108 



BRIDGMAN. 



For some unknown reason the readings on one of these four samples 

 were not satisfactory. The readings on the other three were quite 

 typical of the results by this method, lying on one of several discrete 

 lines. The results obtained with one of these are reproduced in 

 Figure 6. The observed effect for iron is an increase of apparent 

 conductivity of between 5 and 6% under 12000 kg., but the correction 

 for the effect of pressure on the transmitting medium is so large, 5.3%, 

 as to wipe out nearly all the observed effect. The results obtained 



,0 



8 



10 



10^ 



I 2 3 4 5 6 7 



Pressure, Kg. / Crn. X 

 Iron 



Figure 6. Iron. Thermal conductivity on an arbitrary scale against 

 pressure in thousands of kg/cm-. Results obtained with a longitudinal flow 

 specimen. The points lie on several lines of the same slope; the reason for 

 this is explained in the text. 



with the three good samples, after applying all corrections, were 

 —0.6, —0.2, and —0.2% respectively. The discarded data were not 

 inconsistent with these values. We take as the mean —0.3%. 



The pressure coefficient of thermal conductivity given by the above 

 results is —0.063. I found for the pressure coefficient of electrical 

 conductivity between and 12000 kg. +0.05229. Iron is not one of 

 the metals measured by Lussana, so there are no previous values for 

 comparison. 



