Electric Discharge in rarefied Hydrogen and Oxygen. 293 



The space occupied by mercury and the part of the drying 

 vessels occupied by pentoxide of phosphorus are indicated 

 by shading. 



The electrodes A. B were connected by means of the rods 

 a, E to the ends of a battery of small lead cells, whose voltage 

 lay between 1800 and 1900, through resistances including 

 when necessary a megohm. And in the few cases when the 

 battery failed to initiate the discharge, which however it 

 always easily maintained, the discharge was started by an 

 induction-coil whose secondary had its ends connected to 

 wires surrounding but not touching the cathode and anode. 



Method of Experiment. 



Preliminary experiments (made with air) showed that the 

 force in the positive column is independent of the current so 

 long as the pressure is constant, but changes, although less 

 than proportionally, with the pressure, and that the force at 

 other points of the discharge changes with the current as well 

 as with the pressure. 



Nevertheless, by keeping the current and pressure constant 

 the potential difference between the electrodes is made a 

 function of the distance between the electrodes. Hence, if 

 with the distances between electrodes as abscissas and poten- 

 tial-differences of electrodes as ordinates a curve is constructed 

 corresponding to a constant pressure and current, the electric 

 force in the positive column corresponding to that pressure 

 and current can be determined by the final slope of the curve, 

 when the curve becomes straight. 



The chemical reaction in the positive column should be 

 exhibited by a similar method. For I have already shown 

 in previous papers that the chemical reaction, represented by 

 the fall of pressure Ap, is proportional to the quantity of 

 electricity AQ which passed during the discharge, the cir- 

 cumstances ot' the discharge being constant. Accordingly, 



~r should be a function of the distance between the elec- 

 AQ 



trodes when the current and pressure are constant. This 



will certainly be the case if my hypothesis is correct that the 



chemical action is the outcome of the dissociation into atoms 



of the molecules of the gas by the impacts of the gaseous 



ions, for the motion of these ions is wholly determined by 



the electric force. If, then, a curve is plotted whose abscissas 



are the distances between the electrodes and whose ordinates 



Ai> 

 are the values of -rjr for a given pressure and current, the 



chemical reaction per centimetre in the positive column would 



