74 THE STRUCTURK Ol' THE NUCLEUS. 



uotably thiit of carbon disulphide and certain solutions. In these the number of 

 nuclei ultimately reaches a fixed maximum or saturation at an excessively small 

 vapor pressuie, and the final coronas have fixed diameters for fixed supersatuiations. 



THE FLOWER-LIKE DISTORTION OK CORONAS DUE TO GRADED DISTRIBUTIONS OF NUCLEI.' 



7. The fore2;oing general descriptions have brought out a series of coronas 

 obtained very easily with volatile hydrocarbon solvents, benzol, etc., in which the 

 coronas met with are not closed and annular, but of a variety of patterns from oval, 

 with the long axis vertical, to the symmeti-ically open doubly inflected types (lyi'e- 

 shaped, basin-shaped), running continuously into horizontal strata in the limiting 

 case. It will be expedient to examine more closely into these conditions before 

 proceeding to a detailed study of the phenomena in hydrocarbon liquids generally. 



These distortions are due to the non-uniform distribution of nuclei as to size, 

 the largest having sunk deepest and the finer nuclei floating uppermost, in virtue 

 of the precipitation mechanism. When supersaturation is produced by adiabatically 

 cooling the benzol vapor, the condensation begins at the lower strata and then 

 passes upward as exhaustion proceeds and higher degrees of supersaturation are 

 reached. The evolution of coronas is peculiar in this case, and reminds one of a 

 person throwing out his aims lateially and upwards until his hands strike above 

 his head. The sweep of coronal streamers is outward and upward symmetrically 

 with respect to the vei'tical plane through the source of light. If the gradation is 

 not too lapid, they eventually coalesce above it. 



The droplets produced are finer above than below; but it does not follow that 

 there are more particles in the upper layers. The reverse will natui-ally be 

 assumed. The lower particles being laiger, have first received the condensation as 

 already suggested, and have thus grown biggest, as the opportunities for growth 

 came earliest and lasted longest. 



8. It is my purpose in this section to work out the shape of the loci of like 

 color when the nucleation is not uniform as to size or numl)er. The distributions 

 arise from the subsidence of loaded nuclei ; they are, therefore, horizontally 

 stratified. 



In figure 1, let a be the distant point source of light into which the coronas 

 would shrink annularly and symmetrically from Avithout inward, to a limit in a 

 normal case. Let cp be the angle between the horizontal through o and the radius 

 vector, r, to a line of uniform color, ab, in the distorted corona, and let h be the 

 height of the extremity, c, of the radius vector above the datum line through o. If 

 R Ije the distance of /• from the eye of the observer, 2r/7? = s/R is the angular 

 aperture of the corona. 



Let * be the diameter of the particles at the level [)a8sing through c. Then if 

 <5„ and s^/Ji be the corresponding quantities (diameter and aperture) of particles 

 in the datum level, 



^.. = d„.s„ = .00144, (1) 



' Cf. Am. yourn. of Set., (4), pp. 309-312, vol. xiii., 1902. 



