l-'6 



TIIK STUITCTUKK OF THE Nt'CLKFS. 



in benzul (table 28) is on tlie other haml very tletinite as regards the increment of 

 n, due to concentration. (Jt'. figure IT. 



A curious result was incidentally obtained on admitting the air too rapidly 

 thi'ough the filter by accident. Air luiclei were thus introduced simultaneously 

 with the nuclei due to shaking. The exhaustion showed two Lrrouiis of coronal 

 [larticles interpenetratinfj and subsiding through each other, one group subsiding 



SO 'W 60 SO m 



Kl<;s. 17-19. — lX)SS OK NrCE.KI IN TIIK L.M'sE OF M INUTES 1 UK DlKFERENT SOLUTIONS. 



relatively fast and showing intense corona.s, the other, much more slowly with 

 weaker coronas. In onler that the two phenomena may [)reserve their identity 

 throughout, it is necessary that one grou[) of nuclei should regularly exceed the 

 other in size, thus insuring earlier condensation on the larger particle.s. This occur- 

 rence of interpenetrating subsidences was still observed in the next exhaustion, 

 though more faintly. As precipitation pi-oceeded tliiee strata were visible, the 

 upi)er corresponding to clear air, the lower tw<i showing coronas of different .size 

 and intensity and different rates of subsidence. Both fog banks were capped by 

 descending planes. This observation recalls the behavior of snl[)hnr and punk 

 nuclei in Chapter III, § 7. 



F 



