80 BRIDGMAJSr. 



linear expansion the value O.O58 was assumed, and for the cubic com- 

 pressibility Amagat's figure 2.2 X 10~^ with his temperature coeffi- 

 cient of compressibility of 10% for 100°. In order to reduce the part 

 of the table below 100° to relative specific resistances or that above 

 100° to relative "observed" resistances, it would have been necessary 

 to have known the compressibility of sodium over this range of pres- 

 sures and temperatures, and this has not yet been determined experi- 

 mentally. From the difl^erences of the pressure coefficients in the 

 two parts of the table, however, it is possible to get some idea of the 

 magnitude of the compressibility. Thus it will be found that the 

 mean coefficient of "observed" resistance between 5000 and 12000 

 kg. at 100° becomes consistent with the mean coefficient of specific 

 resistance over the same range if the compressibility is 0.00002. 

 Richards found for the initial compressibility at 20° the value 0.000015. 

 The difference between these two values- does not mean an impossibly 

 large temperature effect. 



In Figure 4 the isothermals of resistance against pressure have been 

 drawn for a number of temperatures. The values are taken from the 

 table and have the same discontinuity at 100° as the values of the 

 table. In fact, this discontinuity is quite evident in the figure. 

 The change of resistance with pressure is seen to be large, larger than 

 for any other metal which I have measured except potassium. Under 

 12000 kg. the change of resistance of the solid is of the order of 40% of 

 the initial resistance. The mean coefficient of the liquid is larger, the 

 decrement being about 50% for the same pressure range. The initial 

 coefficient of the liquid varies little with temperature, but the initial 

 coefficient of the solid increases markedh' with rising temperature. 



The pressure coefficient of resistance of sodium has not been previ- 

 ously measured, so there are no other values for comparison, but 

 other observers have measured the temperature coefficient of solid 

 and liquid and the ratio of resistance of liquid to solid at atmospheric 

 pressure. The values of Northrup and Bernini for the coefficient of 

 the solid have been already quoted. It is to be noticed that the 

 values of Northrup and Bernini are for the specific resistance, since 

 their materials were enclosed in a rigid container, whereas my coeffi- 

 cient is of the "observed" resistance, and was obtained on the bare 

 solid. The coefficient of specific resistance is greater than that of 

 "observed" resistance by the linear expansion. Taking as the linear 

 expansion of sodium 0.000069, my value of the coefficient of the 

 "observed" resistance would give 0.00552 for the coefficient of the 

 specific resistance. This value is seen to be much higher than that 



