84 BRIDGMAN. 



The attempt to maintain the external surface at constant tempera- 

 ture by using a sheath was finally abandoned, and the specimen was 

 maintained concentric in the pressure cylinder by an arrangement of 

 springs. The specimen was made larger than when the sheath was 

 used, being now only 0.02 inch smaller than the internal diameter of 

 the pressure cylinder. Between the surface of the specimen and the 

 walls of the pressure chamber were placed six longitudinal strips of 

 german silver 0.002 inch thick, and about 0.42 inch wide. A sectional 

 view through the specimen and the pressure cylinder is shown in 

 Figure 1. The strips are shown bent to a nearly circular figure be- 



FiGURE 1. Section through the pressure cyHnder and the radial flow speci- 

 men, showing the method of keeping the specimen concentric by springs. 



tween the walls of the cylinder and the specimen. It is obvious that 

 these springs will take up any stretch in the internal diameter of the 

 cylinder, and maintain the specimen concentric at all pressures. The 

 springs are so thin, and the thermal conductivity of german silver is so 

 low that any error due to flow of heat away from the specimen through 

 the springs is small, and need not be expected to change greatly with 

 pressure. But as a precaution against any error from this effect the 

 specimen was wrapped with a layer of paper between the springs and 

 itself, thus minimizing any irregularity due to unequal heat loss at 

 different parts of the surface of the specimen arising from inequalities 

 in thermal contact between spring and specimen because of geometrical 

 imperfections. 



As far as I can judge, this method of maintaining the outer surface 

 of the specimen as constant temperature was successful, but there 

 were numerous other sources of irregularity. The chief outstanding 

 mechanical problem now was that of getting the holes into the speci- 



