86 



BRIDGMAN. 



coefficients of this substance. The correction for the thermal ex- 

 pansion of the glass is perceptible as a change of one unit in the last 

 figure in only one part of the table, and the maximum correction for 

 the compressibility is 0.96% of the corresponding resistance. It is to 

 be noticed that the table contains relative specific resistances, and not 

 relative "observed" resistances, as is the case for all metals which are 

 measured in wire form. It is also to be noticed that the table con- 

 tains the resistances of both solid and liquid; the domain of the solid 

 is separated from that of the hquid l^y a heavy Une. 



The data of the table have been plotted in Figure 5. The first and 

 most striking result shown by the table and the figure is the great 

 magnitude of the effect; it is not equalled by any other metal, and is 

 exceeded only by black phosphorus among the substances which I 

 have measured. At 25° the resistance of the soUd at 12000 kg. is only 

 27.5% of its resistance at atmospheric pressure. At the higher 

 temperatures, where change of phase occurs, the resistance at 12000 

 may be as little as 13% of the atmospheric value. At the melting 

 temperature at atmospheric pressure the coefficient of the liquid is 

 greater than that of the solid, but the coefficient of the liquids shows a 

 quite marked decrease with rising temperature. 



TABLE Y. 



Potassium. 

 Relative Values of Specific Resistance on the Melting Curve. 



In Table V are shown the resistances, also in terms of the resistance 

 of the solid at 0° at atmospheric pressure as unity, at the melting 

 points. ^Yithin the limits of error, the ratio of specific resistance of 

 the liquid to that of the solid at the melting point is nearly independ- 

 ent of the pressure and temperature of melting. The experimental 



