808 Mr. H. Carter on the Radioacti 



2 he Direct-current Arc, 



re 



A hand-feed arc-lamp of the type generally used for 

 projection purposes was used. The base-plate of this lamp 

 was screwed to the side of a small table, and in this way the 

 arc was brought over the cover of the electroscope, and no 

 trouble experienced with shadows from the frame and 

 mechanism of the lamp. The distance between the arc and 

 cover was 7 inches. A steel tube 5 feet in length and 

 1J inch in diameter was fitted directly over the arc to assist 

 in removing ionized air from the neighbourhood of the 

 electroscope and to reduce the heating effect. 



A sheet of copper foil 0*06 mm. thick soldered between 

 two pieces of sheet zinc was first used as a cover, but a 

 sheet of thin tinfoil mounted with cardboard was afterwards 

 substituted . 



The method of carrying out the test was as follows : the 

 electroscope was charged and the circulating water turned 

 on some time before the apparatus was required to get rid 

 of soaking effects in the insulation of the electroscope and to 

 allow a steady temperature to be attained. The electro- 

 scope was never allowed to become completely discharged to 

 overcome the soaking effect. After a steady state had been 

 reached, readings of deflexion, time, and temperature were 

 taken at 10 minutes' interval, over periods varying from one 

 to two hours. A thick block of iron was then placed 

 between the cover of the electroscope and the arc, and 

 a blank experiment run with all the other conditions 

 precisely the same as before. By this means any light 

 or heating effects were eliminated, so any difference found 

 between the two rates of leak must have been due to 

 the passage of ionizing rays through the cover of the 

 electroscope. 



As a confirmation, and in order to test the accuracy with 

 which the rate of leak would repeat itself, variations in the 

 method were made as follows : a blank test was run followed 

 by an exposed test, two blank tests were run consecutively, 

 and two exposed tests consecutively. 



A large number of tests were run in this manner and 

 curves of time and deflexion plotted in each case. Examples 

 of these are given below (Table I. and fig. 2). 



