Force of the Electric Discharge through Gases. 433 



was attained where the discharge remained permanently 

 deflected ; for very high pressures of the gas in the discharge- 

 tube there was no steady position (see fig. 18, below); in this 

 case the path of the discharge, after being bent upwards, 

 always fell back to the straight line again, as if it were being 

 blown out. 



Fisr. ft 



The curvature of the discharge, as deflected by convection- 

 currents, was not nearly uniform for pressures in the discharge- 

 tube of 4 or 5 cms. of mercury, but appeared as shown in fig. 6, 

 the curvature being much greater near the negative electrode; 

 as the pressure of the gas was diminished, the curve became 

 more symmetrical until it finally became quite circular. 



The discharge in carbonic acid gas was bent by convection - 

 currents rather more than in air ; but in hydrogen the effect 

 was so small as to be almost imperceptible. 



Rearrangement of Apparatus. 



To get rid of the effect of the convection-currents in the 

 photographs of the discharge, the direction of the magnetic 

 force was made vertical, and that of the discharge horizontal, 

 so that the horizontal component of the deflexion was due 

 entirely to the magnetic force. 



The camera had now to be fixed in a position vertically 

 over the discharge, and with the sensitive plate horizontal. 

 By this means the horizontal component only of the deflexion 

 showed in the photographs, and the deflected discharge then 

 appeared of nearly uniform curvature except for high pres- 

 sures of the gas — 3 to 4 cms. of mercury — when the ™ 

 curvature increased greatly towards the middle of the 

 discharge ; but with the point of maximum curvature 

 always nearer to the negative electrode, as shown 

 in fig. 7, where a is the negative electrode and b the 

 positive. On the other hand, for low pressures of the 

 gas — less than a centim. of mercury — the curvature 

 increased towards the positive end of the discharge. 

 For a pressure of 1 millim. the discharge in air or 

 carbonic acid gas was turned at right angles to the 

 positive electrode by a strong magnetic force, and the 

 positive column was deflected till it pressed against 

 the sides of the tube. 



In hydrogen the deflexion showed similar changes for low 

 pressures, but on a smaller scale; for high pressures the 



'c 



