For nearly a month previous to the signing of the Armistice, the various plants at the Arsenal had shut down or were operated only to an extent sufficient to maintain the machinery and equipment in good working order, on account of the lack of shell into which to fill the gas, so that the above figures do not at all represent maximum productive capacity. These plants will be described in the appropriate chapters. The shell filling plant was really composed of several small plants, each of which was made up of units radiating from a central refrigeration plant which would serve all the units. Each unit could then be fitted with machinery adapted for filling shell of a different size, and for a particular gas. Moreover, an accident in one of the units would in no way impair the working of the remainder. The problem involved in the filling of a shell with toxic material (which is always a liquid or a solid and never a gas under the conditions in which it is loaded in the shell) is similar in a way to that of filling bottles with carbonated water. In the development of plans for the filling plant, many suggestions were obtained from a study of the apparatus used in commercial bottling plants. It was necessary to keep in mind not only the large number of shell to be filled, but also the highly toxic character of the filling material to be used. It was essential that the work of filling and closing the shell should be done by machinery in so far as that was possible, and that the operation should be carried out in a thoroughly ventilated room or tunnel, arranged so that the machinery contained in the tunnel could be operated from the outside. Special care was taken in closing the shell, the closing being accomplished by motors actuated by compressed air, which, in the closing process were driven until they stalled. In this way a uniform closing torque was obtained. The final results secured were admirable, as is evidenced by the fact, reported by the Quartermaster Officer at Vincennes on November 15, 1918, that not a single leaky shell had been found among the 200,000 shell received up to that date. [Illustration: FIG. 7.—A Typical Shell filling Plant at Edgewood Arsenal.] Details of the filling process will be found in the chapter on Phosgene. Besides the ordinary gas filling plants (of which one was completed and two were 80 per cent completed) there was a plant for stannic chloride grenades, one for white phosphorus grenades, and one for smoke shell also filled with phosphorus and a plant for filling incendiary bombs. Shell are designated by their diameter in inches or millimeters. The approximate amount of toxic gas required for filling each type of shell (10.5 per cent void) is as follows: