THE CHEMICAL WARFARE SERVICE 169 



about i p.p.m. (0.007 m S- P er liter). This method is not 

 specific because arsine gives a similar precipitate in less time 

 than does mustard gas, and other compounds such as diphenyl- 

 chloroarsine and butyl mercaptan give positive results. As 

 against this, chlorine, hydrogen chloride, phosgene, chloropicrin 

 and superpalite give a negative test even when present in fairly 

 high concentrations. 



While the copper flame test is not sufficiently sensitive to 

 permit of direct detection of low but toxic concentrations of 

 mustard gas, it has been found possible to modify the method 

 so that one can detect o.i p.p.m. (0.0007 m S- P er liter) or even 

 o.oi p.p.m. under special conditions. The principle has been 

 embodied in a portable field apparatus. The method is really 

 one for halogens and is not specific for mustard gas. Its use- 

 fulness in the field is questionable. 



SMOKE 



One of the most interesting scientific studies made was con- 

 cerned with the theory of smokes. The concentration of the 

 smoke was determined by precipitation in a modified Cottrell 

 apparatus consisting of a central wire cathode surrounded by a 

 cylindrical aluminum foil anode about i/iooo inch in thickness. 

 A 15,000 volt rectified direct current was used and complete 

 precipitation was obtained with fairly concentrated samples of 

 smoke even when drawn through the apparatus at a rate of 

 about five liters per minute. The aluminum foil and adhering 

 smoke were then weighed. Microscopic examination showed 

 whether the smoke particles were liquid or solid. The size 

 of the particles in a smoke can be determined ultra-micro- 

 scopically with fair accuracy by measuring the velocity of a 

 charged particle in an electric field of measured intensity, 

 photographing the path of the particle while the direction of 

 the electric field is reversed regularly by a rotating commutator 

 whose speed is known accurately. When the convection due 

 to the source of light is perpendicular to this motion, a zigzag 

 line is obtained. Since about one-third of the smoke particles 



