The fact that different amounts of the several foodstuffs produce varying increases in metabolism would make it appear that the establishment of definite mathematical relationships between the amounts of food ingested, the character of the food ingested, and the increments would be relatively simple. This, however, on close analysis, proves to be far from the case.

To establish a quantitative relationship between the various foods ingested and the increase in the basal metabolism it is necessary to note first the length of the experimental period to be considered. If the total increment due to the ingestion of food is desired, the period of measurement must be extended until the increment due to the ingestion of food has disappeared and the metabolism has again reached the basal level. For instance, if the basal heat production is 70 calories per hour and the metabolism increases after the ingestion of food to 100 calories per hour for one or two hours, obviously no complete mathematical relationship can be established unless the measurements are continued until the basal value of 70 calories per hour is again obtained. This is somewhat difficult, especially when the stimulus effect is prolonged, as it is with protein.

The first quantitative relationship to be considered is the increment in the metabolism above the basal level, to find how far it is possible to increase the basal metabolism by the ingestion of nutrients. This is in reality a measurement of the absolute maximum increment due to the ingestion of food, somewhat similar to the "peak" effect in the load of a power plant, and has considerable interest as an index of the possible maximum influence of food. With carbohydrates the basal metabolism may be increased to an average maximum of approximately 25 per cent by the ingestion of 100 grams of any one of several sugars. This increment occurs inside of two hours and the metabolism has a tendency to return to the base-line somewhat rapidly thereafter. For a detailed discussion of the differences in effect of various forms of sugars, the section on carbohydrates should be consulted. (See page 171.) Of special significance is the fact that the increment with levulose over that with either sucrose or dextrose, which was earlier reported from this laboratory,1 is not noted when the results of all the experiments are combined, and we find that there is a greater similarity between levulose and dextrose than was at first believed. It will be seen, from table 173, that the levulose has a more pronounced effect upon carbon-dioxide production than the dextrose has, although not so great as that of sucrose. When the comparison is made on the basis of the heat production, it is found that the difference between dextrose and levulose in large part disappears, although sucrose still shows a higher value.

Most of the pure sugars were studied with both 75-gram and 100-gram portions. While no uniform variation was noted in the effect upon the metabolism, it was usually found that the increment with the larger amount was greater than with the smaller amount, although the differences were by no means proportional.

With protein the large increments in metabolism found by all workers were also noted in this research. The heat production increased usually to a maximum above the basal level of approximately 25 per cent, with a possible maximum of 45 per cent. (See tables 215 and 230.) The increment persisted for a long time, often from 8 to 12 hours. Indeed, our experiments were defective in that the experimental periods were in general not sufficiently extended to obtain the entire effect due to the protein ingestion. This prolonged increment is in striking contrast to the increments obtained with carbohydrates, which, while fairly high (25 per cent), nevertheless fell rapidly to a basal value after a relatively few hours.

The experiments with fat are of special interest, though unfortunately the most liable to criticism on the grounds of experimental error and faulty technique. The inherent difficulties in feeding American subjects large quantities of pure oil or pure fat made it impossible for us to carry out any experiments with pure oil, as did Gigon, and we were obliged to confine ourselves to experiments with cream and with butter and potato chips. In practically all of the combinations used, a certain amount of other nutrients was inevitably included, which somewhat complicated the deductions drawn from the experiments. The evidence is, however, sufficiently extensive and the general trend is such as to justify the conclusion that the ingestion of a diet containing a preponderance of fat has a distinct effect upon the metabolism, although this was much less than that found with either protein or carbohydrate.

1Benedict, Trana. 15th Int. Cong. Hygiene and Demography, Washington, 1913, 2 (2), p. 394.

The findings are so irregular that it is probably unjustifiable to use the single highest maximum value found, and we must therefore resort to the average figure, which is not far from a maximum of 12 per cent.

In the experiments with mixed diets, in which frequently the subject took a sufficient number of calories in one meal to provide maintenance for a man at severe muscular work the entire day, i. e., approximately 4,000 calories, the maximum increment reached 40 per cent or more.

An absolute accuracy of ±3 per cent in the measurement of the basal metabolism on any given day is hardly obtainable. Furthermore, variations in the body position have an effect upon the basal metabolism. It has been shown that the basal value obtained with the subject lying quietly without food in the stomach may be increased slightly by having the subject sit upright in a chair.1 On the other hand, Du Bois2 finds that with the subject sitting properly supported in a steamer chair, the basal metabolism is not increased over that found with the subject in the lying position and in certain instances it was even decreased. Standing upright has been found to increase the metabolism approximately 10 per cent.3

The maximum effects obtained with carbohydrates, protein, and mixed diets are, however, very considerably greater than those due to ordinary changes of position. While one might state tentatively that the increments with food, at least at the height of digestion, are such as would be expected when man is performing light muscular work, the inadequate definition of the term "light muscular work" is such as to make this of little significance. Too little knowledge is available at present regarding the increments in metabolism accompanying simple, every-day customs; hence we find ourselves at a loss to compare the increments for these food materials with those accompanying minor muscular activity.

The increments with food are certainly much less than those obtained as a result of walking. On the other hand, since there is an increase in both the respiration and the circulation, it is evident that the ingestion of food and its effect upon metabolism are to be considered much more broadly than as an increment in the gaseous metabolism. There is unquestionably a stimulation in the muscular activity and general muscular tonus, accompanied by a sense of increased vigor, which is wholly out of proportion to the relatively small increase in the metabolism. The increments are, however, sufficiently large to preclude any attempt to measure the basal metabolism during the active stages of digestion. Experimenters have therefore for years wisely insisted upon the post-absorptive condition, that is, 12 hours after the last meal. This is particularly necessary when the preceding diet has contained liberal quantities of protein or an excessive amount of a mixed diet. Our observations, in common with those of many others, show very clearly that the effect of the ingestion of pure carbohydrate or of fat is with normal individuals concluded in a relatively few hours; were it not for the protein in the diet, therefore, one might state that the post-absorptive condition, or the so-called nuchtern condition, could be obtained in a much shorter period than 12 hours. Insistence on the 12-hour period is, in all events, the wiser course. Even with this interval, the injunction should be given to all subjects that excessive protein should not be taken in the last meal prior to the experimental period. (See page 286).

1Emmes and Riche, Am. Journ. Physiol., 1911, 27, p. 406. 2Soderstrom, Meyer, and E. F. Du Bois, Arch. Intern. Med., 1916, 17, p. 872. 3Benedict and Murschhauser, Carnegie Inst. Wash. Pub. No. 231, 1915. Benedict and Carpenter, Carnegie Inst. Wash. Pub. No. 126, 1910.