Condensation, and on the contrast of Colors. 85 



spontaneously produced without change of pressure. The bear- 

 ing of this on the present results is chiefly as an error induced 

 by the subsidence of loaded nuclei. The table contains three 

 independent experiments. 



Omitting farther discussion, one infers from the obviously 

 linear distribution of the orders of coronas in the lapse of time, 

 either that the number of nuclei is not changed by exhaustion, 

 correspondingly greater numbers being produced at low pres- 

 sures to make good the loss, or more probably that the time 

 losses obey similar exponential laws to the exhaustion losses. 

 The following theory is a more rational systematization of the 

 data in accordance with the latter view, for which other cogent 

 evidence might be adduced. 





Table II 



. — Time losses of nuclei. N= 



= ^0 



] Qfolog y 





Esh. 





Coronas. 



; 



No. by 





No. 



Time. 



Colors of successive annuli. 



Table I. 



b 







m 



Nucleation 









•08 



1 



27 



Wh, viol, yl 





4 





2 



61 



Wh, yl, br, gr, rd 





8 





3 



94 



Wh, prp, yl-gr, rd, gr 





10 





4 



124 



Wh, br, gr, rd 





14 





5 



160 



Wh, gr, prp, yl-gr, gr 





17 





6 



190 



Wh, br, gr, rd 





20? 





7 



292 



No color 





— 





Series 2. 















m 



Twice nucleated 









•14 



1 



50 



Yl, rd, gr, rd 





8 





2 



80 



Gr, rd, gr, rd 



11 



-12 





3 



112 



Prp, br, gr, viol, rd 





15 





4 



140 



Wh, br, gr, rd, gr 





20? 





Series 3. 















m 



Nucleation 









•10 



1 



95 



Wh, viol, prp, yl-gr, rd, 



gr 



10 





2 



155 



Wh, prp, br, gr, rd 





15 





3 



185 



Wh, prp, br, gr, rd 





19 





Let n be the order of the corona (exhaustion number), N the 

 number of nuclei producing it for the fixed supersaturation. 

 Without correction for time losses, N=y n , where y=p/p 6 

 under isothermal, and y x ' y =p/p under diabatic conditions, p 

 and p being the pressures before and after exhaustion. 



The data of the preceding section show that N suffers a 

 time loss varying as a+bt in the time t, where a and h are con- 

 stants. Hence the above equation must be corrected to read 

 j\r = yn(i+bt) to ac [ m jt f i^th losses of similar geometrical char- 

 acter. Finally the initial nucleation, ]¥ a =y n ^ is not identical 

 in the different experiments, whence 



JV= 1 o (no+n(1 _60),og y 

 is the final form of the equation. 



