Prof. Owens on Thorium Radiation. 377 



disturb the original electrostatic field, then the relation between 

 the current and electromotive force becomes more complicated. 

 Under the latter conditions, which are probably nearer the 

 truth than the simpler supposition first made, the observed 

 continuous though slight rise of current with voltage beyond 

 the " knee " of the saturation curve, the gas pressure being- 

 constant, can be explained. A certain part of the rise of 

 current beyond the knee when parallel plates are used must 

 of course be attributed to a lack of definite cross-section of 

 the conducting gas ; but the same thing differing only in 

 amount has been observed when concentric cylinders are used. 

 A better arrangement would probably be concentric spheres. 



The number of ions produced in a given volume of a par- 

 ticular gas by the radiations from an active substance is known 

 to vary as the intensity of the radiation and as the density of 

 the absorbing gas. If the intensity of radiation is constant, 

 the number produced and the current for a constant potential 

 gradient will increase with the gas-density up to a certain 

 point. If beyond this point the gas pressure is further in- 

 creased, recombination becoming more and more active, the 

 current should gradually decrease ; and this has been found 

 to be the case not only with thorium, but with other radio- 

 active materials as well. For similar reasons, if the intensity 

 of radiation, potential gradient, and gas density are main- 

 tained constant, the current should at first rise and then fall 

 as the distance between the plates is continuously increased 

 from a sufficiently small quantity. This also has been found 

 to be the case. 



For studying the relation of the conduction-current to the 

 pressure and density of the gas in which the radiations are 

 absorbed, the apparatus shown in fig. 8 was used. A is a 

 short steel cylinder 5 in. in diameter and 6 in. long. The 

 lower end is closed by a stout iron cap fitted tightly in place. 

 On the upper end is screwed a common fitter's flange, to which 

 is bolted the lid or cover C, suitable packing being placed 

 between the two to make a tight joint. F 1 and P 2 are small 

 pipes screwed through the cover and attached, one to a 

 pressure-gauge and one to an exhaust- or pressure-pump. 

 B x is a removable plate to hold the active material ; B-, is an 

 insulated parallel plate 11 cm. in diameter connected through 

 the graduated rod R to the electrometer ; the rod is insulated 

 by a tight ebonite and rubber joint where it passes through 

 the cover. 



Using a thick layer of thorium oxide on the lower plate, 

 and varying the pressure from 60 mm. to nearly 2800 mm. 

 of mercury, the results given in Table VII. and plotted on 



