T1U-: H'l'liUCTURK OK TIIK NUOLKirs. 



63 



correspond to the general difference of behavior, and this is qualitatively the case. 

 Of the determinative loci, .s/iV^ in terms of ?:, in tables .". and 4, the former (sul[)hur) 

 shows a tendency to decrease, the other (punk) a tendency to increase with 

 increasing z. As 6- is observed, it follows that N is respectively too slow and too 

 fast in its variation, instead of being constant with .?. Consequently, I suspect that 

 the mean of tables 3 and 4 is a more cori'ect seijuence for w than either table 

 separately. The same conclusion is reached on comparing a given corona (f<ir 

 instance, the green centered type for z = 14) in the two tables. Obviously, this 

 definite diffraction pattern must, cent. par., correspond to the same number of nuclei 

 in the two cases. Finally, in the experiments of the foregoing chaptei', doubling 

 or quadrupling the initial nucleation by correspondingly lepeating the charges, 

 was found to shift the coronas but a few orders ahead, while the effect should be 

 3 or 6 orders of displacement. 



11. Removal hy subsidence. — The rapid removal of cloud particles toward the 

 end of the tables shown in figure 5, and due to the subsidence of large drops, is note- 



Fig. 5.— Chart Showing the .Accelerated Removal of Cloud Particles in the 

 Final Exhaustions (Numbers =, on Curves). 



Fig. 6.— Chart Showing the Subsidence of Fogs of Different Orders (Numbers z, 

 on Curves) in the Lapse of Time. Fall (/) in cm, Time (t) in Minutes. 



worthy, and the difficulty of coping with the subsidence er roi- ha s induced me to insert 

 a few special subsidence data here. The products, s X aIi/A", were discarded in the 

 last case of table 3 and the two last observations of table 4, because of the magnitude 

 of this error. Since ?■ =: 9 X 10-* X yjv, the reciprocal variable = r^ X 10 8/81 may 

 be computed, as is done for the vanishing coronas in the following table. 



TABLE 5.— SUBSIDENCE OF CLOUD PARTICLES COMPUTED. 



