The complete removal of the carbon dioxide from the air sample and the provision for its quantitative measurement are accomplished by the standard soda-lime bottles and Williams bottles employed in this laboratory for all universal respiration apparatus. If the soda-lime bottle, L, and its accompanying Williams bottle, M (see fig. 8), are weighed prior to the period and again subsequent thereto, the increase in weight of the two indicates the carbon dioxide absorbed during the time, for the air entering L is dried over sulphuric acid in J1 and J2 and the air leaving M is dried to the same degree of humidity. The duplicate set of bottles shown in figure 12, controlled by the valves V3 and V4, gives an admirable check upon this measurement of carbon dioxide. It is of interest that almost invariably the agreement between the two sets is absolute. Occasionally, when there is faulty adjustment of the pressures inside the sampling cans, there is a slight discrepancy, owing to the fact that if the pressure is greater inside of one can than the other, less air is delivered and less carbon dioxide removed.

The determination of the amount of carbon dioxide in the aliquot is thus simple, namely, the weighing of the two absorbing vessels. The carbon dioxide collected in these vessels does not, however, represent an aliquot of that produced inside the chamber. If there were no subject inside the chamber and if outdoor air alone were drawn through this system, there would still be carbon dioxide absorbed in the weighing bottles, for there is a constant amount of carbon dioxide in outdoor air. Some attempts were made to remove this carbon dioxide entering the chamber by blowing the air delivered by the intake blower (see k, fig. 6) through a large scrubber filled with lime. This was found to be impracticable, for the absorbing agents best adapted for the removal of carbon dioxide likewise absorbed considerable amounts of water vapor, became pasty, and obstructed the passage of air. Since the carbon dioxide in outdoor air is such a constant quantity, it was believed that use could be made of this constant and a simple correction applied to the carbon dioxide weighed in the bottles L and M.

1The stoppers are always in place when an experiment is in progress.

While it was stated that the discharge into the sampling cans is approximately 45 to 50 liters per minute, this is not known exactly. Furthermore, the discharge will vary somewhat with the size of aperture used in the opening B (figs. 8 and 9). We can, however, meter the air after it has been discharged from either sampling can, passed through the purifying vessels, and is ready to be discharged into the room. This may be done by passing it through the dry gas meter 0 (see fig. 8), a so-called "3-light gas meter," which has been found satisfactory for the purpose. The meter reads in cubic feet, but metric scales may be obtained, if desired. With the present arrangement of the apparatus, approximately 45 cubic feet of air per half hour are withdrawn from each sampling can. The addition of the meter to the system affects in nowise the discharge into the sampling chamber, meaning simply a slightly greater load for the positive blower, which is already required to force air through two large sulphuric-acid bottles, a system of pipes and valves, a soda-lime bottle, small sulphuric acid bottle, and sodi-umbicarbonate can and, finally, through the meter.

By reading the meter at the beginning and end of the period, a fairly accurate measure is obtained of the total amount of air in each sample. If uncontaminated outdoor air is passed through the entire apparatus for several hours, a measurement may be obtained of the amount of carbon dioxide per 100 cubic feet of air or per cubic meter. This test has been made repeatedly. While slight fluctuations are found, the values are, on the average, very close to 1.48 grams of carbon dioxide per 100 cubic feet with the particular meter used. The correction for the carbon dioxide in the ingoing air is thus made by simply multiplying the reading obtained from the gas meter in cubic feet by the factor 1.48 grams per 100 cubic feet, and deducting the result from the weight of carbon dioxide absorbed. The remainder corresponds to the carbon dioxide in the sample which was produced by the subject inside the chamber.