The large number of experiments with relatively pure carbohydrates permits a comparison of the values for the different kinds of carbohydrate. With dextrose it is seen that the cost of digestion ranges from 2 to 9 per cent, the average for 14 experiments with 11 subjects being 5 per cent. With levulose the total increments ranged from 3 to 9 per cent, and with one somewhat unreliable subject (J. J. C.) rose to 14 per cent. The average for 9 experiments with 8 subjects is thus approximately 7 per cent. With sucrose, the total increment varied from 3 to 9 per cent, with an average for 15 experiments with 9 subjects of 6 per cent. With lactose the total increment ranged from 3 to 6 per cent, with an average for 5 experiments with 5 subjects of 5 per cent.

If we compare the experiments on the basis of the amount of sugar ingested, we find that the average cost of digestion was practically the same for both 100 grams and 75 grams. The only exception was sucrose, which gave an average cost of digestion of 6 per cent for 8 experiments with 100 grams and of 5 per cent for 7 experiments with 75 grams, the average for all experiments being 6 per cent.

It is thus seen that the cost of digestion for the carbohydrates studied in these 43 respiration experiments does not differ materially in the proportion of increase, averaging not far from 6 per cent (5.5 per cent to be exact) of the fuel value of the intake. This figure, 5.5 per cent, is almost precisely the average obtained in 22 calorimeter experiments (5.6 per cent), although in those experiments mixed carbohydrates were taken, such as bananas, popcorn, and rice, rather than pure carbohydrates.

A relatively large number of respiration experiments were made with the protein-rich food, beefsteak, in which the fuel value1 ranged from 234 to 532 calories, or 790 calories if we include the experiment with beefsteak and potato chips. A strict averaging of these experiments is not permissible, owing to the differences in the time relations. Large increments are noted in several instances with several values for the cost of digestion of 16 per cent or more. The average cost of digestion for all of the beefsteak experiments is 13 per cent, more than twice the value observed with the carbohydrates. In the experiment with whole milk, a cost of digestion of 3 per cent was found, while the two experiments with a mixed diet gave values of 6 and 4 per cent, respectively, these being not far from the values obtained for the calorimeter experiments.

1It should be noted that the beefsteak used in these experiments contained a certain proportion of fat, which supplied from 24 to 37 per cent of the fuel value. Consequently, the cost of digestion of the protein itself is not represented by the figures given. In all probability the true value would be higher.

Emphasizing again the fact that in drawing conclusions from the results given in these two tables it should be remembered that the figures given are low rather than maximum values, since in the majority of instances the basal value was not reached before the conclusion of the experiment, we may conclude that the average cost of digestion for the ingestion of pure carbohydrates or a predominatingly carbohydrate meal will be not far from 6 per cent of the fuel value of the food ingested. With fat it is approximately 2 per cent and with a protein-rich diet it averages 12 per cent. With a mixed diet, which more properly corresponds to every-day usage, 6 per cent is doubtless near the correct value.