Canisters are tested for leaks by the apparatus shown at _D_ in Fig. 78. The canister is clamped down tightly by wing nuts against a piece of heavy ¼-in. sheet rubber large enough to cover completely the bottom of the canister and prevent any inflow of air through the valve. Suction is then applied, and a leak is indicated by a steady flow of air bubbles through the liquid in the gas-washing cylinder _E_. A second gas-washing cylinder, empty, is inserted in the line between _E_ and the canister as a trap for any liquid drawn back when the suction is shut off. If a leak is shown, it can be located by applying air pressure to the canister and then immersing it in water. [Illustration: FIG. 78.—Apparatus for Determining Pressure Drop and for Detecting Leaks in Canisters.] METHODS OF CONDUCTING TESTS Three general methods of conducting man tests are followed: (1) Canisters are placed in the brackets outside the chamber or fastened to the wall tubes within the chamber. The subjects of the test remain outside the chamber, and the facepieces of the masks are connected directly to the canisters, in the first case, and to the wall tubes connecting with the canisters, in the second case. The concentration is established and the time noted. Then the men put on the masks and breathe until they can detect the gas coming through the canisters. Reading matter is provided for the men during the test period. When gas is detected, the time is again noted and the time required for the gas to penetrate the canister is reported as the “time to break down” or “service time” of the canister. Ten canisters are tested at one time, and the average of the results for the 10 canisters is taken for that type of canister. Much less accurate results are obtained when the final figure is based on a small number of canisters. This is largely due to the various breathing rates and sensitiveness of different men. (2) The canisters are placed as in (1), but it is only necessary to know if they will give perfect protection for a given length of time. The procedure is the same as in (1), except that the test is arbitrarily stopped at the end of the indicated time, and the number of canisters and the service times of the same noted. (3) When the canisters are of such a type that they cannot be properly tested as in (1), or when it is desired to test the penetrability of the facepiece, the men wear the complete mask and enter the chamber. They remain until gas penetrates the canister or the facepiece, as the case may be, or until it is determined that the desired degree of protection is afforded. The service time is computed as in (1). (4) Maximum-breathing-rate tests are made either by men in the chamber or by the men outside, in which they do vigorous work on a bicycle ergometer. In this test the average man will run his breathing rate up to 60 or 70 liters per min. The concentration of the gas is followed throughout the test by aspirating samples and analyzing them.