the floor of the room.

In this way any difficulty arising from differences in temperature at different points in the room has been avoided.

METHOD OF EXPERIMENTS.

The experiments thus far made have been of forty-eight hours' duration, this period being divided into subperiods of twelve hours each. The animal is placed in the apparatus five or six hours before the beginning of the experiment, which has been conveniently placed at 6 p. m. By this time the apparatus has come into equilibrium, and it is only necessary to shift the current of air from one set of cans and absorption apparatus to another in order to begin the experiment.

The experiments have followed each other at an average interval of from two to three weeks. During the intervening time the animal stands in an adjoining room in a stall which is provided with appliances for the quantitative collection of the visible excreta. An actual experiment requires the services of at least seven men, exclusive of the assistant in charge of the feeding and collection of excreta.

INTERPRETATION OF RESULTS.

A feeding experiment conducted with the aid of the respiration calorimeter is not fundamentally different from one made according to simpler and more familiar methods. In both cases we attempt to compare the results obtained, either from two or more rations under identical conditions or from identical rations under differing but controlled conditions. The difference lies in the extent to which we are able to control the conditions and in the accuracy and minuteness with which it is possible to compare the rations and their results.

The simplest and most obvious form of feeding experiment is that in which the amounts of feed consumed are noted and their effects measured by the increase in the live or dressed weight of the animal or by the weight of milk or wool produced. This method, when skillfully carried out with a considerable number of animals and under the conditions of actual practice, is particularly adapted, and, indeed, may be said to be indispensable, to the study of the economic aspects of stock feeding.

But while this is true, it is also the fact that no considerable or profound knowledge of the principles of feeding can be gained by means of experiments of this class. The factors entering into the problem are too complex. Chemistry has shown that each one of the feeding stuffs consumed consists of a great variety of substances useful, indifferent, and even injurious-mingled in the most diverse and varying proportions, while physiological investigation has demonstrated not only the considerable and irregular fluctuation of live

9043-07- -18

weight from day to day, but especially that a given increase or decrease may be of very varying significance according as it consists of proteid tissue, fat, mineral matter, or simply water. The result of a live-weight experiment, therefore, may be the resultant of any one of many possible combinations of these factors, and no safe conclusion as to its actual cause is usually possible. The history of this class of experiments amply corroborates this conclusion. Great accumulations of experimental data have been made, but relatively few general conclusions have issued from them.

The earliest step in advance was the attempt to separate the factor "food" into its elements. Of these attempts, the one which has secured general acceptance is the familiar one of Henneberg, which groups the chemical ingredients of feeding stuffs into "protein," "carbohydrates," "fat," and "ash," subdividing the carbohydrates into "crude fiber" and "nitrogen-free extract," and distinguishing further between the digestible and the indigestible portions of each group. A great mass of investigation along these lines in the laboratory and digestion stall has materially enlarged our knowledge of feeding stuffs, although much still remains to be done. It is now a comparatively easy matter, by the familiar methods of the digestion experiment, to determine with a fair degree of accuracy the so-called 66 digestible nutrients" consumed in the several periods of a feeding experiment and thus to secure a more rational basis of comparison. To the conventional determinations it is of course easy to add others, such as that of amids, pentosans, etc., and particularly the heat of combustion.

Analyses and digestion experiments as ordinarily conducted, however, afford no direct information whatever as to the effect of the digested matters in supporting the animal or producing gain. It is not even necessary to weigh the animal in a digestion experiment. The conclusion as to the nutritive value of the feeding stuff is simply an inference based on general physiological facts, and its correctness is more than questionable in the light of recent investigation. It is only as we determine the actual changes brought about by a ration in the store of matter or of potential energy contained in the body that we can reach a scientifically accurate determination of the nutritive value of that ration. Unless we do this, no matter how accurately we analyze the feeding stuffs supplied or determine their energy, the second member of the equation is lacking. We stand in urgent need of actual determination by modern methods of the nutritive values of feeding stuffs for different purposes, the results of which we may substitute for the assumptions on which we are now basing our teachings, and it is such determinations for which the respiration calorimeter is designed.

The basis of the method is Henneberg's conception of the schematic body. This is, in brief, that for this particular purpose the animal body may be regarded as composed of water, ash, protein, and fat, each of practically invariable elementary composition, and the effect of a ration is expressed by the gain or loss of ash, protein, fat, and, of course, water, by the body of the animal. This gain or loss may be determined by comparing the amounts of ash, nitrogen, and carbon in the food with those contained in the various excreta-solid, liquid, and gaseous; that is, by a so-called balance experiment. For example in an experiment with a steer the ration consisted of 4,531 grams of timothy hay and 400 grams of linseed meal, and the following figures for daily nitrogen and carbon were obtained:

From the known composition of the proteids and fat of the body it is easy to compute that the loss of 8.2 grams of nitrogen and 225.1 grams of carbon is equivalent to a loss of 49.2 grams of proteids and 259 grams of fat.

Still another method of comparison is afforded when we turn from considering the food as a supply of matter and regard it as the source of energy to the vital machinery.

The potential energy of feed and visible excreta is measured by their heats of combustion, which are readily determined by means of the bomb calorimeter. The production of heat by the animal is determined directly by the respiration calorimeter. Adding to these data the heat of combustion of the methane excreted, which is readily computed from its amount, we have all the data for the construction. of a balance of energy similar to the balance of matter. In the case selected as an example this was: