The Dispersoid Toxic Smoke Candle differs from the B-M candle in that the toxic container is not used. A mixture of smokeless powder and the toxic material (diphenylchloroarsine or D. M., an arsenical obtained from arsenic trichloride and diphenylamine) is filled directly into the container, a cylindrical can 3.5 inches in diameter and 9 inches high made from 27 gauge sheet metal, and packed under a total pressure of 2,500 pounds. The top of the candle is a metal cover, containing the match head scratcher, which is separated from the match head by a Manila paper disc. These are the same as those used in the B-M candle. The candle has a total weight of about 4.25 pounds, of which 3.6 pounds are the smoke mixture, containing about 1.3 pounds of toxic material. In operating the candles, the cover is removed and the match head ignited by friction with the scratcher. The match head burns through the cardboard and ignites the powder. The heat and gas produced by the combustion of the powder vaporizes the particles of the toxic material and carries the vapors out through the orifices at a high velocity whereupon they recondense to form a smoke. The rapid emission of the vapors through the orifice prevents any possibility of their ignition. The time before good emission of smoke takes place after the ignition of the match tip of a candle is 30 seconds. The average time of vigorous smoke emission is from four to five minutes. The result of a field test with the dispersoid candle is shown in Fig. 102. A comparison of a British and a Dispersoid candle is shown in Fig. 103. It should be stated that this may not have been a fair test as only one British candle was available for the comparative test. CHAPTER XVIII SMOKE FILTERS The first types of the Standard Box Respirator contained cotton pads, which sufficed to remove the ordinary smoke of the battlefield and even that from the earlier toxic materials. Improved methods of producing toxic smokes, by means of which smaller particles were obtained, led, early in 1918, to the recognition of the need of improved protection against these smokes. The first attempts to meet this need consisted in improving the filtering qualities of these pads. It was soon found, however, that to make better filter pads would greatly increase the total resistance of the canister. This was highly undesirable, since the resistance of the ordinary canister was already so high as to be very uncomfortable. To overcome this objection, some of the early designs of filter canisters were provided with a mechanical valve, which could be operated by hand, to by-pass the air around the filter when the canister was used against gas alone, or so set as to make the air pass through the filter when smoke was feared. This introduced a factor of uncertainty among the men during a gas attack, since each man must decide for himself whether smoke was present. This reason alone was sufficient for discarding this design.