84 



BRIDGMAN. 



ing the irregularities introduced by the glass capillary were apparent. 

 The individual readings seldom showed irregularities of much more 

 than the sensitiveness of the measurements, but there were consistent 

 departures between readings with increasing and decreasing pressure, 

 and consequent uncertainty as to the correct result. Sometimes an 

 adjustment of as much as 2% in the total resistance was necessary to 

 bring the runs for different temperatures into smooth register with 

 each other. Because of the extreme largeness of the coefficient, a 

 change of 2% in the resistance itself usually means a much smaller 

 percentage change in the decrement of resistance. Since it was the 

 decrement which was measured, the actual measured quantities 



TABLE IV. 

 Resistance of Potassium. 



seldom had to be adjusted by as much as 2%. In making the ad- 

 justments, the observed temperature coefficients of both solid and 

 liquid at atmospheric pressure was accepted as most probably correct 

 and retained with only immaterial smoothing. Also the observed 

 change of resistance on passing from the liquid to the solid at various 

 temperatm'es was accepted as essentially correct. With these as 

 fixed data in the table of resistances, the other measured values for 

 the decrement of resistance were adjusted to smoothness with as little 

 ■change as possible. 



The resistances as thus smoothed are shown in Table IV. The 



