That the writer does not regard the usual distribution of expenditure for food in American families as being either inevitable or ideal may be indicated by the fact that in his own household, consisting of three adults and four growing children, the distribution of money expended for food is about as follows:

*Sherman, Food Products, pages 74-81, 108-111, 139-141, 212-216, 288-295, 346-351, 357, 388-393, 440-444.

Just what prominence should be given to each type of food in the provisioning of a given family or community is a problem calling for consideration of many factors. One important feature of the problem is to ascertain how the normal distribution of expenditure among the various types of food materials affects the relative proportions of nutrients in the resulting mixed diet. The accompanying table permits a comparison between the expenditures for the different types of food and the returns from each in terms of energy, protein, calcium, phosphorus, and iron in the case of the series of 92 family dietaries described on page 389. In individual dietaries the returns will naturally vary according as an economical or an expensive food of its kind is chosen, but in the average of 92 different dietaries, each of a week's duration, the danger of error due to such individual variations is minimized.

If we compare the cost of each type of food with the energy and individual nutrients which it furnishes, we find that because of the differing prominence of the several factors of food value in the various types of food it is often difficult to decide which expenditures were more economical. Thus in the averages just given meat and fish cost one third of the total expenditure for food and furnished about one third of the protein, phosphorus, and iron but only one sixth of the energy and only about one thirtieth of the calcium. Eggs furnished protein, phosphorus, and iron about in proportion to their cost, but less calcium and much less than a proportionate amount of energy. Milk furnished calories and protein about in proportion to cost, twice as much phosphorus, and five times as much calcium in proportion, but only half as much iron.

* Cream, in those cases in which it was purchased, is here included with milk. The amount of cream was small, if any.

By adopting the principle of a score card and assigning weights to the different factors of food value, it becomes feasible to compute a "composite valuation" or "score" for each food or group of foods which may then be compared with its cost. Since the most frequent deficiency in American dietaries is inadequacy of total food or energy value and most dietaries actually observed are of such composition as would furnish enough of each essential element if the total amount of food eaten were sufficient to provide a liberal energy supply, it seems reasonable to assign to the energy value of a diet a weight of about half of its composite valuation. It also seems reasonable to assign the remaining "points" equally to protein, calcium, phosphorus, and iron.*

If then we give to energy a weight of 60 on a scale of 100 and to protein, calcium, phosphorus, and iron each a weight of 10, or to energy 40 and to protein, calcium, phosphorus, and iron each 15, we obtain from the data of the table above the "score values" or "composite valuations" under the designations "I" and "II" respectively in the table which follows:

* In reality this amounts to giving a higher valuation to the protein since this is counted both as protein and as a part of the energy supply as well.

By comparing the composite valuation with the cost it will be seen that if either of these methods of estimating comparative values is at all valid, the money spent in these 92 families for milk and cheese, grain products, and vegetables brought a better relative return in food value and was therefore in this sense better invested than the money spent for meats and fish, eggs, and fruit.

In making any such comparison it must be kept prominently in mind: (1) that the weights assigned to the different factors of food value must necessarily be more or less arbitrarily chosen so that the resulting "composite valuations" or "food values" rest partly on facts and partly on assumptions; (2) that not all the important factors of food value are taken into account in these valuations, "vitamine values" for instance being wholly omitted from the calculation because as yet we have not the data necessary to permit us to give them numerical expression. It is quite possible that when it becomes feasible to state the vitamine values in numerical terms and give them due weight in the composite valuation, the expenditures for eggs and butter may appear more economical than is indicated by the above table. Any comparisons based on the use of such arbitrary weights or valuations as can at present be assigned must therefore be used with much discretion if misconceptions are to be avoided; but if so used they may be found serviceable in guiding the economical choice of food and to some extent in teaching relative food values.