ELECTRICAL RESISTANCE UNDER PRESSURE. 09 



remained. The action tends to cease at the higher pressures. Error 

 from chemical action was eUminated as far as possible by taking the 

 means of readings with increasing and decreasing pressure, and at the 

 higher temperatures by never releasing the pressure to atmospheric, 

 but obtaining the zero reading from an extrapolation of the readings 

 at higher pressures, where the chemical action is slower. The tech- 

 nique in handling sodium was the same as that used for lithium. It 

 is a curious fact that although the chemical action of the mixture of 

 Nujol and paraffine at atmospheric pressure is considerably less on 

 the lithium than on the sodium, at higher pressures the reduction of 

 the action is considerably greater in the case of sodium, so that the 

 zero shift after a run at higher pressures and temperatures was greater 

 in the case of lithium than sodium. 



Runs were made on the effect of pressure on the resistance of the 

 bare wire at 0°, 25°, 50° (partial run), 75°, and 96°. The difference 

 between readings with increasing and decreasing pressure decreased 

 uniformly from zero to the maximum pressure, instead of being al- 

 most entirely confined to the zero reading, as was the case with sodium. 

 The zero shifts were 7% of the total effect at 0°, 5.5% at 25°, 7.8% at 

 75° and 18% at 96°. The run at 50° was not completed because of 

 accident. In spite of the large zero shifts, the mean of the readings 

 with increasing and decreasing pressure ran smoothly, and should be 

 only little affected by the chemical action. 



The temperature coefficient of resistance at atmospheric pressure 

 was obtained from a coil of bare wire similar to that of the pressure 

 measurements. In order to avoid as much as possible the effect of 

 chemical action, four thermostats were kept running simultaneously 

 at 0°, 25°, 50°, and 75°. The coil was immersed in a well of Nujol 

 which had previously come to the temperature of the bath. After a 

 reading at one temperature the coil was transferred in a few seconds 

 to the bath at the next temperature, and readings made after a fixed 

 constant interval. Seventeen minutes proved to be sufficient for 

 acquiring complete thermal equilibrium. Readings were made 

 successively from 0° to the maximum and back to 0° again. The 

 mean of the ascending and descending readings should be free from 

 error from chemical action. The zero shift after the run was 2.8% of 

 the total effect, against 5.4% for sodium. The average coefficient 

 between 0° and 100° was 0.00458. Bernini ^ found for lithium in 

 glass the mean value 0.00457 between 0° and 177.8°. He found the 

 relation between temperature and resistance to be linear. I found 

 the resistance to increase more rapidly at the higher temperatures; 



