130 A CONTINUOUS RECORD OF ATMOSPHERIC NUCLEATION. 



on the outside of the box, the object being to heat the air to room temperature, 

 especially in winter.^ Two thermometers, T, T', with their bulbs respectively 

 in the air and the water within the trough, register the temperature. The end 

 of the influx pipe rises to a height q, near the axis of the trough and opposite to 

 the outlet, C. Finally, the whole inside of the trotigh is lined with a double 

 layer of cotton cloth, ««, supported on a framework of stout copper wire. The 

 trough should be mounted with its longitudinal axis on trunnions in order that 

 the whole interior may be moistened in a single rotation, as shown in figure 

 I a, in detail. The stopcock F, provided with a cotton filter, is often useful in 

 testing. 



The horizontal diameter of the coronas is observed, the point source of 

 light being about 2 meters off on one side and a suitable goniometer about one 

 meter on the other side of the trough. The distances used were 85 cm. and 

 250 cm. With the eye at about i meter from the fog chamber the apertures 

 of coronas are relatively independent of this distance and at the same time large 

 enough for satisfactory measurement. 



An ordinary jet pump suffices for aspiration (with the cocks C and B open) ; 

 and with an added vacuum chamber provided with a vacuum or mercury gauge, 

 for sudden exhaustion (C and B having been closed), care being taken that the 

 connecting tubing beyond B is wide. These details are shown above. 



3. Diffusion from two opposed surfaces. — The high values of nucleation 

 observed during the winter months will not be received without misgiving, since 

 the air during the very cold weather is nearly dry, and after being heated to 

 20° ver}'' far from saturation. Deficient sattiration, however, would decrease 

 the size of the fog particles and, ccet. par., increase the size of the coronas, in 

 this way showing the same result as excessive nucleation. Hence it is necessary 

 to estimate the time which is needed to saturate dry air in the apparatus de- 

 scribed in sj 2. 



Given a rectangular trough at the bottom of which is a sui-face of water 

 and at the top of which a surface of saturated cloth, the diffusion problem is 

 equivalent to the case of an indefinite air plate into which the vapor enters on the 

 two exposed sides. Thus if p be the vapor pressure relative to saturation at a 

 distance x above the surface of the water or a — ::c below the wet cloth, "at the 

 time t, 



P = i-^(sin?r^''/")'^-^ + jsin^e-'^'^/'^^^'V. . .) 



where k is the coefficient of diffusion. 



If diffusion takes place into a partially saturated atmosphere at an initial 

 pressure p^, the factor 4/7T in the last equation is to be replaced by (4/77) 

 (i-:?o). 



' In case of a moderately long (20-30 feet), thin (J inch) influx pipe, experience showed 

 that both the internal coil, p, and the external coil are superfluous. They were there- 

 fore discarded. 



