It is natural to ask next how the variation in the composition of milk affects a nursling injuriously. When its food contains insufficient protein, the infant's flesh becomes soft and flabby, its skin pale, its power of resisting disease is lessened, and its liability to acquire rickets is greater. All this may happen and the child continue to be plump. When its food contains too much protein, particles of milk-curd will be seen in the stools, the infant will have colic, the stools will contain mucus, become numerous, soft, and finally thin. The temperature of an infant often rises when enteritis is fully established as a result of indigestion. An illness thus provoked may become severe or chronic. During the first two or three months of life more than 1 to 2 per cent, of protein is not well tolerated by the average child. After this, protein can be taken in gradually increasing percentage. Some infants.will take comfortably 4 per cent, of albumin after the first month; some others of them will digest it even in their sixth month, although it is more than is found in human milk.

If the milk is not rich enough in fat, the infant will lose flesh and gradually become rickety. Too much fat has a tendency to provoke diarrhea and vomiting. More than 6.5 per cent. is rarely well borne. Carbohydrates, except the amount of sugar of milk in that food, are not needed during infancy, as there is little muscular exertion required and fat seems to supply the needed heat better. Carbohydrates are said to be able to protect tissues from waste, and therefore to lessen the need of protein tood. This is true only of adult tissues. Carbohydrates do not lessen the need of proteins for children during the period of most rapid growth, when they are required to make new cells. An ability to digest starch does not ordinarily develop until several months after birth. Occasionally, however, an infant is seen who apparently utilizes such food from the earliest weeks of its life. It is found, as a rule, that starch and most forms of carbohydrate are inclined to ferment, to become acid in the child's stomach and intestines, and consequently to produce colic and diarrhea. For these reasons carbohydrates are not indicated as a considerable element of an infant's diet. Milk-sugar is the natural and consequently the best form of it for use at this time. It is the least likely to ferment, and is easy of digestion and absorption. From 5 to 6 per cent, is needed during the first two months of life, and about 7 until the tenth month, when a lower percentage can be given, and often a little starch can be added. More than from 7 to 7.5 per cent, is not tolerated by the average infant.

It is evident that a knowledge of the composition of mother's milk often is essential in order to understand the causes of children's illnesses, and to determine what modification of diet is necessary to cure them. A clinical method of analyzing milk is required. An accurate determination of its contents can be made only by elaborate chemical analysis. Resort to this is sometimes necessary in order to determine the exact percentages of fat and protein in it. The proportion of milk-sugar and inorganic matter is almost invariable. By determining the percentage of fat and the specific gravity in a sample of milk it is possible to say what the relative amount of protein in it is. Holt directs that the milk for analysis be taken in the middle of nursing - in other words, that neither the 'fore-milk' nor 'strippings' be used. From fifteen to twenty cubic centimeters of milk are needed. Its specific gravity should be accurately determined by a small hydrometer. Ten cubic centimeters should then be placed in a finely graduated cylindric test-tube. The latter should be stoppered and kept for twenty-four hours at approximately 700 F. During this time the cream will form a well defined ring at the top. It is well known that cream bears to the total fat in milk a ratio of 3 to 5. It is, therefore, easy to determine the percentage of fat in the milk when we know the amount of cream in it and this ratio. The amount of fat can be much more accurately and quickly determined by the centrifuge. This method is used in dairies. The Babcock milk-tester is a cheap centrifuge devised for this purpose. The milk is placed in small flasks with long, graduated necks. Equal parts of milk and commercial sulphuric acid are put into the flasks. The acid transforms the proteins, casein, and fibrin into soluble acid albumin that offers little resistance to the rising of the fat globules. The filled flasks are whirled in the centrifuge with a velocity of from 700 to 800 revolutions a minute. The fat rises quickly in the neck of the flask, which is so graduated that the percentage of fat can be read at once.

The relative proportion of proteins is determined when the percentages of fat and specific gravity are known, because if the proteins remain unaltered and the percentage of fat is low, the specific gravity will be high; but if high, the specific gravity will be low. If the fat remains unaltered and the proteins are high, the specific gravity will be high, but if the percentage is low, the specific gravity will be low. Therefore:

1. If the percentage of fat is from 8 to 10 (high) and the specific gravity from 1033 to 1034 (high), it may be assumed that the percentage of proteins is high.

2. If the percentage of fat is from 2 to 4 (low) and the specific gravity from 1033 to 1034 (high), it may be assumed that the proteins are about normal.

3. If the percentage of fat is from 8 to 10 and the specific gravity low, the proteins may be assumed to be low.

4. If the percentage in fat is low and the specific gravity low, the percentage of proteins must be low.