Tables 1 and 2 show the composition of these known-pure milks, as well as the freezing-point result obtained on each individual sample. The composition of these milks corresponds very well with the range in composition of known-pure milks tabulated by Brown and Ekroth1, Lythgoe2, and others. Also the range in freezing point does not materially overlap the values reported by investigators already mentioned. In a few exceptional instances, a milk may test 0.001° higher than the minimum of 0.540°. In one instance, a sample tested 0.001° below the maximum of 0.560°, fairly confirming the range of ±0.006 reported by the Queensland laboratory.

Table 2 shows also the results of the analysis of two samples of skimmed milk. It is significant that while complete skimming removes all of the fat and produces no material change in the other milk solids, substantially no effect is produced upon the freezing point. result is indicated in the sample of partially skimmed milk.

The same

To a sample of skimmed milk manufactured from a skimmed milk powder was added the theoretical amount of water necessary to restore the powder to its original fluid condition. After beating the mixture up thoroughly, the sample was analyzed, and a freezing-point determination made. As might be expected, an abnormal freezing-point result was obtained, indicating the addition of approximately 12 per cent of water. The sample was left in an ice box overnight, and the next day subjected to freezing-point test, with no material change in the result. Plainly, however, the freezing-point result depends largely upon the amount of water incorporated, and it is possible that an excessive amount of water was used in preparing the artificially made skimmed milk sample.

Combining the theoretical amount of water with a dried skimmed milk, however, could not be expected to restore the product to its original condition. Nevertheless, this point has not been investigated sufficiently to warrant any positive conclusions.

A mixture consisting of equal volumes of distilled water and standard evaporated milk gave a freezing-point test of -0.568°C., indicating that an amount of water insufficient to restore the sample to its original whole milk condition was added. A mixture consisting of 60 per cent water with 40 per cent of evaporated milk gave a freezing-point test of -0.442°, indicating excess water amounting to 19.6 per cent.

These experimental trials show that the freezing-point determination may serve as a criterion indicating the correct theoretical amount of water to be combined with a given sample of evaporated milk. In the case of any sample of dried milk powder or evaporated milk, the original product may be duplicated, providing a determination of the correct percentage of water to be incorporated may be made. Table 3 contains

1 J. Ind. Eng. Chem., 1917, 9: 297.

2 Mass. State Board Health, Monthly Bull., 1910, new ser., 5: 419.

a number of illustrations, based on actual analyses, of the changes which occur in a sample of milk to which skimmed milk or water or both have been added in various ways. These results confirm again the fact that the freezing-point test constitutes a reliable means of determining the addition of water. Inspection of the analytical figures alone might lead to the conclusion that certain samples of so-called adjusted or standardized milk are actually normal. Take, for example, the sample showing 3.0 per cent butter fat and a specific gravity of 1.0298. While comparison with charts showing analyses of known-pure milks makes it obvious that this sample has been adulterated by the addition of skimmed milk, there is no evidence of the addition of water. The application of the freezing-point test, however, indicates about 11 per cent of added water, the amount actually incorporated in the sample when it was made up in the laboratory.

*Reduced to 3.1 per cent fat by adding 20 per cent of skimmed milk.

+ Reduced in fat by adding 15 per cent of skimmed milk and 11 per cent of water.

Table 4.

Freezing-point tests on known pure milks containing known percentages of added water.

"Skimmed milk” made from dried skimmed milk and water.

TABLE 5.

Composition of two samples of milk taken at Farm School Dairy, September 26, 1919.

* First portion drawn from udder, representing only a partial milking. Contains no added water. † Complete milking from udder, supposed to contain added water-approximately 10 per cent.

TABLE 6.

Composition of two samples of milk taken at Farm School Dairy, October 14, 1919.

* First portion drawn from udder, representing only a partial milking. Contains no added water. † Complete milking from four Jersey cows. Supposed to contain 5 per cent added water.

The practical application of the freezing-point determination is further illustrated by the results (Table 4) of tests made on samples of knownpure milk to which various percentages of water were actually added. It is important to bear in mind the fact that the freezing-point result obtained on a sample of milk containing a known amount of added water will depend largely upon the freezing-point test of the original whole milk. For example, in the case of two milks, each known to contain 10 per cent added water, very different results may be obtained, depending upon whether the freezing point of the original milk registered in the neighborhood of -0.540° or somewhere near-0.560°. The narrow range in freezing-point variations, therefore, makes it necessary in estimating percentages of added water to take into account a correction factor which may fairly be given as approximately 2 per cent.

Tables 5 and 6 give the results of analysis of two samples of milk drawn from Holstein cows. One showed a fat content of 3.1 per cent, as well as a very low specific gravity, and solids and ash content. The general composition of the other sample was practically the same, except for the butter fat, which was very low, viz, 1.8 per cent. Judged by ordinary rules, both of these samples might be suspected of adulteration with water. Information accompanying the samples was to the effect that each had been obtained as the first portion drawn from the udder of the cow (about one pint in both instances), representing only a partial milking; in other words, that portion of the milking which would be

discarded as not constituting a representative whole milk. Nevertheless, on application of the freezing-point test to these samples very normal results were obtained, in the one case -0.543° and in the other -0.544°. Thus it appears that the freezing-point test indicates freedom from added water in both samples, while the ordinary analytical results lead to a suspicion that they are adulterated by means of added water. The other two samples included in Tables 5 and 6 were obtained from Jersey cows. In each of these samples the content of butter fat and total solids is very high, the other values being correspondingly fairly normal. So far as the analytical results can indicate there is nothing abnormal in the composition of these milks, but the results of freezingpoint tests indicate substantial amounts of added water, in the one instance approximately 12 per cent and in the other over 6 per cent.

RECOMMENDATIONS.

Practical results obtained by many investigators show that the cryoscopic method is dependable as a means of determining, with a reasonable degree of accuracy, percentages of added water in milk. As a result of a careful study of this question during the past two years and after an extensive amount of experience with the application of the freezingpoint method to a large variety of samples of market milks, as well as milks of known origin and composition, it is recommended

(1) That this method, to be known as the cryoscopic method for the determination of added water in milk, be adopted as a tentative method by this association.

(2) That the cryoscopic method for the determination of added water in milk be subjected to further study with a view to its adoption as an official method.

If these recommendations are adopted in the manner suggested above, a plan of collaborative work can be arranged which will aim to serve the following purposes:

(1) To confirm the reliability of the cryoscopic method as a means of determining percentages of added water in milk.

(2) To develop and standardize the essential details in carrying out the determination.

(3) To arrive at definite conclusions regarding error factors and necessary corrections to be applied under various conditions.

(4) To perfect the details of a standard experimental outfit to be devised for the purpose of making freezing-point determinations on samples of milk, milk products and other food products to which the cryoscopic test may be applied.

REPORT ON MEAT AND MEAT PRODUCTS1.

By RALPH HOAGLAND (Bureau of Animal Industry, Washington, D. C.), Referee.

The official methods for the examination of meats and meat products which were adopted by the association in 19162 are very much in need of further revision in order to bring them up to date. For several reasons such a revision proved to be impracticable. Because of personal interest in two methods, however, the referee undertook a study of: (1) A method for the determination of sugar in meat; and (2) a comparative study of two methods for the determination of moisture in meat. A report of the work follows:

STUDY OF THE PHOSPHOTUNGSTIC ACID METHOD FOR THE
ESTIMATION OF DEXTROSE IN MEAT.

The following method, which has been used successfully by the referee for several years, was sent out to a number of chemists for cooperative work. Reports were received from four.

ESTIMATION OF TOTAL SUGAR.

REAGENTS.

Phosphotungstic acid.-Dissolve 100 grams in water, and make up the solution to

100 cc.

PREPARATION OF WATER EXTRACT.

Weigh 100 grams of the finely ground sample into a 600 cc. beaker, add 200 cc. of water, heat to boiling, and boil gently for 5 minutes. Stir the contents of the beaker frequently during this and subsequent extractions to prevent bumping. When several samples are extracted at the same time a mechanical stirring device is practically a necessity. Remove the beaker from the flame, allow the insoluble matter to settle, and decant the clear liquid on an asbestos mat in a 4-inch funnel. Filter with the aid of suction. Add 150 cc. of hot water to the residue in the beaker, boil gently for 5 minutes, let settle, and decant the clear liquid as before. Repeat the operation, and finally transfer the contents of the beaker to the funnel, wash with 150-200 cc. of hot water, and press the meat residue as dry as possible. Transfer the contents of the filter flask to an evaporating dish, and evaporate on a steam bath to a volume of about 25 cc., but not to dryness. Transfer the extract to a 100 cc. volumetric flask, taking care that the volume of liquid does not exceed 60 cc. Add 25-35 cc. of phosphotungstic acid, shake vigorously, let stand a few minutes for gas bubbles to rise to the surface, make to volume, shake, and either filter or centrifugalize. The use of a centrifuge is to be preferred since a larger volume of liquid is obtained. Test a portion of the filtrate with dry phosphotungstic acid for complete precipitation. If an appreciable precipitate forms, take an aliquot portion of the filtrate, add 5-10 cc. of phosphotungstic acid, make to volume, filter, and test the filtrate for complete precipitation. The filtrate should also show not more than a slight reaction for creatinin by Jaffe's test3.

1 Presented by W. C. Powick.

2 Assoc. Official Agr. Chemists, Methods, 1916, 271.

O. Hammarsten. Textbook of Physiological Chemistry. 1915, 696.