In accordance with this view as to the nature of smoke filtration, the important factors involved are (1) the Brownian motion of the smoke particles, (2) the area and arrangement of the internal surface presented by the filter, (3) the flow of the smoke as a whole, and (4) the attractive forces between the filter surfaces and the smoke particles. The first three of these factors determine how many particles come within the range of the mutual forces of the particle and filter surface, and the fourth factor determines the chance or expectation that the particle will permanently adhere to the surface of the filter. TESTING SMOKE FILTERS All the early tests made on smoke filters used diphenylchloroarsine, because it was felt that the filter must be tested against a toxic smoke. A man test was developed as representative as possible of actual conditions in the field, and the time necessary for a man to detect diphenylchloroarsine smoke in the effluent stream when breathing at a normal rate, using a carefully controlled concentration of smoke produced by detonation, was used as the criterion of the protection offered by the canister. This test was subject to extensive individual variations, due to the varying physiological resistances of different men to diphenylchloroarsine smoke. Further, it was quite inadequate for rapid testing on a large scale. A testing machine was then developed, which gave results comparable with those obtained in the man test. The method used in detecting the gas was physiological, that is, by smell or by its irritating action towards the membranes of the eye. While these are purely qualitative tests, they are much more sensitive than any possible chemical tests. Because of the desirability of having a method which could be controlled chemically, other methods were developed. Ammonium chloride is a solid smoke, consisting of particles of quite variable sizes. It is sensitive to dilution and clogs the pores of the filtering medium quite rapidly. For this reason it was used in the study of the rate of plugging or clogging of the filter (the closing of the pores of the fabric or other material to the passage of air). The smoke is produced by the reaction of ammonia and hydrogen chloride-air streams. The smoke thus generated is passed from the mixing chamber to a larger distribution box and from there through the filter, at a standard rate. The concentration of the smoke may be accurately determined by chemical means or photometrically, using a Hess-Ives Tint Photometer, the Marten Photometer, or a special photometer developed by the Chemical Warfare Service. A comparison of a large number of tests with those of other smokes would indicate that ammonium chloride smoke offers accurate information as to protection sought, but is hardly a desirable smoke for testing on a large scale.