THE STRUCTnKE OF THE NUCLEUS. 115 



(/= .348 -^jm, aud the results uiv inscribed iu table 1. So far as conclusions are 

 concerned, it is quite as satisfactory to use the theoretical data // = 64 s^m X 10«, 

 and these results are also found in the table. 



Curiously enough, the irregular results of the (5 % solution agi'ee better with 

 this curve than the smoother results of the ?> '/u solution ; but the reason, I think, is 

 the quicker evaporation of the smaller [)articles (smaller piessure differences) into 

 the larger particles in the case of the latter. The observed coronal apertures 

 appear too small and the diametei's of the particles (since .sy/ := const.) are thus 

 too large. Possil)ly the smaller nuclei may not be large enough to induce con- 

 densation at the lower pressure differences, but there is evidence below, § 17, 

 against this supposition. 



The .same result uuxy be reached by consulting the column for ii. This num- 

 ber is least for the smallest pressure differences, proving that here many of the 

 pai-ticles have evaporated because of their small size. 



7. faffed of inteuMUj of ■■^huhing. — The exhaustion being ecjuivalent to a 

 pressure decrement of 16 cm. throughout, w =: 73 X 10" ^^ whence ?i = 47 s^ , 

 nearly. If the graph of ii varying with the number of shakes be constructed 

 (figure 4), a curve of surprising smoothness is obtained, which is ultimately nearly 

 linear. In other words, about 2.6 nuclei are produced each time the 500 cub. cm. 

 of HCl solution are vigorously jei'ked once up and down. This would make 130 

 nuclei for 50 jerks. Tlie actual number is about 160, owing to the accelerated 

 action at the beginning of the experiment. 



If each nucleus I'epresents an electron, or a chai-ge of 2.3/10'*' coulombs, 

 1.7 X 10^- shakes would be needed to pi'oduce a coulomb per cub. cm. of aii', or 

 5.6 X 10* shakes to produce an electrostatic unit in each cub. cm. of the par- 

 ticular quantity of air contained in the receiver. As this amounted to about 

 14,000 cub. cm., 4 X lO* shakes would be needed to set free an electrostatic unit 

 within it. It will be shown presently, however, that the productivity in nuclei 

 depends directly on the bulk of licjuid subjected to shaking, and that the number 

 of nuclei obtained in the example computed is abnoimally small. Still the residt 

 that a very small quantity of electi'icity would be in (piestion, in any case, is 

 already apparent. 



S. Time losses. — These nuclei gradually vanish in the lapse of time for j-easons 

 which can not be easily discerned, since they diffuse against gravity and subsidence 

 is out of the question. There may be some species of decay or evaporation, or the 

 nuclei may be absorbed on contact with the walls of the vessel in which they are 

 contained. That nuclei are removable in the latter way is clear from the efficiency 

 of precipitation when used for this purpose. The removal by evaporation should 

 be an exceedingly rapid process, and, in fact, in the case of water, which is the 

 liquid most easily kept pure, the nuclei are phenomenally fleeting. The nuclei 

 from even very dilute and volatile solutions, however, often persist for over a day. 

 They vanish not adventitiously, but in accordance with a definitely prescribed law, 

 however difficult it may be to follow it. They vanish, moreover, most rapidly 

 when the number of nuclei produced is least, and therefore the evanescence can not 



