74 BKIDGMAJS^, 



the results should not have appreciably any larger error than the 

 observed resistances themselves. 



It is obvious that the temperature coefficient of resistance may be 

 found in a way precisely similar to the pressure coefficient. 



Readings were made on the resistance of the liquid as a function 

 'of pressure at 202.5° and 237.4°. The pressure range of the lower 

 temperature run was 8000 kg., since the melting curve restricts the 

 domain of existence of the liquid, but at the higher temperature the 

 pressure range was the entire 12000 kg. In order to avoid chemical 

 action as far as possible, pressure was not released entirely to zero, 

 but the minimum was about 1000 kg., and the results were extrapolated 

 to zero. Measurements were also made on the solid at 171.6° to SOOO 

 kg. (pressure was not raised higher for fear of distorting the capillary), 

 and on the resistance of the solid as a function of temperature at 

 atmospheric pressure down to 0°. 



The most important result is that the pressure coefficient of the 

 liquid is positive like that of the solid, reversing the behavior of bis- 

 muth. At the two temperatures the relation between pressure and 

 resistance was linear within the limits of error of the measurements. 

 At the lower temperature the maximum departure of any observed 

 point from a straight line was 2% of the total effect, and at the higher 

 temperature it was 1.3%. The coefficient is + 0.05927, independent 

 of temperature to the last figure. The coefficient of the liquid is seen 

 to be slightly larger than that of the solid. The correction for the 

 capillary brought the observed value from O.O^TOO to 0.05927. 



The temperature coefficient of the liquid, corrected for the capillary, 

 was such as to give between 202° and 237° an increase of 0.00145 of 

 the resistance at 202° for every degree rise of temperature. This is 

 less than the reciprocal of the absolute temperature, giving, when 

 multiplied by the absolute temperature, 0.689. The coefficient of the 

 solid was found to be markedly higher than the reciprocal of the 

 absolute temperature. This is again an example of the fact that the 

 temperature coefficient of the liquid is in general less than that of the 

 solid. Bernini ^ found for the liquid between 180° and 200° a coeffi- 

 cient equal to 0.00077 of the resistance at 200°, considerably lower 

 than the value above. 



f<. No correction has been applied to the above values for the com- 

 pressibility or thermal expansion of the capillary, since these are 

 not known. We have seen that the temperature correction is in 

 general slight. If the compressibility of the alloy is the same as that 

 of pure iron, which is a not unplausible assumption, the coefficient of 

 the specific resistance will be about 2% less than the value above. 



