432 Mr. J. Henry on the Deflexion by Magnetic 



exhausting the discharge-tube of these gases owing to their 

 deleterious effect on the pump. 



Figs. 2, 3, and 4 show the deflexions obtained in hydrogen, 

 air, and carbonic acid gas respectively; the deflexion in chlorine 

 was much nearer that in air than in hydrogen. 



Fig. 2. Fig. 8. Fig. 4. 



These figures show the discharge deflected to both sides of 

 the straight path, as in the photos. Each plate was exposed 

 to the discharge deflected first to one side, then to the other, 

 by reversing the current in the magnetic-field coils. The 

 figures are the same size as the photos. 



The actual distance between the terminals was Fig. 5- 

 8'8 cms. 



During these experiments I noticed that the cur- 

 vature of the deflected discharge was not uniform 

 but was always greater at the upper electrode, as 

 shown in fig. 5, the discharge being vertical. It did 

 not matter whether b was the positive or the negative 

 electrode. 



The distortion was always much greater for dis- 

 charges passing through gases at high pressures than 

 at low pressures. 



Bending of the Discharge by Convection- Currents. 



To try if the distortion, above referred to, was caused by 

 convection-currents due to the unequal heating of the gas in 

 the discharge-tube, I turned the tube into a horizontal 

 position, and then found that for air (the gas I was using at 

 the time) the discharge was always bent vertically upwards, 

 the deflexion being very large for high pressures of the gas, 

 and vanishing for low pressure. 



When the discharge passed, it could be seen to go, at first, 

 straight between the electrodes, and then to be quickly 

 carried upwards as the gas got heated, until a steady position 



