dupouy), the benzidine (Wilkinson and Peters), and the iodide-starch (Roi and Köhler) methods, can be reversed for the detection of hydrogen peroxide. The tests are conducted as described except that the hydrogen peroxide is omitted and, since the peroxidase may have been destroyed by hydrogen peroxide or by heating, it is usually necessary to add raw milk of known purity. La Wall† in performing the Wilkinson and Peters test first coagulates a mixture of 10 cc. of raw milk and 2 cc. of 4% alcoholic benzidine with 2 to 3 drops of glacial acetic acid, then adds a few cubic centimeters of the suspected sample. A blue zone is formed when from 1.5 to 30 parts of hydrogen peroxide per 10,000 are present.

Hehner-Feder Formaldehyde Method. This is the Hehner method for formaldehyde reversed and slightly modified. Mix 5 cc. of the milk with 5 cc. of concentrated hydrochloric acid and a drop of dilute formaldehyde solution, then heat at 60° C. for 3 to 4 minutes and shake once. If a violet color develops hydrogen peroxide is indicated. Wilkinson and Peters have shown that the reaction is most decisive when about 0.005% of hydrogen peroxide and 0.004 to 0.013% of formaldehyde are present; with other proportions it may fail. Ferric salts, nitrites, and possibly nitrates also give a violet color.

Routine Inspection of Milk for Preservatives. It was Leach's custom in Massachusetts to examine all the samples of milk collected during the months of June, July, August, and September for the commonly used preservatives, in addition to the regular analysis for total solids and fat. The number of samples thus examined amounted to upwards of 500 per month, varying from 10 to 60 per day. The results of such an examination during four years are shown on p. 170. ||

Such a system by no means involves a large amount of time or labor, and is really essential before passing judgment upon the purity of the milk, since, unlike added color, there is nothing in the physical appear

*Milchw. Zentbl., 1, 1905, p. 175.

† Am. Jour. Pharm., 82, 1908, p. 57.

Zeitz. Unters. Nahr. Genussm., 15, 1908, p. 234.

§ Ibid., 16, 1908, p. 515.

Mass. State Bd. Health, Rep., 1902, p. 474.

ance of the milk to suggest the presence of preservatives, nor are they rendered apparent by the taste, if skillfully used.

The methods employed are carried out as follows: *

(1) Formaldehyde. After having been examined for total solids and fat, the milk samples are arranged in order in their original containers, and about 10 cc. of each sample are poured into a casserole and tested in succession by means of the hydrochloric acid and ferric chloride test (p. 165). A large stock bottle, which may be fitted with a siphon if desired, is kept on hand containing the hydrochloric acid reagent. Less than one minute is required in making the formaldehyde test for each sample.

(2) Carbonate and Boric Acid.-These tests have been so simplified as to be, as it were, a side issue in the process of cleaning the platinum dishes used for the determination of total solids. The various residues from the total solids are burnt to an ash in the original numbered dishes in succession, these dishes, after incineration, being arranged side by side on a flat tray. By means of a pipette, one or two drops of dilute hydrochloric acid are introduced into each dish in succession, noting at the time any effervescence that may ensue, which is in itself an indication of sodium carbonate. After every milk ash has been acidulated, a few cubic centimeters of water are added to each dish by means of a washbottle, the dissolving of the ash being hastened by giving a rotary motion. to the tray containing the dishes. A strip of turmeric-paper is then allowed to soak for a minute or so in each dish, after which it is withdrawn from

* Mass. State Bd. Health, Rep., 1901, p. 447.

contact with the solution and allowed to adhere to the side of the dish above the liquid, where it remains until dry. If the paper when dry is of a deep cherry-red color, turning a dark olive when treated with dilute alkali, the presence of boric acid is assured. These methods are, of course, preliminary tests for quickly singling out the preserved samples. Such confirmatory tests as are desired may in all cases be employed.

*

VARIOUS ADULTERANTS.-Cane Sugar is said to be used to increase the total solids of milk, but if present to any marked degree, it could hardly fail of detection by reason of the sweet taste imparted to the milk. Cane sugar in milk may be detected by boiling 5 to 10 cc. of the sample with about o.1 gram of resorcin and a few drops of hydrochloric acid for a few minutes. In the presence of cane sugar, a rose-red color is produced.

According to Richmond, cane sugar may be estimated by first ascertaining the total polarization of the sample as in the estimation of milk sugar (p. 134). The milk sugar is then determined by Fehling's solution (pp. 136 to 138) either volumetrically or gravimetrically. The difference between the anhydrous milk sugar found by the latter, or Fehling method, and that calculated by dividing the polarization by 1.217 will give the percentage of cane sugar present.

Cotton's method † of detecting cane sugar, when present to the extent of 0.1% consists in mixing in a test-tube 10 cc. of the suspected milk with 0.5 gram of powdered ammonium molybdate, and adding to the mixture 10 cc. of dilute hydrochloric acid (1 to 10). Ten cc. of milk of known purity, or 10 cc. of a 6% solution of milk sugar are similarly treated by way of comparison. Both tubes are placed in a water-bath and the temperature gradually raised to 80° C. If cane sugar is present, an intense blue coloration is produced, while the genuine milk or the solution of milk sugar remains unchanged at the temperature of 80°. If the temperature is raised to the boiling-point, however, the pure milk or milk sugar solution may also turn blue.

Detection of Starch in Milk.-A small quantity of milk is heated in a test-tube to boiling, cooled, and a drop of iodine in potassium iodide added. A blue coloration indicates starch.

Condensed Skimmed Milk as an Adulterant.-The use of condensed unsweetened skimmed milk to raise the solids of a skimmed or watered

* Woodman and Norton, Air, Water, and Food, New York, 1914, p. 151.
† Abs. Analyst, 23, 1898, p. 37.

milk above the standard has been noted in Massachusetts. This sophistication is rendered apparent by the abnormally high solids not fat of the sample, which in some instances have exceeded 11%. A solid not fat in excess of 10% is suspicious of this form of adulteration. By fixing a legal standard for both fat and solids not fat, such tampering with milk may readily be checked.

Analysis of Sour Milk. It occasionally becomes necessary for the analyst to deal with samples of sour milk, especially in the summer-time, when the milk has been brought from a long distance. While the process of lactic fermentation results in the formation of traces of volatile acids, unless the sample has become so badly curdled as to render an even homogeneous mixture of the various parts impossible, a fair determination of the solids and fat can readily be made. Experience has proved that, excepting in instances of milk so badly soured as to have become actually putrid, the analysis of sour milk, if carefully made, should not differ materially from that of the same milk before souring.

Care must be taken to secure an even emulsion of the curd and whey. This may sometimes be accomplished by repeatedly pouring the sample back and forth from one container to another. Again, it is sometimes necessary to use an egg-beater of the spiral wire pattern, which preferably should easily fit the can or milk-container. Unless a fine, even emulsion can be secured, it is impossible to make a satisfactory analysis of sour milk. With such an emulsion results can be relied on.

In measuring portions of the thoroughly mixed sample of sour milk for analysis, a pipette should be used having a large opening.

HOMOGENIZED MILK.

This product is prepared from ordinary milk by heating and then passing through the "homogenizer " whereby the fat globules are broken up into smaller globules and the creaming power reduced practically to nil. In the homogenizer the milk is forced, under a pressure varying up to 4cco pounds or more per square inch, into fine jets or sheets which impinge either against each other or against an agate surface thus disrupting the globules. These in normal milk often exceed 10μ and are mostly 5 to 6μ while in well-homogenized milk they are mostly only 1 to 2μ.*

The machine is also used to emulsify oleo, cottonseed, and other oils and low melting-point fats with skim milk thus furnishing a wholesome food

* Baldwin, Am. Jour. Pub. Health, 6, 1916, p. 862.

for calves, hogs, and even human beings, although the temptation to market the products dishonestly has not always been resisted. Homogenized mixtures have also been used in cream, condensed milk, and ice cream.

Analysis of Homogenized Milk.-It has been demonstrated by Richmond * that the Adams paper coil method gives low percentages of fat with homogenized milk while the Röse-Gottlieb, Werner-Schmidt, and Gerber methods are satisfactory. Other constituents are determined as in ordinary milk.

Distinctions from untreated milk are based on the size of the fat globules as above noted, also on physical constants, particularly the viscosity.

PASTEURIZED MILK.

The analyst may be called on to determine whether or not milk has been pasteurized to conform with municipal or state regulations.

Detection of Peroxidase. The following tests show whether the milk has been pasteurized at 80° C., or higher but, as found by Lythgoe,† are of no value when 63°, which is now deemed sufficient, is employed.

Dupouy Paraphenylenediamin Method.‡-Shake 5 cc. of the milk in a test tube with 1 drop of 0.2% hydrogen peroxide solution (containing I cc. of concentrated sulphuric acid per liter) and 2 drops of 2% paraphenylenediamin. If the milk becomes blue immediately it has not been heated to 78° C.; if it becomes gray-blue immediately or within half a minute it probably has been heated to 79-80°; while if it remains white or becomes a faint violet-red it has been heated above 80°.

Wilkinson and Peters Benzidine Method. §-To 10 cc. of the milk add 2 cc. of a 4% alcoholic solution of benzidine and 2-3 drops of glacial acetic acid, or an amount just sufficient to coagulate the milk, and shake. Add cautiously to the mixture 2 cc. of 3% hydrogen peroxide solution, allowing the reagent to run down the sides of the test tube. With raw milk or milk heated below 78° C. an intense blue color appears at once; with milk heated at 80° or higher no color appears.

Other Tests are the original Arnold guaiac test || and its modifications and the Roi and Köhler iodide-starch test.

Analyst, 31, 1906, p. 218.

† Jour. Ind. Eng. Chem., 5, 1913, p. 922.

Dupouy, Rep. pharm. III, 9, 1897, p. 206; Storch, Copenhagen Exp. Sta. Rep., 1898.

§ Jour. Dep. Agric., Victoria, 6, 1908, p. 251.

Jahr. König. Tierärz. Hochsch., 1880-1882, p. 161.

¶ Milch. Ztg., 31, 1902, pp. 17, 113.