88 



TUB STRITCTURE OK THE NUCLEUS. 



i,'iai)lis. If ii smoother iiu\ t- were obtainable, it would be useful for fomt)utation. 

 When the stationary eonditiou is reached, nuclei are lost at the boundary of the 

 vessel at the same rate at w liicli tliey are j)roduoed by the liquid. This process is 



o 



8 



10 



IZ 



14 



10 



Fig. 3. — Graphs Showing the Increase of Nucleation Occurring Spontaneously 

 IN CSj Vapor ; s. Coronal Apertures in the Lapse of Hours ; d. Diameters of Par- 

 ticles (in Centimeters for Identical Supersaturations) after the Lapse of Hours. 



exhaustive and may account for the luaxiiiuun indicated by the curve. Experi- 

 ments showed, howevei', that the curvature of Fig. 3 is exaggerated. 



27. Ethyl alcohol. — The liquids tieated throughout the above paragraphs 

 were non-ionizing solvents; it is important to add, by way of contrast, the corre- 

 sponding behavior of certain ionizing solvents, like the alcohols, acetone, and water. 

 I was long under the impi'ession that a sharp distinction could be drawn between 

 the two classes of solvents, but the behavior of acetone does not bear out this 

 surmise. Table 14 is an example of the coronal display in alcohol vapor. 

 It is particularly to be noticed that, whereas color distortion does occur in the 

 annuli of the coronas, there is no apprecial^le distoition of form. In this respect 

 alcohol resembles water more than any re-agent examined. 



At high exhaustions and a small i-esidue of nuclei, the occurrence of two kinds 

 of condensation maj- be shown. The coronas due to the nuclei are the first to 

 appear at relatively low exhaustions. At the high exhaustion (ja-essure-difference, 

 24 cm.) very large faint coronas are apt to be seen, and to appeal- suddenly, as if by 

 collapse. The nuclei here are probably moleculai'. and it is not impossible that 

 traces of water vapor, if present, may suffer condensation ; but it is the succession 

 of independent condensation to which I wish to advert. 



In both the present and the following experiments the final coronas are 

 regular. This is also true of the coronas obtained from air nuclei in case of 

 alcohol vapor. 



