[bronson] construction OF VERY HIGH RESISTANCES 167 



Even under these favourable conditions, there are many difficulties 

 which are not easily overcome. This is not surprising when it is 

 remembered that the standard testing vessel has a resistance from 10 

 to 100 times as great as the values usually given for the resistance of 

 condensers. In calibrating the standard, the condenser is put in place 

 of the testing vessel of Fig. 1 and the system is charged through the 

 standard vessel. 



In order to eliminate as far as possible any leakage from the 

 electrometer and condenser, the average potential during the time of 

 charging is kept about zero, and the total change of potential of the 

 system is never more than 0.2 volts. The error due to absorption by 

 the ebonite of the standard is not very large, and can be largely elim- 

 inated by applying the potential to the standard a few minutes before 

 taking a measurement. In order to calibrate the standard over a con- 

 siderable range of voltage, for example, from 100 to 0.1 volts, and at 

 the same time not have too great a change in the rate of movement 

 of the electrometer needle, it is necessary to change the capacity of the 

 system. The accurate comparison of these capacities offers the 

 greatest difficulty to the calibration. The writer used a subdivided 

 mica condenser, the sections having capacities from 0.5 to 0.001, micro- 

 farads. Several methods were used in calibrating this condenser. The 

 different methods gave very concordant results for the larger sections, 

 but gave values for the smaller sections which differed in some cases by 

 more than 10 per cent from one another. There is the added dif- 

 ficulty in the use of these small capacities that the potential on the 

 standard is also small. Thus, with one volt on the standard, a change 

 of 0.2 volts in the potential of the quadrants will change the potential 

 on the standard by 20 per cent. 



It has, therefore, been found advisable to use a double method 

 of calibration. If we have some radio-active material whose rate of 

 decay has been accuratel}^ determined (which, of course, originally 

 involved the rate method), we can use this rate of decay as a means of 

 calibration. The active deposit from actinum is probably the best 

 substance for this purpose, as its rate of decay has been carefully de- 

 termined by a number of observers and their results agree very well. 

 The following are some of the best values of the time taken for it to 

 decay to half- value : 



Godlewski 36 minutes 



Meyer and Schweidler 35.8 minutes 



Hahn and Lochur 36.4 minutes 



Debierne 36 minutes 



If we take 36 minutes as the mean value, it cannot be far from 

 the truth. 



