843 



filament current. Fig. 3 sliows ciirnits Nr. 2. Nr. 3 and the ain|ilifier 

 in more detail. 6" is the inductance of Nr. 2. The valves used have 

 been Mullard or Philips receiving valves. The circuit connections 

 in Nr. 1 and Nr. 2 have heen those of the usual Harti,ky circuit a tap 

 off at the middle of the coil leading to the filament and the extre- 

 mities going to the grid and plate batter3- terminals. A condenser 

 was connected across the coil terminals. All the parts of circuit Nr. 1 

 were thoroughly shielded. The coupling to the amplifier A was 

 accomplished by a wire entering the box B through a hole on the 

 side opposite to that shown Fig. 1. This wire was connected to the 

 input of A either directly or through a transformer. The input 

 terminal of A was also coupled by a wire to circuit Nr. 2. The loose 

 end of this wire was stretched in over a rope towards circuit Nr. 2. 

 Circuit Nr. 3 was controlled by a pulley arrangement allowing one to 

 turn its condenser plates without introducing capacity due to the 

 observer's body. The observer was situated at the outstanding corner 

 of B in Fig. 1. The telephones were used on the observer's 

 head. The coupling in this ap|)aratus was sutïiciently loose to make 

 the capacity effect through the phones negligible foi- the small 

 motions of the observer's body during the measurement. The liquid 

 air in the vacuum glass inside C produced sufficient cooling of it 

 to cause unsteadiness. It thus was necessarj' to have a rapid metiiod 

 of measurement. 



For this purpose a system Q (Fig. J) was attached to the rod 

 sliding through P. (2 consists of a horizontal metal rod carrying 

 three collars. The central one is attached to a vertical guide passing 

 through a collar almost directly above P. The motion of this guide 

 can be controlled l)y the lever shown. The collar on the right is 

 connected to P by a vertical rod and the collar on the left carries 

 a glass rod from the lower end of which is suspended a glass tube 

 by means of a thread. The metal rod passing through P ended well 

 inside the tube of Geiman silver supporting 7-" and was thus shielded 

 from the action of the magneti<' field of C. To the lower end of 

 that rod a glass tul)e was fixed with sealing wax and this tube 

 supported the |)aramagnetic specimen. The paramagnetic specimen 

 and the suspended tube could be thus moved in and out of C' 

 simultaneously. Inside the glass tube a single or several copper 

 wires were put and these were selected in sucii a way as to make 

 the number of beats per second in the upper and lower positions 

 the .same. The combination was recorded and a subsetpient caliliration 

 determined the magnitude of the effect. The range of motion was 

 fixed by permanent stops on the sliding rods. This was necessary 



55* 



