After an absorbent has been developed to a given point, and is considered of sufficient value to be used in a canister, the materials are assembled as described in Chapter XII. While the final test is the actual use of the canister, machine tests have been devised which give valuable information regarding the value of the absorbent in the canister and the method of filling. [Illustration: FIG. 74.—Standard Two-tube Apparatus for Testing Absorbents, Showing Arrangement for Gases Stored in Cylinders.] The first test must be that for _leakage_. The canister must show no signs of leaking when submitted to an air pressure of 15 inches of mercury (about half of the normal atmospheric pressure). The second factor tested is the _resistance to air flow_. This is determined at a flow of 85 liters per minute and should not exceed 3 inches. The latest canister design has a much lower resistance (from 2 to 2½ inches). The third test is the _efficiency_ of the canister against various gases. For routine work, phosgene, chloropicrin and hydrocyanic acid are used against the standard mixture of charcoal and soda-lime: Chloropicrin is usually used against straight charcoal fillings, while phosgene and hydrocyanic acid are used against soda-lime. [Illustration: FIG. 75.—Apparatus for Testing Canisters Against Chloropicrin.] Different types of apparatus are required for these gases. They are very complicated, as may be seen from the sketch in Fig. 75, and yet a man very quickly learns the procedure necessary to carry out a test of this kind. The gas is passed through the canister under given conditions, until at the end of the apparatus a test paper or solution indicates that the gas is no longer absorbed but is passing through unchanged. This point is called the “break point,” and the time required to reach this point is known as the life of the canister. This time is also the time to 100 per cent efficiency. Other points, such as 99, 95, 90 and 80 per cent efficiency are determined. These are used in comparing canisters. The canister tests were of two general classes: continuous and intermittent. In the first the air-gas mixture was drawn through continuously until the break point was reached. The results obtained in this way, however, did not give the time measure of the value of a canister in actual use. The intermittent test differs only in that the flow of air-gas mixture is intermittent, corresponding to regular breathing. Special valves were adapted to this work. Canisters must also be tested as to the protection they offer against smoke. These methods are discussed in Chapter XVIII. MAN TESTS