726 Dr. R. W. Bo}de on the Behaviour of 



experiments varied over a wide range, it was necessary in 

 measuring the ionizations to use instruments varying widely 

 in sensitiveness. With the larger amounts of emanation a 

 Kelvin astatic galvanometer, of which the greatest sensitive- 

 ness was I scale-division (millimetre) for 1*3 x 10~ 10 ampere, 

 could be used ; with the smaller quantities a gold-leaf elec- 

 troscope sufficed. 



When using the galvanometer and large quantities of 

 emanation, it was necessary to reduce the pressure in the 

 ionization chamber in order to obtain saturation. But the 

 pressure could not be too far reduced, for a diminution of the 

 pressure caused a decrease in the current, and it was desirable 

 to work with a fairly large deflexion of the galvanometer- 

 needle. (Increased voltage sometimes helped to this end.) 

 In general, approximate saturation with satisfactory deflexion 

 was obtained by manipulating the pressure, the voltage, and 

 the position of the control magnet of the galvanometer. The 

 maximum deflexions in different experiments varied from 

 200 to 400 divisions at a scale-distance of 1*41 metres. 



When using the gold-leaf electroscope the only possible 

 adjustment to secure saturation was to lower the pressure of 

 air in the ionization chamber. 



Some possible causes of error which may enter into the 

 experiments should be mentioned. The emanation produces 

 gases — mostly carbon dioxide — by its action on the marine 

 o-lue with which the mica sheet was secured to the containing 

 tube. If the volume of the tube were very small, as in the 

 case of a capillary, these gases would have an effect in re- 

 tarding by diffusion the passage of the volatilized emanation 

 from the lower to the upper part of the tube. Again, with 

 capillaries, another error due to viscosity would affect the 

 readings in the same direction. The frictional resistance of 

 the gas in the capillary would retard the passage of the 

 volatilized emanation from the lower to the upper part of 

 the tube. Some trial experiments showed the necessity of 

 avoiding these troubles. 



The best conditions of experiment were 



(1) to use well cleaned tubes of not too small a bore, and 

 therefore of not too small a cubical capacity ; 



(2) to perform the experiment as soon as possible after 

 the admission of purified emanation into the tube ; 



(3) to obtain as nearly as possible saturation of ionization. 



From a number of experiments in which these conditions 

 were fulfilled the curves shown in fig. 2 are given as samples. 



