138 



PHENOMENA, ATOMS, AND MOLECULES 



In order to determine this rate of surface heating for various kinds 

 of flames, the device shown in Figure 2 was constructed. It consists 

 essentially of a short rod of tungsten, A, so arranged that heat from the 

 flame can be applied only to the circular surface forming its upper end, 

 while its lower end is in good thermal contact with a large mass of copper, 

 B, which removes the heat from the tungsten rod as fast as it is delivered 

 by the flame. To prevent the flame from striking the sides of the rod, a 

 water-cooled circular plate or guard ring of copper, C, is used. To obtain 

 good thermal contact between the tungsten rod and the copper, the mass 

 of copper B was made by casting it around the tungsten rod which had pre- 

 viously been covered with copper by heating with copper in an atmosphere 

 of hydrogen. The tungsten rod had a diameter of 0.77 cm. and the portion 

 extending out of the copper mass had a length of about 0.8 cm. The hole 

 in the copper plate was 0.95 cm. in diameter. The weight of the copper block 

 B was 8100 grams and the diameter of its cylindrical part was 10 cm. 



The temperature of the copper block was measured at half-minute 

 intervals by means of a mercury thermometer placed in a hole which had 

 a diameter only slightly larger than the thermometer and into which oil had 

 been introduced to make better thermal contact between the thermometer 

 and copper. 



When the atomic hydrogen or other flame was directed against the 

 tungsten tip the rate of temperature rise of the copper block gave a measure 

 of the rate of surface heating. The flame was applied for a definite time 

 which ranged from 2 to 5 minutes, during which time the temperature 

 rose 40° to 80° C. After the flame was removed the thermometer readings 



Fig. 2. — Apparatus for Measuring Rate of Surface Heating Produced by Flames. 



continued to rise for a couple of minutes because of the time needed for 

 the heat to become uniformly distributed throughout the copper block. 

 Readings were continued for several minutes more in order to determine 

 the rate of cooling so that correction could be made for the slight loss of 



