EFFECT OF PRESSURE ON CONDUCTIVITY OF METALS. 79 



in that they carry the determination of thermal conductivity back to 

 the definition. A known steady supply of heat was put into the speci- 

 men, and by means of a thermo-couple the difference of temperature 

 was measured between two points when equilibrium was reached. 

 In terms of the geometry of the configuration the temperature gradient 

 and the rate of heat flow across unit area can be immediately found, 

 and hence the thermal conductivity may be found from its definition. 

 The first method was in theory the simplest, as there were practically 

 no corrections to be applied. This was a radial flow method. The 

 specimen was in the form of a massive cylinder, almost filling the bore 

 of the pressure cylinder. Along the axis of the cylinder was a linear 

 source of heat, and the difference of temperature was measured be- 

 tween two points at different radial distances from the axis. The 

 formula for thermal conductivity in this case is 



Q r. 



where h is the conductivity, Q is the heat input per unit length of the 

 axis, ri and r-i are the radial distances of the two junctions of the 

 thermo-couple, and ^i and 6^ are the temperatures of these two junc- 

 tions. 



The flow of heat under these conditions is radial, except at points 

 near the ends of the cylinder. By locating the thermo-couple midway 

 between the ends of the specimen any end effect may be avoided. 

 The method is therefore unusual in that there is no correction for heat 

 leak. Further there is no correction for the change of dimensions of 

 the specimen under pressure, for it is only the ratio of the two radial 

 distances that enters the formula, and this is not changed by a hydro- 

 static pressure which uniformly changes the dimensions in every 

 direction. There is a correction in the heat input due to the change of 

 length of the heating unit under pressure, but this correction is equal 

 to the linear compressibility, and is so small as to be almost negligible. 

 The heating element was made of nichrome. Assuming the com- 

 pressibility of nichrome to be calculable by the law of mixtures from 

 that of nickel and chromium, its two constituents, the magnitude of 

 this effect is 0.2% under 12000 kg/ cm-. The correction is an addition 

 to the observed pressure effect. There is also to be considered the 

 change of resistance of the heating unit under pressure, but by a 

 suitable arrangement of the circuits this may be made to eliminate 

 itself, as will be seen later. 



I designed this method of radial flow to meet the needs of the high 



