In constructing diets we have first to determine what amount of food is required for each individual in order that the body may be supplied with sufficient energy to keep it in health and vigour; and secondly, to consider what the composition of that food should be.

Under this heading we shall refer to the total quantity of food, expressed in terms of its fuel value and independent of its constituents. The amount of food necessary to keep any person in a natural state of health and activity depends upon a number of circumstances. No definite rule can be laid down, for the energy required will of course depend upon how much is being used in keeping the body warm and in doing work.

If more heat is lost and more work is being done, more food must be supplied, or the body would be obliged to use its own tissues and the weight would fall. The principle of the conservation of energy applies to the human being as it does to the rest of the natural world, animate and inanimate. We have no experimental evidence of exceptions to this rule, know of no subtle method by which one individual can, over any long period of time, do double the work of another upon half the food: although to the unconsidering eye this may sometimes seem to be the case, we shall see that many factors besides the actual external work, have to be taken into account in estimating the energy used by any person; factors of build, of extent of surface, of restlessness or placidity when not actually working, of the skilled action of trained muscles or the energy wasting contractions of unaccustomed movements.

In the first place the food must be proportional to the weight of the body. Other things being equal big people will require more than those of smaller build. In comparing different individuals the weight must, therefore, be taken into account and this is done by speaking of the fuel value required for each unit of weight, expressed as calories per kilogramme. The average fuel value required by adults is about 40 calories per kilogramme, but it is possible to support existence upon much less than this; on the other hand, under special circumstances, much more is required; for instance, when weight is being laid on, or a great amount of work is being done, as much as 110 to 120 calories per kilogramme has been known to be taken.

Since the body cannot produce energy which is not derived from the oxidation of the food or of its own substance, it is clear that there is an irreducible minimum of food which is required if the body tissue is not to be used up. This may be regarded as about 35 calories per kilogramme for a person living a sedentary life; 4 1/2 pints of milk would furnish this amount of energy (1750 calories) for a person of 8 stone, or 1 lb. 2 oz. of bread and 8 oz. of meat. For an individual resting entirely in bed it may be as low as 25 calories per kilogramme, which would be given by 3 pints 2 oz. of milk, or by 5 1/2 oz. of meat and 13 oz. of bread. The extent of surface which the body presents has a considerable influence upon the quantity of food required. Thin and angular people have a very much greater surface relatively to their weight than those of rounded contour, and small people have a greater surface, relatively to their weight, than large. One of the main sources of loss of energy from the body is the radiation of heat from the skin, and this varies directly as the area of surface : a whale, for instance, is materially aided in keeping its temperature constant in icy waters by the fact that its surface is small compared to its bulk. Small mammals, having a relatively great surface, have a more active metabolism and require very much more food per kilogramme than large ones. As an illustration the following figures may be quoted, showing the amounts of urea excreted by a man, a dog, and a rabbit during fasting, when all the energy used is being supplied from the fat and protein of the body.

A fasting man weighing 70 kilos excreted .6 g. urea per kilogramme, dog " 30 " " 1.6g. " "

" rabbit " 1 " " 3.6 g. " " "

The amount of energy derived from the protein is six times as great for each unit of weight in the rabbit as in the man. A better way of estimating the metabolism is to measure the oxidation going on in the body as shown by the carbon dioxide excreted from the lungs. In dogs it has been shown that this is more proportional to the surface than to the body weight. Richet found that dogs varying in weight from 2.2 to 28 kilogrammes excreted the same amount of carbon dioxide per hour for each square centimetre of surface. Now the heat lost from the skin is automatically supplied by the generation of heat in the muscles, and the temperature thus kept constant. In anaesthesia this regulation fails and the temperature falls, since the loss of heat from the surface is no longer co-ordinated with its production in the muscles. Richet accordingly found that in dogs under the influence of chloral the excretion of carbon dioxide was no longer proportional to the surface of the animal but simply to its weight.

These considerations are of special interest in connexion with the feeding of premature and thin infants, who will require more food than plump ones, since a small thin child has a very large surface relatively to its weight. As it grows, and especially as it lays on fat and becomes rounded, the weight increases more rapidly than the surface. Accurately expressed the mass increases as the cube, the surface as the square. Various researches could be quoted to illustrate this point. The amount of milk taken at the breast can be ascertained by weighing the child before and after it is fed. This has been done over long periods, and at the same time analyses have been made to determine the heat value of the mother's milk. In this way it has been found that an ordinary new-born child, excluding the first week when it is losing weight, requires about 100 calories per kilogramme, whilst when the child is six months old it uses 80 or less, the lessened requirement being largely due to the diminution of surface relatively to bulk. The actual amount of milk taken by the older child is, of course, greater; we are here referring to the amount taken for each unit of weight. We may assume, therefore, that a child weighing 8-9 lb. should receive, after the first week or two, a minimum of 100 calories per kilogramme. In English measure this is equal to 2 1/4 oz. of cow's milk to each pound, or a pint for a child of 9 lb. For a child of six months the minimum of 80 calories per kilogramme is equivalent to 1 3/4 oz. of milk for each pound. Artificially fed children have been found to require in some instances more than this heat value if they are to grow at the usual rate. To illustrate the greater quantities required by thin children, Heubner's case may be quoted of a premature child receiving 110-125 calories per kilogramme which failed to increase in weight until it obtained 150 calories per kilogramme. This is equivalent to 68 calories, or about 3 1/3 oz. of cow's milk, to the pound. In an infant whose weight is stationary, the protein tissues continue to grow but the child loses fat. Although the chief reason why thin infants require more food per pound is the loss of heat from their greater surface, there is another factor, namely, the rate of growth; this is greater in the early months of life, but it will only account for a small proportion of the great metabolic activity at this time. If this be so, and the loss of heat from the surface be so great a drain upon the energy of the body in small animals, we should expect that a saving to the metabolism would be effected if the temperature of the external air were raised. This has been found to be the case. If a child be kept in an incubator the loss of heat from the surface is lessened, or entirely prevented, and it has often been possible to keep premature children alive in this way. The premature child, with its small bulk and relatively great surface, cannot deal with enough food to make up the loss of heat from its skin; but it can often assimilate enough for nutrition and growth if that loss be prevented.