A CONTINUOUS RECORD OF ATMOSPHERIC NUCLEATION. 133 



where p is the vapor pressure relative to saturation at a distance x above the 

 surface of the Hqttid at the time t for the diffusion coefficient k. = 23. The data 

 computed suffice to locate the curves which are sufficiently given in figure 3 for 

 X = ^ cm. the middle plane, and :r = 10 cm. the top plane of the trough. When 

 the initial saturation is 1/3 or 2/3, the coefficient 2 /;r is to be modified as stated 

 above. Figure 3 shows the coitesponding results. 



If it were not for convection, therefore, such an apparatus would be un- 

 suitable, for even after waiting 5 minutes, the air at the top, «; = 10 cm., for an 

 initial saturatioti of 2/3 is btit .8 saturated from diffusion alone. 



One might therefore expect to obtain distorted coronas campanulate in 

 outline, small below and large above, whenever condensation is produced 

 within a few minutes after closing the inlet. Yet such is never the case if less 

 than a minute is allowed after inflttx ceases. Granting that 2 to 4 minutes are 

 needed on the average for a particle to pass through the trough, as stated above, 

 if exhaustion is made immediately after closing, the top layer is but 2/3 saturated. 

 Under these conditions there is in fact an unusually large green centered faint 

 corona with a horizontal band of crimson color nmning through it. Half a 

 minute later, however, the figure is quite regular again, showing that the con- 

 vection of light vapor must be very active. After 2 minutes subsequent to the 

 closing of the influx pipe, the air may be regarded sattirated except in the 

 coldest weather. 



6. Absorption and decay of nuclei. — The losses in the influx pipe are 

 difficult to determine because of the variation of atmospheric nucleation. The 

 obsei-ver is left in doubt whether a given difference is due to absorption in the 

 pipe or to causes without. The experiments incidentally made throughout the 

 long experience of Chapter IX showed no serious discrepancy. 



The possibility of loss of nuclei on contact of dry air with the saturated gas 

 in the condensation chamber is an independent question. It is also to be borne 

 in mind that nuclei may possibly be produced by the sudden contact within the 

 chamber. No evidence is forthcoming. 



If the nt:clei after being introduced into the receiver are soltitions, some 

 estimate of their persistence may be formed from my experiments on solutional 

 nuclei, by treating the loss as if it occurred at the boundary of the vessel only. 

 If the nucleation falls oft" from w„ to n in the time t, and k is the absorption 

 coefficient, 



where r and / denote the radius and length of the cylinder in wliich absorption 

 takes place. In case of comminuted pure water, fe = 5 to 10 cm./min., and the 

 nuclei should quite vanish in a few minutes. 



