Starting with the belief that in practice it is unnecessary that the percentage of fat should ever be more than twice that of the protein, Holt has found it easy to construct a table by which a series of formulas can be obtained in which the fat may be as high as twice the protein, and as low as only a little more than one-fourth the protein, with all the variations between these extreme limits. The first step is to secure milks which contain different fat percentages: 7 per cent, of fat, 6 per cent., 5, 4, 3, 2, and 1 per cent. How this is done from one quart bottle of milk which has 4 per cent, of fat, and one which has 5 per cent, of fat is shown in the following table:

1 ounce milk-sugar by weight in 20-ounce mixture adds 5 per cent. 1 ounce milk-sugar by volume in 20-ounce mixture adds about 3 per cent. 1 even tablespoonful in 20-ounce mixture adds 1.75 per cent.

Every ounce of 7 per cent, milk in a 20-ounce mixture will add 1/20 of 7 per cent., or .35 per cent, of fat; every ounce of 6 per cent, milk will add .30 fat. Every ounce of 5 per cent, milk will add .25 fat, etc. Every ounce of each of these milks will also carry with it .175 per cent, of protein and .225 of sugar. In the case of a young infant for whom .70 per cent, of protein is proper, 4 ounces of milk in 20 ounces of food will be required. If this 4 ounces is of 7 per cent, milk the food will have 1.40 per cent, of fat; if it is 6 per cent, milk, the food will have 1.20 fat; if 5 per cent. 1.00, etc. To raise both the fat and the protein proportionally the next increase would be to use 5 ounces of 7 per cent, milk in 20 ounces of food; later 6 ounces in 20; 7 ounces in 20, etc., etc.

To raise the protein above .70 but without increasing the fat, this can be done by using 1 ounce more of milk in 20, but of the series containing the next lower fat; e.g., 5 ounces of 6 per cent.; 6 ounces of 5 per cent.; 7 ounces of 4 per cent., and 9 ounces of 3 per cent., all give approximately the same fat while the protein is raised successively from .70 to .85, 1.05, 1.20 and 1.60 per cent. In calculating the sugar percentage it is necessary only to see how much must be added to that already in the milk to bring it up to the percentage desired, remembering that;

When more than 20 ounces of food is to be made up one should add for 25 ounces one-quarter more of each ingredient; for 30 ounces one-half more; for 40 ounces twice as much, etc. In order to calculate the caloric value of the food quickly it is only necessary to note the number of ounces of the milk used in the formula, multiply this by the caloric value of 1 ounce, and add the caloric value of the sugar or barley used in the formula.

According to Koenig's analysis, the proteins in cow's milk are casein, 2.88 per cent., and whey proteins, 0.53 per cent.; in human milk they are 0.59 and 1.25 per cent, respectively. It is evident from this how large a relative percentage of casein, which is the least easily digested of these proteins, there must be even in low percentage modifications of cow's milk. To make modified milk resemble human milk still more closely, it has been proposed to use the richest cream obtainable by the centrifuge, and to dilute this with whey instead of milk. Thus it is entirely possible to make a modified milk of nearly the same chemical character as human milk. Although 3 2 per cent, cream is needed with which to make the most nearly perfect imitations of human milk, yet with whey instead of milk used as a diluent, even with creams of lower percentage, a mixture can be made that is more digestible than the ordinary forms of modified milk. As yet formulas have not been devised whereby this can readily be done at home. The milk laboratories, however, furnish the combinations upon prescription. Hammar-sten says that cream 200 parts and whey 800 parts, or cream 100 parts, milk 100 parts, and whey 800 parts, will make a mixture that contains the normal proportion of proteins to human milk. Ashby makes a 'humanized milk' of 10 ounces of fresh milk, 20 ounces of whey, and half an ounce of milk-sugar. The composition of this mixture is fat 2.5 per cent., sugar 6 per cent, and proteins 1.75 per cent. This is a low percentage of fat. If 16 per cent, cream, or gravity cream with the approximate composition stated in our table, is diluted one-fourth by the addition of three parts of whey, the following composition will result:

This is a mixture approximating human milk in chemical composition, except that it requires the addition of milk-sugar (4.52 parts) sufficient to make six parts in 100.

By remembering the percentage composition of cream, and knowing that whey contains on the average 0.86 per cent, protein, 0.32 per cent, fat, 0.49 per cent, sugar, 0.65 salts, and 93.38 water (Koenig's analysis1), one can calculate the amount of cream and whey needed to give definite percentages of fat and protein to a mixture.

Whey is best prepared by coagulating with liquid rennet a milk from which the cream has been removed as perfectly as possible. The milk and rennet should be kept at approximately 300 C. (86° F.) until the curd is formed. The latter must be filtered off through clean muslin or absorbent cotton. The filtrate must be quickly heated to 65° C. (say 920 F.) to destroy the rennet. If raised to 700 C. or over (say 100° F.) the whey proteins will also be precipitated.

The experiments of White and Ladd demonstrated that cream and whey mixtures contain as perfect an emulsion of fat as is found in natural milk.

As yet percentage modifications made with cream and whey have not been sufficiently used to make it possible to say what will be the practical results of feeding children upon them.

Freeman says that cream contains 300 times as many bacteria as the lower strata of a bottle of milk and Hess has shown the upper two ounces of cream in a bottle of milk contains the vast majority of streptococci, tubercle bacilli and other bacteria. The abundance of microbes in cream is one of the dangers incident to its use in infant feeding but a danger not generally avoidable.

Many babies need food containing a low percentage of fat as others need a low percentage of protein. Too much fat produces so-called "fat-diarrhea." Milk must often in these cases be stopped altogether for a few days, albumin water or barley water being substituted. Later skim milk or buttermilk must be given. A persistent intolerance of food with a high or even normal percentage of fat is a peculiarity of some infants. Without doubt the high percentage of micro-organisms in milk rich in fat as well as sometimes the inability of individual infants to digest fat well is the cause of the illnesses which such milk provokes in these babies. The same illnesses are moreover found in breast-fed babies and due to the same causes. The necessity of adapting milk to the individual needs of infants as regards both protein and fat must be kept in mind. Moreover, when milk mixtures rich in water are fed it must not be forgotten that babies also require the usual mineral ingredients of mother's milk and these must often be added to milk mixtures.

1 Agricultural Department analysis makes average proteins 1 per cent.