The air, being confined in a space with semi-rigid walls, is subjected naturally to variations in true volume, depending upon the temperature and barometric pressure. If the air inside of the chamber becomes considerably warmer there is naturally an expansion, and were it not for the tension-equalizer there would be pressure in the system. Also, if the barometer falls, there is an expansion of air which, again, in the absence of the tension-equalizer, would produce pressure in the system. It is necessary, therefore, in calculating the true volume of air, to take into account not only the apparent volume, which, as is shown above, is always a constant amount at the end of each period, but the changes in temperature and barometric pressure must also be noted. Since there is a volume of about 1,400 liters, a simple calculation will show that for each degree centigrade change in temperature there will be a change in volume of approximately 4.8 liters. In actual practice, however, this rarely occurs, as the temperature control is usually inside of 0.1° C. and for the most part within a few hundredths. A variation in barometric pressure of 1 millimeter will affect 1,400 liters by 1.8 liters. In actual practice, therefore, it is seen that if the barometer falls there will be an expansion of air in the system. This will tend to increase the volume by raising the rubber diaphragm on the tension-equalizer, the ultimate result of which is that at the final filling with oxygen at the end of the period less is used than would be the case had there been no change in the barometer. In other words, for each liter expansion of air inside of the system, there is 1 liter less oxygen required to bring the apparent volume the same at the end of the period. Similarly, if there is an increase in temperature of the air, there is expansion, and a smaller amount of oxygen is required than would be the case had there been no change; and conversely, if the barometer rises or the temperature falls, more oxygen would be supplied than is needed for consumption. It is thus seen that the temperature and barometer changes affect the quantity of oxygen admitted to the chamber. INFLUENCE OF FLUCTUATIONS IN THE AMOUNTS OF CARBON DIOXIDE AND WATER-VAPOR UPON RESIDUAL OXYGEN.