86 BRIDGMAN. 



which were used in the form of cylinders the heating element has to be 

 put in an axial copper tube. In these cases the axial hole was 0.040 

 inch in diameter, drilled through the specimen, which was then turned 

 concentric with the hole, and the copper tube, which was not less than 

 0.039 inch in outside diameter, was then sweated into place. After 

 some practise it was possible to get fairly good sweated contact. Since 

 the pressure penetrates the interior of the tube, error from unequal 

 compressibility of the two metals at the interior surface must be small. 

 It is obvious that the use of an axial copper tube to contain the heating 

 element can be unobjectionable only for those metals whose thermal 

 conductivity is small compared with that of copper. The method 

 cannot be expected to give good results for copper itself, for example. 



The thermo-couples as well as the heating element were at first 

 placed in two copper tubes cast into the cylinder, but this was later 

 abandoned for a hole cast into the metal itself. In making the hole 

 for the thermo-couple there are two opposing tendencies that must 

 be guarded against. In the first place the wire of the couple must be 

 so good a fit for the hole in which it is placed that the thermal conduc- 

 tion from the ends of the cylinder along the wires of the couple can be 

 neglected. In the second place, if the fit is too close, pressure is not 

 transmitted freely throughout the interior of the hole when pressure 

 is high and the transmitting medium has acquired a certain amount of 

 viscosity. Such viscosity results in the introduction of stresses into 

 the wire, with an effective change in its thermo-electric constants, so 

 that readings of temperature difference are no longer reliable. The 

 attempt was made to meet these two opposing conditions by making 

 the hole larger at the two ends with a narrow neck in the middle where 

 the junction is situated. The cylindrical specimen itself was 2.5 

 inches long. The narrow part of the hole to receive the junction was 

 0.014 inch diameter and 0.5 inch in length. The larger part of the 

 hole at each end was 0.040 inch diameter, and joined to the smaller 

 part by a conical part 0.25 inch long, thus leaving the 0.040 

 part at each end 0.75 inch long. The diameter of the wire of the 

 thermo-couple was 0.010 inch, and this was brought to 0.012 by the 

 enamel. An approximate calculation shows that these dimensions 

 are amply sufficient to prevent any appreciable flow of heat along the 

 thermo-couple wire to the region of the junction, even with the metal 

 of lowest conductivity. The hole was cast in the shape required by 

 filing down a wire to a neck of the dimensions given at the middle, 

 and holding this stretched through the mold in the required location. 

 After the casting had been made, the wire was removed by pulling 



