[Illustration: FIG. 7.--Horizontal cross-section of chair calorimeter, showing cross-section of copper wall at A, zinc wall at B, hair-felt at E, and asbestos outer wall at F; also cross-section of all upright channels in the steel construction. At the right is the location of the ingoing and outgoing water and the thermometers. At C is shown the food aperture, and D is a gasket separating the two parts. The ingoing and outcoming air-pipes are shown at the right inside the copper wall. The telephone is shown at the left, and in the center of the drawing is the chair with its foot-rest, G. In dotted line is shown the opening where the man enters.] [Illustration: FIG. 8.--Vertical cross-section of chair calorimeter, showing part of rear of calorimeter and structural-steel frame. N, cross-section of bottom horizontal channel supporting asbestos floor J; H, H, upright channels (at the right is a side upright channel and to the left of this is an upright rear channel); M horizontal 8-inch channel supporting calorimeter; Zn, zinc wall; Cu, copper wall; J, insulating asbestos.] The respiration chamber used in Middletown, Connecticut, was designed to permit of the greatest latitude in the nature of the experiments to be made with it. As a result, it was found at the end of a number of years of experimenting that this particular size of chamber was somewhat too small for the most satisfactory experiments during muscular work and, on the other hand, somewhat too large for the best results during so-called rest experiments. In the earlier experiments, where no attempt was made to determine the consumption of oxygen, these disadvantages were not so apparent, as carbon dioxide could be determined with very great accuracy; but with the attempts to measure the oxygen it was found that the large volume of residual air inside the chamber, amounting to some 4,500 liters, made possible very considerable errors in this determination, for, obviously, the subject could draw upon the oxygen residual in the air of the chamber, nearly 1,000 liters, as well as upon the oxygen furnished from outside sources. The result was that a very careful analysis of the residual air must be made frequently in order to insure that the increase or decrease in the amount of oxygen residual in the air of the chamber was known accurately at the end of each period. Analysis of this large volume of air could be made with considerable accuracy, but in order to calculate the exact total of oxygen residual in the air it was necessary to know the total volume of air inside the chamber under standard conditions. This necessitated, therefore, a careful measurement of temperature and pressure, and while the barometric pressure could be measured with a high degree of accuracy, it was found to be very difficult to determine exactly the average temperature of so large a mass of air. The difficulties attending this measurement and experiments upon this point are discussed in detail