2 //. ^4. Bum stead — Atmospheric Radio-activity. 



The experiments which I am about to describe were begun 

 before the appearance of the recent papers by Elster and 

 Geitel and by Allan, and were continued in view of the dis- 

 crepancy between their results, and also because indications 

 had been obtained of results not observed by either. A fine 

 copper wire \ mm in diameter w^as suspended horizontally, about 

 8 meters above the ground, between two neighboring build- 

 ings ; it was attached to the negative pole of a Wimshurst 

 machine, driven by a small motor, the positive pole of the 

 Wimshurst being earthed. A parallel spark-gap 5 mm long 

 served to keep the potential-difference constant during an 

 exposure which was usually continued for three hours. The 

 activity of the wire was observed in a cylindrical testing vessel, 

 with the central rod connected to one pair of quadrants of an 

 electrometer ; the rod was protected by an earthed guard-plate 

 in the usual manner. The bottom of the testing vessel was 

 easily removable and bore four vertical brass rods just within 

 the walls of the cylinder, and about these rods the exposed 

 wire could be wound. The electrometer was of special con- 

 struction, with sulphur insulation and adjustable quadrants ; 

 the needle was of silvered paper and the suspension a quartz 

 fiber dipped in a solution of calcium chloride to render it con- 

 ducting, as in Dolezalek's electrometer. With the suspension 

 used and 90 volts on the needle, the sensitiveness was sufficient 

 (250 cm per volt with the scale at one meter;, and the instru- 

 ment was very steady. When the potential on the needle was 

 kept constant, the sensitiveness of the apparatus to a small 

 sample of uranium oxide did not vary appreciably over a period 

 of several days. Accidental motions of the needle were largely 

 avoided by enclosing the connecting wire between electro- 

 meter and cylinder in an earthed brass box, outside of which a 

 small electromagnet served to insulate or short circuit the 

 quadrants.* The exposed wire could be put into the cylinder 

 without disturbing the connections or jarring the electrometer, 

 so that readings could be begun immediately ; the only time 

 lost after the end of the exposure was in taking in the wire 

 and winding it about the frame. After the wire had been put 

 into the testing vessel, measurements of the ionization current 

 were made at intervals of three minutes during the first part 

 of the experiment when the change in activity was rapid ; 

 later, readings were taken every six minutes. 



In the earlier experiments, comparatively short wires were 

 exposed (5 meters) and the decay of the activity of these could 

 be followed with some accuracy for about two hours. The 

 similarity to the behavior of the excited activity due to radium 



* This arrangement has been previously described. This Journal, Feb. 

 1904, p. 100. 



