72 



EXPERnrENTs wrrn ro>nrzED air. 



contact with the charged electrometer. The intervening space was suitably ionized 

 liy a small piece of phosphorus, about as large as a split pea, suspended at tlie center. 

 From such a condenser no ions can escape ; it therefore offers greater theoretical 

 simplicity than the plate condenser as discussed in Chai)ter IV., while the two dis- 

 criminatiiig variables, radius R and voltage V, may be changed at will. Practically, 

 however, the experiments are very difficult not only because of the baffling iriegu- 

 larity of the ionizer, but because it is not easy to lead a charge to oi- from the 



Figs. 3 to 15.— Charts Showing Successive Electrometer Deflections (s), and their Derivatives, in the 

 Lapse of Time (/), for Different Radii (R) of the Condensers. A'M, = A'</(logr)/n'/. 



center, without conduction. Nevertheless the results as a whole show, I think, that 

 decay due to the mutual destruction of ions is virtually absent in the saturated 

 emanation, and that increased dilution effectively generates new ions. 



2. Apptu-afiis. — A series of Konigs resonators seemed very suitable for the 

 present purposes, since they were at hand in a large range of diameters, and figure 

 1 shows the original adjustment. Ji is the biass lesouator ; B the curl of wire 

 making the inner sui-face of the condenser, and supporting the piece of phosphorus 



