It was soon found that better protection was obtained if the absorbents were mixed before packing in the canister. This procedure also simplified the method of packing and was used in canister _B_ and following types. Among other changes introduced in later types were: The integral valve was replaced by a removable check valve plug which enabled the men in the field to adjust the valve in case it did not function properly. The mixture of charcoal and soda-lime was divided into three separate layers and these separated by cotton pads. The pads offered protection against stannic chloride smokes but not against smokes of the type of sneezing gas. The green soda-lime was replaced by the pink granules. In April, 1918, the mesh of the absorbent was changed to 8 to 14 in place of 6 to 14. About July 1, 1918, the authorities were convinced by the field forces of the Chemical Warfare Service that the length of life of the chemical protection of the standard _H_ canister (the type then in use) was excessive and that the resistance was much too high. Type _J_ was therefore adopted, July 27, 1918. In this the volume of the absorbent was reduced from 450 cc. to 300 cc. It was packed in two layers, ⅔ in the bottom and ⅓ in the top. One pad was placed between the layers and one on top. This change gave a lowering of the resistance of 27 per cent (to 2.5 inches) at a sacrifice of 50 per cent of the length of life of the canister, but not of protection during the shortened life. Type _L_ differed from this only in having 325 cc. of absorbent, a change made to decrease leakage about the top cotton pad. [Illustration: FIG. 57.—U. S. Army Canister, Type _J_.] The following table shows the relative efficiency of various canisters: -----------------+----------+----------+----------+----------+------ | | U. S., | British, | French, |German | p. p. m. | Type _H_ | S. B. R. | A. R. S. | -----------------+----------+----------+----------+----------+------ Chloropicrin | 1000 | 770 | 17 | 2 | 43 Phosgene | 2500 | 85 | 54 | 5 | 16 Hydrocyanic acid | 500 | 70 | 90 | | 10 Mustard gas | 100 | 1800 | | 35 | 195 -----------------+----------+----------+----------+----------+------ The figures represent time in minutes till the first traces of gas begin to come through. [Illustration: FIG. 58.—Type _J_ Canister and Contents] MANUFACTURE The following description of the manufacture of the gas mask at the Long Island plant is taken from an article by Col. Bradley Dewey[26]: [Footnote 26: _J. Ind. Eng. Chem._, =11=, 185 (1919).]