Measurement of Electromagnetic Radiation. 



49 



mirror was attached to the rigid glass hook from which the 

 wire was suspended, and this could be seen through a plate- 

 glass window in a torsion-box above the jar. The oscillator 

 consisted of two pieces of brass tube half an inch in diameter 

 with hemispherical ends? each 6 inches long in all. This was 

 placed in a vertical position two or three feet away from the 

 wires, and in the plane of the suspended wire, so that it 

 should not act directly upon it.' The two parts of the oscil- 

 lator were connected by very fine wires with the terminals of 

 a Voss machine. Then it was found that when the halves of 

 the oscillator were separated too far for a spark to pass, there 

 was no deflexion, while when they were gradually made to 

 approach one another, no motion was visible on the scale until 

 sparks passed, and then the motion was evident enough, the 

 light even going in favourable cases off the scale in the right 

 direction. 



However, this was only a preliminary experiment made 

 with materials at hand. It was clear that the material which 

 was most suitable was that which had the highest conductivity 

 for its density, namely, aluminium, and that only the outer 

 surface of the wire could be efficacious, since the oscillations 

 were at the rate of about 500,000,000 a second. Accordingly, 

 thick aluminium wire was rolled into thin ribbon, which was 

 then drawn through successive holes in a draw-plate until a 

 perfect tube, cracked down one side of course, was formed. 



We used resonators made of this material in the next ex- 

 periments, which were made with the view of 

 seeing whether the couple observed was due Fig. 3. 

 equally to the repulsion of the ends and attrac- 

 tion of the centre, or if one preponderated over 

 the other. The fixed cranked resonator was 

 then replaced by a straight one, which was 

 held at first close and opposite to the ends of 

 the suspended cranked resonator, when a 

 strong repulsion was observed, as expected. It 

 was then held in the corresponding position 

 opposite the centre part, when, instead of the 

 expected attraction, feeble repulsion was ob- 

 served ; and finally, on being placed equidistant 

 from the two, the ends were repelled away. 



Since this appeared to show that the elec- 

 trostatic repulsion was much greater than the 

 electrodynamic attraction, we tried the following- 

 experiment, in which the electrostatic action 

 was reduced to practically nothing. A pair of 

 resonators were made of the form shown in 

 fig. 3, and one was weighted with a non-conductor and sus- 



Phil Mag. S. 5. Vol. 31. No. 188. Jan. 1891. E 



o 



