Jan., 1916] Electrical Behavior of Porcelain and Glass 85 



of the same kind as those used by Henderson and Weimar. 

 These were moulded in the form of cyhnders 2 cm. in diameter 

 and 2 cm. in length. The ends were ground plane. The 

 method of mounting and operating upon them is shown in 

 Figure 2. 



The furnace proper consisted of two porous battery jars 

 with bottoms removed, placed end to end. Over this resistance 

 wire was wound. This was covered with asbestos paste and 

 baked on. The furnace was then placed in a section of asbestos 

 steam pipe jacket. A thermo couple (T) projected into the 

 chamber of the furnace. It was a Pt Rh-Pt couple which 

 had been calibrated for other work. 



A thick rod of copper (C) projected from below into the 

 heating chamber. This rod had a copper cylinder (E) screwed 

 to the top. The upper surface of the cylinder was covered 

 with Platinum which was securely fastened to the copper 

 cylinder both mechanically and electrically. This Pt surface 

 served as an electrode which pressed against one face of the 

 porcelain cylinder. Platinum was used because it would not 

 oxidize at the temperatures attained. Above the specimen (S) 

 a rod similarly terminated rested on the upper face of the 

 cylinder. This rod had freedom of motion through guides. 

 Changes in length due to expansion were provided for in this 

 manner; it also insured that the specimen would be subjected 

 to a constant pressure during the experiment. A series of 

 mica vanes placed at intervals about the rod served to reduce 

 heat losses due to convection. 



Such a furnace can be regulated by hand, within reasonable 

 limits. The large heat capacity and low conductivity of the 

 materials entering into the construction serve to iron out small 

 fluctuations of current strength. It was found that a steady 

 temperature of 500° C. could be obtained five or six hours after 

 turning on the current and that this could be maintained 

 constant plus or minus 3° for a period of several hours. With 

 such a device when a potential difference is applied, current 

 flows through the ammeter. The specimen is so mounted that 

 the stream lines of flow, such as we ordinarily consider, are 

 perpendicular to the electrodes and parallel to the axis of the 

 cylindrical specimen. The current we wish to measure is of 

 this character. We will call this the Number One type. There 



