All precautions are taken, however, so far as securing tightness in coupling and installing them on the absorber system are concerned. During this period the other set of absorbers is carefully weighed and made ready to be put in place and tested and about 10 minutes before the experiment proper begins the residual analyses are begun. The series of U-tubes, which have previously been carefully weighed, are placed on small inclined racks and are connected with the meter and also with the tube leading to the mercury valve. The pet-cock which connects the return air-pipe with the drying-tower and the gas-meter is then opened and the mercury reservoir is lowered. The rate of flow of air through the U-tubes is regulated by a screw pinch-cock on the rubber tube leading to the first U-tube. This rate is so adjusted by means of the pinch-cock that about 3 liters of air per minute will flow through the U-tubes, and as the pointer on the gas-meter approaches 10 liters the mercury reservoir is raised at just such a point, gained by experience, as will shut off the air-current when the total volume registers 10 liters on the meter. The pet-cock in the pipe behind the meter is then closed, the U-tubes disconnected, and a new set put in place. A duplicate and sometimes a triplicate analysis is made. When the physical observer calls the time for the end of the period, the switch which controls the motor is opened and the chemical assistant then opens the rear valve of the new set of absorbers and closes the rear valve of the old set, and likewise opens the front valve of the new set and closes the front valve of the old set. As soon as the signal is given that the oxygen connections have been properly made and that the oxygen has been admitted to the chamber in proper amount, the blower is again started. It is then necessary to weigh the U-tubes and disconnect the old set of absorbers and weigh them. If the sulphuric-acid absorbers have not exceeded the limit of gain in weight they are used again; if they have, new ones are put in their place. The first sulphuric-acid absorber is connected to the front valve, then the potash-lime can, and then the last sulphuric-acid absorber; but before connecting the last sulphuric-acid absorber with the sodium-bicarbonate can, a test is made of the whole system from the front valve to the end of the second sulphuric-acid absorber. This is made by putting a solid-rubber stopper in the exit end of the second sulphuric-acid absorber and, by means of a bicycle pump, forcing compressed air in through a pipe tapped into the pipe from the valve at the front end until a pressure of about 2 feet of water is developed in this part of the system. This scheme for testing and the method of connecting the extra pipe have been discussed in detail in an earlier publication.[32] Repeated tests have shown that this method of testing the apparatus for tightness is very successful, as the minutest leak is quickly shown.