r,() THE STRUCrrURE OK THE NUCLEUS. 



stiitfd. 'J'liiis, at the liieaii ilianieter (in cetitim.) d= 0.01, sN/N'= .29 and .In, 

 respectively. lf<^=.0034 and .0052 respectively, all the nuclei will be precipi- 

 tated in a single exhaustion of the value specified (76-58 cm.). 



EXl'ERIJIENTS WITH WATER VAl'OK. 



.*^. Constantft and apertuics uf corunas.- — Followintj tiie unsuccessful investiga- 

 tion with benzol, 1 will now deteimiue as nearly as 1 can the diameters of the cloud 

 paiticles of water and their numbers. It was necessary for this purpose to make a 

 special series of measurements in which the diameters of the normal coi'onas were 

 measured, as far as practicable. These results are given in tables 3 and 4, in which 

 the firet four columns have the usual signification. The time interval between exhaus- 

 tions was uniformly 1 min. 45 sec. The circular source of light was placed 2.5 meters 

 from the spherical receiver (diameter, 'M) cm.). With the light so far distant, the 

 central patches of the coronas following the 7th, Sth, etc., were not white centered, 

 but nearly uniformly colored, showing that the close light in the above experiments 

 fave a ditliise effect not belonging to the coionas pmper. C'oroiias 11, ]'2, and 

 those following If! are truly white centered, the latter being of the normal type, 

 for which alone the measurements of aperture are directly a[>plicable. One may 

 note in passing that sulphur nuclei are sooner emptied out' by precipitation than 

 punk nuclei. Although the former give denser initial fogs, the latter give the 

 greater number of coronas. § 7 shows that sulphur-water particles are the larger, 

 coet. par. 



The column marked s shows the aperture of the corona, being the chord of the 

 radius 20.1 cm. for the goniometer described in §2. In the next column y^l/N 

 is given, a quantity proportional to the diameter, d, of the particles, so that 

 //= (/f, iVl/N. N, the relative number of cloud particles, is found as above, since 

 shaking is never necessary, from A^= A'qIO -**+"> '"'''*', where A^j = 1, b =z A, 

 t = 1.75 min., log y = — .0866, for exhaustions from 76 to 58 cm., nearly. 



As the coionas were advantageously compared with those of lycopodium 

 spoies for which .s' == .45 cm. and d' = .0032 cm., for the same goniometer, the 

 equations used above will again be applicable here. Consequently, since 

 d-=^ .44/.S- X 1 :' , = d^ r T/N, 



, 1.44X10^ 



'"~.rxr' TTa^' 

 The values of sx f'l/JV are given in the tables. They are nearly enough constant 

 foi' measurements of this kind, wlieie the thing to be observed is always a more or 

 less hazy and fleeting outline. C'f. § 11. Their mean value in table 3 is 4.16, and 

 in table 4 is 5.47. The conesponding values of f/,„ or the initial diametei-s of the 

 cloud particles, aie d^ ^= .35 X 10"^ cm. in table 3 and d^ = .26 X 10" •' cm. in table 4. 

 .Ml other diameteis are then found from d^= d^x f^\/N. 



The final datum of the tables, », is the ahsolute numbei' of nuclei per cub. 

 cm., l)rieHy called the nncleation, as found from the value of d and the grams of 



