Transformation of Thorium and Actinium Emanation. 725 



products introduced into the chamber, was observed. Both 

 methods gave the same values for these ratios. These 

 relations were determined under the same experimental con- 

 ditions as the decay of actinium emanation was observed. 

 For a given capacity in parallel with the leaf its deflexion 

 was noted for a certain number of vibrations of the pendulum, 

 taking place between the raising of the keys K x and K 2 . 

 For a larger or smaller capacity the number of vibrations 

 was counted which gave about the same deflexion as before. 

 The natural leaks were subtracted and the ratio calculated. 

 This eliminated any variation in rates of deflexion at different 

 parts of the scale. The scale was calibrated, however, for 

 the sensitivities used and the necessary corrections made for 

 readings over different parts when obtaining the periods of 

 decay. 



Table I. 



Katio. 



Exp. 

 Value. 



5-03 

 3-91 



4-68 

 1-70 



8-70 



Variation of 



individual 



determinations 



from mean. 



Ratios of capacities with 

 "1, "02, &c, in parallel 

 with leaf to leaf, and ! 

 connexions alone. 



•1+ electroscope, &c. 

 -02+electroscope, &c. 

 •02+electroscope, &c. 

 *005+ electroscope, &c. 

 '005+ electroscope, &c. 

 ! '001+ electroscope, &c. 

 •001+ electroscope, &c. 

 9 + electroscope, &c. 

 C 2 +electroscope, &c. 

 Electroscope, &c. 



•6 per cent. 



1-8 „ „ 

 '8 „ „ 



i-o „ „ 

 i-o „ „ 



13610 

 270-6 

 69-2 

 14-79 

 8-70 



The first column of Table I. gives the various ratios 

 employed. The total capacity used at any time is the sum 

 of the auxiliary capacity and that of the electroscope, cup C, 

 chamber, leads, &c. Column II. shows the experimental 

 ratios obtained. The degree of accuracy with which these 

 values could be determined is stated in column III. The last 

 column records the ratios of the various auxiliary capacities 

 plus electroscope, leads, &c, to that of the electroscope, 

 leads, &c, alone. 



Experimental Errors. 

 Over the interval used in these experiments, i.e., from 10 

 to 12 times the half-value periods, it can be easily calculated 



