After the system has been thoroughly tested, the rubber stopper in the exit end of the second sulphuric-acid absorber is first removed, then the tube connected with the pump and manometer is disconnected and its end placed in the reservoir of mercury. Occasionally, through oversight, the pressure is released at the testing-tube with the result that the air compressed in the system expands, forcing sulphuric acid into the valves and down into the blower, thus spoiling completely the experiment. After the testing, the last sulphuric-acid absorber is coupled to the sodium-bicarbonate can. It is seen that this last connection is the only one not tested, and it has been found that care must be taken to use only the best gaskets at this point, as frequently leaks occur; in fact, it is our custom to moisten this connection with soapsuds. If new rubber gaskets are used a leak is never found. SUPPLEMENTAL APPARATUS. To maintain the apparent volume of air through the whole system constant, oxygen is admitted into the tension-equalizer until the same tension is exerted on this part of the system at the end as at the beginning. This is done by closing the valve connecting the tension-equalizer with the system and admitting oxygen to the tension-equalizer until the petroleum manometer shows a definite tension. After the motor is stopped, at the end of the experimental period, there is a small amount of air compressed in the blower which almost instantly leaks back through the blower and the whole system comes under atmospheric pressure, save that portion which is sealed off between the two levels of the sulphuric acid in the two absorbing vessels. A few seconds after the motor is stopped the valve cutting off the tension-equalizer from the rest of the system is closed, the pet-cock connecting this with the petroleum manometer is opened, and oxygen is admitted by short-circuiting the electrical connections at the two mercury cups. This is done by the hands of the observer and must be performed very gently and carefully, as otherwise oxygen will rush in so rapidly as to cause excessive tension. As the bag fills with gas, the index on the petroleum manometer moves along the arc of a circle and gradually reaches the desired point. At this point, the supply of oxygen is cut off, the valve connecting the tension-equalizer with the main system is opened, and simultaneously the needle-valve on the reduction-valve of the oxygen cylinder is tightly closed, preliminary to weighing the cylinder. At this point the motor can be started and the experiment continued.