248 



PROCEEDINGS OF THE AMERICAN ACADEMY. 



heating current was taken from a transformer in steps of 12 volts, and 

 fine adjustment of the current was obtained by a carbon plate rheostat. 

 The pressure was indicated by a mercury manometer, one arm of which 

 was connected with the furnace ; the other was evacuated and sealed 

 off. The arrangement of the apparatus is shown in Figure 1. It was 

 intended to measure the temperature of the crucible by means of a 



SOOA LIME TOWERS 



X 



i 



En3>=-=^=^ 



=<I)=c 



VACUUM FURh4ACE 



CONNECTIONS 



FOR DETERMINING 



EQUILIBRIUM PRESSURES 



Figure 1. 



thermo-electric j unction entering the furnace through the tower pro- 

 jecting above the water in which the furnace was immersed, and two or 

 three such runs were made. The wires of the couple were insulated 

 from each other by different kinds of tubes. One wire was covered 

 with a tube one millimeter in diameter, and this tube and the other 

 wire were placed in a larger tube sealed at the end covering the junc- 

 tion. These tubes were long enough to reach up into the tower when 

 their lower end was resting in the crucible. The hot end of the tube 

 surrounding the wires of the thermo-electric couple was covered with a 

 short graphite tube to protect it from the material in the crucible. No 

 tubes that were available, however, would stand a temperature above 

 1200° C. Even Berlin porcelain was spoilt after a few hours' heating. 

 Carbon seemed to have penetrated the whole mass of the porcelain 

 where it was surrounded by the graphite spiral heater, and as much of 



