True economy in the use of food must be physiological as well as pecuniary economy. The diet must supply amply all the requirements of nutrition (not merely the appetite nor the need for energy and protein) and this must be accomplished without the expenditure of too large a proportion of the income. The majority of families in the United States have had in recent normal times incomes of less than $800 per year, of which not over 45 per cent can be spent for food if other living conditions are to be at all satisfactory. This implies an allowance of approximately one dollar per day for food for the "normal" family of five,* or 20 cents per capita per day.

If this be taken as approximating the average expenditure in normal years,† it would follow that the sum annually spent for food in the United States is in the neighborhood of $7,000,000,000. From such statistical estimates of the value of the different food industries as the writer has been able to find it would appear that this is distributed somewhat as follows:

* If the family of five be reckoned as equivalent in food requirements to 3.7 men, the amount here suggested as available for food would correspond to 27 cents "per man per day" or "per unit."

† No attempt is made in this chapter to quote the fluctuations of prices under war conditions. The economic relationships here discussed will be found to be but little disturbed by a general raising or lowering of the level of prices.

‡ This estimate doubtless includes considerable quantities of nuts not used as such for human food but pressed for oil and the residue fed to farm animals.

§ Including beverages, condiments, and minor unclassified food materials.

Any such estimates as these can be no more than rough approximations since they depend upon data which are by no means complete and accurate for the year in which gathered and are subject to fluctuation from year to year. It also appears impossible to avoid arbitrary assumptions regarding the relations of wholesale and retail values. They are intended, therefore, only to indicate in the most general way the relative prominence of expenditure for the different types of food materials as judged from the statistics of the food industries.

Another statistical estimate may be obtained from the data published by the U. S. Bureau of Labor Statistics, who report that of the total value of food consumed in 2567 workingmen's families the distribution of expenditure was as follows:

These averages are based upon data which were apparently obtained, for the most part at least, by simply asking questions of the housewife regarding the kinds, amounts, and costs of her food purchases and relying upon her memory for the facts. The probable errors in data for individual families would thus be large, but the great number of families included in the inquiry would tend to minimize the errors in the final average.

* Low partly because of purchase of flour rather than bread, partly because oat meal, etc., were often not reported under this head but under "other foods."

A different kind of data bearing on this same problem is found in the dietary studies made under the auspices of the United States Department of Agriculture or of the New York Association for Improving the Condition of the Poor. These dietary studies are accurate records of the kinds and amounts of foods consumed by given groups of people during a period of a week or more. From such studies, chiefly of family groups, 208 have been taken as presumably representative of American food habits generally, and the cost of these dietaries has been studied with reference to the distribution of expenditure under headings corresponding to those used in the case of the above statistical estimates with the following results:

Of the dietaries included in the above average, 92 constituted a series observed during 1914-1915 in connection with the food investigations of the New York Association for Improving the Condition of the Poor. These studies were not entirely confined to New York City nor to families of low incomes. The cost of food per man per day ranged from 12 to 76, averaging 34 cents. The median cost was 31.5 cents per man per day. In one fourth of the families the cost was below 25 cents; in one fourth it was above 40 cents; in one half it was between 25 and 40 cents per man per day.

* Tea, coffee, and other food adjuncts were usually but not always reported under this heading. The reported average is therefore somewhat below the truth.

The average distribution of expenditure in these 92 families was as follows:

When these 92 studies were grouped according to the amount spent per man per day for food, it was apparent that as the scale of expenditure became more liberal a larger proportion of the money was spent for butter and fruit and a smaller proportion for breadstuffs. The distribution of expenditure among other types of food was, however, very similar in the dietaries of low, medium, and high cost.

Each of the three kinds of evidence used in arriving at the above estimates of distribution of expenditure for food may readily be criticized as inaccurate or inconclusive or both. Yet the trend of the data derived from the different kinds of evidence is so consistent that it can hardly be devoid of significance. It can scarcely be doubted that of the money devoted to the purchase of food the average American family spends from 30 to 40 per cent for meats and fish (including poultry and shellfish when used), about 5 or 6 per cent for eggs, about 7 to 10 per cent for milk, from 7 to 12 per cent for butter and other fats, from 10 to 20 per cent for bread and other cereal and bakery products, 3 to 7 per cent for sugar and other sweets, 7 to 10 per cent for vegetables, 2 to 8 per cent for fruit, and less than 2 per cent for cheese and nuts. At the same time it is plain that such a food budget, however prevalent, need not be regarded as fixed. Many people occasionally, and some habitually, put the last and smallest of the items just mentioned in the place of the first and largest by using cheese or nuts as so-called "meat substitute," more properly as an alternative to meat, - a custom which on the whole appears to be growing. The place of each type of food in the diet has been discussed in a general way elsewhere* and space does not permit us to go over the same ground here.