EFFECT OF PRESSURE ON CONDUCTIVITY OF METALS. 119 



—24.7%. It is to be noticed that within the limits of error no differ- 

 ence is to be detected between the cast and the extruded specimens. 



The pressure correction for the transmitting medium was 15.3% on 

 the total conductivity; this means that the final corrected result was 

 three times as large as the observed pressure effect. 



The above results give for the pressure coefficient of thermal con- 

 ductivity — O.O42I, larger than for any other metal except bismuth. 

 I have previously found that the electrical conductivity of antimony 

 also decreases with rising pressure, and at 30° the average coefficient 

 to 12000 kg. is -O.O4IO8. 



Lussana has also measured the effect of pressure on the electrical 

 and thermal conductivity of antimony, and his results are in precise 

 disagreement with mine. He finds that the electrical conductivity 

 increases under pressure, as it does for normal metals. At 25° his 

 coefficient, presumably to 3000 atmospheres, is O.O5874. There must 

 be something vitally wrong here; measurement of electrical resistance 

 under pressure should offer none of the difficulties of thermal conduc- 

 tivity, and there should be no reason for a disagreement as to sign 

 between different observers. The relation between pressure and 

 thermal conductivity Lussana finds to be distinctly not linear. The 

 initial change is at a rate corresponding to a coefficient of O.O525I, and 

 at 3000 the rate corresponds to a coefficient of only O.O5I64. So large 

 a departure from linearity in a metal with as high a melting point as 

 antimony is without precedent. It does not seem improbable that 

 the sign that Lussana found for the coefficients of both electrical and 

 thermal conductivities may be due to a closing of minute fissures 

 between the crystalline grains under pressure, such as Borelius and 

 Lindh found for bismuth.^ The departure of his thermal conductivity 

 from linearity is in accord with this suggestion. 



Petroleum Ether. It has already been explained that it was neces- 

 sary to determine the absolute conductivity and pressure coefficient 

 of this substance in order to obtain the correction due to the effect of 

 pressure on the transmitting medium in the longitudinal flow method. 

 The method adopted for determining these constants for petroleum 

 ether was a radial flow method, and demanded very little change in 

 the apparatus already used for metals. The apparatus is shown in 

 Figure 12. It consists of an inner cylinder of copper held concen- 

 trically within an outer hollow cylinder, which in turn fits closely 

 inside the pressure cylinder, and is maintained concentric with it by 

 the same spring device that was used for the metals. The petroleum 

 ether fills the annular space between the two copper cylinders. The 



