EFFECT OF PRESSUKE ON CONDUCTIVITY OF METALS. 85 



men for the heating element and the thermo-couple. The heating 

 element must be of fine wire in order to have the requisite resistance, 

 and the conditions of the problem demand that there be a cylindrical 

 surface somewhere surrounding the heating element which is at con- 

 stant temperature. That is, a surface at constant temperature is 

 demanded both external and internal. I attempted to satisfy the 

 requirement at the inner surface in the same way as at the outer sur- 

 face by enclosing the heating element in a copper tube. The tube 

 was small, usually about 0.040 inch outside diameter and 0.020 inch 

 inside diameter. The outer surface of the tube was originally tinned 

 and the specimen cast around it. I soon found the same difficulty in 

 making good thermal contact as at the outer surface. If the speci- 

 men were melted after use and the copper tube examined it was very 

 seldom that the surface was found completely wet with the metal of 

 the specimen. But bad thermal contact at the center was not so 

 serious as at the outer surface, because of the much smaller dimensions. 

 Considerable improvement was made by using a tube of silver instead 

 of copper. The difficulty here is due to the exceptional affinity be- 

 tween the silver and the metal of the specimen. If the mold is pre- 

 heated, as is necessary to get a coherent casting, the alloying action 

 between the metal and the silver is in many cases so great as to result 

 in complete eating away of the silver tube. This effect was avoided 

 by an arrangement of the mold by which the silver tube was drawn up 

 into it from below after it was already filled with the molten metal, 

 and the metal immediately chilled from the bottom up, so as to avoid 

 the formation of blowholes. Specimens were made in this way and 

 measured of lead, tin, and cadmium, but the method did not work 

 with bismuth, the alloying here being so rapid that the silver was eaten 

 away before the mold could be chilled. 



The attempt to maintain internal equality of temperature by a tube 

 of highly conducting material was finally abandoned, as it had been at 

 the outer surface, and the heating element was placed directly in an 

 axial hole made by casting the metal around a tungsten wire stretched 

 along the axis during casting. The diameter of the nichrome heating 

 element was 0.005 inch; this was enamelled, bringing the diameter to 

 0.006, and the diameter of the hole was 0.007, so that only a small 

 amount of play was possible; furthermore the wire was always kept 

 from direct contact with the specimen by at least the thickness of the 

 enamel. This method of making the center hole could be used for all 

 the metals that could be cast, including lead, tin, cadmium, zinc, 

 bismuth, and antimony. With other metals of higher melting point 



