Conductivity along\a Glow Discharge and in Flames. 359 



each forming an arc of about 90°, to the ends of the Lecher 

 system, closely fitting round the discharge- tube in the way 

 as shown in fig. 4. It was then found, however, that through 

 the more intense action of the waves on the glow-discharge 

 the latter was disturbed by the presence of the waves. The 

 Faraday dark space appeared in this respect to be a very 

 sensitive part of the discharge. When, namely, the Faraday 

 dark space was brought in the plane of the Lecher wires and 

 the waves were started, the striated positive column suddenly 

 increased in length, thus producing one or two more stria- 

 tions. If thereupon the tube was gradually moved so that 

 the anode approached the parallel wires, the marked tendency 

 of the positive (striated) column was invariably observed to 

 keep the last stria just between the Lecher wires. The first 

 abnormally extended column could in this way be com- 

 pressed like an elastic body till it extended over a length of 

 one or two striations smaller than normal. If the tube was 

 moved still further a point was reached where the com- 

 pression went too far and the first striation suddenly jumped 

 through, the positive column having again its normal length 

 and number of striae. 



The first few big amplitudes of each of the damped wave 

 trains account for this interference. 



The Blondlot exciter was therefore replaced by a three- 

 electrode thermionic tube generating continuous oscillations 

 of a very high frequency, as described elsewhere. At the 

 same time another discharge-tube was blown having a bigger 

 diameter (about 32 mm.) with the aim of getting the stria- 

 tions at greater distances apart. Again two pieces of wire, 

 each forming an arc of 90° round the new discharge-tube 

 (fig. 4), were embedded in a groove in the ebonite support A 

 and connected to the parallel wires. Special care was given 

 to make this arrangement accurately symmetrical with respect 

 to the plane of the Lecher wires. The continuous waves 

 produced by the thermionic generator had a wave-length 

 X = 376 centimetres and were very steady. 



The use of the continuous waves proved to be a great 

 improvement. Though it was still possible for the waves, 

 when the generator was brought very near the oscillator, to 

 disturb the discharge, this disturbance could be eliminated 

 by reducing the coupling between oscillator and resonator, 

 and hence a state could be reached where, while the galvano- 

 meter deflexion amounted to 15-30 centimetres, no direct 

 action of the waves on the discharge could be discovered. 

 In order to make sure that this was not the case, the aspect 

 of the discharge was carefully watched while the waves were 



