A study of the dietary of the institution over a few weeks revealed the fact that it consisted essentially of a soup made of a fairly wide variety of foodstuffs, together with all the bread which the children cared to eat. Cereal was frequently served for breakfast. In Table I the food supply over an interval of one week is given in detail. This is fairly typical of the period covered by this study. All the components of the diet except the bread and some cereal and fruit were incorporated into the soup. The soup was palatable and attractive and was believed to possess a high food value. Actually from the standard of calories and proteins it left little to be desired. During the demonstration 4 to 5 per cent of the total calories of the diet was supplied in the form of meat, which was included in the soup, and went far toward adding to its attractiveness.

Here we had a dietary for children during the period when growth should be proceeding at a rapid rate which consisted essentially of cereals, tubers, fleshy roots, and muscle meats, and a considerable portion of the total food supply was furnished in the form of a refined wheat flour in the form of white bread. Since the diet of bread and soup such as has been described contains no fresh raw food, the question at once arose why scurvy had not appeared among the children. None showed any scorbutic tendencies, and the attending physician assured us that this had not been a problem in the institu tion. The cause of this immunity to scurvy was soon discovered. Certain fruit vendors in near-by markets regularly donated lemons which were still sound but were likely to spoil before the next market day. It had been the custom to slice these lemons and give each child a slice. This one raw fruit was highly appreciated by the children and served as their source of the vitamin C.

The simplest procedure for improving the quality of this diet without involving any additional cookery appeared to be the addition of milk. This plan was adopted. Accordingly one group of 42 children received 1 quart of whole milk each per day during the demonstration period. This was prepared from Merrell-Soule's whole milk powder (Klim). It was prepared by dissolving the proper amounts of whole milk powder in cold water and agitating in a mechanical mixer. Only in very cold weather was the water warmed to some extent. The other group, which we designated as the check group, was continued on the institutional diet. We were thus in a position to compare the children on the two dietaries, since every effort was made in dividing the 84 children into two groups of 42 each to pair them so that any child in one group was quite comparable in age, size, and condition to a child in the other group.

It is to be expected in any demonstration of this kind, where tubercular children, and children who had suffered from faulty skeletal development to a pronounced degree, were included, that certain individuals would be less capable of responding with growth to an improved dietary than would others. The accompanying charts give an accurate idea as to the general results of the comparison.

One of the striking features of the records of all children of both groups is the sharp increase in body weight, and at a more rapid rate than normal, in the first four to six weeks they were under observation. This happened to both the milk-fed and check groups on the

institutional diet. We can not determine with certainty the cause of this sharp increase in weight in both groups, but the most plausible explanation would seem to be the fact that our critical observation of the institution led to a certain increase in the total quantity of food provided to all the children of the institution just at this time. There is strong reason to believe that before our advent the children were being fed not only a diet which was deficient in quality but likewise deficient in amount, so that their condition represented not only the results of malnutrition from a diet of faulty quality but the complication of partial inanition as well.

In the case of the check group, where the quality of the food remained essentially constant before and during the demonstration, the sharp increase in weight of 2 to 5 pounds was not continued. The weight curves of the children of this group suddenly flattened, and in practically all cases they remained nearly or quite stationary. In several there was a gradual loss in weight during the first 12 to 15 months of the demonstration. There was only one child in the check group which made any appreciable gain during the first 15 months of the demonstration.

The milk-fed group showed marked contrast with the check group in several respects. In the first place, the sharp rise in body weight which was observed in both groups during the first few weeks was continued in most children throughout the demonstration, which for a considerable number of children covered a period of 21 months. One child of 5 years, who weighed but 28 pounds in the beginning, increased its body weight by 70 per cent at the end of eight months on the milk, 90 per cent at the end of the first year, and reached a weight which was somewhat over the normal curve as given by the Holt, Boas, and Burke table. Several of the children increased 50 per cent or more in weight during the first year.

There was likewise a marked change in the behavior of the milkfed group as contrasted with the check group. The latter were apathetic and very tractable. The discipline of the institution was strict and these children were all quite obedient. Those in the milkfed group, on the other hand, soon caused annoyance to their teachers by their restlessness and desire for activity. They also were frequently guilty of infractions of the rules.

The results of the supplementing of the institutional diet with milk were so striking that after 15 months it was decided to supplement the diet of the check group with a quart of reconstituted milk per day, in order to demonstrate whether these children, at least a considerable number, possessed the capacity to grow which they were not manifesting on the institutional diet. As was anticipated, during the next six months a considerable number of the check group grew at rates comparable to what had been observed in the original milk group.

All things considered, the results of this demonstration constituted, we believe, a most satisfactory demonstration of the validity of the view that a dietary selected from cereals, tubers, fleshy roots, and meat does not prove satisfactory for the physical development of young children. It shows further that milk is as effective a supple

mentary food for such a type of diet as has been repeatedly shown to be the case with experimental animals.

In making up these tables I selected three periods (two of two weeks each and one of one week) and gave the diet for the milk group and the diet of one of the check groups only. In the last set of tables I selected a period when supplementary feeding (sirup, sugar, and bread) was being given to the check group.