CHAPTER V. 



THE DIFFUSION OF VAPOR INTO NUCLEATED AIRJ A CORRECTION. 



1. Apparatus and manipulation. The apparatus with which experiments 

 of the present kind are made is conveniently described by aid of the accompany- 

 ing diagram. The appurtenances necessary in practice are given in my report 

 on the "Structure of the Nucleus" (Smithsonian Contributions, No. 1373, 1903), 

 to which reference has frequently been made. A , Figure i , is a tall glass vessel 

 about one meter high, either cylindrical or rectangular in section, in the latter 

 case with opposed plate -glass sides. The liquid, L, whose vapors are to be 

 tested, is placed in the bottom. The wide tube, c, is used for sudden exhaustion, 

 while a vacuum gauge, g, registers the pressure differences. The tubes a and 

 b to the top and the bottom of A serve for the admission either of filtered 

 air or of nucleated air. They are used together, one for influx and the other 

 for efflux, in connection with the suction of an aspirator. 



When the diffusion of the necessarily heavy vapors from L is to be meas- 

 ured, the air in A is first cleansed of vapor by a current of nucleated air from 

 a to b. Thereafter the stopcocks are closed at a stated time. If now at a 

 subsequent time a sudden exhaustion is made in A through c, for a stated 

 pressure difference, dp' , shown at g, the progress of the diffusion may be com- 

 puted from the height of the fog-bank after an allowance is made for the rise 

 due to the exhaustion. 



On the other hand, if the aspirating current is of filtered air and moves in 

 the direction from b to a, over the surface of the volatile liquid, the receiver, A , 

 should become uniformly saturated to a high degree throughout. If nuclei are 

 added at a stated time below near the surface of the liquid, the corresponding 

 height of the fog-bank seen on exhaustion at a later time should indicate the 

 rate at which the nuclei diffuse, if they diffuse more slowly than the residual 

 concentration of vapor. This method for nuclei, which I pursued with entire 

 confidence, leads, however, to erroneous results, as the present paper will show; 

 for the diffusion of the nuclei is a much more rapid process than the accompany- 

 ing complications of vapor diffusion. 



2. Equation. To state the case specifically, let p be the vapor pressure 

 relative to the saturation pressure at the temperature S , at a time t after 

 diffusion of vapor commences and at a height x above the surface of the liquid 

 in the receiver, A . Then from well-known principles it may be shown that 



d) 



-r 



where k is the coefficient of pressure diffusion 



51 



