116 



BRIDGMAN. 



The readings on the hquid went very smoothly. ^Yithin the Hmits 

 of sensitiveness there was no difference between points obtained with 

 ascending or descending pressure, and except for one point, all of the 

 observed points lay on smooth curves within the sensitiveness of 

 reading, which was about 0.2% of the total effect. The observed 

 values, smoothed for temperature and pressure, are shown in Table 

 XIII in terms of the resistance at the melting point at atmospheric 

 pressure as unity. The A^alues tabulated are "observed" values, 

 that is, the\' have not been corrected for the thermal expansion or 



TABLE XIII. 



Bismuth. 



Relative Values of Observed Resistance of the Liquid in Glass Capillary. 



The resistance of the hquid at atmospheric pressure and 271.0° is taken as 

 unity. 



compressibility of the glass capillary. It did not seem best to do 

 this because of the uncertainty in the values for the glass at the tem- 

 peratures and pressures of the measurements. The glass used for the 

 capillary was an ordinary soft soda glass. The best value for the 

 cubic compressibility is probably 2.7 X 10~^, taking Amagat's values 

 for the compressibility and temperature coefficient of compressibility, 

 and for the linear thermal expansion the best value is probably 8 X 10"^. 

 The most important result shown by the table is that the pressure 

 coefficient of resistance of the liquid is negative like all normal metals. 



