84 THE STnUCTtTRK OF THE NUCLEUS. 



The last three data show a uniform rate of about .014 cm./sec. The first 

 datum corresponds to a laiger rate (.023 cm./sec), meaning, probably, that the 

 initial couche was uot sufficiently lowered. The results are furthermore in accord 

 with a rate of diffusion independent of the shape of the vessel. {Cf. Chapter VI.) 



'20. Repetition. — The occuireuce of apparently spontaneous uucleation in 

 benzol and not in the other liquids, induced me to investigate the origin of tliese 

 particles again, using the more perfect filter already described. Beginning with 

 .-lii' nuclei, the results were: 



Exhaustion Xo., 12 3 456 



Corona, Strong. Weaker. Vanishing. .\bsent. Absent. .Absent. 



After 30 minutes this cleaned charge was again tested, but no coronas were in 

 evidence. Hence, nuclei were not spontaneously produced. The e.vperiment wjis 

 carefully repeated later in the day, with the same lesults. The vessel, e.vniniiied 

 ne.\t day, showed, nevertheless, that nuclei had arisen from somewhere during the 

 night. They were lemoved in two e.\haustions. It was thus necessary to add new 

 tests, with a speciall)' cleaned vessel and long waiting (15-20 houi-s) between the 

 e.Yi)eriments. These finally placed the absence of spontaneous nucleation in benzol 

 beyond a doubt. What does occur is the diffusion of the residual band overlooked 

 on exhaustion, and brooding immediately over the liquid surface. Thus a fine 

 sheet of fog, due to air nuclei near the surface of the liquid, will diffuse sufficiently 

 in 30 minutes to produce a full corona ou exhaustion, and 5-10 exhaustions may 

 be required to remove the nuclei. 



21. Shalcen nuclei. — The following incidental example of the persistence of 

 the nuclei produced by shaking benzol may be inserted here for comparison. The 

 subject will be examined at length below with other apparatus: 



Lapse of time after shaking, 5'" Corona, strong 



44'" strong 



30'" strong 



60'" distinct 



iSo" distinct 



15'' absent 



These nuclei are not as remarkably i)ersistent as the cases examined below, 

 l)Ut they cannot be particles of pure li(juid, as this would instantly evaporate. 



22. Toluol. — This li(Miid behaves similarly to benzol. The followiiiir table 

 contains results .showing the coronas produced by air nuclei on exhaustions from 

 TO to 60 cms. : 



