﻿in Air and Hydrogen at Atmospheric Pressure. J 77 



II. It was suggested that perhaps unsteady air-currents 

 were influencing the readings, but from the result above 

 noted and from the fact that all the heating was from the 

 top and that the only opening into the measuring-chamber 

 (0, fig. 1) was small and symmetrically placed, it was con- 

 cluded that this was hardly a possibility. This conclusion 

 was confirmed by the next step. 



III. The fact that the strip gave off considerable heat 

 suggested that perhaps the chamber was being heated up, 

 and that the variation in the apparent value of the charge 

 was due to this change of temperature. Then during the 

 early part of the time a strip was being used, the proportion 

 of the charged particles going to the outer ring became 

 gradually larger, indicating either a decrease in the charge, 

 or an increase in the rate of diffusion, and thus lending colour 

 to the above suggestion. However, the fact that this pro- 

 portion always came to a fairly constant maximum value, and 

 that there was no such variation when the negative ions were 

 being tested, rather pointed the other way. 



To test this the variation in the temperature in the 

 chamber " K " was measured. The thermometer used con- 

 sisted of fine platinum wire wound on hard rubber pins, 

 which were screwed into holes drilled as near as possible to 

 the inner edge of the central brass ring, as in figs. 1 and 2, 

 in which the dotted line represents the wire. Fig. 1 gives 

 vertical section ; fig. 2 gives horizontal view of brass ring in 

 question, 



The middle of the wire was soldered to the brass ring at 

 (a) so that it would not disturb the potential gradient 

 appreciably. 



The thermometer showed a maximum variation during a 

 Phil May. S. 6. Vol. 23. No. 133. Jan. 1912, N 



